Handbook of Renewable Energy 9783642394874

Citation preview

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Advanced Generator Design for Wind Turbines David A. Rivkin* College of Science and Technology, Sustainable Methods Institute, San Francisco, CA, USA

Abstract This chapter will provide an overview of prior generator design and a discussion of the limitations and challenges in overall wind turbine system design including global sourcing and environmental concerns and conclude with a discussion of advanced generators specifically designed for use on large wind turbines, based on electrostatic generation rather than electromagnetic design. This chapter will discuss both the mechanical and electrical issues related to the design of electrical generators for wind turbines. The mechanical issues will include the mass as a pendulum and its impact on tower design and control system design, aerodynamic impacts of generator dimensions, fixed versus dynamic generator torque control, and material issues including international barriers to supply and environmental concerns in production and deployment. The electrical issues will include AC grid synchronization, HVDC integration, transformer losses, and energy storage compatibility. The three major generator designs, induction, permanent magnet, and electrostatic, will be discussed related to the pros and cons of each generator design in overall wind turbine design.

Keywords Generator; Electrostatic; Electromagnetic; Power control; Induction; Synchronous; Permanent magnet

Introduction Throughout the history of wind turbines and water power systems, they have been used to power industry and provide for the common good. Until the advent of the modern wind turbine, they were entirely mechanical power systems, turning drive shafts and gears and powering water pumps, flour mills, and factory equipment that drove the industrial revolution. Today’s wind turbines and hydropower systems are almost exclusively electricity-generating systems to fit our electricity-oriented world. Much of the world has substantial electricity distribution networks or “grids” to which large and small wind and hydro systems are connected. Electrical power generators are a key component for wind turbines, hydropower including tidal power, and many other applications. Because wind turbines have unique needs when it comes to electrical generators, we will focus our perspective on this industry. Motors and generators are very similar but generally perform opposite functions. Motors convert electric energy into mechanical energy, so generators convert mechanical energy into electric energy. Both are generically called electrical machines because they can perform either task. In this chapter, you will focus on generators. You will not read specifically about motors, but in general, you can think of a motor as a “reverse” generator in terms of how it works.

*Email: [email protected] Page 1 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

In this chapter, we will first review the major types of rotating electrical generators including recent advancements. They are: 1. 2. 3. 4. 5. 6.

Fixed-speed induction generators Synchronous generators Doubly fed induction generators Permanent magnet generators Electrostatic induction generators Fully rated converter-based generator system

Following this review of the overall designs, we will discuss timely topics related to these generator types: 1. Mass and size of generators and their impact on wind turbine design including: (a) Pendulum nature of wind turbines and the impact on tower design and control system design (b) Aerodynamic impacts 2. Fixed versus dynamic generator torque control 3. Material issues and supply chain issues 4. Environmental concerns in production and deployment 5. AC grid synchronization 6. HVDC integration 7. Transformer losses 8. Energy storage compatibility

Fixed-Speed Induction Generators We will begin our review with fixed-speed induction generators (FSIGs) because this generator is one of the most common types of generators in use on wind turbines installed to date.

Fixed-Speed Induction Generator Overview Fixed-speed induction generators are one of several different types of generators used in wind turbines, hydroelectric facilities, natural gas, and even coal power plants. They are the most common generators found in the currently installed wind turbines. Induction generators are also known as asynchronous generators. Such induction generators can be divided into two main types: 1. Squirrel cage generators 2. Wound-rotor generators Fixed-speed induction generators are very common for three main reasons: 1. FSIGs are relatively simple to design and manufacture, reducing cost. 2. Wind turbines do not like to change RPMs even with changes in wind speeds since wind turbines are large and massive, making their rotational inertia very high. Changing rotational inertia requires significant energy input and also leads to vibration and stress on the entire structure. 3. Easy connection to the grid.

Page 2 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 1 Squirrel cage generator rotor (a) and stator (b) with support rings and conducting bars highlighted

Fig. 2 Wound rotor with laminations and windings highlighted

The so-called squirrel cage generator with rotor and stator is shown in Fig. 1. This is because the rotor does not contain coil windings; instead, it is made using two support rings connected by conducting bars. The bars conduct electric currents and are covered by an electrical insulation to protect them from arcing. The rotor’s appearance resulted in the moniker given. The so-called “wound-rotor” generator is shown in Fig. 2. The moniker is given because rather than conducting bars, the generator has large copper coil windings to transmit electrical currents. Laminations, typically of iron, help focus fields to increase efficiency. The stator of the wound rotor is very similar with laminations and windings. As you may have noticed, both squirrel cage and wound-rotor FSIGs are very similar to AC motors. They are nearly identical in fact, except in their operation and the lack of a second set of windings, the starter windings since these are not needed in a generator.

Fixed-Speed Induction Generator Design Internally, most induction generators use windings that are made of copper cables laid into grooves in the stator and rotor (in wound-rotor designs), though some advanced designs use high-temperature superconducting materials to reduce cable losses. These generators must be cooled using liquid nitrogen however, which can be costly just for operation, on top of the cost of the superconductors. The cables are insulated to ensure protection from the environment and to ensure cable to cable shorts do not occur which would be a catastrophic failure for a generator. The stator’s windings are typically placed at 120
, 240
, and 360
. The stator is a round metal casing, often made of iron, that contains copper windings, for the wound rotor or conducting rods for squirrel cage, as mentioned previously. The rotor rotates inside the stators hollow, round cavity using bearings or slip rings.

Page 3 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

The rotor is a conductive metal cylinder, again either the squirrel cage or copper windings and laminations, that spins inside the stator and has the drive shaft at its core. This drive shaft in a generator is called an armature, and it can be connected to gears, as in the case of most wind turbines, or directly, as in the case of hydro stations. Externally, the design of induction generators is similar to those of all generators as well as large motors. Stator housings are generally steel and designed to protect the interior components from the environment, including moisture, pests, and dust as well as impact damage. Generators come in standardized sizes referred to by NEMA standards or by manufacturers of wind turbines, hydroelectric generators, etc. Generators often must fit inside a given enclosure, such as a wind turbine nacelle, and must be mounted in a way that is defined by the systems. There are two common designs for enclosing a generator for environmental protection: 1. “Open drip-proof” 2. “Totally enclosed, fan-cooled” (TEFC) The open drip-proof design is the most common for wind turbines and other systems where the generator is itself enclosed in a larger housing. This housing is sufficient to protect inner components from environmental hazards and may even be environmentally controlled. However, recent analysis has shown that TEFC designs provide additional protection from damage and can operate where less protection is provided, enhancing usefulness of the generator to a variety of customers. TEFC generators are becoming more common in modern generator designs. Fixed-speed induction generators are relatively easy to design, having been designed for over 100 years now, but they are also very large and heavy due to all the copper windings. The mass of copper involved also makes them very costly and susceptible to supply chain disruption. A typical 2.5 MW generator is typically 3 m in diameter and 3 m long weighing over 4 t.

Fixed-Speed Induction Generator Characteristics Induction generators function in the following manner: 1. The stator’s windings, previously mentioned, producing a rotating magnetic field in the stator. This rotating field spins at synchronous speed. 2. The rotating stator magnetic field induces electromagnetic currents in the rotor. 3. The rotor spins at a slightly different speed than the stator’s rotating field. The rotor’s spinning action inside the stator and the magnetic field that the stator produces allow for the conversion from mechanical energy into electricity. Electricity is derived from the relative motion between the rotor and the stator’s magnetic field. 4. The interaction between the stator field and the induced rotor field causes high voltages in the terminals, typically 440 or 880 V by design. Each generator design has unique performance and operational properties which must match the design of input mechanical power source. Note: A fixed-speed induction generator would not work: 1. In an off-grid application, as there is no field generated in the stator without power from the grid 2. With a mechanical power source that does not have RPM control mechanisms, e.g., a variable-speed wind turbine

Page 4 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 3 Schematic of a single-phase induction generator equivalent current

Slip An important characteristic of induction generators is called “slip.” Slip is defined as “the ratio between synchronous speed and rotor operating speed.” The formula for determining slip is: s¼

ns  n ns

(1)

In this equation, s = slip, ns = synchronous speed, and n = drive operating speed. When the values of this equation yield a positive result, the machine functions as a motor. When the slip results are negative, the machine functions as a generator. This again shows how induction generators and motors are closely related. The prevalent characteristics of induction generators are shown in Fig. 3, which is a schematic diagram. It represents the equivalent current of induction generators.

Synchronous Generators In this section, we will review the unique features of synchronous generators, a special type of induction generator, including their design, construction, and operation.

Synchronous Generator Overview At the present time, synchronous generators are the most commonly used generators for utility-connected power generation, be it wind turbines or hydroelectric or nuclear or other thermal, steam turbine-based power generation. As power electronics and control system technology continues to improve, this is already changing to other forms of generators. However, the large installed base of synchronous generators warrants their in-depth understanding. Figure 4 illustrates a basic generator. Imagine the armature spinning due to mechanical energy coming from a power source, be it a wind, hydro, gas, or steam turbine. The spinning magnetic field of rotor relative to stator creates a voltage, as Faraday’s law predicts.

Synchronous Generator Design Synchronous generators are a bit more complex than asynchronous generators, and they have more design considerations and alterations possible. Synchronous generator design is similar to that of asynchronous generator design. The stator and rotor are usually constructed of a minimally conductive steel alloy. The stator and rotor windings are most often made of copper or a copper alloy, though some high-performance generators are made of low-temperature superconducting materials.

Page 5 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 4 Cutaway of an induction generator

Coil Usage With balanced three-phase currents, the kind of power that is used for most of the grid, a coil sometimes replaces the stator windings, and it is aligned with the a-phase axis. Its magnetic field rotates at synchronous speed. This type of generator design is most common in studying steady-state designs. In real-life applications, the coil design is rarely possible because currents change in magnitude and phase in the three-phase windings. Also, the stator field vector changes with the use of the coil. These challenges complicate control and reduce efficiency. Damper Windings In synchronous generators, damper windings function by combining currents with the air-gap flux which produces torque. The torque then dampens rotor fluctuations after any kind of transient disruption, e.g., wind gusts. The use of various current dampening techniques, including damper windings, facilitates this effect. Salient-pole (designs where the poles “stick out”) and wound-rotor designs both contain solid copper bars running through the rotor. The copper bars provide circulatory paths for damping currents. Salientpole generators contain the damper bars inlaid into pole faces. Some generators have end rings that link poles. This provides even more pathway for damping current flow. The damper windings in cylindrical pole generators are a little different. Damper windings are embedded into a rotor’s slot wall which contains the field winding and the damper winding on top. These are held in place by a specialized wedge. All wound-rotor generators contain end rings that connect the dampers. Sometimes the grid to which a generator is connected experiences an unsymmetrical fault, such as a voltage drop or frequency instability. This causes the air-gap flux to develop two components: positive sequence and negative sequence. Positive sequence is defined as flux resulting from currents in the direction of rotation. Negative sequence is flux resulting from currents flowing opposite to the rotational direction. The negative sequence counteracts the rotational direction of the rotor leading to a high relative speed, and this produces a large torque aspect. Damper winding currents interact with negative sequence in the air-gap flux which produces a counteracting torque resulting in acceleration that is controlled, and the generator’s rate of increase is limited, maintaining performance and synchronous performance.

Page 6 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Synchronous Generator Characteristics Direct Current Commutators are conductive segmented rings with brushes that contact one segment at a time. In each revolution, brush–segment pairings reverse from segment to segment. The induced voltage would be made of half-sine waves that all have the same sign, i.e., direct current (DC) with a sinusoidal current and voltage. This should not be confused with a direct current that is constant in magnitude as one might find from a battery, for example. Alternating Current If the cables or rods carrying electricity are connected by slip rings, where one cable or rod is connected to the armature coil and one connected to the other coil, alternating current (AC) is generated. The slip rings contain brushes that allow the current to pass from one ring to the other. The voltage generated by the armature’s spinning depends on its position in the magnetic field, and the resulting current and voltage show a full-sine wave, both positive and negative. Like asynchronous generators, synchronous generators contain two main parts: 1. Field winding, which is located on the rotor 2. Armature winding, located on the stator A direct current flows through the field winding. The magnetic field rotates as the rotor turns. Low-speed generator rotors contain a concentrated winding and a nonuniform air gap. This kind is called a salient-pole generator. High-speed generator rotors have distributed windings and a uniform air gap. This kind of generator is called a wound-rotor or cylindrical pole generator. Both types of synchronous generators produce sinusoidal magnetic fields in their air gaps. Salient-pole generators’ poles are shaped to create a sinusoidal air-gap flux. Rotor windings in wound-rotor generators are distributed over two-thirds of the rotor surface. The produced flux aggregates into a sinusoidal shape. Moreover, stator winding placement helps to create voltage with sinusoidal waveforms. In this chapter, you will focus on the workings and modeling of salient-pole generators. In spite of these differences in the details of their designs, the functionality for both types of synchronous generators is effectively the same. Three pairs of stator windings with axes are spaced 120
apart. When the rotor turns, its magnetic field spins in the air gap at synchronous speed. The spinning field cuts the stator’s three voltages. Currents are then induced in the three pairs of stator windings. If the stator windings are connected to identical loads, a three-phase current results. Of course, the three-phase currents are also displaced by 120
. The displaced currents will also create three magnetic fields. Thus, the air gap contains a combination of magnetic fields. Both fields produced by the stator currents as well as the rotor currents are present. When t = 0, the a windings are at their maximum current level. Conversely, the b and c windings are at their half-maximum negative currents. At t1, the c windings reach their maximum negative, while the a and b currents coincide at their half-maximum positive currents. At t2, the b windings reach their maximum current. You can see that if the timescale continued, the c currents would reach their maximum at t4. Continuing, at t1, the a- and b-phase currents make two magnetic fields. The fields’ magnitudes are proportional to the number of ampere-turns along the a and b axes. Likewise, the current in the c phase is proportional to the number of ampere-turns relative to the c axis. Given these interactions, you can deduce that the stator magnetic field at t1 shifts by P3. Within each phase, the stator’s magnetic field has rotated by P3. This is also known as rotating at synchronous speed.

Page 7 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Synchronous Generator Control Synchronous generators have unique control issues which should be understood. Power systems must operate within certain parameters, and maintenance of system stability is critical to compliant function as required by all power regulatory bodies around the world. The functional controls of synchronous generators are divided into two main functions: 1. Reactive power and voltage control 2. Active power and frequency control A discussion of the control of synchronous generators can be found in many other texts and is far to vast for discussion here, particularly when one includes a discussion of overall systems control with optimization of power production using synchronous generators for wind turbines.

Doubly Fed Induction Generators Doubly fed induction generators (DFIGs) are similar to fixed-speed induction generators, but there are some important differences to be aware of, including designs that permit variable-speed operations. DFIGs revolve around mathematical principles that would require an entire chapter to describe properly so we will only touch on them here.

Doubly Fed Induction Generator Overview Doubly fed induction generators provide the most efficient way of converting available wind power. They are, however, complex to design due to the complex mathematics required in many aspects of their design. Wind turbines with DFIGs are a good choice when an area has highly variable winds over the course of a year due to the DFIGs’ ability to provide efficient power generation across a broad range of RPMs, hence different wind speed. Doubly fed induction generators use a wound-rotor configuration, similar to that previously described in the induction generator section of this chapter. In such a wound-rotor configuration, slip rings deliver current into or out of the rotor winding. Controllable voltage is driven to the rotor at slip frequency. Controllable voltage is key to the DFIGs’ ability to provide variable-speed operation. The rotor windings are connected through a variable-frequency power converter which are based on two alternating current/ direct current (AC/DC) voltage source converters made using insulated gate bipolar transistors (IGBTs). The voltage source converters are linked via a DC bus. The grid is not connected directly to the rotor output, so the grid frequency does not impact the generator. Electric crowbars, circuits that intentionally short out when there is an overvoltage situation, protect the electrical components of these generators. Electric crowbars are like very large fuses or circuit breakers in their operation. Doubly fed induction generators supply network power in two ways: 1. Generator stator, if the generator operates at above-synchronous speed 2. Voltage source converters, if the generator is operating at below-synchronous speeds

Doubly Fed Induction Generator Design

Doubly fed induction generator design is largely the same as that of fixed-speed induction generators. The housings in a DFIG are typically made of steel and come in standard sizes just as induction generators do and as described in that section of this chapter. Windings are copper cables laid into grooves in the stator and rotor in wound-rotor designs. The cables are insulated to prevent short circuits and protect Page 8 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 5 Steinmetz per-phase equivalent circuit model

them from the environment. The exterior is made to protect the insides, no different than any other induction generator. The differences come in when one looks at the control of the variable-frequency input for excitation and how this variability impacts the electromagnetic field and related components to ensure optimal field–winding interaction across such a range of frequencies. Such designs often require sophisticated electronics and electromagnetic design software.

Doubly Fed Induction Generator Characteristics When a doubly fed induction generator is in operation, it is in steady state, a state defined by the Steinmetz per-phase equivalent circuit model. This model describes generators as well as motors. The Steinmetz equivalent circuit is expressed in terms of the following: • • • • • • • • • •

Stator resistance (Rs) Stator leakage reactance (Xs) Rotor resistance (Rr or R0 r) Rotor leakage reactance (Xr or X0 r) Rotor slip (s) Rotor power loss from windings (Pr) Electromechanical power (Pem) Magnetizing reactance (Xm) Air-gap power (Pgap) The imaginary number √1 (j)

Figure 5 shows the Steinmetz per-phase equivalent circuit model using the notations above. The DFIG slip’s controllable range is critical to the optimal operation of the generator. This value determines the size of necessary converters for doubly fed induction generators. The normal slip and speed range for most DFIGs is from 0.7 to 1.2 pu. The calculations can be found in other texts to derive the slip torque and other critical factors in the design of such a generator and are beyond the scope of this introduction.

Permanent Magnet Generators Permanent Magnet Generator Overview Permanent magnet generators have become more prominent in wind turbines in recent years. There are two main reasons for the move:

Page 9 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

1. Higher efficiency 2. Direct drive compatible at low RPMs Permanent magnet generators are not common in other utility-scale production since more systems provide high-RPM power to the generator. Steam and gas turbines operate from 10,000 to 50,000 RPM which is ideal for directly driving induction generators, but wind turbines operate around 15 RPM, and the need for a great many poles requires a smaller field generation area which permanent magnet generators offer.

Permanent Magnet Generator Design Permanent magnet generators are some of the oldest designs and what most of learn about in primary education. For a permanent magnet generator, a conductive coil spins inside alternating north and south magnetic fields generated by high field strength, typically so-called “rare-earth” metal-based permanent magnets. See Fig. 6 for a simplified schematic of this design concept. Rare-earth metal permanent magnets have field strengths that are many times that of iron magnets, which were used for over 100 years in generators. Iron magnets, due to the low field strength, are too large and heavy to warrant their use in modern power generators, particularly in wind turbines. The rotor typically contains iron cores which help concentrate the magnetic field onto the coils. Very large generators, those sometimes reaching 10 m in diameter for modern large wind turbines, mount the coils on I-beam struts which are welded to the armature. Permanent magnet generators for wind turbines are usually very large, many meters in diameter, containing many poles in order to generate power at 50 or 60 Hz from a 15 RPM direct drive shaft from the hub directly connected to the armature. A simplified diagram of such a design can be found in Fig. 7. Permanent magnet generators are relatively simple to design and over one hundred years of experience to work from. However, they are a material and supply chain challenge as will be discussed later. These issues make them relatively expensive to manufacture. Some permanent magnet generators move the location of the magnets to the rotor and the coils to the stator. This allows for a brushless design. This can be more reliable but can reduce the number of poles which can make the design not useful for the intended application (e.g., a wind turbine).

Permanent Magnet Generator Characteristic Like the synchronous generators mentioned in a previous section, permanent magnet generators can use commutators or slip rings to conduct the power from the coils to the output terminals, giving us DC or AC

Fig. 6 Permanent magnet generator, simplified schematic showing single north–south magnet in red and blue, respectively, armature in dark gray, and conducting coils in yellow Page 10 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 7 Simplified schematic of a multi-pole permanent magnet generator design

current. However, as previously mentioned, a brushless design can also be implemented, and these give only AC current output. Unlike induction generators, no connection to the grid for energy to energize magnetic fields is required since the magnets provide the field which induces current in the coils. The magnetic fields also do not rotate in a standard design, so there is no slip. Permanent magnet generators are capable of functioning at a variety of RPMs making them useful for wind turbines with variable-RPM designs, when used with AC–DC–AC power electronics to ensure the frequency is matched to the grid. Otherwise, different RPMs would result in different frequencies output which could cause system faults on the grid.

Electrostatic Induction Generators Electrostatic Induction Generator Overview Electrostatic induction generators are a relatively new technology to become viable for high-voltage power generation. While the concept has been around for nearly 100 years, electrostatic induction generators were not at all viable until the nanotechnology age came. Today, most generators work on the electromagnetic principal. These generators are large due to both the copper windings needed and large permanent magnets. Even superconducting material-based generators are large compared to the proposed technology. The electrostatic induction generator is based on electrostatic forces. Unlike electrostatic generators, such as the Van de Graaff type, this generator operates by inducing charge flow into thin plates in the stator by holding a large charge in the rotor plates, resulting in voltage potentials of over 100KV in a similar design to electromagnetic generators, i.e., rotating rotor and fixed stator with standard armature. Like Van de Graaff generators, this generator can create high voltages, but unlike the Van de Graaff, it can also produce a large current resulting in a high power density. Electrostatic induction generators are also ideal for lower RPM inputs due to its small size, thereby removing the need for gearboxes to step up the RPMs, unlike other generators. They can also be designed for very high-RPM systems, such as those connected to gas turbine engines and everything in between. Comparing an example 5 MW electromagnetic (EM) generator with a 5 MW electrostatic induction generator, the EM generator would be approximately 10 times the size and 20 times the mass. Electrostatic induction generators are lightweight due to being made of primarily aluminum and plastics. There are no rare-earth or toxic materials used in the electrostatic induction generator. All materials are commonly

Page 11 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 8 Schematic of an electrostatic generator design

available everywhere, avoiding sourcing issues which we will discuss further in later sections of this chapter.

Electrostatic Induction Generator Design Electrostatic induction generators, being essentially rotating capacitors, do not produce a pure sine wave, and the voltage follows current, not synchronous with it. They are best suited for use with relatively low-cost high-voltage rectifiers to produce high-voltage DC, rather than AC. This can then be electronically converted by inverters to any desired frequency or even multiple frequencies and voltages as needed for shipboard operations. The electrostatic induction generator is also very efficient since it has very low resistance due to the short conductor lengths (less than 1 m in length) and ability to have very thick conductors since there are no windings, just straight and circular conductors to connect the stator plates together and to the power output ports. These voltages are very high, typically over 100KV, which requires the generator to have extra care in safe design. Electrostatic induction generators, being highly modular with multiple thin stacked rotors and stators, can allow for scaled torque by simply energizing more or less rotors. This allows for higher efficiencies for a great range of power needs and the ability to start the generator with less stress on the input power source, which is particularly important to gas turbines. This modularity can be used to allow wind turbines to start at lower speeds due to the low torque of just one plate set being energized, and more plate sets can be energized as power from the turbine hub increases. This modularity also allows for easy field repairs and does not require the entire generator to be swapped out in case of localized damage without having to remove the entire generator from the nacelle. In the past, electrostatic generators of this type were not viable since high voltages on charged plates would simply arc across the rotor–stator gap. Such work was given up on in the 1960s due to material limitations. The electrostatic induction generator is now viable for compact, lightweight, high-energy density, power conversion due to the innovation of functionalized composite nanostructured charge retention materials. These materials have linear performance and high dielectric strengths, acting as very strong insulators unlike any other material known. By allowing very high voltages on the rotor plates, a high electromotive force is applied to the stator plate, producing high currents very efficiently. Since there are no lengthy copper windings, resistance is low in spite of using aluminum, with conduction paths being on the order of a meter for a 5 MW generator design (Figs. 8 and 9).

Electrostatic Induction Generator Characteristic As mentioned previously, electrostatic induction generators operate at very high voltages, typically over 100 KV, much higher than typical generators which operate at 440 or 880 V typically. Since the generator is mostly polymers, only the capacitive plates and conducting bars and cables need to be highly insulating in order to prevent arcing and safe for operation. Also as mentioned previously, as a capacitor, the performance is based on the RC constant of the system and, for optimal performance, should be designed to match a particular RPM. Deviations from this design Page 12 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 9 Detailed schematic showing relationship of rotor and stator disks

point RPM will result in lower efficiency, though newer designs are looking to broaden the RPM performance range. Electrostatic generators, when designed and operated at a specific RPM, are extremely efficient, exceeding 98 %. Losses are minimal in the generator itself due to the short path lengths from capacitive plate to output terminals. Energy transfer from mechanical to electrical form is very efficient as well due to the very close proximity, typically only less than 100 mm, of the inducing plates on the rotor and the induction plates on the stator. The field densities are extremely high at these distances. It is important to note that the charge retention materials are low-k materials, typically with a dielectric constant of 1.5, which means little energy is lost in the dielectric. And again as previously mentioned, the voltage and current are not synchronous like electromagnetic generators nor are they sinusoidal; therefore, power electronics, which will be described in the next section, are necessary to convert the power from an electrostatic induction generator into something useful for motors and the grid. With costs, size, and mass considerably lower for this type of generator, this easily makes up for the cost of the power electronics.

Fully Rated Converter-Based Generator Systems This section is less a review of a specific generator technology as much as it is about generator systems, their control, and how they can be effectively connected to the grid. “Fully rated converter-based generators” (FRC generators) are the most flexible generators for use on the grid, as they can be conventional, synchronous or asynchronous, permanent magnet, or electrostatic. Working as a system, the generator’s rated power determines the power converter rating. The system as a whole is designed for maximum energy transfer from the mechanical source to electricity for the power grid.

Page 13 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

FRC Generator System Overview We will focus on direct-drive FRC generators in this chapter, i.e., those with no gearbox between the input power shaft such as a wind turbine rotor and the generator. FRC generators can be controlled with two different control methodologies: 1. Load angle method 2. Vector control strategy FRC generators are named such because they contain fully rated voltage source converters (VSCs), meaning they can control the voltage output at the full output voltage and power using power electronics. FRC generators are ideal for many applications because they feature variable-speed operations. Variablespeed operations are useful for a wind turbine designer who wants to make turbines capable of extracting the most power in changing wind speeds. Because of these benefits, fully rated VSCs fulfill current grid requirements such as: 1. 2. 3. 4.

Fault ride through (FRT) requirements Active power (the amount of power generated for load consumption) and frequency control Voltage and frequency operating ranges Reactive power (reactive power is the resultant power loss derived from power generation) control and voltage regulation

Most FRC generators are similar in design and functionality to induction generators; however, their construction allows for the extra flexibility offered, and custom designs are easier to manufacture for customers with specific needs. In the case of wind turbines, all rotor power is transferred through a power converter in the use of an FRC generator; hence, the operational characteristics and dynamics of the turbine are isolated from the grid. Such power isolation permits generator frequency variation which is necessary for variable-RPM operations. The selection of a power converter arrangement determines which operation and power control strategy is best for a given application. There are multiple configuration options for power converters. Most fall into two categories: 1. Generator-side converter which may consist of either: (a) A diode-based rectifier or (b) Pulse width modulation, variable-speed control (PWM-VSC) component 2. Network-side converters (usually PWM-VSC based)

FRC Generator Design Synchronous FRC generators can be magnetized in two ways: 1. Electromagnetic, being induction generator in design 2. Permanent magnets Most turbines in operation today rated for more than just a few kilowatts use four-pole (standard) generators operating between 750 and 1800 RPM. For wind turbines, where the main drive shaft rotates much lower, a gearbox is typically required to increase the RPMs going to the generator so that it may function optimally. Alternatively, a direct-drive generator can be used. See the schematic diagram of a permanent magnet direct-drive generator in Fig. 7 from the previous section. Page 14 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Low-RPM Direct-Drive Generators In direct-drive wind turbine configurations, for example, the generator receives mechanical energy directly from the turbine rotor assembly’s main shaft. The main shaft typically rotates at 15–60 RPM at optimal wind speeds for any given design of wind turbine. The generator must be specifically designed for this low-speed operation, typically resulting in many poles which significantly increase the size and mass of the generator itself. Many FRC-based turbines use a single-stage, low-ratio gearbox allowing the use of smaller generators with fewer poles. Direct-drive generators have two main benefits: 1. Reduction in drive train power losses. Gearboxes reduce power transfer from the rotor to the generator mostly due to the friction in the gears. Because there is no gearbox to maintain, costs are lower and revenue is higher. 2. Quiet operation. Direct-drive turbines are quieter than conventional ones primarily because of the lack of gears. The lack of a gearbox increases torque on the generator; therefore, direct-drive generators must be designed for very high torque. A generators’ size and power losses depend on the torque rating, not the power rating. For example, a 500 kW, 30 RPM generator and a 50 MW, 3000 RPM generator both theoretically have the same torque rating. Due to high torque rating and generator size, direct-drive generators are very large particularly in diameter. They are also less efficient. Direct-drive generator designs also depend on small pole pitch, meaning the angular distance between poles, to help maximize generator efficiency and increase the frequency of the output current. A modern direct-drive generator may have 20–60 poles or more. Permanent Magnet Versus Electromagnetic Excitation Excitation is the interaction between static and rotating magnetic poles inside the generator. As mentioned in a previous section, early generators used permanent magnets for excitation. Early permanent magnets were heavy, being made of iron. As technology advanced, electromagnetic excitation became viable, and most manufacturers started using electromagnetic excitation. Synchronous generators self-excite. They use copper windings that generate electromagnetic fields. Synchronous generators can also use permanent magnets which create a true magnetic field without the electric field component. Wound-rotor generators have an advantage. The excitation current is adjustable. This permits delinking output voltage control from load current levels. So most constant-speed generators use wound rotors. This is especially true in direct-grid connections. But most synchronous generators connect to a grid with electronic converters. Then delinking output voltage control from load current is less beneficial. Nowadays, wound-rotor generators are heavier than permanent magnet-based versions. Generators with small pole pitch are also bulkier. Copper windings contribute to power losses. That is because each pole needs windings. As the number of poles increases, the number of windings increases. As winding quantities increase, power losses grow. Permanent magnets do lose some power. Permanent magnet losses are lower than copper winding losses. Permanent Magnet Synchronous Generators Permanent magnet-based generators have advantages over other designs: 1. No field current supply is needed, simplifying design, lowering material and assembly costs, and allowing for off-grid systems such as in very remote locations. Page 15 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

2. No reactive power compensation, simplifying design and costs. 3. Permanent magnets do not require slip rings, reducing system costs and increasing reliability. In most permanent magnet synchronous generators, a grid-side converter (PWM-VSC) controls generator operations. There are two methods for PWM-VSC control: 1. Load angle methods 2. Voltage-oriented reference frame current controller The maximum power/speed characteristic defines the power reference used in the selection of PWM-VSC controller. FRC Induction Generators FRC induction generators (FRC-IGs) operate at variable frequencies which is ideal for wind turbines. The characteristics are based on three measurement sets and calculations: 1. Torque/rotor speed 2. Active power/reactive power 3. Slip/reactive power In wind turbines, optimal energy extraction from the wind occurs if the generator speed varies along with wind speed. The PWM network-side converter control signal changes to allow this change in power frequency from the generator. At low wind speeds, generator frequency is low, and at high wind speeds, the generator frequency is high. Since reactive power absorbed by the generator is constant at all frequencies, slip permits stabilization of reactive power absorption, thereby managing the power loss.

Generator Considerations In this section, we will briefly introduce topics for consideration in selecting a generator for a given topic, namely, utility-scale (5MW) wind turbines. We will compare induction (as a group), permanent magnet, and electrostatic generators.

Mass and Size of Generators and Their Impact on Wind Turbine Design Including Pendulum Nature of Wind Turbines and the Impact on Tower Design and Control Systems Design Let us first look at the equation for the torque of a stationary pivot point inverted pendulum, which is what a tower is: T ¼ mgl sin y In this equation, T is the torque, m is the mass, l is the length of the pendulum, and y is the angle of displacement. In the case of a wind turbine, this is the torque on the tower for a given angle of bend. In order to manage this torque, the tower must be constructed strong and stiff enough to take this torque and not bend much further and certainly not fail. Over long periods of time, this torque can fatigue the tower causing it to collapse if not sufficiently strong.

Page 16 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Wind turbines are very tall structures with a large mass that of the blades, hub, nacelle, and most notably the generator set atop of the tower. Even without these masses atop the tower, the load of the wind will cause the tower to bend and oscillate. With a large mass atop the tower, this becomes a very large pendulum swaying back and forth in the wind. Without controls to limit this sway, it is possible for the spinning blades of the wind turbine, which themselves bend, to hit the tower, resulting in catastrophic damage to the blades and even the tower itself. In order to limit this sway, engineers implement blade pitch control to increase and decrease load on the turbine synchronized such that, when the tower moves forward into the wind, the load is increased to push backward on the tower and, when the tower moves away from the wind, the load is decreased to minimize the backward push on the tower, hence dampening the oscillation as much as possible. The control system often turns one blade that is approaching the tower such that the load is less so that it will not bend and hit the tower. All of these activities decrease power generation relative to wind energy available. This loss of efficiency can be significant when every watt to be converted is the goal of designers. A generator with a large mass, like induction and permanent magnet generators, has a significant impact on this pendulum like behavior. Electrostatic generators are 1/5 to 1/10 the mass and therefore reduce the torque this amount due to the linear nature of the equation. Aerodynamic Impacts For wind turbine, the overall size of a generator has a noticeable impact on the aerodynamics of the turbine and the design of the nacelle for aerodynamic efficiency. Most induction generators, particularly those with gearboxes, fit within the nacelle such that the nacelle is not much larger than the hub. The hubs’ size is dictated by the size of the blades and the hydraulic control systems needed to control blade pitch, so this has not been much of an issue in the past. In this configuration, the generator can be positioned as close as possible to the hub, with some of its mass directly over the tower structure. However, direct-drive generators, particularly permanent magnet direct-drive generators, are much larger in diameter than the hub. In many designs recently implemented, the aerodynamics of the generator have been completely ignored since it is impossible to put this large, disk appearing generator inside the nacelle itself. Instead, the generator is positioned as close as possible to the tower, counterbalancing the blade. Here, none of the generator sits directly over the tower structure, but rather cantilevered opposite the hub. This requires extra design and materials to accomplish. Electrostatic generators are very small in comparison to either of these technologies, and they are often much smaller than the hub or the diameter of the tower. In this case, the nacelle requires no extension past the dimensions of the tower itself. In fact, current designs could fit the generator inside of the hub with room for maintenance personnel and other equipment. Of course, this would not be a practical design since the generator needs to be mounted to the nonrotating portion of the turbine (i.e., the tower).

Fixed Versus Dynamic Generator Torque Control For many applications, torque control is not a significant issue, such as in hydroelectric power plants where the flow of water is controlled and consistent in any turbine–generator pair. However, other systems would benefit from a variable torque on start-up, such as steam and gas turbines, where system start-ups can have low initial torque and can stall the system because of high-generator torque requirements. Wind turbines are even more complex with relation to torque. A low start-up torque would benefit the system both due to potentially lower wind speeds and hence hub torque and due to start-up low rotational inertia. Once the turbine is in rotation, the rotational inertia varies little compared to the difference from start to full-speed operation as the rotational speed in operation changes little by design. Standard induction motors, particularly synchronous generators, do not offer variable torque, nor do permanent magnet generators. DFIGs have some ability to control generator torque directly, but rather Page 17 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

through the variation in RPM. This happens as a feedback system rather than as an intelligent controller parameter of the generators’ operation. As mentioned previously, electrostatic induction generators are ideal for variable torque. While this variability is stepped based on the number of rotor disks energized, this is typically sufficient torque control for wind turbines. Review the section on electrostatic induction generator design and characteristics for more details.

Material Issues and Supply Chain Issues With generators being such a critical component in current renewable energy production and conversely the increased dominance of renewable energy facilities as the customers of generators, the “green” interests of the renewable energy industry are having an impact on generator manufacturing and design. Costs Currently, generators are still large and massive, consuming great quantities of copper and iron. These materials are quite expensive, particularly copper, at this time on the global market. This has resulted in generator prices increasing, in some cases 200 % over 5 years. Such costs negatively impact the cost of energy for a given renewable energy system and the investment in that system. Material Access Permanent magnet generators have even more complex cost and supply chain issues. Since the magnets are made from rare-earth metals, and these metals are found in a limited number of locations around the world, specific countries have instituted export prohibitions of rare-earth metals and permanent magnets made from them in order to promote local generator and wind turbine firms. Rather than allowing open trade and competition on a global scale, these governments have controlled what they see as a strategic asset. This has not only affected the price of such magnet and the resulting generators but gone so far as to completely shut down all wind turbine production that did not have sources within the countries controlling the materials. Environmental Impact The environmental impact of generators is a complex analysis. From raw material acquisition to production, from operation to decommissioning, generators have environmental impacts that are the subject of extensive study and reports. We will summarize and discuss some lesser considered issues here. Raw Material Acquisition Induction generators of all types are primarily composed of iron, steel, and copper. This is some 95 % of their constituency. Copper is also a highly recycled material, but demand for copper exceeds the ability to find sources from recycling most of the time. While steel is the most recycled material and has great availability, the kind of iron used in induction generators typically comes from strip-mining. Copper mines are often strip-mines today as well. These mines are true ecological disaster areas due to the devastation to all life in their areas and runoff that often contains hazardous materials, including many copper salts and oxides. Permanent magnet generators also have a great deal of copper, but less than induction generators, and have some iron as well. Due to the very large diameters of many of the designs for wind turbines, large steel beams make up much of the volume of the generator. This steel may be from a recycled source. The iron and copper are similar problems in their raw material sourcing as mentioned for induction generators. Permanent magnet generators use rare-earth metals, which are not as rare as their name might infer, but there are a limited number of locations around the world where they are found. Typically, these are in Page 18 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

countries with little or no environmental protection laws and/or enforcement. For rare-earth metals, stripmining is very common. The metals, their oxides, and salts are very toxic to all living things. With toxic runoff and even toxic dusts, these mines are major ecological disaster areas. Electrostatic induction generators have several raw material sourcing advantages, including the fact that they are much smaller and less massive and they consume less raw materials. Electrostatic induction generators are composed primarily of fiberglass-reinforced polymers, such as acetal or nylon 6/6, for their structural materials. While these materials, in virgin form, are currently made from petroleum, there is much research into making these polymers from bio-renewable sources. Both of these materials also have growing recycled supply chains. The other main component of electrostatic induction generators is aluminum. In much of the world, aluminum is the second most recycled material, and since it is made from common sand, strip-mining is neither necessary nor common. The remaining materials of the electrostatic generator are also made from common sand, natural resins, polymers, and/or common carbon sources. Very little of these materials are used, and they are generally highly available on the commodity market and relatively low in toxicity. In order to increase performance and reliability, electrostatic generators typically fill the rotor–stator gap with a low-viscosity, high dielectric lubricant. Because of the availability recently of plant-based oils for lubrication that meet these needs, electrostatic induction generators are manufactured with the plant-based oils that are biodegradable and even edible, rather than petroleum. All generators have bearings and these require lubrication. Lubricants have traditionally been petroleum based, but this is changing as many firms now produce lubricants derived from canola and other commonly available plant sources. Production Process Converting raw materials into finished products is almost always an energy intensive process. Recycled materials certainly cut the energy needs significantly, for example, steel, copper, and aluminum. In addition to energy for activities such as smelting furnaces or electrochemical plating, the chemicals used can be very toxic to workers and the environment. This field is very extensive, and “green” processes and chemistries are active areas of research, so we can only touch on the topic here. Induction generators may be one of the lowest in terms of toxic materials use of the conventional generators. As stated before, copper, iron, and steel are all commonly available and easy to work with materials. The smelting process, for virgin metal production, does consume the most power of any industry however. The insulation used on the copper wire for the windings can be made from a great variety of materials, many of them from petroleum. Newer materials from plant resins are making inroads in the industry. Permanent magnet generators have many of the same process issues as induction generators plus the conversion of raw rare-earth metal oxides and salts to the metal alloy forms that make them powerful magnets. This requires both chemicals that may be toxic to the environment and workers, toxic waste products that need to be managed safely, and a great deal of energy. Electrostatic generators use the lowest energy to produce again partly due to their size but also material selection. The aluminum components are precision cut typically by die and the wastes are 100 % recycled. The structural polymers are injection molded, resulting in the least energy used and minimal waste which again is 100 % recycled. The change retention materials are made using the leading green chemistry methods, including the use of acids that are used in salad dressings.

AC Grid Synchronization Most power generation systems, particularly of significant size and power, are meant to provide electricity to distant locations and many customers via the power grid. In order to achieve this, the generator must Page 19 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

connect to the grid and operate within a strict set of guidelines. The most common connections for power generation are that of mid-voltage alternating current (AC). These voltages can be anything from 880 V to 38KV depending on the nearest transmission line near the power generation source. There are some higher voltage generation facilities such as hydroelectric dams that can connect at 115KVAC. Many wind farms produce power only at 880 V, but newer turbines and wind farms are connected to the grid at 38KV. Induction generators operating at fixed speeds (RPMs) have the best and easiest low-voltage grid integration and synchronization. As mentioned in the sections on induction, synchronous, and doubly fed induction generators described, these generators have various methods for controlling feedback so that the generator produces power compatible with the grid that is powering its excitation. Permanent magnet generators and induction generators that are part of FRC-based systems and electrostatic induction generators use power electronics to convert the power produced to be AC grid compatible. These electronic systems are typically computer-controlled systems that use phase-locked loop controllers to achieve grid synchronization of the alternating current.

HVDC Integration High-voltage DC grids are not a new concept, having been considered and initially tested in the 1930s and 1950s, mostly in Europe and Russia. Early systems operated at 100KV, which today is not considered a very high voltage. Today’s HVDC networks operate at very high voltages, typically 600KV, 800KV, and 1GV, and higher voltages are currently being developed. There are many advantages to HVDC networks for long-distance grid lines. These include a lower cost for cabling since there are two, not three, conductors, but the main reason is efficiency. At higher voltages, the current required for a given amount of power is less, per Ohm’s law shown in Eq. 2. Ohm’s law f or power : P ¼ V  I

(2)

One might think that induction and permanent magnet generators that use commutators or FRC systems are suited for HVDC grid connection since they can produce DC power; however, since transformers only work with AC and the generators produce only low voltages, the upconverting of voltage makes this not a simple solution. Typically, such generators are not directly connected to a HVDC grid because these generators will feed low-voltage AC to an AC to HVDC substation. Because they operate at the same voltages as HVDC, electrostatic induction generators rated for these voltages can be rather easily connected to the HVDC grid. In fact, the power electronics needed for power conversion are simpler in a HVDC configuration than in an AC one, basically being high-voltage rectifiers.

Transformer Losses Most utility transformers are designed with cost and reliability considered before efficiency; therefore, it is helpful to a power generation system to be able to operate at grid voltages, rather than requiring step-up transformers. Induction and permanent magnet generators are not able to operate without such transformers to increase the voltage to grid voltages, 115KV typically. Electrostatic generators are able to operate without transformers since they operate at grid voltages. It should be noted here that both conventional (induction and permanent magnet) FRC-based systems and electrostatic induction generators, which require a power converter, have losses in the power conversion. In the case of electrostatic induction generators, this loss is considerably less than that of a transformer used for an equivalent power synchronous generator, for example. Obviously an FRC induction-based generator system would have the most cumulative losses between the induction generator

Page 20 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

itself, the power electronics, and transformer. However, low installed cost can be a reason for selecting such a system.

Energy Storage Compatibility Energy storage systems, both local and grid tied, have become the topic of much discussion in the energy industry. Be they supporting renewable energy time shifting or conventional energy production or any of the many other applications, energy storage may be a major contributing factor to improving overall energy production-to-use efficiency. The non-determinant nature of wind in particular can benefit from energy storage systems. Nearly every energy storage technology that is appropriate for grid-scale energy storage, particularly battery- or capacitive-based systems, is a DC power system. Some systems are low voltage, like all batteries, and some are high voltage, like GigaCapacitor-based systems. The conversion from AC to DC and the voltage change that most impacts the efficiency of the overall generation-storage system. Even low-voltage, for example, 440 V, induction generators are much higher voltage than a 12 or 24 V battery, and the energy storage system is more likely to be connected to a mid-voltage grid, not directly to the generator. These inefficiencies make the costs less attractive. Imagine a wind turbine with a transformer upconverting to 38KV, crossing a vast distance, and then being down-converted by another transformer and rectified for an energy storage system. The losses would be significant. However, the GigaCapacitor and other high-voltage technologies may be more appropriate for grid connection, but that is a separate discussion. Electrostatic induction generators can ideally pair with high-voltage DC energy storage due to their very similar nature and designs. Only a HV rectifier at the generator would be needed between an electrostatic induction generator and a matched high-voltage GigaCapacitor, for example.

Conclusion Today, there exist great many choices of technologies for generators and their complete systems for both grid-tied and off-grid electrical productions for many renewable energy sources. Wind power is the most demanding in terms of complex requirements and highly adaptable generator systems. Today, electromagnetic and electrostatic forces are used in generators to induce electrical current. Designs, materials, and processes will continue to improve providing new capabilities for power generation.

References Anaya-Lara O, Jenkins N, Ekanayake J, Cartwright P, Hughes M (2009) Wind energy generation: modeling and control. Wiley, West Sussex Arrillaga J (1998) High voltage direct current transmission, 2nd edn. The Institution of Electrical Engineers, London Blaadjerg F, Chen Z (2006) Power electronics for modern wind turbines. Aalborg University, Copenhagen Burton T, Sharpe D, Jenkins N, Bossanyi E (2001) Wind energy handbook, 1st edn. Wiley, West Sussex Plantier K, Smith KM (2009) Electromechanical principles of wind turbines for wind energy technicians. Texas State Technical College Publications, Waco Rivkin D (2012) Whitepaper: introduction to electrostatic power generators. SciEssence Intl, Portland Rivkin D (2013) High dielectric strength nano-composite: charge retention for high voltage electrostatic applications. STI Phys Lett 101:201–233 Page 21 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_2-1 # Springer-Verlag Berlin Heidelberg 2015

Rivkin D, Randell M, Silk L (2014) Wind power generation and distribution. Jones & Bartlett Publishing, Burlington. ISBN 978-1-4496-2450-7 Spera D (ed) (2009) Wind turbine technology: fundamental concepts in wind turbine engineering, 2nd edn. ASME Press, New York. ISBN 978-0-7918-0260-1 Wind Energy Basics (nd) American wind energy association. From http://www.awea.org/ Woodward Corporation (1991) Speed droop and power generation (2011) http://www.matsuda-gov.com/ topic/droop&powergene.pdf

Page 22 of 22

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Risk Assessment in Hydroenergy Projects: Learning from Experts and Data Serhat Kucukali* Department of Civil Engineering, Çankaya University, Ankara, Turkey

Abstract The purpose of this chapter is to assess the potential risks of hydroenergy project by using expert judgments and multi-criteria scoring technique. The risk assessment tool analyzes technical, economic, environmental, social, and regulatory risks. The risk assessment framework includes three activities: (i) identification of potential risks and determination of their relative importance, (ii) risk analysis, and (iii) risk evaluation by using a evidence-based multi-criteria scoring technique. A survey was conducted with the experts with over than 10 years’ experience in the planning and construction of hydroenergy projects in order to determine the relative importance of external risks. The most concerned risks are identified as site geology and environmental issues. Applicability of the proposed tool is tested on multiple hydroenergy projects in Turkey. The findings of case studies showed that the perception of inadequate understanding of the potential risks can lead to cost overrun or project failure. The risk assessment tool can give a competitive advantage in the field of hydroenergy system deployment and it can reduce predevelopment time and costs.

Keywords Hydroenergy; Risk assessment; Expert judgments

Introduction Hydropower, as a capital-intensive and site-specific technology, involves various external risks such as economic, environmental, social, and regulatory. These risk factors can have a negative effect on project outcome or can lead to failure in projects. Investors and project developers need to understand these challenges in order for their businesses to be successful (Kucukali 2011). As there is no way to eliminate all those risks, they have to be assessed and mitigated as much as possible. Hydropower plant (HPP) project life cycle includes three main phases: planning, construction, and operation (Fig. 1). Each phase has different risk-reward characteristics. The first two phases do not generate revenue; in contrast they bring about many costs and uncertainties (Lawrence and Dickson 2010). Gordon (1983) identified the main factors which may lead to cost overrun as the rate of inflation and site geology. The International Energy Agency (IEA) (2012) reported that depending on the nature of the project, the main risks potentially affecting hydroplant financing may include construction risk, hydrologic risk, off-taker risk, regulatory risk, and life cycle risk. Accordingly, Binquet (2010), based on his 35 years of experience in the design and construction of hydropower projects worldwide, pointed out that the hydropower projects that did not involve tunnels and caverns like Birecik in Turkey and Ita in

*Email: [email protected] Page 1 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014 High Risk

Planning Phase

Low Risk

Feasibility studies Project appraisal and investment decision Engineering design Environmental Impact Assessment (EIA) Licensing and permitting Land leasing and acquisition Financing

Speculative Return

Construction Phase Construction Connection to grid Testing and commissioning

Operation Phase Plant operation Electricity generation

Reliable Return

Fig. 1 Hydropower project life cycle phases

Brazil completed on time and with budget identical to what was expected, whereas projects that involved long tunnels or located near active faults experienced considerable cost overruns and delays like the Dul Hasti hydropower project in India. Another expert from the World Bank, Palmieri (2012), noted that the unforeseen geological conditions (unexpected rock types along tunnel alignment, fault areas, and karstic cavities) are major contributors to cost and schedule overruns in hydropower projects worldwide. Another important challenge for hydroenergy development is environmental issues. While standard environmental impact assessments may have been enough in the past, more detailed guidelines on the evidence required are now provided by international financial institutions such as the World Bank (2013), European Bank for Reconstruction and Development (EBRD 2013), and European Investment Bank (EIB 2013). Mitigating the various negative environmental impacts of hydroschemes is vital to complete licensing procedures promptly and to secure funding from international financial organizations (Ford 2008). Moreover, Gronbrekk et al. (2010) reported that investors and project developers are stressing the importance of managing environmental risks in hydropower projects especially in developing countries. The most common risk assessment models, such as Monte Carlo simulation and tornado chart, are based on complex numerical calculations (Mund 2004). However, it is very difficult to deal with some types of risks such as environmental and regulatory risks by using those numerical methods (Pike 2010; Kucukali 2014a). So, using expert judgments and linguistic expressions can be a better way to cope with such type of risk elements. For instance, Mermet and Gehant (2011) stated that the statistical treatment of data cannot replace the expert judgments in the operational risk management process of hydropower plants. Mermet and Gehant (2011) identified the major sources of risks for the operation and maintenance of hydroelectric power plants by (i) making discussions with the operational staff, (ii) analyzing the investment plans and technical and incident reports, and (iii) conducting technical visits to the hydropower schemes. Figure 2 presents another expert (senior engineer economist of Electricite de France) risk assessment approach during the operation phase of hydropower plant, and their associated impacts. This simple and action-oriented approach also provides adequate mitigation measures that may be offset major risks (Branche 2011). The objective of this study is to assess the potential risks of hydroenergy projects in an integrated approach by using expert judgments and multi-criteria scoring technique based on measurable relevant data and documented evidence. The methodology addressing the decision circumstances at a planning phase of an investor investigates the risk-reward characteristics of alternative hydroenergy projects in order to select the most appropriate one.

Page 2 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 2 An expert risk assessment approach for hydropower plants in operation (Branche 2011)

Identification of Potential Risks

Case Studies Expert Judgments Literature Review

Analysis of Risks

Root-Cause Analysis Risk Driven Factors Relevant Parameters

Evaluation of Risks

Data-Based Scoring Spider Diagrams Risk Index

Fig. 3 Risk assessment framework

Risk Assessment Framework Every economic activity implies risks. A risk is defined as the uncertainty to reach objectives due to unexpected events and regulations (Mermet and Gehant 2011). The proposed risk assessment framework includes three activities: (i) identification of potential risks and determination of their relative importance, (ii) risk analysis, and (iii) risk evaluation (Fig. 3). The framework is compatible with the ISO 3100 (2009) risk management principles and guidelines in which communication and dialogue has an important role. The risk assessment activities are described in the following sections.

Identification of Potential Risks In the context of the study, a survey was conducted with the experts with over than 10 years’ experience in the planning and construction of hydroenergy projects in order to determine the relative importance of external risks. The external risks, which are partly under the control of companies, were considered in the model. A total of ten classes of risk factors were determined based on the expert interviews, field studies, and literature review as follows: site geology, land use, environmental issues, grid connection, social acceptance, macroeconomic, natural hazards, regulatory uncertainties, access road, and revenue. The survey was done based on a qualitative basis and 15 experts were asked to mark the level of identified risk factors via e-mail as low, medium, high and very high (Table 1). The same risk identification approach was also used by Ernst and Young (2010), Fraser (2010), Kucukali (2011), and Mermet and Gehant (2011). Fraser (2010) stated that the success of this type risk identification process is a result of its simple preparation and effectiveness of purpose and achieved good results with the application of this risk assessment approach to Hydro One, which is the largest electricity delivery company in Ontario since 1999.

Page 3 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Table 1 The expert ratings on potential risks in hydroenergy projects. The survey was done based on a qualitative basis where low = 1, medium = 2, high = 3, and very high = 4 Site geology 2 3 3 4 3 3 3 3 2 3 3 4 3 3 2 44

Land use and permits 1 2 2 2 3 1 4 4 4 3 3 3 3 3 3 41

Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp. 6 Exp. 7 Exp. 8 Exp. 9 Exp. 10 Exp. 11 Exp. 12 Exp. 13 Exp. 14 Exp. 15 Total score Relative 12.0 % 11.2 % weight

Environ. issues 2 3 3 3 3 2 3 3 3 4 3 3 3 4 4 46

Grid connection 1 3 3 2 2 1 2 4 2 2 1 3 1 3 3 33

Social acceptance 3 3 3 2 3 2 3 2 2 4 2 2 2 3 4 40

Macro economic 2 1 3 1 2 2 4 4 2 2 1 2 2 2 3 33

Natural hazards 2 1 2 1 2 2 3 3 2 4 2 3 1 2 2 32

Regulatory uncertainties 2 2 3 1 2 1 3 3 3 1 2 2 1 2 3 31

Access road 2 1 2 1 2 2 2 3 2 3 1 2 1 2 2 28

Revenue 3 1 3 2 3 2 4 4 4 1 4 2 1 2 3 39

12.5 %

9.0 %

10.9 %

9.0 %

8.7 %

8.4 %

7.6 %

10.6 %

Road Access

7.6%

Regulatory Risks

8.4%

Type of Risk

Natural Hazards

8.7%

Macroeconomic

9.0%

Grid Connection

9.0%

Revenue

10.6%

Social Acceptance

10.9%

Land Use and Permits

11.2%

Site Geology

12.0%

Environmental Issues

12.5% 0%

5%

10%

15%

Relative Importance

Fig. 4 Relative weights of potential risks in hydroenergy projects. Weights reflect the relative importance of each criteria and the values obtained from the expert ratings

The participants to the survey have considerable experience in the hydro energy sector. For instance, Expert 14 is the head of the project finance department of a well-known bank in Turkey that has international awards for the implementation of sustainability principles successfully, and Expert 15 is the chief risk manager in Europe’s largest renewable energy company. The survey was done based on a qualitative basis, and 15 experts were asked to mark the level of 10 identified risk factors via e-mail as low, medium, high, and very high (Table 1). The relative weight of each risk factor was obtained by dividing its total score to overall score of total ten risk factors. The survey results are presented in Fig. 4 and the most Page 4 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

important risks appeared to be environmental issues and site geology. The findings are consistent with the findings of Gordon (1983) and survey results of Gronbrekk et al. (2010). Detailed information for the identified risks is given below. Site Geology This parameter is related with the geotechnical properties of the construction site, and it also represents the geological risks in the construction site. Rock Quality Designation (RQD) is used to represent the quality of site geology (Waltham 2009). The geological risks that are taken into account in the study are distance from active faults, large-scale fault and fractured zones, and large-scale landslide area (Kucukali 2014b). In order to avoid delays and extra costs during the construction phase, favorable geological conditions at the potential site are desirable. Grid Connection Distance Construction of a new transmission line is costly and could be complicated since it involves obtaining appropriate permits and may also require land acquisition (Kucukali 2014b). Therefore, the closer the hydropower plant is located in the existing transmission lines, the lower the costs of integration to the grid will be. So, the distance to the grid connection should be short, and the risk score is directly proportional with the grid connection distance. Access Road A successful hydropower project is dependent on topographic and suitable site conditions which permit proper sizing and better arrangement of the principal features (Zipparro and Hasen 1993). Once the hydroturbines are delivered to the site, the access road has to be able to bear heavy trucks with trailers and a heavy mobile crane. The demand to the access road depends on the installed capacity of the power plant and sizes of the hydroturbine. Environmental Issues The hydropower scheme’s possible impacts on the natural environment through habitat destruction and biodiversity loss must be taken into account. Sensitivity regarding environmental aspects must be ensured, so that critical habitats, threatened species, and spawning areas can be protected. Since, most of the developing countries have weaker regulatory regimes and institutional frameworks, it can be a good strategy to use internationally acknowledged guidelines to ensure environmentally sustainable hydropower projects (Gronbrekk et al. 2010). For example, the European Bank Reconstruction and Development (EBRD) (2013) follows five environmental criteria during the Environmental Impact Assessment (EIA) of small hydropower (SHP) projects. Those criteria are environmental flow, water quality, fish passage and protection, watershed protection, and threatened and endangered species (Ford 2008). EBRD’s document is a kind of guideline that enables one to develop SHP plant with respect to natural environment (Kucukali 2014a). Social Acceptance Social acceptability is regarded as a key aspect to be considered in addressing the potential for deployment of hydroenergy. The project should not stop or reduce local communities’ ability to use either the river or surrounding lands to provide a livelihood, i.e., by fishing, as leisure amenity or to utilize the land around the river where they may rely on the river for irrigation purposes (EBRD 2013). A public consultation process should be carried out before the development of any project. Furthermore, Social Risks can be mitigated by establishing cooperation with local community and sharing the project benefits.

Page 5 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Land Use and Permits The hydropower projects must have all the necessary permits that are defined under national laws and regulations. Water Utilization Right Agreement, Land Use Permits and Acquisition, Grid Connection Agreement, Construction Permit, etc. are the main approvals which a hydropower plant must obtain during the project planning phase. Private lands involve more complex permitting procedure, and therefore, they have higher risk scores compared to forestlands (Table 2). Macroeconomic Risks Macroeconomic risks are not related to the project in particular, but to the economic environment in which it operates such as interest rate, inflation, and currency exchange rates (Yescombe 2002). In this chapter, macroeconomic risks are considered in relation with the macroeconomic issues and country-specific conditions. Some national banks use local currency as a strategy to finance construction costs of the hydropower projects. Increase in exchange and interest rates can have negative effects on project economic performance. Many economists stated that there is a link between the price instability and current account deficit of a country (Fratzscher et al. 2010). Accordingly, current account balance of a country as a percentage of gross domestic product (GDP), which reflects the price stability of a country, is selected as the relevant parameter to evaluate macroeconomic risks. Regulatory Uncertainties The surveys conducted by Gronbrekk et al. (2010) identified the highest risk as political and regulatory changes for renewable energy projects in developing countries. Similarly, Ernst and Young (2010) identified the most important business risk for 2010 as regulation and compliance. For hydropower development, regulatory issues include: the rules of the electricity market where the project will operate, the track record of the regulating agency, mechanisms in place for feed in tariffs, changes in laws and regulations, enactment of new laws and regulations, and predictability of policy framework (IEA 2012). The hydroelectric generation is influenced by the actions of regulators. The rules under which regulators operate will likely change as the past experiences shows. The hydropower projects which may be exposed to regulatory risk can result in revenue loss and O&M cost increase. Methods for mitigating the risk include increased and more effective communications with the responsible government agencies. Regulatory uncertainties are related to the level of political stability of a country. If the country has robust legislation for hydroenergy and environmental issues, the risk is regarded as minimum and has low-risk score (Table 2). Natural Hazards The natural hazards considered in this study are earthquake, climate change, and flooding. Those natural hazards can cause widespread damage and interruption of hydroelectricity generation. The probability of each natural hazard is evaluated. The highest risk of the natural hazards is assessed in the tool. For example, in Turkey, there are various types of active faults: the North Anatolian Fault is the biggest active fault, and the East Anatolian Fault is the second biggest. So the earthquake is always a natural hazard risk for all projects. Revenue In the case of hydroenergy, revenue result from the electricity sold and electric generation (E) is proportional with the discharge and effective head as follows: E ¼ r  g  Q  He 
 t

(1)

Page 6 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Table 2 Scoring procedure of the potential risks in hydroenergy projects Score = 4 RQD is poor–very poor: RQD = %0–50 or the project is located near an active fault or landslide areas 2. Grid connection The distance of the The distance of facility to The distance of facility to The distance of grid connection is facility to grid facility to grid connection grid connection is is less than 5 km between 5 and 15 km between 15 and 25 km connection is higher than 25 km 3. Access road Access road is able to Access road is partially Access road is not able to A new road must be bear heavy lorries with able to bear heavy lorries bear heavy lorries with built that is higher trailers and a heavy with trailers and a heavy trailers and a heavy than 5 km in length mobile crane mobile crane mobile crane under the scope of the project 4. Environmental The project has a detailed The project has an The project has no The project is located issues Environmental Impact Environmental Impact Environmental Impact close to an Assessment Report. Assessment Report Assessment Report or the environmentally Biodiversity issues are report does not evaluate protected area (i.e., investigated with field relevant parameters natural parks, Natura measurements sites) A public consultation 5. Social A robust public A public consultation The facility limits acceptance consultation process has process has been carried process has been not local communities’ been carried out. No out. The locally affected carried out ability to utilize the major objections from community has been surrounding lands local communities were notified and adequate provide a livelihood raised mitigation measures have been taken 6. Land use and Forest Property of treasury Private property: Private property: permits Agricultural land Residential land 7. Macroeconomic The country’s current The country’s current The country’s current The country’s current account balance (% of account balance (% of account balance (% of account balance (% of GDP) is > %3 GDP) is between 0 % and GDP) is between 3 % GDP) is < %3 3% and 0 % The country has legal The country has instable The country has no 8. Regulatory The country has stable regulation for framework for and immature legal uncertainties and mature legal hydroenergy deployment of framework for framework for deployment deployment of hydroenergy. The deployment of frequency of annual hydroenergy. The hydroenergy. The legislation change is frequency of annual frequency of annual between 0.25 and 0.5 legislation change is legislation change is >0.50 3,000

Page 8 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

For example, if the PRS value of a hydropower project falls into high-risk category, the project is expected to have unit investment cost higher than 3,000 $/kW.

Application of the Proposed Methodology: Case Studies from Turkey Hydropower Plants in Turkey

Turkey is situated through a mountainous topography (average altitude = 1,132 m) and owns an annual average runoff value of 186 billion m3 which favors ideal locations for hydropower development (Kucukali 2014b). By September 2013, 69 storage-type and 365 run-of-river-type hydropower plants have been in operation in the country with a total installed capacity of 15.5 and 5.6 GW, respectively (TEIAS 2013). Hydropower has the highest share with 93.8 % among renewable energy sources in Turkey in terms of installed capacity. Turkey has been divided into 26 river basins (Fig. 5); however, 97 % of its economically feasible hydropower potential is distributed into 14 river basins (Table 4) which are

Fig. 5 River basins of Turkey Table 4 Economically feasible hydropower potential of Turkey’s major river basins River basin Euphrates Tigris East Black Sea Coruh Seyhan East Mediterranean Kizilirmak Ceyhan Antalya Yesilirmak West Mediterranean Sakarya West Black Sea Aras Total

Energy potential (GWh/year) 37,823 16,562 13,194 10,973 6,957 6,749 6,420 5,996 5,345 4,984 3,240 2,585 2,149 2,692 125,669

Power potential (MW) 9,555 4,890 3,900 3,247 1,788 1,856 2,116 1,779 1,437 1,257 881 1,191 642 868 35,407

Drainage area (km2) 120,917 51,489 24,022 19,894 20,731 22,484 78,646 21,222 22,615 36,129 14,518 56,504 29,682 27,548 546,401

Runoff (Tm3/year) 33.48 21.81 14 6.46 7.06 12.27 6.28 7.21 7.76 5.54 11.24 6.03 10.04 5.54 155

Page 9 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

mostly situated in the mountainous areas. Euphrates River, with its 127,304 km2 of drainage area and elevation range between 500 and 5,000 m, itself is responsible for 30 % of the country’s hydropower potential. Turkey’s largest hydropower plants were constructed on the Euphrates River, namely, Ataturk (P = 2,400 MW), Karakaya (P = 1,800 MW), and Keban (P = 1,330 MW). Moreover, the Black Sea region, which has steep and rocky mountains that extend along the coastline, has a considerable hydropower potential. The Eastern Black Sea region is of particular importance in terms of hydropower potential due to its higher capacity factor, which is the ratio of actual operating hour per year to that of the total. The projects developed by private sector for this basin account for 20 % of the total projects in the country. The Turkish Renewable Energy Law (Law No.5346), which comes into force in May 2005, defines a reservoir area less than 15 km2 for hydropower plants and does not have a limit for installed capacity. Even though it accelerated hydropower plant investments to an extent, these features specified in Law No.5346 cause private sector to shift their investments from small hydropower plants (SHP) to large hydropower plants. For example, Koprubasi Dam, which is situated on the Western Black Sea Basin, having a height of 108 m, a reservoir area of 5.9 km2, and a 79 MW of installed capacity, lies within the frame of Law No. 5346. However, this creates a conflict between Turkey and the EU policy. In European Union member states, countries limit the installed capacity and give the sector extra payments if they build SHP plant (Kucukali and Baris 2009). For example, in Clean Development Mechanism, for hydropower projects above 20 MW, member states must ensure that relevant international criteria and guidelines will be respected during the development of such project activity (Branche 2010). In Turkey, a company that intends to have an SHP license must sign the Water Usage Rights Act with the General Directorate of State Hydropower Works (DSI). In this context, the company must meet the requirements stated in this act. In applications to get an SHP license, river basin plan prepared by DSI is taken into account. In Turkey, the Environmental Impact Assessment (EIA) report was not required for hydropower plants with less than 50 MW installed capacity before 17 July 2008. However, a regulation was issued on this date stating that hydropower plants having an installed capacity between 0.5 and 25 MW have to undertake an EIA. However, this regulation did not have an expected effect since many of licenses for hydropower plants were granted before the enactment of this regulation. After the enactment of Turkish Renewable Energy Law in May 2005, hydropower plant operators received a guaranteed price of 73 $/MWh for the generated electricity under the scope of the law for 10 years without a limitation for the installed capacity. Table 5 clearly shows the influence of the Renewable Energy Law on the development of small hydropower plants in Turkey. The hydropower potential increased 15 % in 2007 as compared to 2006. Moreover, the construction of hydropower plants increased by a factor of four in 2007 as compared to 2006, and the planned plants are almost doubled. By June 2012, hydropower has the highest share with 89 % among renewable energy sources in Turkey in terms of installed capacity. 290 hydropower plants have been in operation with a total installed capacity of 16,265 MW, and 589 power plants have been in the construction stage under the scope of the Renewable Energy Law (EMRA 2012).

Table 5 Progress in hydropower plants after the enactment of Renewable Energy Law in Turkey (DSI 2006; Tutus 2008) In operation (2006) Number of projects 142 Installed capacity 12,788 (MW) Energy (GWh/yıl) 45,930

In operation (2007) 148 13,306

Under construction (2006) 40 3,197

Under construction (2007) 158 6,564

Planned (2006) 573 20,765

Planned (2007) 977 22,260

47,590

10,518

23,620

73,851

79,177

Page 10 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 6 Locations of the Koknar, Sifrin, and Kulp IV hydropower schemes

Table 6 The design characteristics of assessed hydropower plantsa Name 1. Koknar 2. Sifrin 3. Kulp IV a

Q (m3/s) 2.50 3.40 20

H (m) 373 224 75

P (MW) 8.20 6.74 12.68

Lt (km) Tunnel is absent 1 1.88

Lg (km) 5 33 30

River basin West Black Sea Euphrates Tigris

Q design discharge, H effective head, P installed capacity, Lt tunnel length, Lg grid connection distance

Table 7 Cost breakdowns of Koknar, Sifrin, and Kulp IV hydropower projects Cost item ($) Project design Civil works Electromechanical equipment Hydromechanical equipment Grid connection Land use and permits Financial Other Total cost Unit investment cost ($/kW)

1. Koknar 600,000.00 13,000,000.00 3,800,000.00 400,000.00 300,000.00 3,600,000.00 1,300,000.00 – 23,000,000.00 1,551

2. Sifrin 354,000.00 16,865,500.00 5,013,806.00 1,984,469.00 5,423,667.00 600,000.00 6,120,000.00 845,500.00 37,206,942.00 2,971

3. Kulp IV 1,180,000 25,700,000 7,490,000 4,800,000 5,400,000 2,600,000 5,700,000 – 52,870,000 4,162

Case Studies from Turkey The proposed risk assessment technique is applied to real-time hydropower projects in Turkey, namely, Koknar, Sifrin, and Kulp IV. The locations of those projects are shown on map in Fig. 6 and the characteristics are presented in Table 6. All of the assessed projects are diversion type where the river flow is diverted into an artificial channel or a tunnel. Since the installed capacity of a hydropower plant is linearly correlated with the discharge and head, hydropower operators in Turkey have a tendency to maximize those parameters by reducing the amount of environmental flow and increasing the head by selecting diversion-type HPP (Kucukali 2014a).

Page 11 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Table 8 Analysis of the investment cost of Kulp IV hydropower plant (P = 12.7 MW, unit investment cost = 4,163 $/kW) Estimated Actual Description cost ($) cost ($) Project design 1,090,000 1,180,000 Civil works 12,500,000 25,700,000 Electromechanical 6,790,000 7,490,000 equipment Hydromechanical 8,900,000 4,800,000 equipment Grid connection 2,600,000 5,400,000 Land use and permits Financial

Cost increase (%) 8.3 105.6 10.3

Share of total cost (%) 2.2 48.6 14.2

46.1

9.1

107.7

10.2

2,200,000

2,600,000

18.2

4.9

3,100,000

5,700,000

83.9

10.8

Reason Additional project designs Poor geology (serpentine) at the tunneling site Under estimated costs of Technical equipment demand The prices of DSI are very high compared to the market Technical demands by TEIAS and length of the power supply line were increased The cost of the forest usage permit was not taken into account Increase in interest rates because of financial crisis

Table 9 Analysis of the investment cost of Sifrin hydropower plant (P = 6.74 MW, unit investment cost = 2971 $/kW) Estimated cost ($) 300,000 8,174,345

Cost Actual cost ($) increase (%) 354,000 18 16,865,500 106.3

Share of total cost (%) 0.95 45.33

5,013,806

5,013,806

0

13.48

Reason Revision of the project Unforeseen geotechnical conditions and project revision –

1,984,469

1,984,469

0

5.33

–

1,080,000

5,423,667

402.2

14.58

Land use and permits 225,000 Financial 6,120,000 Extra costs 750,000

600,000 6,120,000 845,500

166.7 0 12.7

1.61 16.45 2.27

The length of the power supply line was increased Unforeseen expropriation costs – –

Description Project design Civil works Electromechanical equipment Hydromechanical equipment Grid connection

Table 10 Analysis of the investment cost of Koknar hydropower plant (P = 8.24 MW, unit investment cost = 1551 $/kW) Description Project design Civil works Electromechanical equipment Hydromechanical equipment Grid connection Land use and permits Financial Extra costs

Estimated cost ($) 600,000 10,000,000

Share of Actual cost ($) Cost increase (%) total cost (%) 600,000 0.00 2.61 13,000,000 30.00 56.52

2,800,000

2,800,000

0.00

12.17

Reason – Landslides occurred during the excavations –

1,000,000

1,000,000

0.00

4.35

–

300,000 200,000 2,500,000 600,000

400,000 300,000 3,600,000 1,300,000

33.33 50.00 44.00 116.67

1.74 1.30 15.65 5.65

Increase in material prices Additional permission costs Additional loan costs Challenging site and weather conditions

Page 12 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 7 The photo of the serpentine at the tunnel route of Kulp IV hydropower scheme. This unforeseen rock type along the tunnel alignment caused the increase of the construction cost by a factor of two

Although the installed capacities of the evaluated projects are very close and in the range of 8.2–12.7 MW, the unit investment costs of the projects vary between 1,551 and 4,162 $/kW (Table 7). Those results confirm the site-specific characteristics of the hydropower projects. For all assessed projects, civil works dominate the total cost in the order of 45–56 % which are consistent with the published data (IRENA 2012). Tables 8, 9, and 10 present the investment cost analyses of Kulp IV, Sifrin, and Koknar hydropower plants. All of the three projects experienced cost overruns (100 % for Kulp IV, 57 % for Sifrin, and 28 % for Koknar) mainly because of the unforeseen geological conditions. For example, the estimated cost of civil works for Kulp IV increased twofold because of the unexpected rock type (serpentine) along tunnel alignment (Fig. 7). Those findings are in agreement with the judgments of Binquet (2010) and Palmieri (2012). Table 11 shows the application of the risk assessment tool to the Koknar HPP. The project risk scores are attained based on the principles and procedures presented in Table 2, and the justifications of risk scores are given in Table 11. For the given project, the key risks appeared to be country related such as macroeconomic risks and regulatory uncertainties (Table 11). For instance, Turkey’s current account balance as a percentage of GDP is 7.5 % at the end of 2013 which indicates that the country has high macroeconomics risks which can lead to an increase in currency and interest rates. Therefore, the macroeconomic risk factor has a score of 4 for the project. Turkey does not have stable and sufficient regulations for the devolvement of hydropower and for environmental issues. Electricity Market Law (Law No. 4628) has changed 8 times since 2001 and EIA Regulation (Regulation No. 25318) has changed 13 times since 1994. The frequency of annual legislation change is 0.72. For example, the Environmental Impact Assessment (EIA) report was not required for hydropower plants with less than 50 MW installed capacity before 17 July 2008. The regulation was changed on this date stating that hydropower plants having an installed capacity between 0.5 and 25 MW have to undertake an EIA. So, the regulatory risk factor has a score of 3 for the project. Another regulatory risk for hydropower development in Turkey is the implementation of European Union Water Framework Directive. Since Turkey is a European Union (EU) candidate, its laws and policies are expected to be consistent with those of EU. European Union adopted the Water Framework Directive (WFD) in October 2000 which constitutes that a good ecological status for all aquatic ecosystems must be sustained. In this respect, EU countries must prepare their integrated river basin management plants and establish their monitoring networks. Turkey has begun to adopt the Water Framework Directive in its legislation system by 2012 with the enactment of Regulation on Protection of Water Basins and Preparation of Management (Official Gazette no. 28444, 17 October

Page 13 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Table 11 Risk scorecard for the Koknar HPP No 1 2 3

Risk factor Site geology Grid connection Access road

4

6

Land use and 1 permits Environmental 2 impact Social acceptance 2

7

Macroeconomic

4

8

Regulatory uncertainties

3

9

Natural hazard

3

5

10 Revenue

Score 1 1 1

2

Evidence There is no tunnel construction under the scope of the project Grid connection distance of the scheme is 4.9 km Existing roads are capable for the construction of the scheme, and there is no need to build new roads The scheme is located in a forest land The project has Environmental Impact Assessment Report, but the biodiversity issues are not investigated in detail, supported with field studies A public consultation process has been carried out. The locally affected community has been notified, and adequate mitigation measures have been taken Turkey’s current account balance as a percentage of GDP is 7.5 % at the end of 2013 (Economist 2013) Turkey does not have stable and sufficient regulations for deployment of hydroenergy and environmental issues. For instance, Electricity Market Law has changed eight times since 2001 and EIA Regulation has changed 13 times since 1994. The frequency of annual legislation change is 0.72 The most important natural hazard for the project site is severe earthquake, and its occurrence probability is between 10 % and 50 % (Linnerooth-Bayer et al. 2005) There is no weir and dam at the upstream of the scheme. For the energy yield estimate, sufficient and reliable stream flow data are used, and calculations are done by experts in this field

Project risk score (PRS) = 0.47

2012). In the near future, it is highly possible that small hydropower plants may sustain their current operational conditions and be largely developed in Turkey if they generate electricity with respect to the environment and meet the ecological requirements (Kucukali 2014a). For the energy yield estimate of the Koknar HPP, sufficient and reliable stream flow data are used and calculations are done by experts. Also, there is no weir and dam at the upstream of the facility (Fig. 8). So, the revenue risk has a score of 2. There is no tunnel and new road constructions under the scope of the project which yield the score of site geology and access road as one. By using Eq. 2, the cumulative effects of ten risk factors were evaluated and the PRS of Koknar HPP is computed as 0.47 which means that the project involves low risk. Based on the PRS of the facility, unit investment cost falls into 1,000–2,000 $/kW range which is in agreement with the observed unit investment cost value of 1,551 $/kW. The risk assessment tool is also applied to other hydroenergy projects. For Sifrin and Koknar HPPs, PRS values are calculated as 0.67 and 0.78, respectively. Again there is good agreement between the forecasted (Table 2) and actual unit investment cost values of the projects (Table 7). Figure 9 shows the risk profiles of assessed hydroenergy projects in a spider chart. This graphical representation enables one to compare risk-reward characteristics of alternative projects in an easy-to-read profile.

Conclusions In this study, a practical risk assessment framework is proposed for hydroenergy projects based on an integrated approach. The proposed risk assessment technique is applied to three different hydropower projects in Turkey. Project risk score (PRS) values calculated are 0.47, 0.67, 0.78 for Koknar, Sifrin, and Kulp IV hydropower schemes, respectively. The Koknar HPP has the lowest project risk score (PRS)

Page 14 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 8 Photos of the Koknar hydropower plant. (a) power house and (b) stream flow measurement gauge

Koknar

Site Geology Revenue

Grid Connection

Sifrin Kulp IV

Natural Hazard

Access Road

Land USe and Permits

Regulatory Risks

Environmental Impact

Macroeconomic Social Acceptance

Fig. 9 Risk profiles of Koknar, Sifrin, and Kulp IV hydroenergy projects. For each identified risk factor, the level of risk is scored on a scale from 1 (low) to 4 (very high)

mainly because the large underground works under high rock cover such as tunneling were absent in this project. The calculated PRS values of the assessed hydropower projects are in agreement with the unit investment cost values of HPPs. By applying the proposed methodology, investors and project developers can understand the challenges and risk-driven factors at the initial stage of a project in order for their businesses to be successful. Page 15 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

Equally, potential investors and financiers can understand and quantify potential risks such as environmental issues and geological risks before deciding to proceed. The risk assessment tool can give a competitive advantage in the field of hydroenergy system deployment, and it can reduce predevelopment time and costs. Additionally, project developers and investors can develop a strategy in project planning stage to cope with the project risks and complex permitting procedure. For now, the investigated sites cover Turkish hydropower plants. However, the developed tool can be applied to other hydropower projects worldwide by adjusting the relevant parameters. The main advantages of the proposed methodology are its ease of use and simplicity.

References Binquet J (2010) Allocation of risks in EPC contracts. In: Proceedings of hydro 2010 conference, Lisbon, CD Branche E (2010) Impact of carbon credits on hydropower project financing. In: Proceedings of hydro 2010 conference, Lisbon Branche E (2011) Hydropower: the strongest performer in the CDM process, reflecting high quality of hydro in comparison to other renewable energy sources. In: Proceedings of hydro 2011 conference, Prag, CD DSI-General Directorate of State Hydraulic Works (2006). The Statistics of Turkey Hydropower Plants. DSI, Ankara EBRD (2013) Eligibility criteria for small hydro power projects. http://www.midseff.com/downloads/ eligibility_criteria.pdf. Accessed Dec 2013 Economist (2013) Trade, exchange rates, budget balances and interest rates. http://www.economist.com/ news/economic-indicators/21595022-trade-exchange-rates-budget-balances-and-interest-rates. Accessed 25 Jan 2014 EIB (2013) The EIB statement of environmental and social principles and standards. http://www.eib.org/ attachments/strategies/eib_statement_esps_en.pdf. Accessed Dec 2013 EMRA (2012) The database of the power plants that obtained license in Turkey. http://www2.epdk.org.tr/ lisans/elektrik/lisansdatabase/verilentesistipi.asp. Accessed June 2012 Ernst and Young (2010) Business risk report. EYG no. AU0583, London Ford N (2008) Europe seeks small hydro. http://www.waterpowermagazine.com/news/newseuropeseeks-small-hydro. Accessed Dec 2013 Fraser JR (2010) How to prepare a risk profile. In: Fraser J, Simkins BJ (eds) Enterprise risk management. Wiley, Hoboken, pp 171–188 Fratzscher M, Luciana J, Lucio S (2010) Asset prices, exchange rates, and the current account. Eur Econ Rev 54:643–658 Gordon JL (1983) Hydropower costs estimates. J Water Power Dam Constr 35:30–37 Gronbrekk W, Barton H, Khoury RH (2010) International sustainability tools for hydropower role, relevance and industry reporting trends. In: Proceedings of hydro 2010 conference, Lisbon Hall DG, Hunt RT, Reeves KS, Carroll GR (2003) Estimation of economic parameters of U.S. hydropower resources. Idaho National Engineering and Environmental Laboratory, Idaho Falls IEA (2012) Hydropower technology roadmap. International Energy Agency, Paris IHA (2013) Hydropower sustainability assessment protocol. http://www.hydrosustainability.org/ Protocol-Assessments.aspx. Accessed 15 Apr 2013 IRENA (2012) Renewable energy technologies: cost analysis series-hydropower. The International Renewable Energy Agency, Bonn Page 16 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_5-1 # Springer-Verlag Berlin Heidelberg 2014

ISO 31000 (2009) Risk management-principles and guidelines on implementation. International Organization for Standardization, Geneva JICA (2011) The study on optimal power generation for peak demand in Turkey. IDD: JR 10–134 Kucukali S (2011) Risk assessment of river-type hydropower plants by using fuzzy logic approach. Energy Policy 39(10):6683–6688 Kucukali S (2014a) Environmental risk assessment of small hydropower (SHP) plants: a case study for Tefen SHP plant on Filyos River. Energy Sustain Dev 19:102–110 Kucukali S (2014b) Finding the most suitable existing hydropower reservoirs for the development of pumped-storage schemes: an integrated approach. Rene Sust Energ Rev 37:502–508 Kucukali S, Baris K (2009) Assessment of small hydropower (SHP) development in Turkey: laws, regulations and EU policy perspective. Energy Policy 37:3872–3879 Lawrence S, Dickson P (2010) Clean energy infrastructure. In: Underhill MD (ed) Handbook of infrastructure investing. Wiley, Hoboken Linnerooth-Bayer J, Mechler R, Pflug G (2005) Refocusing disaster aid. Science 309(5737):1044–1046 Mermet SR, Gehant B (2011) Risk management for hydroelectric power plants. In: Proceedings of hydro 2011 conference, Prag, CD Mund J (2004) Applied risk analysis: moving beyond uncertainty in business. Wiley Finance, Hoboken Palmieri A (2012) Managing financial risks for uncertainty. In: Proceedings of hydro 2011 conference, Prag, CD Pike R (2010) Scenario-building techniques for improved risk management. In: Reuvid J (ed) Managing business risk: a practical guide to protecting your business. Kogan Page, Hoboken, pp 17–22 TEIAS (2013) Power plants in Turkey. http://www.teias.gov.tr/Eng/DispatchReports.aspx. Accessed Sept 2013 Tutus A (2008) Dams and hydropower plants. In: The Status of Energy Sector in Turkey and in World Symposium, METU, Ankara, Turkey Waltham T (2009) Foundations of engineering geology. Spon Press, Oxford Williamson SJ, Stark BH, Booker JD (2011) Low head pico hydro turbine selection using a multi-criteria analysis. In: World renewable energy congress 2011, Linköping, CD World Bank (2013) Environmental strategy. www.worlbank.org Yescombe ER (2002) Principles of project finance. Academic Press, California Zipparro VJ, Hasen H (1993) Davis’ handbook of applied hydraulics. McGraw-Hill, New York

Page 17 of 17

Agronomic Management of Straw and Its Energy Use in a Long-Term Sustainability Perspective Massimo Monteleone, Mitra Kami Delivand, Pasquale Garofalo, and Anna Rita Bernadette Cammerino

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consequential Life Cycle Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Description of the Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cropping Systems and Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wheat Cultivation and Energy Inputs of the Cropping Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Energy Analysis of Post-harvesting Operations and Electricity Generation . . . . . . . . . . . . . . . . . . . Life Cycle Energy Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GHG Life Cycle Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Direct and Indirect Land Use Impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 3 4 7 8 10 10 12 13 14

Abstract

There are some features in the energy valorization of agricultural residues and in their carbon footprints calculation that are still very difficult to detect, and, therefore, they are frequently neglected. A general assumption for the environmental assessment of bioenergy systems is that crop residues are “carbon neutral.” Nevertheless, these residues play a critical role in sustaining soil organic matter with an inherent carbon sequestration value. This chapter is focused in assessing the actual energy demand and carbon equivalent emissions associated with an “expanded” life cycle of a straw-to-energy system. Three

M. Monteleone (*) • P. Garofalo • A.R.B. Cammerino STAR-AgroEnergy Research Group – Department of Agriculture, Food and Environment, University of Foggia, Foggia, Italy e-mail: [email protected]; [email protected]; [email protected] M.K. Delivand Engineering and Renewable Energy, HTW-Berlin, Berlin, Germany e-mail: [email protected]; [email protected] # Springer-Verlag Berlin Heidelberg 2015 W. L. Filho (ed.), Handbook of Renewable Energy, DOI 10.1007/978-3-642-39487-4_11-1

1

2

M. Monteleone et al.

farming systems, according to different management practices and agroenergy inputs, were compared. A crop-system simulation model was applied to Mediterranean climatic conditions to analyze the long-term soil organic matter and the annual N2O emissions from the soil, converting the simulation results into GHG emissions and considering the emissions of the whole straw-to-energy process. Wheat rotation under no-tillage soil management showed the highest net energy gains and the maximum GHG emission savings, thus offering an optimal trade-off between wheat grain yield and straw-to-energy conversion, under the constraint to protect and maintain soil resource and its quality. Keywords

Soil organic carbon • Life cycle GHG emission abatement • Soil N2O emission • Land use change • Agroenergy farming

Introduction The European Renewable Energy Directive (2009/28/EC), also called RED, has defined the framework for the promotion of energy from renewable biomass sources to achieve the target of a 20 % share of renewable energy and a 20 % reduction of greenhouse gas (GHG) emissions. To boost these bioenergy forms, a specific preference for “advanced” biofuels (those obtained from residues, waste, and by-products) is assigned by the RED; moreover, biofuels obtained from this kind of feedstock provide a “double counting provision” in achieving the national targets of each EU member state, together with a zero GHG emission (Whittaker and McManus 2012). Consequently, this is expected to trigger an increased interest in the use of straw for energy purposes rather than for conventional, non-energy target use, such as agricultural soil amelioration or animal beddings. The retention of straw in the soil may have several advantages, such as sustaining organic matter, recycling favorable nutrients, allowing carbon sequestration, improving soil aggregation and structure, enhancing erosion control, and increasing water infiltration, retention, and drainage (Limon-Ortega et al. 2008; Bakht et al. 2009; Del Gado 2010; Johnson et al. 2010; Kludze et al. 2013; Powlson et al. 2008). However, these positive effects are greatly variable depending on the local climate conditions, soil types, and crop management. Therefore, the potential trade-offs of the straw-to-energy systems and the actual amounts of straw that can be removed from arable land without threatening the long-term sustainability of the soil need to be addressed (Lal 1997; Blanco-Canqui and Lal 2009; Lafond et al. 2009). Furthermore, no direct conflict with food production should arise, and the possible negative impacts due to straw removal could be avoided or significantly mitigated by integrating best-practice guidelines in the agricultural area under study, both in the crop cultivation systems and in the management of residues (Faaij and Domac 2006).

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

3

Consequential Life Cycle Assessment The sustainability of agricultural-based bioenergy systems may not be met without understanding the long-term consequence of a change in the content of soil organic matter, when crop residues are removed for energy purposes. The same applies to the N2O gaseous emissions from the soil that could be significantly affected by straw management (i.e., if straws are released on the soil or, alternatively, partially removed). Indeed, the boundaries of the straw-to-energy system should be properly expanded in order to include also the effects conventionally considered external to the bioenergy system and related to the upstream phases of farming (when both straw and grain are produced). The consequential approach to life cycle assessment (CLCA), different from the “attributional” one, is mainly focused on the marginal effects of possible decisions (Curran et al. 2005; Finnveden et al. 2009). A system expansion is also applied to solve the multi-functionality processes, i.e., the processes with more than one single output (Ekvall and Weidema 2004). Despite the fact that there is uncertainty on the full consequences of a decision change, the CLCA proved to be a useful tool to evaluate the environmental suitability of future energy scenarios and policies (Brander et al. 2009; Whittaker et al. 2011; Whittaker and McManus 2012). CLCA could be specifically applied in order to investigate the direct and indirect effects on land use change (LUC). The EU bioenergy regulation framework currently “in progress” (i.e., the next amendment to the RED) or other voluntary or compulsory sustainability schemes at international level (Sánchez et al. 2012) are considering LUC as one of the most significant issues related to biofuel sustainability.

Description of the Case Study A geographical location close to Foggia (41
270 Lat. N; 15
040 Long. E) at 90 meters above the sea level in the Southern part of Italy is the reference area for conducting the cropping simulations. The area is characterized by an accentuated thermo-Mediterranean kind of climate (according to the UNESCO-FAO bioclimatic classification, 1963), showing an average annual rainfall of 550 mm, an average daily evaporation rate in summer of 10 mm, a maximum temperature of 40
C in summer, and a minimum temperature of 0
C in winter. The soil considered is a vertisol of alluvial origin, Typic Calcixeret (Soil Taxonomy 10th ed., USDA 2006), a soil depth of 1.35 m, and a silt-clay texture (12.9 % sand, 43.7 % clay, 43.4 % silt) with a bulk density of 1.24 t m3, pH 8.5, field capacity at 42 % (v/v), and a permanent wilting point at 24 % (v/v). The geographical area under study is illustrated in Fig. 1. This agricultural area is characterized by a significantly extended wheat cultivation and, therefore, by the highest straw spatial density of the whole region. For this reason, the province of Foggia is considered as one of the most suitable areas in Italy to install a biomass power plant supplied by straws. In Fig. 2, the approximate location of a 25 MWe combined heat and power

4

M. Monteleone et al.

Fig. 1 Geographical position of the study area (province of Foggia, South of Italy)

(CHP) plant is illustrated with respect to the basin of cereal cultivation and spatial straw availability.

Cropping Systems and Simulations Considering that the majority of the processes involved in the performance of a cropping system are very site specific, i.e., highly dependent on climate and soil characteristics as well as on agronomic practices, the use of a simulation model could be very useful and almost essential if a long period of time is under consideration. A calibrated crop simulation model is able to estimate the annual amount of available straws and the long-term fate of soil organic carbon (SOC) and N2O emissions from the soil with respect to the systems of interest. CropSys is the simulation model employed to carry out this case study.

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

5

Fig. 2 Geographical location of the biomass power plant with respect to the straw availability of the agricultural area corresponding to the biomass supply basin (left) and also considering the straw spatial density of the same area (right)

CropSys is a multiyear, multi-crop, daily time-step crop growth simulation model, designed to serve as an analytical tool to assess the effect of management and environmental conditions on crop productivity (Sto¨ckle et al. 2003, 2009). CropSys is able to simulate a very wide range of processes, according to several environmental and management conditions, such as soil-water balance, soil-plant nitrogen budget, canopy cover and root growth, crop phenology, total biomass production, crop grain yield, residues left on the soil and their decomposition rate, soil organic matter (SOM) and soil organic carbon (SOC), etc. Previous work allowed a proper calibration and validation of the model, specifically on wheat cultivation systems representing environmental conditions of Southern Italy (Garofalo et al. 2009). The model, as compared to experimental data, showed a very good performance after a parameters’ calibration set (Garofalo et al. 2009). Seasonal cropping cycles were performed over a consecutive 50-year period (1960–2010) applying the actual meteorological data recorded in Foggia. Soil N2O emissions depend on the microbial nitrification and denitrification processes, which can be affected by several factors such as climate or soil characteristics, and are also under the influence of agricultural practices (Signor et al. 2013). In CropSys a specific output of the model is the amount of the gaseous N2O released from the soil related to both nitrification and denitrification processes. The absence of a previous experimental data set for N2O emissions did not allow to assess the actual reliability of the simulations; however, the results obtained from modeling showed consistency with similar works by other authors (Gabrielle and Gagnaire 2008; Gan et al. 2012). The development of conservation tillage systems such as reduced tillage or no-tillage is intended to improve soil structuring, increase water retention and availability for crop growth, and especially to minimize soil erosion (Anderson 2011). On the other side, conservation tillage often requires an increased use of herbicides and pesticides (Farine et al. 2010; Shrestha et al. 2006).

6

M. Monteleone et al.

The three alternative wheat cropping systems compared in this case study are characterized by the following management operations: (i) straw retention in a conventional tillage and single wheat cultivation system (W0); (ii) similarly to W0, conventional tillage was applied but straws were partially removed from the soil to be used for energy purposes (W1); and (iii) partial removal of straw in a no-tillage, wheat-wheat-triticale rotation system (W2), which should be considered an innovative crop management for the region. For all the three wheat management systems, the initial content of soil organic matter (SOM) was 41.72 Mg ha1 at the first 0.30 m of soil depth, corresponding to 1.2 % in weight (Monteleone et al. 2015). This starting value of SOM is a rather common condition of the agricultural Mediterranean soils, with a climate characterized by a semiarid summer. The grain annual yield was assumed fixed and equal to 3.0 Mg ha1 (corresponded to the actual average value of the region), while the amount of collectable straw depended on the height of the cutter blade of the harvester and on the technical efficiency of the baler equipment. The straw harvest index (straw to stubble ratio) and the baler harvesting efficiency were assumed 0.80 and 0.60, respectively. The consequent overall straw collectable coefficient was equal to 0.48. The harvested herbage amounted to approximately 10 Mg ha1 on dry basis. Fertilizing nitrogen (N) and phosphorous (P2O5) were applied at rates of 120 and 100 kg ha1, respectively. The trend of the SOC over a long-term simulation for the three management systems is illustrated in Fig. 3. The results indicate that in the no-tillage and crop rotation management (W2), the quantity of SOC remained almost constant compared to the initial SOC, whereas in the conventional tillage practice with straw removal (W1) a significant reduction of SOC could be estimated. As was expected, a sharp increase in SOC was observed in case of straw retention (W0).

Fig. 3 Simulated long-term changes of soil organic carbon (Monteleone et al. 2015) – W0, conventional single crop-straw retention; W1, conventional single crop-straw removal; W2, no-tillage rotating crop-straw removal

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

7

Wheat Cultivation and Energy Inputs of the Cropping Systems The energy analysis considered all forms of auxiliary fossil energy inputs following the physical material flows of the productive process (Alluvione et al. 2011). Material flows were transformed into energy flows by applying the methodology reported in the RED (Part C, Annex V) and those conversion coefficients used in the BioGrace project (http://biograce.net/). This choice has been made to harmonize the calculations based on the European Union. Direct energy inputs were referred to the energy content of fuels and lubricants. Indirect energy inputs accounted for the amount of energy spent for the production, transportation, and maintenance of all agro-technical inputs. Some of these inputs were completely consumed in the course of the cropping cycles, while others were only gradually employed (e.g., implements and machineries); in the latter, their partial energy consumption or energy amortization should be carefully determined with respect to their productive life horizon. The cultivation of wheat followed the common farm practices in the area (i.e., province of Foggia) with respect to the mechanical operations and agrochemical applications for the three cropping managements. W0 and W1 consisted of a first moldboard plowing at a soil depth of 0.35 m in late August, followed by two consecutive harrowings at a soil depth of 0.15 m in November to prepare the seedbed. In W2, wheat was cultivated for two continuous years and then in rotation with triticale, as an herbage crop, in the third successive year. In no-tillage system (W2), the seed was sown by a direct drill seeder in a narrow trench with sufficient coverage without other soil preparation. In the South part of Italy, wheat is generally sown from November through December and harvested in June. The winter herbage of triticale was cultivated similarly to wheat but mowed much earlier, at the dough development stage, approximately 20–25 days after flowering. In W0, straws were left on the soil after being chopped and uniformly distributed by a cutting-shedding device installed on the combine harvester. Differently, in W1 and W2, straws were removed through baling. Prior to baling, straw was left on the field to dry; the typical moisture content achieved is around 15 %. Collection of the straw was assumed to be performed by a rectangular baler (2.00  1.20  0.85 m; 400 kg weight). The remaining crop residues after straw harvesting, referred to as “stubbles,” were incorporated back into the soil by plowing in W1 or left on the soil surface in W2. Energy inputs of the three mentioned cropping systems (W0, W1, and W2) were divided into three categories: (i) diesel and lubricants; (ii) mechanization including tractors, machineries, and equipment; and (iii) agronomic material inputs such as seeds, fertilizer, herbicides, etc. A summary of the primary energy inputs for cropping systems based on conventional and conservation tillage, respectively, is reported in Table 1. The total amount of energy required for wheat cultivation, according to the conventional tillage system (W0 and W1), is equal to 14.6 GJ ha1. No-tillage (W2) allowed a 10 % of reduction in the total energy input (13.1 GJ ha1) due to direct sowing, but a net increase in the energy associated with herbicide application was observed. Fertilizers, pesticides, and seeds (63 % in W0 and W1,

8

M. Monteleone et al.

78 % in W2, respectively) accounted for the highest energy consumption; diesel together with lubricants accounted for 34 % (W0 and W1) and 20 % (W2) of the total energy budget, while the remainder is negligible (2–3 %) and was allocated to the machinery manufacturing.

Energy Analysis of Post-harvesting Operations and Electricity Generation The post-harvesting operations consisted of straw chopping and incorporating into the soil for W0 and straw bailing and transporting to the CHP plant for W1 and W2. Specifications regarding the technical assumptions on the projected CHP plant are shown in Table 2. Logistic information about the straw supply basin and straw haulage are reported in Table 3. As shown in Table 2, an electrical efficiency of 31.25 % was taken for the combustion process, considering an average straw moisture content of 15 % and a Table 1 Energy inputs for the wheat cultivation system performing conventional tillage (A) or no-tillage (B). All the values are expressed in MJ ha1

(A) Operations Plowing Harrowing (I + II) Basal dressing Sowing Top dressing Herbicide application Combined harvesting Total Partitioning (%) (B) Operations Basal dressing Sowing Top dressing Herbicide application Combined harvesting Total Partitioning (%)

Diesel and lubricants

Tractors and machinery

1,193.0 1,303.3 288.6 525.2 288.6 144.3

108.7 132.5 29.2 62.3 29.2 11.6

1,171.3

80.4

4,914.3 33.6

454.0 3.1

9,252.4 63.3

14,620.7 100.0

288.6 620.5 288.6 288.6

29.2 79.8 29.2 23.2

3,165.1 1,162.6 4,507.1 1,342.0

3,482.9 1,862.8 4,824.9 1,653.8

1,171.3

80.4

2,657.5 20.3

241.8 1.8

Materials

3,165.1 909.1 4,507.1 671.0

Total 1,301.7 1,435.8 3,482.9 1,496.7 4,824.9 826.9 1,251.7

1,251.7 10,176.9 77.8

13,076.2 100.0

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

9

Table 2 Assumptions about the functional parameters of the CHP plant in the study area Functional parameters of the biomass plant Power capacity Power capacity used for self-consumption Gross electrical efficiency Annual operation Gross electricity produced Net electricity produced Low heating value of straw (15 % moisture content) Theoretical equivalent power output (15 % moisture content) Annual straw demand at the plant (15 % moisture content)

Units MWe MW % h year1 MWh year1 MWh a1 GJ t1 MWh t1 Mg year1

Values 25 2 31.25 8,000 200,000 184,000 14.62 4.06 157,582

Table 3 Assumptions for the straw supply to the CHP plant in the study area Parameters of straw supply Unitary biomass availability (on dry weight) Total surface of arable land required Total area of the supply basina Straw spatial density Maximum distance of supply Weighted average distance of supply Unitary load capacity of truck Number of round transport trips Cumulative transport distance

Mg ha1 year1 ha

Single wheat crop system (W1) 1.22 109,790

Rotation wheat crop system (W2) 2.41 55,502

km2 Mg km2 km km

3,640 36.81 50.78 34.61

1,896 70.68 36.04 12.66

Mg N

22.4 7,038

22.4 7,038

km

487,116

178,228

a

The farmer willingness to conferring the straws is assumed to involve 50 % of the arable surface under wheat cultivation

corresponding average LHV(wet) of 14.62 MJ kg1. The annual amount of dry straw needed to supply the biomass plant is approximately 134,000 t, while the annual net electricity obtained from straw combustion is equal to 662,000 GJ (considering an internal energy consumption of 8 %). Straw transportation distances were consistently estimated with the use of a GIS software on the basis of the road network map of the area. The environmental performance of the energy conversion process was assumed to correspond to the average European mix (at medium voltage) equal to 128.25 kg CO2 eq. GJ1 (reported in “BioGrace,” List of Standard Values, 2011). The following analysis is focused on the electricity generation from straw, without considering the use of heat generated from straw combustion in a combined heat and power (CHP) installation.

10

M. Monteleone et al.

Life Cycle Energy Flows The overall energy balance of the pre- and post-harvesting operations with respect to the energy systems based on straw is presented in Table 4. It demonstrates that using straw for energy conversion contributed significantly to fossil displacement and efficiently compensated for the energy consumed along the straw logistic operations. Straw removal in a single crop system (W1) can allow a net energy gain of approximately 5.5 GJ ha1, while straw removal combined with crop rotation (W2) improved the net energy savings to 8.6 GJ ha1. Considering the W2 system, the very strong energy penalty of 4.4 GJ ha1 can be explained by the reduction of cereal yields due to the establishment of an herbage crop in rotation with wheat (on one-third of the available cropped surface). In order to compensate the nutrient loss from straw removal, another energy penalty was estimated both with respect to W1 and W2 (according to the actual mineral content of straw in N, P, and K and also considering a fertilizing efficiency of 0.7). Nevertheless, the higher spatial concentration of straw in the crop rotation system (as an average value over the 3-year rotation) significantly increased the overall energy credits due to power generation, from 6.0 GJ ha1 in W1 to 11.9 GJ ha1 in W2 (i.e., the gross energy gained from one unit hectare of arable land almost doubled). Considering the indirect energy consumptions due to the compensation for the reduced grain and for the lost fertility value of soil, in total (Table 4), W2 would save almost 1.6 times more energy than W1.

GHG Life Cycle Analysis The same conceptual approach applied in the energy flow analysis also pertained to the GHG balance calculations. The RED methodology was followed and the BioGrace conversion coefficients were applied to assess the GHG emissions associated with the straw-to-energy systems in the study area. The overall GHG balance was performed (i.e., considering both debts and credits) along the whole energy Table 4 Breakdown of the energy flows with respect to the three considered cropping systems (W0, W1, and W2). Units are all expressed in MJ ha1 Energy fluxes Total energy for cultivation Straw chopping (W0) or baling (W1 and W2) Straw transportation Compensation for the straw fertilizing loss Compensation for reduced grain yield due to rotation Delta (with respect to W0) Net electricity generation Energy balance (straw to electricity)

W0 14,620.7 125.2

W1 14,620.7 229.9 68.0 383.7 556.4 6,033.3 5,476.9

W2 13,076.2 359.3 49.3 255.8 4,358.7 3,353.4 11,934.7 8,581.3

Note: W0 conventional tillage, single crop, straw retention; W1 conventional tillage, single crop, straw removal; W2 conservation tillage, crop rotation, straw removal

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

11

value chain (from the farm field to the electricity delivery). Some of these factors could be considered “hidden” features in the agricultural energy and carbon assessment, still difficult to detect and to account for in the ordinary LCA. On this respect, using the simulation model, it was possible to estimate the amount of N2O released from the soil as well as the changes in SOC due to different straw managements. Both direct and indirect emission factors were considered. Emissions due to straw logistics (i.e., baling and transportation) are an example of direct emission components, while changes in the soil nutrient content or the equivalent amount of emissions due to changes in wheat grain yield can be considered indirect factors. In the two considered straw-to-energy systems (W1 and W2), the emission credits consisted of the avoided emissions consequent to the displaced fossil fuels to deliver the same electricity production. The results of the final budget are reported in Table 5. The overall GHG savings on the basis of the land unit, for both W1 (373 kg CO2 eq. ha1) and W2 (1,086 kg CO2 eq. ha1), have shown favorable results as compared to W0. The same can be observed based on the energy unit delivered (223 kg CO2 eq. MWh1 in W1 and 760 kg CO2 eq. MWh1 in W2). Therefore, the shift from conventional single crop system (W1) to a no-tillage rotation system (W2) could result in about 2.9-fold larger GHG savings per unit hectare of arable land and 3.4-fold greater GHG abatements per unit MWh of power delivered. The effect of straw management and crop rotation on GHG fluxes is evident considering the change in the soil organic carbon content. Straw left on the soil (W0) increased the soil organic matter resulting in a carbon credit of 374 kg CO2 eq. ha1. A systematic straw removal under single crop cultivation (W1) created a carbon debt of 154 kg CO2 eq. ha1, while straw removal coupled with a partial

Table 5 Life cycle GHG emissions and savings with respect to the three considered cropping systems (W0, W1, and W2 are specified in Table 4) GHG fluxes Cultivation Straw chopping (W0) Straw baling (W1 and W2) Straw transportation Straw fertilizing loss Change in soil organic carbon (SOC) N2O emissions due to wheat cultivation Reduced grain yield Total GHG emissions Delta (with respect to W0) Fossil fuel displacement Total balance

W0 W1 kg CO2 eq. ha−1 1,378.7 1,378.7 10.1 22.4 5.9 39.7 −374.0 154.0

W2 1,184.5 35.0 2.2 26.4 −146.7

451.9

263.6

409.3

1,466.7

1,864.3 397.6 −770.9 −373.4

394.8 1,905.6 438.9 −1,525.0 −1,086.1

W0 W1 W2 kg CO2 eq. MWh−1 822.6 822.6 357.3 6.1 13.4 10.6 3.5 0.7 23.7 8.0 −223.2 91.9 −44.2 269.6

157.3

123.5

875.2

1,112.4 237.2 −460.0 −222.8

119.1 574.8 −300.4 −460.0 −760.4

12

M. Monteleone et al.

herbage removal (W2) restored the soil carbon credit conditions by the amount of 147 kg CO2 eq. ha1 proving to be effective in preserving soil organic matter. N2O gas released from the soil had a prominent effect on the GHG budget; the amount of the equivalent CO2 losses is substantially in line with what is reported by other authors (Gabrielle and Gagnaire 2008; Gan et al. 2012) regarding the impact of straw retention on the soil enhanced N2O emissions. Furthermore, the carbon debt attributed to W2 as a compensation for the reduced grain yield due to crop rotation was equal to about 395 kg CO2 eq. ha1.

Direct and Indirect Land Use Impacts As already stated, the reference land use considered in the case study was agricultural land where wheat straw is left on the field (W0). When residues were collected and used as energy feedstock, the main GHG implications in comparison with W0 were factors related to direct land use change (Cherubini and Ulgiati 2010) such as: • Changes in soil organic carbon. A decreasing SOC led to a loss in the carbon stocks, thus to increased GHG emissions. • Possible changes in crop grain yields due to the lower mineralization of N in soil (together with P and K). This would require a corresponding increase of synthetic fertilizer application to compensate for the nutrients removed with straw (Gabrielle and Gagnaire 2008). A higher fertilizer supply involves a correspondingly higher energy costs and GHG emissions. • Changes in N2O emissions from the cultivated soil. The amount of N2O released from the soil was related to the amount of N fertilizer applied. It is generally observed that increasing straw removal decreases N2O emissions. This could imply that the straw returned to soil would increase the soil nitrification/denitrification potential and its capacity to release N2O (Cai et al. 2001). Indeed, the stimulating effect of residue incorporation on N2O emission could be related to the accelerated soil microbial activity. Microbial nitrification increases under conditions of higher soil pH and temperature, together with a soil moisture not exceeding field capacity. These conditions are frequently found under Mediterranean types of climate. It is worth to note that straw incorporation into the soil had greater influence on N2O emissions than that resulting from the compensation of straw fertilizing losses (Cherubini and Ulgiati 2010). Furthermore, in the analyzed case study, other relevant factors altering the energy balance and GHG emissions were identified. These factors were mainly related to indirect land use changes. Comparing the no-tillage cropping system (W2) with conventional system (W1) revealed that in W2, the average available straw was about 2.41 Mg ha1 (two-thirds wheat and one-third triticale),while in the W1 it was 1.22 Mg ha1. Considering that straw availability per unit of land surface was almost two times higher in W2 compared to W1, in fact, the total

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

13

surface of the biomass supply basin was almost halved. The consequences of the decrease in the agricultural land surface were: • The same amount of electrical energy was obtained from half of the surface area of land. • The intensive straw availability reduced the energy cost and GHG emissions due to logistics operation and transportation by around 40 % on land unit basis and 20 % on energy unit basis. • Since the agricultural surface area dedicated to wheat grain production was reduced by 33 %, it follows that the overall grain productivity would be reduced. However, it should be noted that the total energy consumptions and GHG emissions related to the cultivation of crop would also reduced by approximately the same magnitude. Therefore, the reduced grain production assigned to W2 could be assumed to be compensated, and the missing grain production could be produced somewhere else. Indeed, in the province of Foggia, extra agricultural land should be easily available due to a large buffer of “set-aside,” marginal, and “fallow” land.

Conclusions Considering the environmental assessment of a bioenergy chain in the agricultural area of Southern Italy (province of Foggia), an optimized wheat cropping system, specifically tailored for both grain production and straw utilization, could create a positive energy budget and a significant reduction in GHG emissions without altering the farming ecological conditions and the carbon stock in the soil. There was clear evidence that for the area studied the use of straw for energy conversion was a worthy environmental choice. Several environmental improvements could be emphasized in adopting a no-tillage cultivation system coupled with a wheat sequence in rotation with a herbage crop (triticale) one year out of three (W2). The positive effects showed by W2 as compared to a single wheat cultivation system with conventional soil tillage (W1) can be summarized as follows: • A reduced agricultural surface allocated to straw collection due to the higher biomass density in W2; this reduction was about 50 % of the original surface area in W1. • A reduced straw transportation impact in consequence of the higher availability of straw, as mentioned above. • The electrical generation was obtained at reduced GHG emissions and improved energy savings. • The long-term simulation model showed a quite stable maintenance of SOC and reduced CO2 emissions due to the organic matter mineralization. Over a 50-year period, SOC would not decrease beyond the initial carbon contents in the alternative W2 cropping system.

14

M. Monteleone et al.

• N2O released from the soil was significantly reduced in W2 as compared to the conventional reference system where straw was left on the soil (W0). This was not observed in W1 because N2O emissions were significantly correlated to the total amount of residues left on the soil. Overall, it appeared that energy from cereal residues can be optimally coupled with grain production, not only because adverse impacts do not descend but also because substantial improvements in the environmental conditions can be favorably determined. Apart from the clear environmental advantages, coupling both the production of grain and straw in wheat cultivation is a very positive integration to farm income. The sale of straw can account for about 15–20 % the value of grain on the market (250–300 € t1), thus significantly contributing to increase the economic viability of wheat cultivation. This condition is particularly relevant considering that specific subsidies on wheat have been removed in compliance with the new CAP measures (the European Common Agricultural Policy). This additional income, therefore, represents a form of integration essential for inland farms in marginal rural areas where no alternative crop than low-input wheat is achievable. According to the LCA methodology applied, it was evident that in order to obtain the best environmental performance, the use of straw for energy generation in parallel with the optimization of the cropping system is a key factor in long-term sustainability. It was also demonstrated that a wheat cultivation system based on no-tillage and a rotation with a herbage crop (triticale, one year out of three) provided evidence of a win-win strategy: preserve soil quality and obtain renewable energy. Acknowledgments This study is part of the research carried out with the EU project “Strategic & Technological Advancement in Research on AgroEnergy” (STAR*AgroEnergy), which promotes and integrates approaches to renewable energy generation and knowledge-based bioeconomy according to sustainability criteria (http://www.star-agroenergy.eu/). The project is funded by the European Commission, Seventh Framework Programme (FP7), REGPOT 2011–14, Grant Agreement N
286269.

References Alluvione F, Moretti B, Sacco D et al (2011) EUE (Energy Use Efficiency) of cropping systems for sustainable agriculture. Energy 36:4468–4481 Anderson S (2011) Cropping systems, effects on soil physical properties. In: Glinski J, Horabik J, Lipiec J (eds) Earth sciences series. Encyclopedia of Agrophysics: Springer Reference (www. springerreference.com). Springer, Berlin/Heidelberg, doi:10.1007/SpringerReference_330183 2012-05-23 11:44:32 UTC Bakht J, Shafi M, Jan MT, Shah Z (2009) Influence of crop residue management, cropping system and N fertilizer on soil N and C dynamics and sustainable wheat (Triticum aestivum L.) production. Soil Tillage Res 104:233–240 Blanco-Canqui H, Lal R (2009) Crop residue removal impacts on soil productivity and environmental quality. Crit Rev Plant Sci 28:139–163

Agronomic Management of Straw and Its Energy Use in a Long-Term. . .

15

Brander M, Tipper R, Hutchison C et al (2009) Consequential and attributional approaches to LCA: a guide to policy makers with specific reference to greenhouse gas LCA of biofuels. Ecometrica Press. Technical Paper TP‐090403‐A Cai Z, Laughlin RJ, Stevens RJ (2001) Nitrous oxide and nitrogen emissions from soil under different water regimes and straw amendment. Chemosphere 42:113–121 Cherubini F, Ulgiati S (2010) Crop residues as raw materials for biorefinery systems- A LCA case study. Appl Energy 87:47–57 Curran MA, Mann M, Norris G (2005) The international workshop on electricity data for life cycle inventories. J Cleaner Prod 13(8):853–8632 Del Gado JA (2010) Crop residue is a key for sustaining maximum food production and for conservation of our biosphere. J Soil Water Conserv 65(5):111–116 Ekvall T, Weidema BP (2004) System boundaries and input data in consequential life cycle inventory analysis. Int J Life Cycle Assess 9(3):161–171 Faaij APC, Domac J (2006) Emerging international bio-energy markets and opportunities for socio-economic development. Energy Sustain Dev 10(1):7–19 Farine DR, O’Connell DA, Grant T, Poole ML (2010) Opportunities for energy efficiency and biofuel production in Australian wheat farming systems. Biofuels 1(4):547–561 Finnveden G, Hauschild MZ, Ekvall T et al (2009) Recent developments in life cycle assessment. J Environ Manage 91(1):1–21 Gabrielle B, Gagnaire N (2008) Life-cycle assessment of straw use in bio-ethanol production: a case study based on biophysical modeling. Biomass Bioenergy 32:431–441 Gan Y, Liang C, Campbell CA et al (2012) Carbon footprint of spring wheat in response to fallow frequency and soil carbon changes over 25 years on the semiarid Canadian prairie. Eur J Agron 43:175–184 Garofalo P, Di Paolo E, Rinaldi M (2009) Durum wheat (Triticum durum Desf.) in rotation with faba bean (Vicia faba var. minor L.): long-term simulation case study. Crop Pasture Sci 60:240–250 Johnson LMF, Karlen DL, Andrews SS (2010) Conservation considerations for sustainable bioenergy feedstock production: if, what, where, and how much? J Soil Water Conserv 65 (4):88–91 Kludze H, Bill Deen B, Weersink A et al (2013) Estimating sustainable crop residue removal rates and costs based on soil organic matter dynamics and rotational complexity. Biomass Bioenergy 56:607–618 Lafond GP, Stumborg M, Lemke R, May WE, Holzapfel CB, Campbell CA (2009) Quantifying straw removal through baling and measuring the long-term impact on soil quality and wheat production. Agron J 101:529–537 Lal R (1997) Residue management, conservation tillage and soil restoration for mitigating greenhouse effect by CO2 enrichment. Soil Tillage Res 43:81–107 Limon-Ortega A, Govaerts B, Sayre KD (2008) Straw management, crop rotation, and nitrogen source effect on wheat grain yield and nitrogen use efficiency. Eur J Agron 29:21–28 Monteleone M, Garofalo P, Cammerino ARB, Delivand MK (2015) Straw-to-soil or straw-toenergy? An optimal trade-off in a long term sustainability perspective. Appl Energy 154:891–899 Powlson DS, Riche AB, Coleman K, Glendining MJ et al (2008) Carbon sequestration in European soils through straw incorporation: limitations and alternatives. Waste Manag 28:741–746 Sánchez ST, Woods J, Akhurst M et al (2012) Accounting for indirect land-use change in the life cycle assessment of biofuel supply chains. J R Soc Interface 9(71):1105–1119. doi:10.1098/ rsif.2011.0769, Epub 2012 Mar 30 Shrestha A, Lanini T, Wright S, Vargas R, Mitchell J (2006) Conservation tillage and weed management. University of California. Division of Agriculture and Natural Resources. Publication, Oakland, California 8200 Signor D, Cerri CEP, Conant R (2013) N2O emissions due to nitrogen fertilizer applications in two regions of sugarcane cultivation in Brazil. Environ Res Lett 8:9

16

M. Monteleone et al.

Sto¨ckle CO, Donatelli M, Roger Nelson R (2003) CropSyst, a cropping systems simulation model. Eur J Agron 18:289–307 Sto¨ckle CO, Kemanian A, Nelson R (2009) CropSyst Version 4.15. Biological systems engineering department. Washington State University, Pullman Whittaker C, McManus M (2012) The renewable energy directive and cereal residues. In: Proceeding of the 20th European biomass conference and exhibition, Milan, pp 2049–2057 Whittaker C, McManus MC, Hammond GP (2011) Greenhouse gas reporting for biofuels: a comparison between the RED, RTFO and PAS2050 methodologies. Energy Policy 39:5950–5960

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Quality Management of Forest Biomass Supply for Energy and Biorefining Tanja Ikonena, Miina Jahkonenb, Karri Pasanena* and Lauri Sikanena a Bio-based Business and Industry, Natural Resources Institute Finland (Luke), Joensuu, Finland b Joensuu, Finland

Abstract Forest bioenergy value chains must be economically viable in order to create profitable and sustainable feedstock supplies for growing biomass markets. The competitiveness of renewable, biomass-based energy production strongly depends on the operational costs, policy environment, and quality of the feedstock supply. The objective of this chapter is to address the key factors affecting feedstock supply quality from end users’ points of view in heat and power plants and in state-of-the-art biorefining processes. Also the importance of international and national policies concerning sustainability and profitability of the use of forest biomass for energy production is discussed.

Keywords Forest; Biomass; Bioenergy; Quality; Moisture; Impurities; Particle size; Profitability

Introduction In addition to traditional forest industries as well as to heat and power generation, forest biomass has recently become a raw material for several high-technology and added-value products such as biodiesel and bioethanol. At the same time quality requirements concerning forest biomass used in these addedvalue products and processes have tightened up. Companies operating in the forest biomass supply chains have had to adapt both to the changes in the business environment and the changed requirements concerning the feedstock quality, traceability, and delivery schedules. In the future, the management of supply chain costs in all phases of logistic system will hold the key position (Kärhä 2011; Ikonen et al. 2013a). Cost competitiveness of forest bioenergy is based on the competitive edge of the companies operating in the forest bioenergy supply chains. Since energy produced from forest biomass must also be as cheap as or cheaper than energy produced from competing fuels, more attention must be paid to procurement solutions and quality of the feedstock. In the tightening competition for resources and in the more specialized production, quality becomes a success factor, which cannot be ignored. Feedstock management for energy generation and biorefining differs from traditional wood supply to sawmills and pulp mills.

*Email: karri.pasanen@luke.fi Page 1 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Wood-Based Bioenergy Production and Biorefining Wood as a Fuel Compared to the fossil fuels, forest biomass has many advantages from an environmental perspective, such as renewability and low CO2, SOx, and NOx emissions (Khan et al. 2009). Forest biomass-based energy is considered to be carbon neutral, and thus, it has a key role in greenhouse gas reduction (Eriksson et al. 2002). Forest-based bioenergy production also increases employment in rural and remote areas and enables the use of local energy resources. Woody biomass consists mainly of three elements: carbon, oxygen, and hydrogen (e.g., Hakkila 1989; Alakangas 2000; Khan et al. 2009), of which carbon and oxygen are combustible (Hakkila 1989). The chemical composition and the nutrient content of wood have a remarkable effect on the quality of the fuel and energy production process (Alakangas 2000). Fresh, unseasoned wood consists of water and dry mass. About 50 % of the fresh mass of wood is water (e.g., Thörnqvist 1985; Hakkila 1989; Jirjis 1995; Kärkkäinen 2007; Jahkonen et al. 2012b) (Fig. 1). The moisture content varies depending on the season, tree species, and tree components. It is typical for the seasonal variation of a tree that the moisture is higher during the active period (summer) and lower during dormancy (winter). The moisture depends also on the density of wood; denser parts of the tree tend to have lower moisture content (Kärkkäinen 2007). The main organic compounds of wood are carbohydrates, mainly cellulose, hemicellulose, and lignin, of which lignin has the highest heating value (Khan et al. 2009). Cellulose and hemicellulose contain mostly oxygen and less carbon and hydrogen (Hakkila 1989). In addition, wood contains other inorganic compounds and minerals such as nitrogen, sulfur, chlorine, and alkali. The heating value means the energy, which is produced by combusting a mass unit of wood (Kärkkäinen 2007). The energy produced by combustion of wood is indicated as calorimetric heating value, aka gross calorific value (MJ/kg or kWh/kg). Part of the energy in combustion is used to evaporate the water in the wood. The net calorific value, aka the lower calorific value, is the energy which indicates the amount of thermal energy after the water from the wood is evaporated. The net calorific value is then lower than the gross calorific value, and the difference between them increases when the moisture content of wood rises. Woody biomass has a lower heating value than, e.g., coal or oil (Alakangas 2000;

Fig. 1 Moisture content of an unseasoned tree is about 50 %. The other half is dry mass, which consists mainly of carbon, oxygen, and hydrogen Page 2 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Table 1 Forest chips have a lower heating value and price (* VAT 0 %) than fossil fuels (Statistics Finland 2014) Energy source Hard coal Natural gas (liquefied) Forest chips Milled peat

Default net calorific value (as fired) GJ/t MWh/t 25 6.9 49.3 13.7 9.8 2.7 10.1 2.8

Price* $/MWh 38.17 63.03 28.69 25.45

Table 2 Lower heating value with different tree species (Nurmi 1993) Tree species Pine (Pinus sylvestris) Spruce (Picea abies) European white birch (Betula pendula) Downy birch (Betula pubescens) European aspen (Populus tremula) Speckled alder (Alnus incana)

Lower heating value, dry (MJ/kg) Delimbed energy wood harvesting 19.33 19.02 19.15 19.19 18.65 19.00

Whole-tree harvesting 19.53 19.29 19.21 19.30 18.65 19.18

Khan et al. 2009, Table 1), partly because woody biomass contains usually some water. The water content of unseasoned wood is approximately 360 kg/m3 and that of seasoned wood approximately 100–170 kg/m3 (Hakkila and Fredriksson 1996). In addition to lower heating value, woody biomass is also cheaper than fossil fuels (Table 1). The characteristics and the structure of different tree species and tree components vary remarkably, and thus, the heating value and combustion of different components have some special features (Table 2). For example, conifers contain plenty of resins and lignin, and their heating value per mass unit is higher than the heating value of deciduous trees. For the same reason, the heating value of logging residues is slightly higher than the heating value of stem wood (Hakkila 1989). Wood contains also some other extractives, but the amount of them is relatively small compared with the amount of extractives of bark and foliage (Kärkkäinen 2007).

Forest Biomass Supply Chains Carefully planned supply chains are crucial for the utilization of renewable forest biomass. Forest biomass supply from forests is a logistically challenging process. Harvesting sites are often sparsely located, and logistics needs to be planned and optimized dynamically for all the time changing layout of harvesting sites. The selection of raw material and technology affects a lot for the characteristics of the biomass ending to energy generation or biorefining processes. All supply chains include the basic steps: purchasing, harvesting, and transportation. In addition, management and storing are needed to some extent for all supply chains (e.g., Laitila 2008; Kärhä 2011). All steps of the supply chain can be done by the same company or, on the contrary, all steps can be done by independent subcontractors. Stumps and logging residues are “coming on market” especially in northern Europe as a side product of clear cuttings or late thinnings. That feature can be utilized in supply chain planning. The challenge is that residues and stumps need storing in order to decrease the moisture content, whereas pulpwood and logs are transported away as fresh as possible. Freshness is a positive quality factor for sawlogs and pulpwood but a negative quality factor for solid woodfuels.

Page 3 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

In Fig. 2 alternative options of supply are described for four different energy wood assortments. The unique step of energy wood supply is the baling of logging residues. It can be used when transportation distances are long, truckloads tend to remain too loose, and energy content is too small per load. One must remember that baling requires an extra machine (baler) and is a relatively expensive operation. The use of baler can be very effective and profitable but it must be pre-evaluated by careful calculations. The location of chipping (comminution) is a fundamental decision in the supply chain. Chipping at roadside (landing) is very flexible and relatively effective, but the chipper needs to be relocated regularly according to roadside storages. Chipping at the plant or at the terminal is very effective, but it requires a large constant stream of material to be chipped to reach high efficiency. Quite often transportation of unchipped material is also relatively inefficient and is favoring roadside chipping. The supply chain forwarding material with ordinary forwarders, chipping at roadside, and transportation of chips by chip trucks is often called as “an idiot-proof all-purpose supply chain.” One can hardly fail totally with it and more efficiency can be later achieved by experience-based reengineering. Roadside chipping for stems is illustrated in Fig. 3.

Bioenergy Production and Biorefining Processes Wood-based bioenergy can be heat, electricity, or gas and oil products based on the various production technologies. Heat generation is the most simple; wood can be burned for heat in many different installations from open campfires and fireplaces up to fully automated boilers attached to district heating networks. Electricity can be generated, for example, by heating water by wood until it is steam. The steam can then be used to run the steam turbines which generate electricity. In the wood gasification process, the wood is gasified by reacting the material at high temperatures (>700  C), without combustion. The produced gas can be burned in a piston engine or a gas turbine. However, about 60 % of the generated energy is heat, and therefore, electricity generation alone is not very efficient. Combined heat and power (CHP) generation means the system where the generated heat is utilized in the heating systems through the district heating network, usually the heat, is sold for households or used in industrial processes. The electricity is produced in the balance with heat to maximize the efficient biomass utilization. Biorefining processes can vary a lot. Bio-oil and bioethanol production needs close to similar feedstock than energy generation, while some other processes can have remarkably different requirements for the raw materials. In this chapter, quality is monitored from pyrolysis oil and bioethanol production points of view. As a whole, the variety of value chains in biorefining is already versatile (Fig. 4). The pyrolysis process requires very dry raw material, preferably close to 10 % of moisture. From the transportation point of view, effective natural drying before transportation is a good practice. On the other hand, volatile compounds are very important for pyrolysis oil production. Currently there is no knowledge on how much volatiles can be lost if logs are dried, for example, at roadside storages. The assumption is that gentle natural drying of roundwood is a better option than aggressive artificial drying by kilns and tumblers. Rotten material and impurities are not welcome for pyrolysis production. Fresh material has more volatiles and pure material supports the grinding, which is necessary to achieve a small enough particle size especially for flash pyrolysis. Wood-based ethanol can be produced by two alternative technologies (Suokko 2010). The traditional way is to transform cellulose and hemicellulose to fermentation-ready sugars by acidic hydrolysis. After that, fermentation and distillation are well-known ethanol production methods. The alternative and more modern way is to gasify wood thermally to CO and methane. These gases can be fermented to ethanol by

Page 4 of 17

Fig. 2 Alternative steps of forest energy supply chains (Laitila 2006)

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Page 5 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 3 Example of bioenergy supply chain where chipping is at the roadside

Fig. 4 New and traditional value chains of forest biomass use

using microbes. This method is more effective because a greater share of the energy content of the biomass can be transferred to ethanol. The method is successfully demonstrated in semi-large scale but not yet used industrially. From a quality management point of view, gasifying needs dry wood with all possible volatile compounds present. Traditional hydrolysis and fermentation need cellulose more than anything else.

Page 6 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Quality in Forest Biomass Supply Chains Quality Management For small- and medium-scale forest bioenergy companies (SMEs), it is difficult to achieve the same economies of scale as fossil fuels or large companies operating in the forest bioenergy supply chains. Therefore, improving the company’s cost efficiency by focusing on intangible factors, such as quality or optimization of the deliveries, has a significant value. Intangible assets have an effect on competitiveness especially when a company cannot compete with tangible assets (e.g., machinery, raw materials) or production prices. Customers and competition in the market have turned quality into a strategic factor, through which it is possible to gain competitive advantage in the market (Savolainen 2010). Companies that can guarantee even quality of the produced forest biomass, reliable deliveries, and moderate price development in the medium or long term are the most competitive in the supply chains (Ikonen et al. 2013b). Especially small- and microscale companies operating in the forest bioenergy supply chains can reduce their operational risks by improving the quality of the forest biomass and deliveries and gain competitive edge compared to the other companies in the bioenergy sector. Nowadays, there is a wide agreement that nontechnical factors such as know-how and quality are crucial also for successful bioenergy projects and companies (e.g., Leskinen et al. 2010). The quality of the forest biomass produced must fulfill the terms set by the customer (Lecklin 2006). When the customer is satisfied with the product or the service of the company, operations are of good quality. From the end user’s point of view, the quality of the forest biomass can be divided into quality of the fuel and reliability of the fuel deliveries. Lillrank (1998) divides the elements of quality into technical quality and relative quality. Technical quality refers to the quality variables in the production process, which can be influenced and measured (Lillrank 1998). Forest biomass’ technical quality can be defined as the suitability of a certain type of wood for a specific purpose. Technical quality factors measured in the forest biomass supply chain are, for example, particle size distribution and moisture content of the delivered forest biomass. Relative quality refers to the value perceived by the end customer and it is achieved through technical quality. Relative quality variables include, for example, produced forest biomass’ suitability for the plant. Forest biomass must meet the process requirements, and therefore, consistency of the biomass is often more important than achieving the highest possible quality. Large variations within the fuel may disturb the balance of the production process. The main phases in the biomass supply chain are as follows: purchasing of the energy wood, harvesting and forwarding to the roadside, roadside storage, comminution, long-distance transport to the plant, and receiving and combustion of the forest biomass at the plant (Fig. 5). When the quality experienced by the end customer is based on the quality of several companies operating in the supply chain, the significance of the quality management and quality know-how of a single company and its employees for the whole supply chain’s quality is emphasized (Leskinen et al. 2010; Blumer et al. 2013). Understanding the significance of quality and finding the quality management techniques and best practices, through which to improve quality in the production and supply, are critical. For example, through supply chain optimization, which consists of the right harvesting method and storage time, competitiveness of the whole feedstock supply chain and the quality of the forest biomass can be improved (Röser 2012). There are no guidelines on how to execute quality management, but different quality management systems and techniques can be applied as part of the company’s management system. Quality management makes also quality criteria and requirements transparent for every party involved in the supply chains (Loibneggar 2011). Implementing quality system is not mandatory, but it is a useful tool in practice. The best known quality management systems are based on ISO 9000 standard, but since they Page 7 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Fig. 5 Energy wood supply chain consists of many phases, and quality at each level affects the quality of the product

are often too heavy for small companies, quality management can be executed also without standardization (Vihtonen 2004). Standards are used to improve both the quality of solid biofuels produced and the operating environment of the forest biomass supply chain. By applying standards, it is also possible to describe a fuel more accurately, which is useful for both producers and consumers of solid biofuels. Standardization of solid biofuels is subdivided into parts including standards on fuel specification and classification of solid biofuels based on the origin and source (EN 14588) and standards determining the fuel quality (EN 14961) (Loibneggar 2011). In 2014, EN 14916 standards will be replaced with the international ISO standards (17225 series) for solid biofuels.

Quality Factors Moisture The costs and quality of the forest biomass produced and supplied vary depending on the local conditions such as supply chain structure, efficiency, and demand variation. In boreal regions and areas where the demand is primarily for heat, the demand strongly correlates with seasonal temperature changes, and winter demand can be seven times larger than demand in the summer. In the winter, the demand for highquality forest biomass is also highest. In temperate zones demand variations are smaller than in boreal regions (Ikonen et al. 2013a). The harvesting conditions, roadside landing capacity, transportation distance, and type of forest biomass harvested will influence on the quality and thus the costs of the forest fuel produced (Kärhä 2011). According to an interview study of forest bioenergy operators in Eastern Finland (Jahkonen and Ikonen 2014), work phases at the forest-end of the supply chain are affecting the most on the quality of the forest biomass. The selection of the right harvesting technology and working methods is the basis for highquality and cost-efficient feedstock supply. For instance, a wrong decision about machinery or careless harvesting may cause serious tree and soil damages causing economic effects that might overcome the revenues gained from energy wood sales. Poor quality may also affect the forest owner’s willingness to sell energy wood in the future. Moisture is one of the most significant factors affecting the quality of energy wood at every phase of the forest energy supply chain. Quality requirements that determine the optimal moisture content of the forest biomass depend strongly on the size and the type of the plant and technology used in the combustion process. Small-scale plants usually require more homogenous and drier fuel, whereas larger plants are able to combust even fresh energy wood. The forest biomass used at the small-scale plants should contain less than 40 % moisture (Erkkilä et al. 2011). Considering the use of energy wood as a whole, a substantial improvement in the profitability of the energy wood business can be achieved by delivering and combusting the energy wood consignments at the right time and at the right moisture content Page 8 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015 25

2,5 2,25

Net calorific value, MJ/kg

1,75 1,5

15

1,25 1

10

0,75 5

Energy density, MW/m3

2

20

0,5 0,25 0

0 0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

Moisture content %

Fig. 6 Net calorific value (MJ/kg) and energy density (MWh/m3) of energy wood (MC from 0 % to 70 %). The range used in energy density calculations was 350–450 kg/m3 (Jahkonen et al. 2012b)

(Jahkonen et al. 2012b). Hence, the knowledge and prediction of the moisture content of the forest biomass is a prerequisite for delivering energy wood at the right time. In road transportation, the total weight of the load may exceed the maximum allowed net payload because of the too high moisture content of the forest biomass, or as an opposite, the truckload space cannot be optimally utilized due to the weight limitation. Reduction of moisture content increases chip trucks’ competitiveness (Andersson et al. 2002; Tahvanainen and Anttila 2011). In a case study done in Eastern Finland, lowering the moisture of chips from 45 % to 35 % before truck transportation, total transportation costs could be reduced by 0.8 €/m3 for 70 km transportation (Tahvanainen and Anttila 2011). If customers at the plant are paying by energy content of the fuel, the quality of the forest biomass and moisture content are the key factors for the profitability of the companies supplying the chips (Leskinen and Maier 2010). In the case of high moisture content and poor quality, the entrepreneur will incur financial losses. Based on the conditions of the delivery contract, the buyer can also discard poor quality chips and oblige the entrepreneur to take back the delivered chip lot. In energy production, the more water there is in the biomass, the more energy is needed to evaporate the water and the less energy is gained from the biomass. The net calorific value of the forest biomass decreases linearly as the moisture content of the forest biomass increases, and to gain the same amount of energy than with drier biomass, more fuel has to be used (Fig. 6). High moisture content causes problems also in other biorefining processes using forest biomass, such as production of pyrolysis oil, biodiesel, and other state-of-the-art products. For these processes, moisture content of the forest biomass is an even more critical quality factor than it is in heat and combined heat and power (CHP) production. However, low moisture content in the forest biomass also increases the risk for overheating and disturbances in the process. The limit for too low moisture content depends on the boiler type, fuel mix, and process control setup. During winter, icing and blocking due to the high moisture content complicate conveying, mixing, and handling of the fuel by forming ice bridges and lumps in storage bins and conveyors. Process disturbances and service breaks raise the costs in the energy production process. In addition, the combustion of wet

Page 9 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

wood material increases carbon monoxide, carbon dioxide, and fine particle emissions (Hakonen and Laurila 2011). Since freshly harvested wood is usually too wet for energy use, it must be seasoned for periods from weeks to a few years before comminution and transportation. The moisture content of unseasoned logging residues is 44–60 % (e.g., Kärkkäinen 1976; Verkasalo 1987; Hakkila et al. 1995; Nurmi 1999a; Jahkonen et al. 2012b). The moisture content of pine stumps is on average 53 % and that of spruce stumps on average 51 % (Laurila and Lauhanen 2010; Erkkilä et al. 2011). Storing of energy wood either on the site or in roadside storage is the cheapest and the most commonly used method to improve the quality of energy. Woody biomass is usually first seasoned in small piles at the logging site and after that forwarded on the roadside into bigger storage piles. The aim of the storage is to improve fuel quality and enhance the deliveries during the highest fuel demand in the winter (Nurmi 1999b). Forest biomass can be stored at the roadside, terminal, or plant (Nurmi 1999b). Companies can use stored biomass to even out seasonal employment, balance raw material supply, and reduce idle time. The storage site should be carefully selected in order to ensure accessibility of chippers and trucks and to provide the best possible drying conditions. A good roadside landing is in a sunny, windy place, which is easily accessible, and the base of the pile is clean from impurities. In addition, the drying depends on the location and weather conditions of the storage site and the size and form of the storage. Under favorable conditions, energy wood’s moisture content can be reduced to 20–30 % during the storage in field conditions (e.g., Jirjis 1995; Nurmi 1999a; Nurmi and Lehtimäki 2011; Jahkonen et al. 2012b). However, due to varying circumstances during the drying process, there are many challenges in maintaining the quality of the biomass during storing. Especially in large buffer storage, the uneven moisture distribution and uneven drying within the storage pile may reduce the uniformity and the overall quality of fuel chips. Degradation and dry matter losses of biomass may also occur. Degradation is due to colonization by fungi and mold. These microorganisms, via metabolic activity, generate heat, which in turn accelerates oxidation, moisture adsorption, hydrolysis, pyrolysis, and other chemical processes resulting in dry matter loss, which can cause considerable losses of raw material (Andersson et al. 2002; Stupak et al. 2008; Asikainen et al. 2002). This may have an effect on profitability and profits gained from forest biomass sales and use. Fresh logging residue chips stored in a large pile for 7 months lost approximately 12 % dry matter, mainly during the first few weeks (Thörnqvist and Jirjis 1990). Due to the climatic conditions, the low moisture content achieved during the summer is challenging to keep during autumn and winter, and in many cases the storage pile remoistens in many cases due to the intercepting rain and snow (Jahkonen et al. 2012b). By covering the pile with cover material such as cover paper, the moisture content in the pile and the amount of intercepting rain and snow can be reduced. Covering is more effective for logging residues than for stem wood because the surface area per volume in logging residue is higher. Covering lowers the moisture content of the stem wood by 3–6 % (Nurmi and Hillebrand 2007) and of logging residues as much as 10–15 % (Hillebrand and Nurmi 2001). Achieving higher quality usually increases total costs throughout the whole forest bioenergy supply chain. Increased working time in supervision, for example, may incur higher costs. The use of terminal systems increases feedstock quality by controlling the moisture content and impurity levels, but these are not very costcompetitive quality improvement methods, because terminal storage systems have high capital and maintenance costs. Particle Size The particle size of forest biomass affects mainly the steadiness of combustion processes, but in addition it influences handling and conveying of the fuel. Smaller boilers and automatic feeding systems require usually smaller chips than large power plants (Kofman 2013a). An optimal length of a wood chip is usually 30–40 mm. Under 3,15 mm particles are classified as fine particles. Uniform particle size is considered to be optimal from the end user’s point of view. Page 10 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

The bulk density of the wood chips depends on the particle size and shape, moisture content, comminution, and loading techniques and settling. Solid volume factor indicates how many solid cubic meters one bulk cubic meter contains (Alakangas 2000; Hakkila 2000). Possibilities that affect quality through particle size are mainly technical and methodological, e.g., selection and maintenance of comminution device (grinder/chipper), type and maintenance of screens, cutting speed and angle, knife angle and sharpness, and selection of discharge system (Kofman 2013b). Comminution increases the surface area of the biomass, which provides a larger area for microbes that might decrease the quality of the wood biomass (Jirjis 2005). Therefore, the timing of comminution has to be selected carefully in order to maintain the quality of the fuel. Selection of comminution technique and equipment affects the quality of produced fuel. Practical operators in forest bioenergy supply chain in North Karelia saw that forest chips comminuted with a chipper are better and more uniform than chips comminuted with a grinder (Jahkonen and Ikonen 2014). The particle size of a wood chip depends on the tree species, size of base material, moisture content, operator experience and training, and speed and angle of feeding (Kofman 2013b). Usually particle size distribution of chipped roundwood is more uniform than that of chipped logging residues (Alakangas 2000). However, particle size in one consignment is never uniform, and smaller chips tend to fill the spaces between larger chips during loading and transportation (Hakkila 2000; Garstang et al. 2002). Small particles are more common in chips produced from logging residues. According to Nati et al. (2010), chips made of logging residues contained also more oversized particles than chips produced from roundwood, and therefore, a standard-sized mesh screen is recommended when chipping logging residues. According to Garstang et al. (2002), harvesting and comminution should be optimized to provide uniform chips with minimal amount of fines in order to ease air flow in the boiler and in the storage pile and guarantee good quality fuel. At the end-use facility, long particles and branches may cause bridging and block the conveyors especially during winter if the fuel is wet. Fine particles block air movement both in a storage pile and in a boiler. Air movements in storage piles are important in order to promote drying of the inner parts of the pile and to prevent self-heating inside the pile (Garstang et al. 2002). Oversized particles also tend to burn with too high temperatures causing sintering of sand in the boiler. Impurities Purity is the definite prerequisite for using forest biomass in energy production. Impurities (e.g., soil, stones, metal) in the forest biomass increase the ash content of the produced fuel, lower calorific value in the fuel (Jahkonen et al. 2012a), and cause operating problems in the supply chain. Especially when using stumps impurities such as soil and stones can be a problem (Erkkilä et al. 2011). The average ash content of stumps after a long storage period is according to Laurila and Lauhanen (2010) 1.7 %, but because the ash content depends on the amount of impurities, it can increase up to 14 % (Alakangas 2000). At comminution, stones and other contaminants wear and even break the chipper blades. Blade wearing decreases the quality of the produced chips (Nati et al. 2010). In energy production, impurities may break chip conveyors or the structures of the boiler and block air intakes. In fluidized bed boilers, soil reduces the quality of the bed sand and increases the need for boiler maintenance and sand exchange. High ash content may also increase costs through the disposal of the processed ash. In addition, impurities in the raw material causes extra costs at the plant, because the soil extracted from the comminuted energy wood has to be transported from the plant site, e.g., to a landfill site. Impurities can end up into forest biomass in every phase of the supply chain. According to energy wood operators, the most critical phase affecting the forest biomass quality is forwarding, and it is important to pay special attention to all work phases from harvesting to roadside storage (Jahkonen and Ikonen 2014). If the pile is frozen to the ground, it is more likely that the driver will pick contaminants from the soil when Page 11 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

loading the wood. Impurities can end up in energy wood also during roadside storages while piles are exposed to snow, soil, and other debris from the traffic and snow plowing. In stump harvesting, splitting of the stumps facilitates drying but also improves the quality by reducing the soil and stones attached between roots. For stumps, the amount of soil attached depends on the form and shape of the stump, soil type, stump harvesting method, and duration and conditions of storage (Jahkonen et al. 2012a). In comminution, impurities can be reduced by pre-grinding. It speeds up the final comminution and facilitates the lack of space in the terminals. The contaminants from the soil increase the amount of silica and thus the risk of operational difficulties at the power plant (Saidur et al. 2011). Together with potassium phosphate silica may form silicates which can melt or soften in lower temperatures in a boiler causing also blocking of the air intake (McKendry 2002). In fluidized bed boilers chlorine, potassium, and silicon cause also agglomeration problems (Khan et al. 2009). Woody biomass contains also some sulfur, which reacts together with alkali forming alkali sulfates (Khan et al. 2009) and thus causing corrosion problems in the boiler. In addition, high percentages of chlorine have proved to be a major problem, especially when combusting logging residues. Even though chlorine is an essential substance in wood, it is an unwanted component in the combustion process especially in fluidized bed boilers and can therefore be considered as an impurity. Chlorine found especially in needles and leaves has many harmful effects causing corrosion of superheater tubes at high temperatures (Alakangas 2000; Khan et al. 2009). The amount of chlorine can be reduced by seasoning logging residues until the needles/leaves have lost their green color (Alakangas 2000). Volatile Organic Compounds Volatile organic compounds (VOCs) are important from an energy production point of view, because they make wood highly reactive and are beneficial for heating value in general (Saidur et al. 2011). VOCs of biomass can be subdivided into gases and tars (Khan et al. 2009). The content of volatiles in wood is 80–90 % (Kytö et al. 1983). As a result of easy and rapid volatilization, some VOCs are released from wood already during drying at field conditions and in mild temperatures. The volatile compounds emitted are primarily monoterpenes. The less volatile compounds include fatty acids, resin acids, diterpenes, and triterpenes (Granström 2005). The high VOC content of biofuels makes woody biomass easily flammable even at lower temperatures (Khan et al. 2009). The combustion process of woody biomass is expected to proceed rapidly due to the high amount of VOCs, which has to be taken into account in the combustion process. In combustion, VOCs generate large amounts of inorganic vapors (Jenkins et al. 1998), which can be used either directly or indirectly for energy production. VOCs can be used also as a raw material for special energy products, e.g., pyrolysis oil. In addition to these beneficial facts, VOCs and combustion of wood itself may also cause environmental damage, not only through greenhouse gas emissions but also through other toxic emissions. The most severe problem caused by the VOCs is the formation of ozone in the lower atmosphere. The amount of harmful substances formed as a result of incomplete combustion can be regulated through combustion and drying techniques and fuel moisture control (Jenkins et al. 1998). Biofiltration has been proved an efficient method to reduce VOC emissions (Leson and Winer 1991). It is necessary to use longer hightemperature zones in order to achieve a complete combustion that is efficient to reduce pollutant emissions (CO, PAH) (Khan et al. 2009). The combustion technique of small-scale heat production is incapable of providing sufficient high temperatures in early stages of combustion, when the volatiles are released, and thus, the formation of pollutants is difficult to avoid in small-scale heat production. Also small particle emissions are mainly caused by small wood-burning devices such as fireplaces and sauna stoves. Emissions from stoves and small furnaces can be controlled most effectively by fuel quality and the

Page 12 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

right burning technique. Large (over 0.5 MW) wood-heated boilers have better controlled burning conditions and devices are equipped with filters and scrubbers.

Origin of Wood as a Quality Factor Forest certification systems, such as FSC and PEFC, aim to guarantee that the management of commercial forests is economically, ecologically, and socially on a sustainable level and that wood originating in forests fulfills the requirements set for sustainability. In many countries having a strong forest sector and high share of private forest ownership, such as in Finland, more than 90 % of the forest area is under certification scheme. However, globally only 10 % of the forest area is under certification system. With the energy use of forest biomass, new requirements have been set for the sustainability and source verification of forest raw material. These new requirements are not being fulfilled in all respects by the forest certification systems, and the sustainability requirements cannot unambiguously be verified by them. The questions concerning land use and indirect land use change in particular have come up when assessing the sustainability of bioenergy production because many plants used in bioenergy production can also be used as food plants, and therefore, their area of cultivation can be used for food production. Another central issue relating to land use is carbon-rich land areas, such as forests, and their role as carbon sinks in climate change mitigation. The Directive 2009/28/EC of the European Union and the Council of Europe on promoting the use of energy from renewable sources, the so-called RED directive, sets national binding goals for renewable energy production and determines the sustainability criteria for liquid biofuels and bioliquids as well as the requirements concerning the verification of the criteria. The European Commission has also made a presentation on expanding the sustainability criteria to the use of solid and gaseous biomasses in energy production as well. Regarding forest biomass and forest chips produced from it, the sustainability criteria concern the greenhouse gas emission reduction of liquid biofuels and bioliquids, the biological diversity of the raw material’s line of production (i.e., forest), and change in land use and drying of peatlands. Verification, or achieving the sustainability criteria, has to be proven according to either a national system of a member state, a voluntary (certification) system approved by the European Commission, or a bi- or multilateral agreement between the EU and the Third World. For the present, for example, the PEFC or FSC forest certification system is not on the list of the voluntary systems approved by the EU. The origin of raw material and sustainability requirements can be seen as quality factors for forest biomass, because producers of liquid biofuels and bioliquids are required to verify those requirements through a sustainability system. Without a sustainability system and verification, the greenhouse gas emissions of renewable forest biomass will be considered to be equal with emissions of fossil fuels. As a consequence of the these requirements set for producers of end products, there will be more pressure also toward the beginning of supply chains to ensure the sustainable procurement of raw material. Recently, companies have introduced applications where the chain of custody can be traced by ICT solution made for the supply management of forest biomass. This sector is developing rapidly because of the pressure and demand from the liquid biofuel business.

Conclusions Forest biomass-based energy production is one of the most important sources of renewable energy especially in northern Europe. Due to the increasing production and limited availability of raw material as well as a demand for efficiency and economic viability, the quality aspects will play the key role in the successful realization of high targets set for renewable energy production. By increasing the quality both Page 13 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

in forest biomass and feedstock production and in supply chain management, the cost competitiveness of forest biomass can be increased compared to that of fossil fuels. The quality can be divided into quality of the feedstock and reliability of deliveries. The technical quality factors, such as moisture content and particle size, are measurable physical properties, which can be directly influenced by decisions made during the supply chain. In the end, these technical properties define the forest biomass’ suitability for certain combustion or biorefining processes. The relative quality, on the other hand, reflects the end user’s specific needs and requirements for feedstock and may differ from one end use to another, even with the same technical quality specifications of the forest biomass. Improved quality of the forest biomass increases the end users’ paying capability due to the higher value of the forest biomass and improved reliability in deliveries and lower maintenance costs. Also, forest owners are able to sell wood with better prices which increases the availability of forest biomass in the market. In addition to heat and electricity production, forest biomass is nowadays used also in biorefining processes producing high-value products such as pyrolysis oil and other biofuels and bioliquids. Biorefining of woody biomass sets new requirements also on the quality of raw material as the production processes are often more sensitive on variation of different quality factors. Quality is dependent on decisions made concerning raw materials and technologies, but first of all quality is dependent on motivation. The making of quality needs to be rewarding and clearly reasoned for all operators and stakeholders. In addition to the quality of products and supply chains, there is also a strong need to ensure the quality of renewable energy production from an environmental and climate change perspective. International climate and energy policy aims to lower the greenhouse gas emissions from energy production and secure the sustainability of raw material production. The origin of raw material and compliance with sustainability criteria have to be proven through sustainability systems. This will expand the quality management aspect from a technological and economical perspective toward the ecological aspect. The rural development and employment opportunities, on the other hand, give a social dimension on the supply chain quality management.

References Alakangas E (2000) Suomessa käytettävien polttoaineiden ominaisuuksia (Properties of fuels used in Finland). Technical Research Centre of Finland, VTT research notes 2045. 172 p. + 17 p Andersson G, Asikainen A, Björheden R, Hall P, Hudson J, Jirjis R, Mead D, Nurmi J, Weetman G (2002) Production of forest energy. In: Richardson J, Börheden R, Hakkila P, Lowe AT, Smith CT (eds) Bioenergy from sustainable forestry – guiding principles and practice. Kluwer, Dordrecht/Boston, pp 49–123 Asikainen A, Björheden R, Nousiainen I (2002) Costs of wood energy. In: Richardson J, Börheden R, Hakkila P, Lowe AT, Smith CT (eds) Bioenergy from sustainable forestry – guiding principles and practice. Kluwer, Dordrecht/Boston, pp 125–157 Blumer YB, Stauffachera M, Langb DJ, Hayashic K, Uchida S (2013) Non-technical success factors for bioenergy projects – learning from a multiple case study in Japan. Energy Policy 60:386–395 Erkkilä A, Hillebrand K, Raitila J, Virkkunen M, Heikkinen A, Tiihonen I, Kaipainen H (2011) Kokopuun ja mäntykantojen korjuuketjun sekä varastoinnin kehittäminen. Tutkimusraportti, VTT. 52 p Eriksson H, Hall J, Helynen S (2002) Rationale for forest energy production. In: Richardson J, Björheden R, Hakkila P, Lowe AT, Smith CT (eds) Bioenergy from sustainable forestry: guiding principles and practice. Kluwer Academic Publishers, Dordrecht, pp 1–17 Page 14 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Garstang J, Weekes A, Poulter R, Bartlett D (2002) Identification and characterisation of factors affecting losses on the large scale non-ventilated bulk storage of wood chips and development of best storage practices. First Renewables Ltd, London, UK. 89 p (+ appendices) Granström K (2005) Emissions of volatile organic compounds from wood. Dissertation. Karlstad University Studies 2005. 6 p Hakkila P (1989) Utilization of residual forest biomass, Springer series on wood science. Springer-Verlag, Berlin Heidelberg. 568 p Hakkila P (2000) Metsähakkeen energiatiheys (Energy density of wood chips). Puuenergia 1/2000. Puuenergia ry, Helsinki, pp 24–25 Hakkila P, Fredriksson T (1996) Metsämme bioenergian lähteenä. Metsäntutkimuslaitoksen tiedonantoja 613. 92 p Hakkila P, Kalaja H, Saranpää P (1995) Etelä-Suomen ensiharvennusmänniköt kuitu- ja energialähteenä. Metsäntutkimuslaitoksen tiedonantoja 582. 99 p Hakonen T, Laurila J (2011) Metsähakkeen kosteuden vaikutus polton ja kaukokuljetuksen kannattavuuteen. Seinäjoen ammattikorkeakoulun julkaisusarja B, Raportteja ja selvityksiä 55. 31 p Hillebrand K, Nurmi J (2001) Hakkuutähteen laadunhallinta. In: Alakangas E (ed) Puuenergian teknologiaohjelman vuosikirja 2001. Puuenergian teknologiaohjelman vuosiseminaari, Jyväskylä, 5.–6.9.2001. VTT Symposium 216:285–295 Ikonen T, Asikainen A, Prinz R, Stupak I, Smith T (2013a) Economic sustainability of biomass feedstock supply. IEA Bioenergy Task 43: 2013:01. Technical report. 59 p Ikonen T, Jahkonen M, Pasanen K, Tahvainen T (2013b) Laadunhallinta ja keskeiset laatutekijät metsäenergian toimitusketjuissa. Metlan työraportteja/working papers of the Finnish Forest Research Institute, 275. 41 s (in finnish) Jahkonen M, Ikonen T (2014) Toimijoiden näkemykset metsähakkeen toimitusketjun laadusta PohjoisKarjalan alueella. Metlan työraportteja/working papers of the Finnish Forest Research Institute, 280. 20 p Jahkonen M, Jouhiaho A, Lindblad J, Rieppo K, Mutikainen A (2012a) Kantoharalla ja kantoharvesterilla korjatun kantopuun lämpöarvo ja tuhkapitoisuus. Metlan työraportteja/working papers of the Finnish Forest Research Institute, 239. 20 p Jahkonen M, Lindblad J, Sirkiä S, Laurén A (2012b) Energiapuun kosteuden ennustaminen. Metlan työraportteja/working papers of the Finnish Forest Research Institute, 241. 35 p Jenkins BM, Baxter LL, Miles TR Jr, Miles TR (1998) Combustion properties of biomass. Fuel Process Technol 54(1–3):17–46 Jirjis R (1995) Storage and drying of wood fuel. Biomass Bioenergy 9:181–190 Jirjis R (2005) Effects of particle size and pile height on storage and fuel quality of comminuted Salix viminalis. Biomass Bioenergy 28:193–201 Khan AA, de Jong W, Jansen PJ, Spliethoff H (2009) Biomass combustion in fluidized bed boilers: potential problems and remedies. Fuel Process Technol 90:21–50 Kärhä K (2011) Industrial supply chains and production machinery of forest chips in Finland. Biomass Bioenergy 35(8):3404–3413 Kärkkäinen M (1976) Puun ja kuoren tiheys ja kosteus sekä kuoren osuus koivun, kuusen ja männyn oksissa. Silva Fennica 10(3):212–236 Kärkkäinen M (2007) Puun rakenne ja ominaisuudet. Metsäkustannus Oy, 468 p Kofman PD (2013a) Fuel characterization and standardisation. Course material, COST Training School – Forest Fuel Quality Assurance, Norwegian Forest & Landscape Institute, Campus Ås, Norway, 9th June – 15th June 2013

Page 15 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Kofman PD (2013b) Influencing factors on chip and hog fuel quality. Course material, COST Training School – Forest Fuel Quality Assurance, Norwegian Forest & Landscape Institute, Campus Ås, Norway, 9th June – 15th June 2013 Kytö M, Äijälä M, Panula E (1983) Metsäenergian käyttö ja jalostus. In: Alakangas E (ed) Properties of wood fuels used in Finland. Technical Research Centre of Finland, VTT Processes, Project report PRO2/P2030/05 (Project C5SU00800), Jyväskylä 2005, 90 p. + app. 10 p Laitila J (2006) Cost and sensitive analysis tools for forest energy procurement chains. Metsanduslikud Uurimused For Stud 45:5–10 Laitila J (2008) Harvesting technology and the cost of fuel chips from early thinnings. Silva Fennica 42(2):267–283 Laurila J, Lauhanen R (2010) Moisture content of Norway spruce stump wood at clear cutting areas and roadside storage sites. Silva Fennica 44(3):427–434 Lecklin O (2006) Laatu yrityksen menestystekijänä. Talentum, Helsinki Leson G, Winer MA (1991) Biofiltration: an innovative air pollution control technology for VOC emissions. J Air Waste Manage Assoc 41(8):1045–1054 Leskinen L, Lähtinen K, Tanskanen J, Sikanen L, Asikainen A (2010) Koneyrittäjät tunnistavat metsäenergiayrttämisessä liiketoiminnan kehittämismahdollisuuksia. Teoksessa: Rieppo, K. (toim.) 2010. Kasvun eväät metsä- ja puualan pienyrityksille. Työtehoseuran julkaisuja 406. s. 47–52 (in finnish) Leskinen L, Maier J (2010) Metsäenergian taustalla vahvat verkostot. Teoksessa: Rieppo, K. (toim.) 2010. Kasvun eväät metsä- ja puualan pienyrityksille. Työtehoseuran julkaisuja 406. s. 18 (in finnish) Lillrank P (1998) Laatuajattelu. 1. painos. Otava, Helsinki Loibneggar T (2011) Roadmap for implementing standards. Woodheat Solutions, 38 p. Available: https:// ec.europa.eu/energy/intelligent/projects/sites/iee-projects/files/projects/documents/whs_roadmap_for_ implementing_standards_en.pdf McKendry P (2002) Energy production from biomass (part 1): overview of biomass. Bioresour Technol 83:37–46 Nati C, Spinelli R, Fabbri P (2010) Wood chips size distribution in relation to blade wear and screen use. Biomass Bioenergy 34:583–587 Nurmi J, Hillebrand K (2007) The characteristics of whole-tree fuel stocks from silvicultural cleanings and thinnings. Biomass Bioenergy 31:381–392 Nurmi J, Lehtimäki J (2011) Debarking and drying of downy birch (Betula pubescens) and Scots pine (Pinus sylvestris) fuelwood in conjunction with multi-tree harvesting. Biomass Bioenergy 35:3376–3382 Nurmi J (1993) Pienikokoisten puiden maanpäällisen biomassan lämpöarvo. Helsinki. Acta Forestalia Fennica, 236. 30 s Nurmi J (1999a) Hakkuutähteen ominaisuuksia. Metsäntutkimuslaitoksen tiedonantoja, 722. 32 p Nurmi J (1999b) The storage of logging residue for fuel. Biomass Bioenergy 17(1):41–47 Röser D (2012) Operational efficiency of forest energy supply chains in different operational environments. Doctoral thesis, University of Eastern Finland, Dissertationes Forestales 146. 83 p Saidur R, Abdelaziz EA, Demirbas A, Hossain MS, Mekhilef S (2011) A review on biomass as a fuel for boilers. Renew Sustain Energy Rev 15:2262–2289 Savolainen T (2010) Managing competiveness. Lecture notes. University of Eastern Finland Business School. 16 p Statistics Finland (2014) Energy prices [e-publication]. ISSN=1799-800X. http://www.stat.fi/til/ehi/ index_en.html

Page 16 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_16-1 # Springer-Verlag Berlin Heidelberg 2015

Suokko A (2010) Lignoselluloosaetanolin ja synteesikaasusta fermentoitujen polttonesteiden teknologiatarkastelu. VTT Tiedotteita 2533. 88s. ISBN 978-951-38-7576-3 (nid.), ISBN 978-951-38-7577-0. http://www.vtt.fi/publications/index.jsp Stupak I, Asikainen A, Röser D, Pasanen K (2008) Review of recommendations for forest energy harvesting and wood ash recycling. In: Röser D, Asikainen A, Raulund-Rasmussen and Stupak I (eds) Sustainable use of forest biomass for energy, a synthesis with focus on the baltic and Nortic region. Managing Forest Ecosystems 12. pp 155–196. Springer, Dordrecht Tahvanainen T, Anttila P (2011) Supply chain cost analysis of long-distance transportation of energy wood in Finland. Biomass Bioenergy 35(8):3360–3375 Thörnqvist T (1985) Drying and storage of forest residues for energy production. Biomass 7:125–134 Thörnqvist T, Jirjis R (1990) Bränsleflisens förändring över tiden vid lagring i stora stackar (Changes in fuel chips during storage in large piles). Department of Forest Products, report no. 219. Swedish University of Agricultural Sciences, Uppsala Verkasalo E (1987) Metsähakkeen kosteuden ja kuivamassan mittaus kuormaotantamenetelmällä. Folia Forestalia 694. 35 p Vihtonen T (2004) Laatujärjestelmien taloudelliset vaikutukset ja toimivuus maatalous- ja elintarvikealojen pienissä ja keskisuurissa yrityksissä. MTT:n selvityksiä 58. Maa- ja elintarviketalouden tutkimuskeskus. Helsinki. 84 p. + liitteet

Page 17 of 17

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

Social Science Explanations for Host Community Responses to Wind Energy Stewart Fast* Department of Geography, Queen’s Institute for Energy and Environmental Policy, Queen’s University, Kingston, ON, Canada

Abstract A more complete understanding of the social reasons for host community opposition to wind energy is important for both practical and political reasons. Host community responses are varied and can range from hostility and active protest to relatively routine integration and even pride. The underlying reasons for these varied responses are complex and social science research has not coalesced around one generally agreed upon explanatory framework. However, many authors have put forward factors influencing social acceptance of wind energy and persistent themes can usefully be identified within this broad literature. This chapter reviews the approaches used to study the topic including survey research, economic modeling, political science, interviewing, discourse analysis, and q-method while highlighting key findings. These findings are synthesized into two categories of explanation: place-based and trust building. The chapter concludes with recommendations on practices for mitigating the community conflicts that accompany wind farm development and notes several key questions which remain a focus for study from social scientists.

Keywords Public opinion; Host community; Wind energy; Social acceptance

Introduction The responses of residents living in proximity to wind farms are diverse and in some cases projects fail because of opposition from local communities. Over the last several years, it has become increasingly important for wind developers and policymakers to consider social variables in planning for wind energy (Raven et al. 2009). For developers, favorable relations with host communities is an additional prerequisite to consider when siting projects beyond the traditional focus on level of the wind resource, access to the electrical grid, and financing. For policymakers, community opposition creates new policy actors that can derail planned energy transitions, disrupt social cohesion, and bring into question the legitimacy of societal decisions to pursue wind energy. Thus, better understandings of the responses of host communities to wind energy are important for both practical and political reasons. It is a complex topic and one that has attracted a considerable amount of study from social scientists. The aim of this chapter is to bring forward key findings from this literature, and offer two broad categories of explanations for host community responses. The chapter proceeds with a review of the approaches and methods used by social scientists to study host community responses to wind energy. Each approach is associated with a number of key findings and discussion of these findings

*Email: [email protected] Page 1 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

is interspersed with comments on methodological strengths and weaknesses. A synthesis section organizes the key findings into two underlying explanations for host community response: trust building and place based. The chapter concludes with recommendations on how to maintain and build community support for wind and suggests future trends in social science research for this topic.

Approaches to Studying Host Community Responses to Wind Energy Quantitative Methods Statistical Survey Research: The “Social Gap” and the “U” Shape of Public Opinion A conventional approach to finding out what a group of people think about an issue is to conduct a survey. Empirical statistical survey research is widely used in studies of host community responses to wind projects with two related objectives. First is to quantify the level of support or opposition, and second is to identify potential variables associated with support or opposition and thus develop and assess theoretical explanations for support or opposition. The level of support for wind projects is relatively consistent among jurisdictions. As demonstrated in Fig. 1 national and regional level polling typically shows high levels of support for wind energy at the 70–95 % range. However, a high level of public support for the general idea of more wind energy can obscure a much different story when it comes to the level of host community support for actual wind projects. Bell et al. (2005) citing figures from the UK, drew attention to this paradox of broad public support for wind energy against the fact that only a small fraction of projects proceeded to completion in the UK, largely because of high levels of local level opposition. They labeled this the “social gap” – a term that signifies the gap between public support for the goal of more wind energy but a lack of on-the-ground results. The “social gap” concept has come to be widely, although not uncritically, used in framing social acceptance research (Aitken 2010; Batel et al. 2013; Bell et al. 2013). Efforts to understand the social gap have taken various approaches. Survey research provides data to assess if wind project opponents (or supporters) have statistically consistent personal or social characteristics. These studies have found no clear trend linking attitudes to such characteristics. For example, age and gender influences are contradictory, e.g., sometimes older people are more opposed than younger people, sometimes the opposite is true (e.g., Devine-Wright 2005; Ladenburg 2010; Molnarova 100% 90% 80% 70% 60%

Support for wind energy (questions differ ∗)

50%

Australia, 2011, Qdos

Itally, 2011, Eurobarometer

France, 2011, Eurobarometer

Spain, 2011, Eurobarometer

Germany, 2011, Eurobarometer

Sweden, 2007, Eurobarometer

Netherlands, 2011, Eurobarometer

EU, 2011, Eurobarometer

UK, 2013, BBC/COMRes

UK, 2011, Eurobarometer

Denmark, 2009, Danish Wind Energy Assoc

United States 2013, Gallup

Great Lakes Region (Canada,US) 2014, UMichigan

Canada, 2012, Oracle

40%

* For example, respondents to the Eurobarometer poll (2011) were asked “ to what extent are you in favour or opposed to use of wind energy in your country”; the Gallup poll (2013) in the United States asked if people were in favour or opposed to the statement ” the US should put more emphasis on producing domestic energy from wind resources”

Fig. 1 Public opinion poll results for questions of wind energy support Page 2 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

et al. 2012). The same is true with other characteristics such as income. Wealthy individuals are associated with wind energy support in some studies (e.g., Devine-Wright 2005; Ek 2005) but the trend is reversed in other studies (e.g., Ertor-Akyazi et al. 2012; Fast and McLeman 2012; Firestone et al. 2009). Surveybased research has also identified various physical, environmental, and social characteristics of turbine development linked to host community opposition (Table 1). The multiple and sometimes contradictory results regarding possible variables for host community responses requires more sophisticated analytical techniques to tease out confounding influences. One of the leading researchers in this area is Wolsink (2000, 2007) who has combined multiple survey data sets and offers two significant findings. One, the visual impact of wind turbines is the strongest predictor of attitudes. This has led to efforts by some to quantify and distinguish landscapes perceived to have aesthetic value and thus to be avoided for wind energy development (Molnarova et al. 2012). Second, is a theoretical prediction that public opinion follows a “U shape” (Fig. 2) over time (Gipe 1995). In other words, attitudes towards wind power are high when wind energy remains an abstract possibility, dip during construction when the impacts of the project are the most uncertain for residents, and finally rise again after residents have become familiar with the project. Wolsink (2007) demonstrates empirical evidence for this in the Dutch context. However, more longitudinal studies combining preproject, planning phase, and postproject surveys of the same group of residents are needed to clarify this finding. One potential methodological issue is that residents opposing the project gradually move away thereby artificially inflating levels of support in postconstruction sampling. Surveys and Hypothesis Testing: Information-deficit; NIMBY; Place-attachment Survey research can be combined with social-psychological theory or other conceptual models to propose and test theories of attitude formation. One notable finding from this research area is the repudiation of the information deficit model. This model operates under the expectation that educating the public will enhance public attitudes and facilitate a kind of “social engineering” towards public support of wind energy. Survey research has proven this to be a faulty theory. High levels of knowledge about wind energy do not predict wind turbine support (Baxter et al. 2013; Graham and Rudolph 2014; Wolsink 2007). Instead, the literature suggests that opponents and supporters demonstrate comparable levels of knowledge about relative differences between wind and other forms of energy sources. Where opponents and supporters differ is the relative importance placed on the characteristics of wind generated electricity. Whereas a policymaker or wind developer might stress the greenhouse gas savings achieved by wind energy, for example, residents will downplay this aspect in favor of other impacts such as landscape impact or local economic impacts (Haggett 2011; Mulvaney et al. 2013; Warren et al. 2005). It is also important to note that the information deficit model assumes there is no scientific controversy over the information disseminated to the public. Yet, the evidence around issues such as turbine noise profiles and possible health impacts (Ellis et al. 2009) as well as property value impacts (Gulden 2011; Vyn and McCullough 2014) is evolving (Walker et al. 2014a). Thus, the information deficit model is erroneous on two counts: opponents are not ignorant of wind energy facts; and sometimes there is no one correct set of information with which to educate the public. It is important to point out at that the “public” is a broad concept. It can refer to both residents of host communities and to the general populace. The focus of this chapter is on research carried out on host community dynamics, however the target of education and information campaigns carried out under the information deficit model/social engineering model is larger. For example, the architect of the Feed-inTariff program, one of the most successful policy instruments to build wind energy adopted in 98 countries (REN 2014), described the real barriers to renewable energy as mental and political not technical or economic (Scheer 2007). Governments adopt FIT programs and other policies to change the sociopolitical landscape and facilitate market and political acceptance of wind energy. Residents of host communities Page 3 of 15

21

20–26 (This is also a bi-directional variable but not to the same extent as visual impact, a few surveys show host community residents expect property values to rise)

15

14–18

6–18

16–38

10–15

11–62 (This is a bi-directional variable. Survey research also shows residents see a positive impact on the landscape at about the same range)

(Corscadden et al. 2012; Devine-Wright 2005; Eltham et al. 2008; Jones and Eiser 2009; Musall and Kuik 2011; Shamsuzzoha et al. 2012; Slattery et al. 2012)

Lack of consultation on placement or revenue sharing

Impact on property values

Social/Economic/Political characteristics

# of turbines (fewer)

Interference with TV/radio/radar transmission Impacts on birds and other wildlife

Visual impact on resident (e.g., strobe effect/flicker) Noise

Visual impact on landscape

Physical or environmental characteristics

% of residents expressing concern

Table 1 Summary of descriptive statistics from a range of surveys on host community concerns over wind energy development

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

Page 4 of 15

high medium Low

Level of local support

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

no planned turbines

approval process

built turbines

Fig. 2 The predicted “u-shape” of local acceptance before turbines are planned, during project approval, and after construction. See text for details

represent a separate type of social acceptance (Wustenhagen et al. 2007); these residents are a unique and more engaged segment of the public. The information deficit model reflects a tendency for policymakers, wind developers, and other experts to see recalcitrant local public who oppose wind energy projects as ignorant and uninformed. It is linked to another prominent explanation for host community reactions to wind energy. The not in my backyard concept (NIMBY) concept describes the situation in which an individual is in favor of wind energy as long as it is nowhere near their own residence. It has become unfashionable for social scientists to give credence to this explanation. Instead researchers strive for less biased understanding of opponents positions’ and not dismissing opponents’ actions as selfish or deviant (Aitken 2010; Devine-Wright 2010). Nevertheless survey results often show the NIMBY position is present at low levels in host communities (Fast and McLeman 2012; Walker et al. 2014b; Wolsink 2000) and some (e.g., Cohen et al. 2014) argue that NIMBY is a rational reaction for some individuals which should be recognized via policies which attempt to quantify any costs incurred by those living in proximity to a wind energy project and arrange for compensation. Others point out that it is entirely consistent for an individual to have simultaneously different attitudes and beliefs around wind energy locally and wind energy more generally. Indeed Batel and Devine Wright (2014) argue that such multiple representations are a typical psychological pattern when individuals and groups become familiar with new technology. Another hypothesis advanced by survey research is that of place attachment as a determining variable for attitudes towards wind energy projects. Devine-Wright (2009, 2011a, b; Devine-Wright and Howes 2010) seeks to advance psychological conceptions of place attachment and place identity as alternatives to NIMBY explanations. He notes that change to a place in the form of wind energy development can result in feelings of anxiety and loss and elicit place-protective actions. In survey research, the level of place attachment is operationalized by including questions to participants such as level of agreement with statements like “I feel at home in [place name]”. These are then treated as variables to include in more detailed analysis. In order to demonstrate to readers the range of statistical techniques applied in survey research within this field, and also to provide an extended example of the development of one proposed explanation from the social science literature for host community responses, the study by Devine Wright and Howes (2010) is summarized briefly below. Devine-Wright and Howes (2010) drew participants from two communities both within 15 km of a series of wind farms in Wales. The sample of 590 questionnaires was distributed and 457 returned in the two communities which have a combined population of approximately 30,000. As is common in survey

Page 5 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

research, respondents were asked to assess their agreement on a scale of 1–5 (Likert scale) with a number of statements on perceptions of the wind project such as it “will tackle climate change”, “will reduce property values”. Another element of the study aimed to assess place meanings by asking respondents to quickly write down phrases or words they associated with their place. These free-association meanings were organized into thematic codes such as “holiday resort”, “pleasant living”, “coastal features” which lead to a distinction that emerged between the communities; one was perceived as beautiful and pleasant while the other was perceived as rundown. The use of a Likert scale rating permits mean ratings to be calculated and significant differences between groups to be assessed. In this study, residents of the community perceived to be rundown were significantly more likely than those of the neighboring community to agree that the wind project would tackle climate change for example. Another common statistical test to perform is to look for correlations between variables. In this study, place attachment was significantly correlated to perceived negative outcomes such as “will create an eyesore”, but only for one of the communities. This lead the authors to conclude that wind energy developments may lead to placeprotective attitudes in some locations but others may be indifferent or have fatalistic attitudes. Another form of statistical analysis not conducted in the study reviewed above but often employed in this field is regression analysis (e.g., Culley et al. 2011; Ek 2005; Wolsink 2000); this procedure is used to test how well variables or groups of variables explain a dependent variable (i.e., position on wind turbines) while controlling for responses to other questions. For example, in Baxter et al. and (2013) responses to questions on aesthetic appeal of turbines and to questions about an innate preference for wind generated electricity predict support (R2 = 0.6) but including other variables such as responses to questions about health impacts increases the predictive power of regression model. In this way, various hypotheses about the contributing factors to peoples’ attitudes on wind power can be tested. Economic Modeling: Choice Experiments and Property Value Studies Two types of economic modeling studies have contributed to understandings of host community responses. Choice experiments are specialized surveys that attempt to model the decision making preferences of individuals in order to place a value on non-market benefits and costs. The second type of modeling work is in property value studies that attempt to isolate the contribution of wind projects to local property values by analyzing property sales data together with variables such as distance and view of turbines. Ek and Persson (2014) provide a good example of a choice experiment. They asked 1,500 respondents to choose between twelve sets of attributes including placement of turbines (e.g., offshore, forest, open landscape, mountainous area) ownership, consultation, use of revenue (0.5 % to general revenue, 0.5 % earmarked to nature conservation), and the fee that each respondent would be willing to pay to subsidize the project (0.3, 07 and 1.3 euro cents). The choices are assessed in a regression model that calculates the probability of choosing one scenario of attributes over another. When combined with the willingness to pay responses a financial measure of the preference function can be developed. Ek and Persson (2014) found the best case scenario was municipally owned offshore wind farms with extended consultation with revenues earmarked for nature conservation. The willingness to pay associated with this scenario was 0.6 euro cents versus 0.5 euro cents for a privately owned mountainous wind farm with the same consultation and revenue options. Additional willingness to pay survey and choice experiments show varied results. Bollino (2009) surveyed Italian citizens finding 47 % expressed a willingness to pay for renewables that averaged 8.5 Euros per month. Zografakis et al. (2010) found a figure of 1.5 Euros per month for Greek residents. The same type of experimental model can be used to quantify the costs that residents are willing to bear, and thus the compensation that could be offered to encourage acceptance of projects. Kreuger et al. (2011)

Page 6 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

show that the visual disamenity costs of offshore wind turbines for residents living on the ocean in eastern US decreases the further the turbines are placed offshore. Choice experiments and willingness to pay models are well positioned to offer recommendations on the level of subsidy that decision –makers should pursue when designing renewable energy policy. However in terms of explanatory value for assessing host community responses they face some of the same problems as other survey research, namely, no consistent trends have been revealed across social groups (i.e., willingness to pay is heterogenous), and there is still a “social gap” between measuring a general preference for wind energy among a population and the response to specific wind farms by a local population. The other type of economic modeling study which has contributed to a better understanding of host community responses to wind projects concerns the impact of such projects on local property values. This literature encompasses several different measurement tools including price comparison surveys, simple statistical tests (t-tests or other), and detailed regression analysis (see Vyn and McCullough 2014 for a review). Hedonic regression analysis are generally favored and typically assess distance to turbines as the variable to isolate for any price effect (“hedonic” refers to the pleasure or utility of a property ultimately measured in its price). North American studies of this type have generally found no significant effects (Hoen et al. 2009; McCullough 2009; MPAC 2014; Hoen et al. 2013). However, price comparison studies (Lansink Appraisals and Consulting 2012) and contingent valuation surveys (Landenburg and Dubgaard 2007) sometimes find property value reduction associated with turbines. In his nuanced review of the property value literature from a homeowner’s perspective, Gulden (2011) points out that this uncertainty in the literature provides evidence to both opponents and supporters of the claim that turbines decrease property values. When combined with more anecdotal evidence disseminated by real estate agents, there are often multiple conflicting claims circulating in places hosting wind energy facilities. This in turn fosters substantial skepticism and concern on the part of residents and contributes to host community opposition to wind projects (Cohen et al. 2014; Groth and Voigt 2014).

Qualitative Methods Qualitative research methods are also well placed to assess many aspects of host community responses to wind energy projects. The types of questions asked in qualitative research generally fall into two categories; questions of social structures and questions of individual experiences (Winchester 2010). Social structures are practices that govern social interaction. They can be relatively obvious such as the landowner-lessee arrangement between wind farm developers and landowners leasing their land for wind turbines. Or they can be more subtle such as a cultural aesthetic to see beauty in certain rural landscapes. Key questions in social science have to do with how structures have the effect that they do. Individual experience questions are related to this, but instead focus on the differences in how people experience places and processes. Research of this origin often purposely attempts to give voices to people who might otherwise be silenced. The observation made above that the NIMBY explanation for host community responses is unfashionable in social science is rooted in this tradition of giving voice to different individuals in order to strive for less biased understandings of individuals and of social structures. There are strong methodological and philosophical reasons to recognize the researcher and research process as inevitably biased which means there are major differences between the types of evidence and reasoning found in qualitative social science versus natural science and quantitative social science (for more see Sayer 2000). Generally, qualitative research is more interpretive and less hypothesis testing than quantitative research. It uses a wider variety of conceptual frameworks and theory. The following section reviews some of the qualitative research methods used to study host community response to wind energy development and highlights the major findings to come out of this literature. Page 7 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

Interviewing and group interviews Interviewing the actors in host communities for wind energy projects is a valuable method for exploring host community responses. The major methodological issues that come with interviewing include who is selected for interview, what questions are asked (e.g., the bias of the interviewer) and once enough material has been gathered, what is the correct interpretation of that interview material. The answers to these questions depend on the number and type of actors involved, the historical background to the wind project, and other contextual factors. As a result, interview-based research tends to involve rich and detailed descriptions of case studies and often incorporates comparisons between case studies. This allows for nuanced explanations of why a certain proposal or policy effort is or is not successful in gaining local acceptance. Cass et al. (2010) provide a good example of the type of insight that can be derived from interviewbased research. Their study examines the practice of wind developers proving a “community benefit” during development of wind projects. These benefits can take multiple forms including a regular payment or lump sum to residents; benefits in kind such as the construction of a boardwalk in a conservation area; or simply the employment opportunities a wind project provides to local residents. Cass et al. (2010) interviewed planners, wind company representatives and members of the public. Their interviews revealed tensions between these actors around issues such an impression that payments might be seen as a “bribe” to achieve local compliance, and that some benefits expected by residents, such as cheap electricity were outside of a developer’s controls. Interviews are often combined with survey research in order to explain the reasoning behind survey responses and to triangulate data in order to be more certain of findings. Walker and Baxter (2014a) combine survey results showing differing levels of support for two different wind projects with interviews of survey respondents to show how perceptions of the projects is tied to psychosocial stress and feelings of unacceptable community conflict. Slattery et al. (2012) quote from survey participants to bolster their findings that economic impacts are the predominant determining factor of wind energy acceptance in Iowa and Texas. Mulvaney et al. (2013) combine survey and interview research to show that support for wind projects in Indiana can partially be explained by a desire from residents to combat urban sprawl and to keep periurban areas as farmland. In their case study, wind turbines were seen as a form of development that permits farmers to keep farming. Group interviews are another method of exploring host community responses to wind energy. Interviewing a group of people has the advantage of potentially bringing concerns and questions to a discussion which may not occur to an interviewer in a one to one setting. Group interviews/focus groups were used by Parks and Theobald (2013) to investigate the expectations of local residents and planners for public participation during the siting of a turbine. They found that planners dismisses residents’ proposals as unworkable not because of the content of the proposal but because of the timing or form in which they were delivered. Their focus groups also revealed a lack of trust in the host landowner because they were assumed to operate from a profit motivation. This kind of mismatch between the types of issues deemed important by residents and planning officials has also been seen in other studies of wind energy planning (e.g., Fast 2013). Political Science and Planning Studies : Institutions and the Role of Public Participation Political scientists have tackled host community responses by drawing attention to differences in the organizational structure of the political economy. Toke et al. (2008) compare four types of institutional variables (planning rules, financial support mechanisms offered by the state, organizations concerned with landscape protection, and ownership patters of wind farms) to develop hypotheses of why wind power is more developed in certain countries. For example, they suggest wind power is more developed in Germany than the UK because Germany has more stable state support mechanisms and a tradition of local energy activism which the UK lacks. Ferguson-Martin and Hill (2011) perform a similar analysis of Page 8 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

the Canadian provinces and add other institutional variables. They show that liberalized electricity markets and an existing electricity grid architecture with high levels of hydropower which is able to complement the intermittent production of wind are additional variables contributing to higher levels of wind energy development. Loring (2007) developed variables to compare public acceptance of 18 wind energy projects in England, Wales and Denmark according to level of public participation, and the degree of network stability for key actors. She found that high levels of public participation (e.g., community groups initiate and manage the project) are contributory to public acceptance. Furthermore, she suggests that organized and active opposition groups contribute to project failure. Finding like this highlight the importance of approval processes and opportunities for participation in decision-making about the location of wind projects. Ellis et al. (2009) notes that these issues are connected to how governments deal with dissent. Thus, host community responses to wind development are connected to issues of governance and legitimacy. From a governance perspective, there is a tension between normative expectations for meaningful consultation with citizens of host communities and pragmatic pursuit of renewable energy targets. Political scientists and planning scholars have long recognized that public trust in experts is on the decline (Fischer 1993). In this context, the fact that many citizens of host communities are skeptical of the rationale for wind projects is not surprising. Community-led and locally owned wind power is frequently seen as the optimum form of organization for gaining project acceptance (Loring 2007; Walker and Devine-Wright 2008). Yet, this form of ownership is rare in many countries and as Toke et al. (2008) points out, it requires a history of local energy activism. Without this tradition, the capacity for community led ownership models can be weak (Fast and Mabee 2015; Winfield and Dolter 2014). Discourse Analysis: Wind Farm Versus Wind Turbine Discourse analysis covers a spectrum of methods concerned with the analysis of written text or spoken language. “Discourse” here has a specialized meaning that is different than “discussion”. Discourses are a combination of ideas, concepts, and categories through which meaning is created (Hajer and Versteeg 2005). When actors use a certain discourse they favor certain interpretations over others. The use of the term “wind farm” is a good example. Groups opposing wind energy do not use the term “wind-farm” rather it is wind-turbine or “industrial wind turbine” (Barry et al. 2008; Hagget and Toke 2006). This language choice marks a purposeful effort to undermine the notion of “farm” as natural, safe, and rural and replace it with wording that breaks that connection (Hagget and Toke 2006). Thus, discourse analyses seeks to criticize the assumptions that are created by certain discourses, and bring into question what processes or outcomes are made necessary or made to seem natural by the discourse. Barry et al. (2008) contrast wind energy supporters and opponent discourses and identify stark differences. For example, supporters’ discourse assumes that climate change is a danger that should be dealt with by all means possible including renewable energies. Opponents’ discourse, on the other hand, assumes that turbines are unnatural and contribute to local environmental damage. The latter discourse is more likely than the first to accept the possibility of other claims such as wind energy development is driven by greed. Discourse analysis has also been used to assess broader public and media coverage of wind energy policy. Winfield and Dolter (2014) identify three discourses “ecological modernist”, “economic rationalist” and “market fundamentalist” each decreasingly supportive of wind energy subsidies. The authors evaluate nine studies on the economic impact of Ontario’s FIT program and show that different results are associated with the different discourse employed. Each brings with it different assumptions that are hidden to readers of the studies (e.g., expected price of conventional fossil fuels, including of externalized

Page 9 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

costs). These assumptions have major influences on the conclusions of the reports. Expert reports such as these in turn shape the discourses of proponents and opponents in host communities. Q-method: Unexpected Similarities and Divergences Q-method is a statistical method of measuring an individual’s views on a topic. Its use is growing in host community wind energy studies (Brannstrom et al. 2011; Ellis et al. 2007; Fisher and Brown 2009) because it is well suited to elucidating the gradations that exists between positions of support and opposition. Q-method identifies commonalities in the way people think about a topic. It is often referred to as a method for assessing human subjectivity. The “q” in q method refers to the acronym given to participants (q) to distinguish it from the “r” used to denote the degree of correlation between traits of people (cases). The method requires participants to assign statements (including possibly their own statements) a rank which permits a factor analysis of the affinity of an individual to arrangements of statements (also called “subjectivities”, “discourses”, “accounts” by different authors) and the similarity and divergence of the participant’s views with the views of with other participants. The outcome of a q-study is a coherent picture of the viewpoints on a wind project held by people in a host community. Ellis et al. (2007) identified four main factors/discourses in a dispute in Northern Ireland. These included what they called the “rationalizing globally” position in which statements such as “awareness of climate change and the need for action” ranked highly, and the “developer skeptic” position for which the same statement ranked highly but this position also favored the statement “developers are more interested in profit than saving the planet”. These kinds of distinctions reveal unexpected similarities between people that initially appear to be strong opposites. The reverse is also true; Q-method studies have shown sharp differences between what might be considered monolithic groups of support. For example Brannstrom et al. (2011) found two types of supporter perspectives among host communities in Texas. They differed sharply on whether or not hunting was still feasible in fields with turbines. Such distinctions are often lost in survey research and interview research.

Synthesis: Place-Based and Trust-Based Explanations The preceding sections have highlighted the diverse methodological approaches used to develop better understandings of host community response to wind energy projects. This field of study is rich and contains a range of findings. However the diversity of disciplines associated with questions of social acceptance and host community responses complicates efforts to synthesize results. Thus, it is useful to organize the major findings from across the field of study in some fashion. Table 2 summarizes major findings and categorizes them into two general types of explanations for host community responses: place-based and trust-building explanations. Place-based explanations are associated with the emotional attachment residents have to their home region. These emphasize landscape aesthetics, property value concerns, traditions of local energy activism, and other factors that are associated with particular places. Trust-building explanations on the other hand stress the importance of residents’ trust in planning processes or trust in key actors such as wind development companies. These include the trust created by community organized owned energy projects, the power of discourses to frame wind energy as a necessary step to combat climate change and the extent to which dissenters are marginalized as NIMBYs in planning processes. The organization of this table is inevitably subjective. It is one of many possible ways to organize this field of study.

Page 10 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

Table 2 Place-based and trust-building explanations can be used to organization social science research of host community responses to wind energy developments Finding Social gap between high levels of general public support for wind energy and lack of on the ground-uptake “Not in My Backyard” (NIMBY) attitudes are poor explanations for all forms of community opposition Opposition to wind energy is not due to information deficit Property values close to wind farms tend not to be significantly negatively affected (but not always) Locally owned and community driven wind projects show greater public acceptance Profit motive of wind companies engenders distrust Mismatch between planning officials and residents in terms of appropriate time and place to express concerns over projects Actors reproduce discourses which privilege selective scenarios

Reference (to illustrate, not an exhaustive list) Bell et al. (2005, 2013)

Methods used Survey

Place- based or trust –building Place –based

Wolsink (2000), Devine-Wright (2009)

Survey

Trust-building

Haggett (2011), Wolsink (2007) Hoen et al. (2013), Vyn and McCullough (2014)

Survey

Trust – building

Hedonic economic modeling, contingent valuate surveys Political science/ institutional analysis, interviewing Q- method, interviewing, discourse analysis Interviewing,

Place –based

Discourse analysis, q method

Trust building

Loring (2007), Toke et al. (2008) Cass et al. (2010), Ellis et al. (2007) Parks and Theobald (2013), Ellis et al. (2009) Barry et al. (2008), Hagget and Toke (2006)

Trust-building

Trust-building Trust- building

Conclusions and Future Challenges This chapter has shown that place making and trust building are key mechanisms for maintaining and building support for wind energy projects. The next question becomes how to encourage these activities in policy and wind project development. The policy recommendations offered in the social science literatures are often quite abstract and are a subject of considerable debate and study themselves. In concluding this chapter, two specific recommendations from the literature are worth highlighting. First, Raven et al. (2009) focus on the need for individual project managers to treat community relations as an ongoing negotiation. Trust is built when expectations are fulfilled and thus careful attention has to be paid to the claims made by wind companies and the concerns that arise from residents. For example, it is not enough for wind companies simply to state to residents that property values will not be affected by a wind project. Trust will only be built by concrete actions. These could include posting a bond to be activated if property values dip below a threshold or restructuring compensation so that all property owners abutting turbines are compensated and not only the ones with turbines on their land. The second recommendation concerns the structuring of approval processes. The tendency has been to develop planning and zoning regimes that treat all opposing residents as NIMBYs (Ellis et al. 2009). Streamlined and central authority planning fails because the place-specific issues that matter to host communities are not considered. Regional decision-making processes, especially those that require local stakeholders to consider energy choices (e.g., if not wind energy than what are other acceptable energy sources) are more suitable. Finally, there are pressing research gaps to be filled in order to round out a more complete picture of host community responses to wind energy. These include validation of the predicted “u-shape” of public

Page 11 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

opinion. There is some evidence that host communities that start out opposed to projects become more supportive over time but until more longitudinal studies are published this remains a hypothesis. Secondly, the social science literature is overwhelmingly dominated by western European and North American studies for this topic. More studies are needed of public opinion and host community responses in jurisdictions with different political economic systems. China has the most installed wind energy capacity, India is close behind. Jurisdictions such as these offer important contexts for theory testing and development. Lastly, it has become commonplace to suggest community ownership of wind projects is required for public support. Yet, this form of ownership is still rare. Why is that? These are only some of the questions that will continue to drive social science research over the coming years.

Acknowledgements Financial support from Robert-Gilbert Postdoctoral fellowship at Queen’s University is gratefully acknowledged.

References Aitken M (2010) Why we still don’t understand the social aspects of wind power: a critique of key assumptions within the literature. Energy Policy 38:1834–1841 Barry J, Ellis G, Robinson C (2008) Cool rationalities and hot air: a rhetorical approach to understanding debates on renewable energy. Global Environ Polit 8(2):67–98. doi:10.1162/glep.2008.8.2.67 Batel S, Devine-Wright P (2014) Towards a better understanding of people’s respones to renewable energy technologies: insights from social representations theory. Public Underst Sci. doi:10.1177/ 0963662513514165 Batel S, Devine-Wright P, Tangeland T (2013) Social acceptance of low carbon energy and associated infrastructures: a critical discussion. Energy Policy 58:1–5 Baxter J, Morzaria R, Hirsch J (2013) A case–control study of support/opposition to wind turbines: perceptions of health risk, economic benefits, and community conflict. Energy Policy 61:931–943 Bell D, Gray T, Haggett C (2005) The ‘Social Gap’ in wind farm siting decisions: explanations and policy responses. Environ Polit 14(4):460–477 Bell D, Gray T, Hagget C, Swaffield J (2013) Revisting the ‘social gap’: public opinion and relations of power in the local politics of wind energy. Environ Polit 22(1):115–135 Bollino CA (2009) The willingness to pay for renewable energy sources: the case of Italy with sociodemographic determinants. Energy J 30(2):81–96 Brannstrom C, Jepson W, Persons N (2011) Social perspectives on wind-power development in west Texas. Ann Assoc Am Geogr 101(4):839–851 Cass N, Walker G, Devine-Wright P (2010) Good neighbours, public relations and bribes: the politics and perceptions of community benefit provision in renewable energy development in the UK. J Environ Policy Plan 12(3):255–275 Cohen J, Reichl J, Schmidthaler M (2014) Re-focussing research efforts on the public acceptance of energy infrastructure: a critical review. Energy Econ 76(1):4–9 Corscadden K, Wile A, Yiridoe E (2012) Social license and consultation criteria for community wind projects. Renew Energy 44:392–397. doi:10.1016/j.renene.2012.02.009

Page 12 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

Culley MR, Carton AD, Weaver SR, Ogley-Oliver E, Street JC (2011) Sun, wind, rock and metal: attitudes toward renewable and non-renewable energy sources in the context of climate change and current energy debates. Curr Psychol 30(3):215–233. doi:10.1007/s12144-011-9110-5 Devine-Wright P (2005) Beyond NIMBYism: towards an integrated framework for understanding public perceptions of wind energy. Wind Energy 8(2):125–139. doi:10.1002/we.124 Devine-Wright P (2009) Rethinking NIMBYism: the role of place attachment and place identity in explaining place-protective action. J Community Appl Soc Psychol 19(6):426–441. doi:10.1002/ casp.1004 Devine-Wright P (ed) (2010) Renewable energy and the public: from NIMBY to participation. Earthscan, London Devine-Wright P (2011a) From backyards to places: public engagement and the emplacement of renewable energy technologies. In: Devine-Wright P (ed) Renewable energy and the public: from NIMBY to participation. Earthscan, London, pp 57–70 Devine-Wright P (2011b) Place attachment and public acceptance of renewable energy: a tidal energy case study. J Environ Psychol 31(4):336–343. doi:10.1016/j.jenvp.2011.07.001 Devine-Wright P, Howes Y (2010) Disruption to place attachment and the protection of restorative environments: a wind energy case study. J Environ Psychol 30(3):271–280. doi:10.1016/j. jenvp.2010.01.008 Ek K (2005) Public and private attitudes towards “green” electricity: the case of Swedish wind power. Energy Policy 33(13):1677–1689. doi:10.1016/j.enpol.2004.02.005 Ek K, Persson L (2014) Wind farms – where and how to place them? A choice experiment approach to measure consumer preferences for characteristics of wind farm establishments in Sweden. Ecol Econ 105:193–203 Ellis G, Barry J, Robinson C (2007) Many ways to say ‘no’, different ways to say ‘yes’: applying Q-methodology to understand public acceptance of wind farm proposals. J Environ Plan Manag 50(4):517–551 Ellis G, Cowell R, Warren C, Strachan P, Szarka J (2009) Wind power: is there a planning “problem”? Plan Theory Pract 10(4):521–547 Eltham DC, Harrison GP, Allen SJ (2008) Change in public attitudes towards a cornish wind farm: implications for planning. Energy Policy 36(1):23–33. doi:10.1016/j.enpol.2007.09.010 Ertor-Akyazi P, Adaman F, Ozkaynak B, Zenginobuz U (2012) Citizens’ preferences on nuclear and renewable energy sources: evidence from Turkey. Energy Policy 47:309–320. doi:10.1016/j. enpol.2012.04.072 Fast S (2013) A Habermasian analysis of local renewable energy deliberations. J Rural Stud 30:86–98 Fast S, Mabee W (2015) Trust-building and place-making: the influence of policy on host community responses to wind farms. Energy Policy 81:27–37 Fast S, McLeman R (2012) Attitudes towards new renewable energy technologies in the Eastern Ontario Highlands. J Rural Community Dev 7(3):106–122 Ferguson-Martin C, Hill S (2011) Accounting for variation in wind deployment between Canadian provinces. Energy Policy 39:1647–1658 Firestone J, Kempton W, Krueger A (2009) Public acceptance of offshore wind power projects in the USA. Wind Energy 12(2):183–202. doi:10.1002/we.316 Fischer F (1993) Citizen participation and the democratization of policy expertise: from theoretical inquiry to practical cases. Policy Sci 26(3):165–187 Fisher J, Brown K (2009) Wind energy on the Isle of Lewis: implications for deliberative planning. Environ Plan A 41:2516–2536 Gipe P (1995) Wind energy comes of age. Wiley, New York Page 13 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

Graham K, Rudolph D (2014) Geography, communities and energy futures: alternative research paths. Scott Geogr J 130(3):143–151 Groth T, Voigt C (2014) Rural wind farm development: social, environmental and economic features important to local residents. Renew Energy 63:1–8 Gulden W (2011) A review of the current evidence regarding industrial wind turbines and property values from a homeowner’s perspective. Bull Sci Technol Soc 31(5):363–368 Hagget C, Toke D (2006) Crossing the great divide - using multi-method analysis to understand opposition to windfarms. Public Adm 84(1):103–120 Haggett C (2011) Understanding public responses to offshore wind power. Energy Policy 39(2):503–510. doi:10.1016/j.enpol.2010.10.014 Hajer M, Versteeg W (2005) A decade of discourse analysis of environmental politics: achievements, challenges, perspectives. J Environ Policy Plan 7(3):175–184 Hoen B, Wiser RH, Cappers P, Thayer M, Sethi G (2009) The impact of wind power projects on residential property values in the United States: a multi-site hedonic analysis. J Real Estate Res 33(3):279–316 Hoen B, Brown J, Jackson T, Wiser R, Thayer M, Cappers P (2013) A spatial hedonic analysis of the effects of wind energy facilities on surrounding property values in the United States. Ernest Orlando Lawrence Berkeley National Laboratory, Berkley Jones CR, Eiser JR (2009) Identifying predictors of attitudes towards local onshore wind development with reference to an English case study. Energy Policy 37(11):4604–4614. doi:10.1016/j. enpol.2009.06.015 Krueger A, Parsons G, Firestone J (2011) Valueing the visual disamenity of offshoer wind power projects at varying distances from the shore: an application on the Delaware shoreline. Land Econ 87:268–283 Ladenburg J (2010) Attitudes towards offshore wind farms-the role of beach visits on attitude and demographic and attitude relations. Energy Policy 38(3):1297–1304. doi:10.1016/j.enpol.2009.11.005 Landenburg J, Dubgaard A (2007) Willingness to pay for reduced visual disamenities from offshore wind farms in Denmark. Energy Policy 35:4059–4071 Lansink Appraisals and Consulting (2012) Case study diminution in value wind turbine analysis, Lanksink Appraisals and Consulting Loring J (2007) Wind energy planning in England, Wales and Denmark: factors influencing project success. Energy Policy 35:2648–2660 McCullough R (2009) Assessing the impact of the melancthon phase 1 wind project on nearby agricultural property values: a hedonic approach. University of Guelph, Guelph Molnarova K, Sklenicka P, Stiborek J, Svobodova K, Salek M, Brabec E (2012) Visual preferences for wind turbines: location, numbers and respondent characteristics. Appl Energy 92:269–278. doi:10.1016/j.apenergy.2011.11.001 MPAC (2014) Impact of industrial wind turbines on residential property assessment in Ontario 2012 assessment base year study. Municipal Property Assessment Corporation, Toronto Mulvaney K, Woodson P, Prokopy L (2013) A tale of three counties: understanding wind development in the rural Mid-Western United States. Energy Policy 56:322–330 Musall FD, Kuik O (2011) Local acceptance of renewable energy-a case study from southeast Germany. Energy Policy 39(6):3252–3260. doi:10.1016/j.enpol.2011.03.017 Parks JM, Theobald KS (2013) Public engagement with information on renewable energy developments: the case of single, semi-urban wind turbines. Public Underst Sci 22(1):49–64. doi:10.1177/ 0963662511400962 Raven RPJM, Mourik RM, Feenstra CFJ, Heiskanen E (2009) Modulating societal acceptance in new energy projects: towards a toolkit methodology for project managers. Energy 34:564–574 Page 14 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_21-1 # Springer-Verlag Berlin Heidelberg 2015

REN (2014) Renewables 2014: global status report. Renewable Energy Network, Paris Sayer A (2000) Realism and social science. Sage, London Scheer H (2007) Energy autonomy: the economic, social and technological case for renewable energy. Earthscan, London Shamsuzzoha AHM, Grant A, Clarke J (2012) Implementation of renewable energy in Scottish rural area: a social study. Renew Sustain Energy Rev 16(1):185–191. doi:10.1016/j.rser.2011.07.146 Slattery MC, Johnson BL, Swofford JA, Pasqualetti MJ (2012) The predominance of economic development in the support for large-scale wind farms in the U.S. Great Plains. Renew Sustain Energy Rev 16(6):3690–3701. doi:10.1016/j.rser.2012.03.016 Toke D, Breukers S, Wolsink M (2008) Wind power deployment outcomes: how can we account for the differences? Renew Sustain Energy Rev 12(4):1129–1147. doi:10.1016/j.rser.2006.10.021 Vyn R, McCullough R (2014) The effects of wind turbines on property values in Ontario: does public perception match empirical evidence? Can J Agric Econ. doi:10.1111/cjag.12030:1-28 Walker G, Devine-Wright P (2008) Community renewable energy: what should it mean? Energy Policy 36:497–500 Walker C, Baxter J, Oulette D (2014a) Adding insult to injury: the development of psychosocial stress in Ontario wind turbine communities. Soc Sci Med 46(3):730–745 Walker C, Baxter J, Oulette D (2014b) Beyond rhetoric to understanding determinants of wind turbine support and conflict in two Ontario, Canada communities. Environ Plan A 43(5):730–745 Warren CR, Lumsden C, O’Dowd S, Birnie RV (2005) Green on green: public perceptions of wind power in Scotland and Ireland. J Environ Plan Manage 48(6):853–875 Winchester H (2010) Qualitative research and its place in human geography. In: Hay I (ed) Qualitative research methods in human geography. Oxford University Press, Oxford Winfield M, Dolter B (2014) Energy, economic and environmental discourses and their policy impact: the case of Ontario’s Green Energy and Green Economy Act. Energy Policy 68:423–435 Wolsink M (2000) Wind power and the NIMBY-myth: institutional capacity and the limited significance of public support. Renew Energy 21:49–64 Wolsink M (2007) Wind power implementation: the nature of public attitudes: equity and fairness instead of ‘backyard motives’. Renew Sustain Energy Rev 11(6):1188–1207. doi:10.1016/j.rser.2005.10.005 Wustenhagen R, Wolskink M, Burer J (2007) Social acceptance of renewable energy innovation: an introduction to the concept. Energy Policy 35(5):2683–2691 Zografakis N, Sifaki E, Pagalou M, Nikitaki G, Psarakis V, Tsagarakis KP (2010) Assessment of public acceptance and willingness to pay for renewable energy sources in Crete. Renew Sustain Energy Rev 14(3):1088–1095. doi:10.1016/j.rser.2009.11.009

Page 15 of 15

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Wind Scalability and Performance in the Real World: A Performance Analysis of Recently Deployed US Wind Farms A Performance Analysis of Recently Deployed US Wind Farms G. Bothun* and B. Bekker Department of Physics, University of Oregon, Eugene, OR, USA

Abstract Real world performance, costs, and supply chain issues determine the rate at which wind turbines can be constructed, deployed, and installed. To study this issue for the case of the USA, a sample of ~600 individual US wind farms that have come into operation as of January 2010 has been assembled and analyzed using standard statistical procedures. Results for wind farm turbine composition and overall wind energy installation rates are: (a) individual unit turbine capacity ranges from 1–5 to 3 MW, although the bulk of the installations are 2.0 MW; (b) starting in late 2012, turbines of size 2.5–3.0 MW are being installed but there are logistical transport problems that come into play on this scale; (c) as of July 1, 2014, the Horse Hollow development in Texas has the largest individual wind farm nameplate capacity of 736 MW and 10 other locations have aggregate capacity that exceed 500 MW; (d) over the period of 2006–2012, cumulative wind capacity growth was sustained at a rate of 23.7 % per annum and by January 1, 2013, total installed US wind energy was 60,000 MW; (e) if this growth rate could be sustained over a 10-year period beginning at the end of 2012, then the USA would achieve a total nameplate capacity of ~475 GW for wind, which would be approximately 40 % of the entire US electrical generating nameplate capacity; (f) due to considerable uncertainty over the future of the federal production tax credit (PTC), continued growth in 2013 did not occur and this flattening has now placed US wind energy growth on a significantly more subdued trajectory; and (g) a small extension of the PTC in 2014 allowed for an additional 4,800 MW to come on line bringing the total US installed capacity to 66,000 MW as of January 1, 2015. This is approximately 45 GW less than would have been predicted based on the previous growth rate and clearly shows the dependence of growth rate on the continuation of the PTC. Statistical analysis of real world performance and costs shows that: (a) average state measured capacity factors (CF) have a mean of 0.31 and standard deviation of 0.05; this can be compared to the average CF for European countries of 0.245 +/ 0.04. The difference in these averages is statistically significant; (b) approximately 15 % of the sample has CF > 0.4 and these sites are primarily located in Kansas, Oklahoma, and Texas; (c) recently constructed wind farms with actual published project costs indicate an average of $2 per nameplate watt; (d) CF weighted costs average about $7 per generated watt but there is considerable variation around this average; (e) CF weighted costs lower with increasing nameplate capacity which indicates an investment strategy; (f) considerations of supply chains and transport logistics indicate that the delivery of single turbine blades to the turbines that require blades of length >55 m will likely be the limiting factor that governs the rate at which new large-scale wind farms can deployed.

*Email: [email protected] Page 1 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Keywords Wind energy supply chain; Wind turbine growth; Wind farm capacity factors; Growth of wind energy in the US

Introduction The feasibility of wind power as a renewable form of electrical energy generation has been clearly borne out in the US and Europe over the last decade or so. Individual wind farms and wind farm complexes are most properly characterized by their nameplate capacity and their capacity factor (CF). Nameplate refers to the maximum power for which an individual wind turbine is rated. For instance, a 2.5 megawatt (MW) wind turbine will produce 2.5 MW if the wind is continuously blowing. Since wind is an intermittent source then, on a daily basis, the average electricity production will be less than the nameplate. CF is a measure of the percentage of time a wind farm is able to produce at its nameplate capacity. Canonical and planning values for CF are usually taken to be 1/3 (i.e., the wind blows 8 h a day). However, CF must be properly measured in the real world for a reliable evaluation of wind energy’s potential contribution to the overall portfolio of electricity generation for the US. To make these measurements, we have assembled a data base of approximately 600 US wind farms and wind farm complexes that have come online since 2010. The primary data sources used in this study are, (a) The US Energy Information Administration (EIA), (b) market reports from the American Wind Energy Association (AEWA), (c) various white paper studies do from the National Renewable Energy Laboratory (NREL), and (d) informal e-mail and telephone conversations with various wind farm operators who have responded with various levels of information giving. In addition, much pertinent information comes from local newspaper articles related to the opening or commissioning of a facility. Oftentimes, these articles contain a “fun facts” section which conveys information, e.g., total project costs that cannot be found via conventional sources as this is not a reporting requirement. The gathering of as much real world information as possible is critical to a proper assessment of the performance and scalability of wind power in the real world as the real world is often much different than the ideal world. For instance, annual reporting of wind farm output and capacity factors for the WA state in the year 2012 are significantly compromised due to a policy decision in the real world. Spring runoff in the Pacific Northwest for 2012 was unusually high and so the hydroelectric capacity of the 31 dams on Columbia River system was fully realized. As the Columbia River System electrical grid could not handle any increased load (due to the lack of energy storage facilities), the Bonneville Power Administration ordered that the regions wind farms were to be turned off for the entirety of May 2012. Although this story was well covered via local newspapers (e.g., Sickinger 2012), this physical incident is not readily noted in the EIA databases Data is reported to the EIA in the form of Plant Identification number and this reported data set consists of approximately 600 unique plant IDs that have come online since 2010. In the cases of large complexes or aggregates (e.g., the Altamont Pass area in CA), there can be a number of individual plant IDs. A wind farm complex is therefore defined as all aggregated individual plant IDs within the radius of 15 km from a geographic center. Approximately 40 such complexes satisfy this definition, but many of these involve

Page 2 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 1 Cumulative build out of wind power in the US 2001–2014; Y-axis is in units of MW

only two components. This data is then statistically analyzed using simple Gaussian procedures in order to derive basic statistics and distributions for parameters like wind farm outputs and capacity factors. The goal of this study is to combine this statistical data, with real world interrogation of some facilities, to more accurately report on the current US wind situation and to assess overall feasibility and scalability with respect to future deployment scenarios.

The Current State of Wind Energy in America Current Growth Rates In the USA, the three main considerations for development of wind farms at certain locations are: (a) Land areas with reliable wind power (b) Access to the existing transmission line infrastructure (c) An available power purchase agreement (PPA) The weighted combination of these three factors determines the annual rate of wind turbine installation. Currentlywithin the USA, there are now several individual wind farms or wind farm complexes that would be considered as utility scale (e.g., >500 MW nameplate power) and this is a sign of a healthy wind industry in which the combinations of a, b, and c have continuously worked together for sustainable wind deployment. As discussed below, this is no longer the case as of 2013 and the growth rate of US wind energy deployment has dramatically lowered. In most cases, the PPA is driven either at the Federal level (via the production tax credit mechanism – PTC) or at the State level via the adoption of renewable portfolio standards (RPS). At the time of this writing (April 2014), in the USA there is 12,700 MW of new wind currently under construction and 3,300 MW of PPAs were signed in 2014. Over the 7-year period, 2006–2012, installed wind capacity grew from 11,450 to 60,012 MW. This 5.25 factor of growth corresponds to an annual growth rate of 23.7 % or a doubling time of 3 years. This is an astonishing high growth rate, which, if sustained for the next 10 years (2013–2022) would yield a total name plate capacity of 475 GW for wind, which would be approximately 40 % of the entire US electrical generating nameplate capacity. It is precisely this high relative growth rate that promotes policy optimism with respect to the viability of wind energy has a significant component of America’s renewable energy portfolio. Figure 1 below shows the cumulative build out of wind power from 2000 through 2014 in units of megawatts (MW). The black exponential trend line is a fit of the 2001–2012 data (the blue points) with Page 3 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

26 % annual growth. If sustained, this growth rate predicts approximately 110 GW of installed wind nameplate by the end of 2014. The two red points show the actual 2013 and 2014 data. The growth rate from 2013 to 2014 was 7.5 % or 3.5 times lower than the previously sustained rate. Thus, the USA has lost significant momentum in its cumulative wind energy install rate due mostly to uncertainty over the future of the PTC as discussed below.

The Role of the Production Tax Credit The overall waveform shown in Fig. 1 indicates sustained growth over the period 2001 through 2012. However, sustained growth can only occur under conditions of scalability. In general, wind output growth physically means that the number of turbines built and deployed per year must increase. To support this increase, the number of turbine components built per year must also increase and each of these new turbines must have access to transmission line infrastructure. In addition, PPAs must be in place. Historically, these have been in for the form of the Federal Production Tax Credit (PTC). In general, the federal PTC reimburses wind development at the rate of 2.3 cents per kilowatt hour (KWH). AWEA and DOE reports that in 2013, the average metered rate for wind energy in the USA was in the range 5–7 cents per KWH, so the PTC is a significant form of revenue return for wind energy providers. However, in the early months of 2012, the US congress made noises to suggest that the PTC would expire at the end of 2012. This led to an extreme rush to complete various existing projects and indeed Q4 of 2012 witnessed an additional 8.4 GW of wind power installed and deployed. This value can be compared to the values of 4.1, 4.1, 3.3, and 3.5 for the Q4 build outs in 2008, 2009, 2010, and 2011. Thus, it is clearly standard operating procedure for the wind industry to assume that the PTC will expire at the end of any given year and there is a rush to complete projects before that expiration. This lingering uncertainty over the PTC is not helpful to maintain the waveform shown in Fig. 1 and at worst, threatens to mostly throttle the previously established momentum (see also Jenkins 2013). This PTC uncertainty basically explains why 2013 manifests an extreme loss of momentum as only about 1 GW (1.5 % increase from 2012) of new wind energy came on line (and that was almost entirely in Q4). Using the growth rate previous established, approximately 12 GW of new wind resource should have come on line in 2013. This effect is best expressed in a press release issued by the AWEA in late 2013: The supply chain had slowed down during the months preceding the threatened expiration. As a result of the slowdown and the months needed to region momentum, the industry saw a 92 percent drop in installations, down from a record 13,131 MW in 2012 to just 1,087 MW in 2013.

This indicates the momentum is still very dependent on the continuation of the PTC. As it turns out, the PTC did not, in fact, expire at the end of 2012 but has expired at the end of 2013. A complication arose in that a new provision was included in the American Taxpayer Relief Act of 2012 (enacted in January 2013) allowing eligible projects that were under construction before January 1, 2014, to qualify for the PTC. This allowed for a restart of many stalled projects which has therefore helped the 2014 build out to recover from the 2013 abyss. In late 2014, the US congress authorized a short-term extension to the PTC. However, at the time of this writing, it is not clear that the PTC will achieve a more permanent incentive status. While the industry can still survive without the PTC, its absence clearly and strongly affects overall production and deployment of this renewable resource. A simple numerical exercise shows in the year 2021 ~480 GW of nameplate capacity would have been reached under the older growth rate, where the new growth rate (7.5 %) would produce ~100 GW of nameplate capacity.

Page 4 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Table 1 Growth of unit turbine capacity in US wind farms Time frame 2000–2005 2005–2010 2010–2015 2015–

Unit capacity (MW) 1.5 1.8 2.3 3.0

# Needed for 10 GW ~6,700 ~5,500 ~4,400 ~3,300

# Blades needed for 10 GW ~20,000 ~16,500 ~13,000 ~10,000

Required daily construction rate ~18 ~15 ~12 ~9

128

96

64

32

0

200.00 0

560.00 400.00

760.00

Wind Farm Name Plate Capacity in MegaWatts

Fig. 2 Distribution of individual plant ID nameplate capacities

The Role of Increases in Individual Turbine Capacity One of the factors that have aided the build-out of wind energy is the evolution towards larger unit capacity wind turbines. As of the end of 2014, there are approximately 48,000 individual turbines that comprise the current 66 GW of name plate capacity. This means an “average” turbine capacity of 1.38 MW. However, many projects that started in 2014 are making use of individual turbine capacities of 2.5 MW. For the year 2014, 2,500 turbines were installed at average capacity of 1.83 MW. Additionally, 80 % of these installed turbines had rotor blades larger than 50 m. Near future planned projects may move to unit capacities of 3 MW, using the Vestas V90 or V112 units. These increases in unit capacity represent a great gain in efficiency. Independent of unit capacity, turbine installation still requires labor, cranes, and access roads. The physical process of installation is no different for a 1.5 MW turbine than a 3.0 MW turbine other than the need for more heavy lifting cranes. However, at unit capacities of 2.5–3 MW, individual turbine length is 45–55 m. This makes the delivery of blades to the wind site highly problematical as the blade length exceeds the radius of curvature of many roads and highways. Thus, shipping by rail is the only viable option. As described in detail in section “Logistical and Supply Chain Challenges,” annual production and delivery of turbine blades is the likely logistical limiting factor for wind energy build-out. Table 1 summarizes the overall evolution of unit capacity increase and its associated economy of scale as it relates to the task of building out 10 GW of new capacity per year. Very likely, 3.0 MW is the limiting size that can be achieved for land-based turbines due to the difficult logistics of transporting single blades larger than 60 m in length. In addition, overall tower heights must remain less than 152 m to be in compliance with existing FAA regulations (see Cotrell et al. 2014).

Page 5 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

8

6

4

2

0 100.00

350.00

800.00 600.00

1050.00

Wind Farm Complexes (Capacity in MegaWatts)

Fig. 3 Distribution of nameplate capacities for wind farm aggregates Table 2 Listing of individual wind farm complexes with nameplate >500 MW Complex name Alta Wind Energy Center Horse Hollow Capricorn Ridge Fowler Ridge Sweetwater Flat Ridge Cedar Creek Buffalo Gap Majestic Shiloh Aggregated Meadow Lake

# 7 1 1 1 5 3 2 3 2 4 4

Total capacity 1,020 736 663 600 586 570 551 523 511 504 501

State CA TX TX IN TX KS CO TX TX CA IN

Nameplate Capacities of Current US Wind Farms Across the US, there are approximately 900 wind farms that provide the current 66 GW of US name plate capacity. This reduces to an average wind farm size of only 60–70 MW but it would be quite misleading to use this value as a “standard” wind farm capacity for the USA. Most wind farms that have been built within the last 5–10 years are individual aggregate capacities much larger than this value. Figure 2 shows the distribution of individual wind farm capacities, as defined by unique plant identification number (several different plant IDs often exist in the same geographic area that defines a large-scale wind complex) for the 608 operations that are in the analysis sample. Clearly, given the skewed nature of this distribution, the concept of the “average wind farm” has no meaning. In this distribution, the median wind farm is of size ~120 MW and 2/3 of the distribution is less than 180 MW. Many of these individual plant ID operations can be aggregated together into one geographic wind complex (for example, the Altamont Pass area in CA). The distribution of those aggregate capacities is shown below for 37 complexes (Fig. 3). Table 2 lists individual complexes and aggregates that are above nameplate capacity of 500 MW. In principle, these large-scale facilities are what can eventually help to replace coal and NG fired electricity.

Page 6 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014 132

99

66

33

0 0.12

0.25

0.33

0.45

0.53

Wind Farm Capacity Factors

Fig. 4 Distribution of wind farm capacity factors of 2013

Real World Capacity Factors Electrical output of wind farms is dependent on their nameplate infrastructure capacity convolved with the percentage of time that the wind blows at sufficient speeds to rotate the turbines. This percentage is known as Capacity Factor (CF). As measured in the real world, capacities factors have a very large range of values. For instance, the 21 MW Pakini Nui wind farm (in Hawaii) has an extremely high CF of 0.635 due to very reliable trade winds at that location. In contrast, the 275 MW wind farm located on Vansycle Ridge in WA experienced a CF in 2012 of only 17 % suggesting this was a poor site location for wind, at least for that year. Figure 4 shows the distribution of CF for the year 2013 for 594 individual plant IDs that were certified to be in operation for the full year. This distribution is well fit by a normal distribution with mean = 0.32 and standard deviation = 0.07. This real world mean is close to anecdotal value of 1/3 that is often used informally for CF. However, in this normal distribution approximately 15 % of the sample would have CF > 0.4; these wind farms are generally located in the large flat areas of the Midwest. On the other hand, there are approximately 100 locations which are operating at CF < 0.25 and these facilities may be suffering from the micrositing issues discussed below. For some of the complexes and aggregates shown in Table 2 the results are quite mixed: • The Altamont Pass area has an integrated capacity factor which is rather poor, at 0.24. Within the aggregate, individual “arrays” range from 0.195 to 0.30. • Similarly, the Horse Hollow project (still under construction) did not perform well in 2013 with a CF of only 0.23. The Capricorn Ridge location, near the Horse Hollow project, performed much better with CF = 0.35. • The Flat Ridge complex in KS performed extremely well with a measured CF of 0.42. This was followed by the Sweetwater TX location with CF = 0.37. Both of these locations are sited in large, flat, open country. Table 3 shows State average CFs for states with at least 10 separate sites that can be measured to form a reliable average, for 2011, 2012, and 2013 (2014 data not available at the time of this writing): These state average CFs are fairly constant from year to year with the notable exceptions of KS, MN, OK, and WY whose 2013 CFs are at least 5 % lower compared to the 2011 and 2012 performance. This might be due to some large-scale shift in wind patterns, which is one of the many potential manifestations

Page 7 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Table 3 Average capacity factors for the most wind farm active states State CA CO IA ID IL KS MN ND NY OK OR PA TX WA WI WY

2011 CF 0.25 0.28 0.32 0.29 0.29 0.39 0.33 0.39 0.23 0.40 0.29 0.27 0.34 0.30 0.25 0.36

2012 CF 0.26 0.33 0.34 0.34 0.30 0.36 0.35 0.39 0.24 0.41 0.26 0.25 0.35 0.28 0.26 0.35

2013 CF 0.30 0.31 0.31 0.32 0.33 0.30 0.24 0.38 0.25 0.36 0.27 0.29 0.33 0.26 0.29 0.26

Table 4 Average capacity factors European countries Country Denmark Finland Germany Ireland Italy Netherlands Portugal Switzerland United Kingdom a

2011 CF 0.28 0.28 0.19 0.32 0.18 – 0.26 0.20 0.27

2012 CF 0.23 0.24 – 0.28 – 0.20 0.28 0.20 0.27

2013 CF 0.27 0.26 0.18a 0.31 0.21 0.22 0.29 0.20 –

The Fraunhofer Institute reports 2013 German wind CF was 0.166

of climate change. Averaging over the 3 year period, the best performing states (green shading) are ND, KS, and OK while the worse performing states are NY, WI, and OR (orange shading).

Comparison with European Wind Farms Table 4 summarizes average CFs for countries in Europe that have substantial wind facilities which have reported data. These data come from the annual wind reports compiled by the International Energy Agency (IEA). Not all countries report annual values. In general, the CFs are lower in European countries than in individual states in the USA and it is noteworthy that none of the reported data in Table 4 exceeds 0.32 whereas 18 out of the 48 entries in Table 3 (38 %) exceed this value. For the US states listed in Table 3, the average 3-year capacity factor is 0.31 +/ 0.05 while for the listed European countries the average is 0.245 +/ 0.4. Application of the statistical Z-test shows that the difference in these means is significant at 5.3 standard deviations. Thus, statistically, CF’s in the US are significantly higher than those in Europe. This difference is driven mostly by the lack of large, windy, open, and flat areas in Europe as seen in the US in the form of the States of Kansas, Oklahoma, and Texas. The case of Germany is especially noteworthy in terms of the cost effectiveness of renewable energies. A detailed study (Burger 2014) for the year 2013 showed that both wind and solar installations in Page 8 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Germany have quite low capacity factors (16.6 % and 9.5 %, respectively). These low capacity factors can aid in pushing the return on energy investment to below unity – meaning that it takes more energy to produce the infrastructure than is returned to the grid.

Micrositing Issues

Micrositing of individual wind turbines is designed to minimize the amount of disturbed air flow, either by terrain or adjacent wind turbines, to an individual wind turbine. By nature, micrositing is a constrained optimization problem due to the nonlinear nature of turbulence and its associated wake effects. Many different approaches exist to model and simulate these effects (e.g., Gaussian particle swarm – Chan et al. 2012; application of the greedy algorithm – Song et al. 2014). In general, micrositing becomes an important issue when wind farms are built in hilly or mountainous areas, or are located on ridge lines. The typical wind farm or wind farm complex in the US, however, is generally not sited in these locations. While the actual site-specific optimization remains complicated and difficult to employ, the following guidelines are generally used in actual US installations: • For large-scale wind farms (see Fig. 5 below), turbines should be placed 2–3 rotor diameters apart in the plane which is perpendicular to the prevailing wind and 10 rotor diameters apart in the parallel plane. The windfarms in the US Midwest and Texas all conform to this guideline and generally show the highest CFs. • For ridgeline installations, the turbines should be set back from the edges to minimize the impacts of updrafts and other vertical components. The wind farms in OR and WA are mostly located on ridge lines and the complicated air flow caused by this terrain may be responsible for the lower CFs at these sites. • Wind farms should not be placed on steep slopes because of enhanced turbulence as the wind flows over the terrain. Only the Altamont Pass complex in CA is guilty of being constructed in such terrain and this is likely directly responsible for the relatively low integrated CF of this complex, as noted earlier. Figure 5 below shows a Google Earth Image of part of the Flat Ridge complex in KS. The rotor blades are approximately 50 m in length and the individual turbines are clearly well spaced with respect to that dimension. Extant data already show that these locations have the highest CFs in the USA and their relatively wide open domain facilitates the optimum deployment of individual wind turbines, making the problem of micrositing much less severe. Given the available large land area in the flat and windy Midwest, it seems quite clear that this region is optimal for growth.

Wind Farm Construction Costs As wind farms are not required to report their capital construction costs, it is not directly possible to then know this for many wind farms. However, many times these costs are revealed through some “Fun Facts” article in a local newspaper related to the commissioning or opening of the project. Capital costs for some wind farms also appear on some State agency web sites. Reliable information was found for a small sample of wind farms of widely varying nameplate capacity. Figure 6 shows the distribution of these costs, in units of $ per Watt as shown for a sample of 29 wind farms with reported costs built from 2008 through 2012. While N = 29 is low, if this method of obtaining costs does represent random selection, then the derived mean and standard deviation is a good indicator for the entirety of wind farms constructed during this period.

Page 9 of 20

Fig. 5 Microlayout of individual wind turbines Flat Ridge complex KS

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Page 10 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

12

9

6

3

0 1.39

2.05

2.49

3.14

3.58

Wind Farm Construction Costs per Nameplate MW

Fig. 6 Distribution of wind farm construction costs in units of $ per nameplate watt

The average of this real world derived cost is $2.2 + 0.45 per watt. A subsample of 12 projects completed in 2012 is slightly lower at $2.06 +/ 0.33. These real world values are consistent with the 2012 DOE (Wiser and Bolinger 2013) report that cites a capacity weighted average of $1.93 per watt. Figure 6 reveals there is a fairly large range around the average. This suggests that wind installation costs have strong local components. While the sample size is too small to convincingly demonstrate that costs are coming down with time (as claimed by the DOE report), there is some anecdotal evidence to support this. • The Flat Ridge complex in KS shows that the 2009 cost installing 40 2.5 MW turbines was $2.82 per watt while the 2012 costs of installing 294 1.6 MW turbines was significantly lower at $1.70 per watt. • A similar low cost is seen for the Bison Wind Energy Center in ND where 96 3 MW turbines were installed in 2012 at cost of $1.71 per watt. • The most expensive is the 2008 Smoky Hills project in KS, where 56 1.8 MW turbines were installed at cost of $3.47. The 2012 sample can be updated with 11 new projects completed in 2013 that have published costs. That combined sample of 23 produces a real world cost average of $2.00 +/ 0.37, again indicating that these costs are declining in the real world and has a mean value consistent with DOE expectations. Moreover, the data do reveal an economy of scale in that cost reduction occurs when nameplate capacity increases. In general, capacity per wind farm increases due to either a greater number of turbines installed or an increase in unit turbine capacity. Since the same access roads, cranes, and labor is needed to install a 1.8 MW turbine or a 2.5 MW turbine, there is a natural expectation that the $ per watt costs should decrease in response to this. Plotting this cost against nameplate capacity for the 2012 + 2013 sample yields the following: In this data, the average costs for the 13 facilities less than 200 MW is 2.25 +/ 0.37 $ per watt compared to 1.77 +/ 0.31 for the 7 facilities greater than 200 MW. Using the Z-test shows that the difference between these means is 3.1 standard deviations which is significant; the data are consistent with lower costs as nameplate capacity increases. In addition, what is readily apparent in Fig. 7 is that all these large scale facilities have reported real world costs that are less than 2$ per watt. The leader is the Flat Ridge Aggregate which, as of 2013, would have production costs of $1.4 per watt. However, there is an important caveat: this computed cost index

Page 11 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014 3.4 Nameplate 2.8

2.2

1.6

1.0

0

150

300

450

600

Fig. 7 Relation between construction cost ($ per watt on the Y-axis) and wind farm nameplate capacity

does not take into account any costs associated with acquiring new transmission capability. Those costs can significantly increase the overall cost of building wind turbines to produce and export electricity.

The Generation Cost Index Utilizing the CF can produce a cost index per wind farm. This cost index simply divides the capital costs of constructing the nameplate capacity of a given wind farm by the CF for that wind farm. This creates a cost per generated watt or the generation cost index (GCI). Figure 8 shows this cost index for some example newly constructed projects to show the real world range with most projects coming in at cost index of $6–8 per generated watt. However, there are clearly examples of projects with rather large cost index, suggestion that those have turned out to be relatively poor sties for turbine siting. Owing to its high CF (see above), the Flat Ridge project in KS is clearly the most cost effective current installation. When CF is factored into the costs, an important and reasonably linear relation appears in the expected sense that sites with higher CF are more cost effective. Figure 9 shows this trend for the 2012/3 sample. While there are some outliers, the data show a decreasing GCI as capacity factor increases. In addition, the scatter around the relation significantly narrows once CF > 0.3 is achieved. As can be seen, sites with CF > 0.4 are 2–2.5 times more cost effective than sites with CF ~ 0.25. The overall mean value of the GCI is $6.97 +/ 2.4 with most of that deviation coming for wind sites with CF < 0.3. Thus, from an investment efficiency point of view, it is much better to populate KS and OK with wind farms rather than CA and OR.

Permanent Job Creation and Wind Farm Construction

One final aspect of wind farm construction and installation which remains relevant to the real world involves the number of permanent jobs created. As might be expected, this is difficult to ascertain through any formal reporting channels so once again, the limited data set is culled from various PR based publications regarding certain wind farms. Figure 10 below shows the relation between the number of permanent jobs created and the nameplate capacity. These data are consistent with a rubric that there is one permanent job created per 10–15 MW of nameplate capacity. Thus a 300 MW nameplate facility would be expected to create 20–30 permanent jobs.

Logistical and Supply Chain Challenges The logistics of transporting wind turbine components, particularly the turbine blades is one of the biggest challenges faced in the scalable deployment of new onshore wind farms. For example, the latest update to

Page 12 of 20

0.285

0.285 0.651 0.477 0.285 0.285 0.285

0.285 0.285 0.287 0.285 0.285 0.285 0.285 0.285 0.285

2.000

Fig. 8 Example CGI costs for some recent wind projects

0.000

Bison wind energy...

Cedar creek wind farm 2 Caney River Wind... Blue Sky Green Wind...

Chestnut flats

Lakefield Wind Project Kibby wind farm High Lone some Mesa.... Flat ridge 2 Flat ridge 1

Smoky hills 1 Shilo II Red Mesa wind farm

Cassia Wind

High Lone some Mesa... Hatchet Ridge

4.000

3.13

Cost index

8.000

7.85

8.69

8.70 8.36

Capacity Factor

6.64 6.14

6.000

7.80 7.01 7.36

6.66 6.92

10.000

Capacity factor , Cost Index per wind farm

12.17

12.000

11.24

11.24

12.17

14.000

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Page 13 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014 12.00

2012-13 Sample 10.00 8.00 G C I

6.00 4.00 2.00 0.00 0.150

0.200

0.250

0.300

0.350

0.400

0450

0.500

Capacity Factor

Fig. 9 Relation between generated cost index and capacity factor

35

Permanent Job Creation 30 N u m 25 b e 20 r o 15 f J 10 o b s 5

0 0

50

100

150

200

250

300

350

400

450

500

Nameplate Capacity

Fig. 10 Number of permanent jobs created vs nameplate capacity

the Bison Wind Farm in North Dakota, scheduled for completion in December of 2014 is utilizing individual turbines of capacity 3 MW (Siemens 3-MW D3). Individual blade lengths for that turbine are 56.5 m. It is this blade length that is likely the fundamental constraint on the maximum size wind turbine that can be deployed on land. While larger tower heights are needed for larger capacity turbines, those towers are typically composed of three sections that can be individually shipped an assembled on site. Turbine blades, on the other hand, cannot be made in sections (yet) and assembled on site; they need to be shipped as one single unit. According to Cotrell et al. (2014), the maximum length blade that can be transported given radius of curvatures on various rail routes as well as overhead obstacles is 62 m. In addition, many wind turbine components are manufactured internationally so that port delivery is a requirement. For instance, the Port of Longview, just north of Portland, OR, has become one of the

Page 14 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Table 5 Break down of major component suppliers for wind turbines Component Blades Gearboxes/mainshafts Generators Towers Turbines/nacelles

# 12 3 2 17 15

Locations CO (3), ND, SD, TX, IA, IL, AS, IN, PA, NJ OR, IL, PA CO, NC CA, CO(2),TX(4), NE, KS(2),OK(2), MN, IL, WI, MI, MA CA(3), NV, CO, TX(2)KS, IA, AS, MI, FL, SC, NH

nation’s top entry points for wind energy equipment made in Denmark and Germany. To maximize efficiency of port to site component transport, the port designed a specialized system to transport these components from the dock directly to railcars, eliminating the need for trucks as an intermediate transport vehicle. As one example of scale, for a 150 MW wind farm (using 2 MW turbines) it is estimated by Freight Rail Works that approximately 700 trucks and 150 railcars are required for full turbine component delivery that required eight port deliveries. Based on this scale, the required logistics to support, for example, 10,000 MW of annual installation would be ~45,000 trucks (125 trucks per day) and 10,000 railcars distributing component parts from 600 port deliveries. Hence, at some level, annual installation will be limited by logistics even if the component manufacturing supply chain can keep up with the demand rate. For example, an aspirational goal set by the DOE is that by 2030, 20 % of America’s electricity could come from wind power alone. Reaching that level would require installation of at least 7,000 new turbines which requires more than 50,000 shipments of turbine components by rail annually by 2018. As of 2013, there are 550 separate manufacturing bases that sell to the US wind industry. Most of these are minor players, but there are approximately 50 Tier 1 facilities. These Tier 1 facilities are devoted to individual major components, such as blades or towers. Table 5 shows the breakdown of these facilities as laid out in a 2013 NREL report (James and Goodrich 2014). In 2012, tower production equated to 7.4 GW of power, Blade production 8.1 GW, and Nacelle production 13.1 GW. This indicates that the component supply chain is not in equilibrium and has weak link production points. In 2012, that weak link would have been tower production. In 2012, the 12 blade production plants combined production was 12,500 individual blades (AWEA 2013, 2014). This production is consistent with building 4,200 2 MW wind turbines (whose combined annual nameplate would be 8.4 GW). A useful footprint of blade production is supplied by the Vestas plant in Windsor CO. This is a 400,000 square foot facility that produces 1,800 blades per year requiring 650 workers. Hence, to reach an annual install goal of 10,000 MW with 2 MW turbines would require the production of 15,000 blades per year or the equivalent of 8 Windsor plants. To therefore reach these kinds of goals, the US public–private partnership would have to make a major investment in new production line facilities to avoid blade production being the limiting factor in wind energy build out. The production plant in Brighton CO is dedicated to producing the 57.5-m blades for the new Vestas 117 3.3 MW wind turbine to be deployed in Denmark. These blades are then transported by rail from Colorado south to the Gulf ports (no mountains are encountered along this path) for direct shipment to Denmark. The current production capacity of this factory is 1 blade per day (or 120 turbines per year). These issues illustrate a tension that is emerging in the construction of future wind farms. On the one hand, it is clear that an economy of scale is reached when that future wind farm utilizes turbines of capacity 2.5–3.0 MW. On the other hand, delivery of turbine blades may significantly slow down the overall construction of the plant such that an equivalent nameplate capacity plant consisting of say 1.6 MW turbines may be easier and quicker to build as blade delivery is less complicated. In any case, supply chain and transport logistics must evolve to be a strong component of future wind farm siting. Page 15 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 11 Rail transport of 55 m blades using custom railcars

Although transportation logistical difficulties most definitely exist, they can be overcome, but likely not at a scalable rate. Here are two example success stories: (a) In Gloucester MA, the Blackburn Industrial Park has now installed 3 2.5 MW wind turbines each with blade length of 49 m. The blades were shipped to Gloucester Harbor. The final mile of transport to the site was done with trucks navigating surface streets to reach the site, thus indicating that it is indeed possible to maneuver a 160 ft long item via surface streets. Obviously this was a slow process that required partial street closures and police directing traffic. But this had to be done only nine times, to populate the three turbines. However, clearly this scenario wouldn’t sale if there were 300 turbines that were located at that site. (b) In July of 2012, Transportation Technology Services (TTS) were challenged with shipping a 55 m blade from the Windsor plant in Colorado to a new wind energy project in New England, some 1,800 miles away. To meet this challenge, TTS had to design, build, and test a custom railcar loading mechanism as well as new support pivoting structures that would allow the blade to negotiate curves without being stressed. The image below shows this unique design in action (Fig. 11). As of late 2014, it appears that the most common turbine being installed is one with unit capacity of 2.5 MW which require blades that are 50 m (164 ft) long. Hence, logistical problems are here and in some cases have required the construction of new, temporary roads, so the blades can be delivered. This, of course, adds to the total project costs. In addition, some newly planned wind farms (e.g., the potential Steens Mountain Oregon facility) are to be sited at altitude which makes the delivery of 50 m blades by rail impossible and by truck quite problematical. As a result of this challenge, novel new transport systems are being developed. For instance, Goldhofer (a German transportation company) has designed a flatbed combination trailer than can transport a turbine blade up to length 62 m on standard European highways. Aeroscraft is designing a unique Hydrogen dirigible (airship) that would lift and deliver individual blades to remote locations at a cost which would be less than current Helicopter rates, particularly for delivery at high altitude.

Summary: Realizing the Potential of Wind As previously discussed, there are three basic factors that need to come together in some optimal way to support sustained growth in wind farm deployment.

Page 16 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

• A favorable incentive/PPA climate • Access to transmission line infrastructure at windy locations • A robust supply and transport chain that can continue to deliver components to various remote sites Clearly the last 2 years have dramatically showed the effect of the PTC as an incentive for wind energy build out. The future of the PTC is currently unknown and is the subject of much speculation. For instance, in the longer term, Bloomberg New Energy Finance projects that the US wind market may be able to sustain approximately 6,200 MW per year of new wind capacity from 2017 onwards. This rate is significantly below that which is illustrated in Figure 1 and indicates that while the wind industry can survive the permanent expiration of the PTC, it is not likely to thrive and therefore not likely to meet various aspirational goals. Moreover, the current “Shale Boom” in the USA has most definitely impacted investments in wind energy, as without the PTC, wind energy generation and natural gas generation have similar costs (see Trembath and Jenkins 2012). Momentum is also highly dependent on continued build out of new transmission. AWEA estimates that 10,000 MW of new transmission capability was completed in 2013, but that new capability is not dedicated to just wind, but to all potential sources. Annual build out of new transmission must be considerably larger than this value to help foster the build out of wind energy. For example, the commitment and forward thinking by the State of Texas to simultaneously develop 18,500 MW of new transmission line infrastructure while building out large-scale wind farm projects (e.g., Horse Hollow, Capricorn Ridge) has allowed them to be the clear leader in wind-based electrical power in the nation. As of July 1, 2014, TX had a combined capacity of 12.75 GW more than twice the leading competitor of CA at 5.38 GW. In fact, this capacity in TX exceeds the combined capacities of OR, WA, CO, KS, OK, and ND (the state with highest wind energy density) despite continuing build out in those six states. This is entirely because of the commitment to transmission lines, largely financed through public bonding in TX. In contrast, ND with a significantly lower population and much lower infrastructure has had a very difficult time exporting its abundant wind energy due to transmission line limitations. This demonstrates the important of extant or new transmission line infrastructure and the continual build-out of wind energy capacity additions. The wind energy projects in TX, for instance, clearly would have not been possible if there were not a parallel effort to increase transmission capability. Similarly, a private industry project is underway in Michigan to increase infrastructure in the “thumb” region to support more wind farm development. This project involves 140 miles of new 345 kV lines with total capacity of 5,000 MW. This incentive for this project comes from the 2008 renewable energy standard enacted by Michigan that requires utilities to get 10 % of their power from renewable sources by 2015. Overall, the performance of recently constructed wind farms in the USA is generally favorable and there are a few large sites that have experienced real world capacity factors larger than 40 %. This suggests a national strategy to continue to invest in those locations that exhibit high real world CF. However, national strategies and individual State’s RPS may be in conflict or even compete against one another. One example of this is provided by both OR and WA in that each State has aggressive RPSs and even though they share the Columbia River hydroelectric resource, this is excluded from their RPSs. This has led to a significant effort at wind farm construction that turns out to be sited at locations with relatively low CF. Given the previously discussed logistical transport issues it is unclear what unit capacity wind turbine leads to the best build out scenarios. Obviously it makes little sense to plan a wind farm at a windy location if you can’t easily transport the turbine blade to that location. Currently there are many obstacles. These obstacles include FAA blade tip height restrictions, trucking of large diameter towers, hoisting larger nacelles onto increasingly taller towers, and trucking longer and larger blades. And overall weight issues for trucking the components.

Page 17 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

More specifically, the FAA height barrier is currently 152 m high. Planned wind farms from 100 to 140 m in height would need this restriction lifted in order to operate with less risk. Any structure over the 152 m height triggers FAA lighting requirements and a review process that could prevent construction either entirely or cause expensive delays. The FAA restrictions impact blade manufacturing from considering production of any wind turbine exceeding 152 m. This restriction also excludes 320,000 km2 or roughly 1,000 GW of deployable land in areas where smaller wind turbines will not work due to lower wind speeds. One fifty two meter or larger wind towers are needed to effectively operate in these regions. Nacelle hoisting is another barrier to be addressed. Currently, single crane picks are extremely large in order to carry out the task of raising hubs to larger and larger towers. There is a risk of crane shortages as there is approximately less than 20 cranes nationwide capable of working with 152 m and greater wind turbines projects involving 3 and 5 MW nacelles, the latter size needed for offshore facilities. These massive cranes also encounter logistical challenges due to limited access to complex terrain and a need for wider access roads due to weight and size. As a result of these concerns, smaller, less efficient, wind turbines, and shorter towers are used more frequently. For instance, there is a fleet of approximately 90 smaller cranes capable of working with 140 m wind turbines that are being used for current wind projects. Hypothetically, if this issue of crane size and availability is addressed, it will nearly double the deployable land available to 614,000 km2 or approximately 200,000 new MW of available wind power achieved by increasing hub height from 96 to 140 m. In so doing, states which have lower average wind speeds now become viable for wind power production with infrastructure improvements and readily available cranes for these higher hub installations. Overall wind energy build out in the USA has enjoyed a very successful growth curve up until 2012 but now faces a very uncertain future. What was once a promising component of the national renewable electricity generation portfolio is now being complicated by real world national politics and real world logistical issues. The overall potential in wind power blowing across the US is enormous and it is easy to construct hypothetical scenarios that could build as much as 1 TW of nameplate capacity in on- and offshore US wind. To realize this future, however, investments need to be made in component production and component delivery transport systems. Without these investments, current wind farms may end up, in 20 years, as rusted vertical monuments, testifying to the inability to deploy a scalable solution.

Conclusion The generation of electricity by wind remains a viable endeavor for the USA. The US aspirational goal of 30 % wind by the year 2030 would result in about 350,000 MW of wind name capacity. As, of the end of 2014, the US stands at 66,000 MW – 20 % of what is needed. Given the significant loss of wind energy build out momentum in 2013/14, it will make achieving this aspirational goal quite problematic. Indeed, to achieve this would now require annual capacity additions of 17,500 MW – much larger than any year to date. Moreover, the scalability of wind build out is now being challenged by the following two main issues: 1. The production tax credit – this has expired at the end of 2013. Projects that were not under construction prior to January 1, 2014, are ineligible for this credit. As a result, 2014 will likely see little added wind production. Indeed, as of July 1, 2014, 4,350 MW of new utility-scale generating capacity was brought on line but more than ½ of that was in the form of new natural gas fired electricity. Only 675 MW of new wind generation occurred during this first half of 2014 and most all of that occurred in Texas Page 18 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

Fig. 12 Offshore wind resource potential for NW US Coast

2. Supply chain logistics – while it’s clear from the data that large unit wind turbine capacity leads to more cost effective wind farms, this is balanced by the logistical problems associated with the delivery of individual turbine blades. For instance, if one were to build a 900 MW wind farm, it is very likely in the real world that such a farm could be built significantly faster by installing 600 1.5 MW wind turbines rather than 300 3 MW wind turbines On the bright side, the potential in offshore wind for the USA is large and tapping that resource now seems like the most likely pathway to reach the US aspirational goal. For the UK, the measured average CF for OFF shore wind from 2011 to 2013 is 37 %, significantly larger than the UK onshore CF of 27 %. Similarly, the USA has several large-scale offshore wind sites that would have CF’s in the range of 40–45 %. These promising sites include the North–south axis of Lake Michigan, the East–west axis of Lake Superior, and, most importantly, the nearshore coastlines of WA, OR, and Northern CA. The west coast of the USA is not subject to large-scale cyclonic disturbances, unlike the US Southeast coast and so there is little concern for large-scale damage of an offshore complex. Figure 12 below shows the most current NREL wind resource map for this region. The red areas indicate incident power density at 50 m height of 600–800 W per square meter and the blue areas correspond to 800–1,600 W per square meter. Opening up this resource, at individual wind turbine capacity of 5–8 MW, is likely the only way that the USA can reach the 30 % by 2030 aspirational goal. The build out of wind energy in the USA has clearly lost momentum in recent years. Much of this momentum loss can be attributed to continued investment in natural gas, via hydraulic fracking, as de facto US energy policy. During the period 2002–2012, wind energy was growing at a substantial rate and continued investment, especially keeping the PTC, turned on, allowed for a very bright future in wind energy. Now that future is in peril. Given the proven cost effectiveness of wind as a viable alternative to fossil fuel-based generation of electricity, the USA soon needs to make a more informed decision,

Page 19 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_25-1 # Springer-Verlag Berlin Heidelberg 2014

balancing the need for capacity additions with the need for total greenhouse gas reductions, in order to sensibly more forward.

References American Wind Energy Association (2013) 2013 US wind industry annual market report American Wind Energy Association (2014) 2014 US wind industry annual market report Burger B (2014) Electricity production from solar and wind in Germany in 2013. Fraunhofer Institute report Chan W et al (2012) Wind farm micro-siting by Gaussian particle swarm optimization with local search strategy. Renew Energy 48:276 Cotrell J et al (2014) Analysis of transportation and logistics challenges affecting the deployment of larger wind turbines: summary of results. NREL report James T, Goodrich A (2014) Supply chain and blade manufacturing considerations in the global wind industry. NREL report Jenkins J (2013) Can the American wind energy industry survive without the PTC? The energy collective Sickinger T (2012) BPA braces for strong spring runoff, excess power and wind cuts. Oregonian Newspaper Article 6 Apr 2012 Song M et al (2014) Optimization of wind farm micro-siting for complex terrain using greedy algorithm. Energy 67:454 Trembath A, Jenkins J (2012) Gas boom poses challenges for renewables and nuclear. Breakthrough Institute report Wiser R, Bolinger M (2013) 2012 wind technologies market report. Department of Energy report

Page 20 of 20

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_30-1 # Springer-Verlag Berlin Heidelberg 2014

Biogas Production via High-Solid Anaerobic Codigestion of the Silage of Winter Triticale with Mixed Waste of Fruit and Vegetable and with Biogas Biscuit Meal of EKPO-EB Type Katerina Kasakova* and Jiri Rusin Institute of Environmental Technologies IET, VŠB – Technical University of Ostrava, Ostrava, Czech Republic

Abstract This paper presents the results of semicontinuous mesophilic high-solid anaerobic codigestion of the silage of winter triticale with mixed waste of fruit and vegetables and with biogas biscuit meal of EKPOEB type. The digestion was carried out in a partially mixed horizontal fermenter (active volume 0.3 m3). This test was separated to three phases. During 60 days by the start-up phase, the intensity of biogas production 2.49 mN3m 3d 1 with the average methane content of 45 vol.% at the average load of 2.55 kgVSm 3d 1 was achieved. During the second phase (31 days) with dosing of biogas biscuit meal of EKPO-EB type and mixed waste of fruit and vegetables, the intensity of biogas production 4.05 mN3m 3d 1 with the average methane content of 57 vol.% at the average load of 4.29 kgVSm 3d 1 was achieved. During the last phase (38 days) with dosing of only EKPO-EB type, the intensity of biogas production 8.76 mN3m 3d 1 with the average methane content of 57 vol.% at the average load of 10.49 kgVSm 3d 1 was achieved. It was confirmed that single-substrate batch is not optimal for starting the process. The batch from several different substrates is more appropriate for the nutrient microorganisms need.

Keywords Anaerobic digestion; Winter triticale; Biowaste; Biogas; Horizontal fermenter

Introduction Anaerobic digestion is one of the best methods for processing biological waste mainly from agriculture and food industries. In practice, it proved to dozens of different technologies for wet anaerobic digestion and the last time a variety of technology of high-solid digestion (Sch€afer et al. 2006; Kumar et al. 2010). Most of these technologies are designed for biogas on a large scale. The space for reaction of high-solid biogas stations is most frequently conceived like a gastight box or garage (Brummeler and Koster 1989; Radwan et al. 1993; Kalyuzhnyi et al. 2000; Mumme et al. 2010). For those interested in the production of biogas on a smaller scale isn’t yet on the market many technologies. For this reason, VSB – Technical University of Ostrava in cooperation with CERNIN Ltd. –develops the technology of mobile potentially low-cost biogas plant suitable for low- and high-solid digestion. The core of technology is newly developed fermenter consisting mainly of pliable plastic bags. The use of plastic bag like anaerobic fermenter is very common, especially in developing countries (Jewell 1979; Wanjihia 2010). In America Woods End Laboratories (William and Brinton 2006), the

*Email: [email protected] Page 1 of 11

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_30-1 # Springer-Verlag Berlin Heidelberg 2014

production of biogas was tested as an early stage before composting in the bags AG-Bag. Previous studies (Linke et al. 2002; J€akel 2004; Linke 2004) show that this method is economically viable only if it is applied on a large scale and a good profitability can be expected only in the hottest months of the year. The technology works the most effectively in the period from June to September. Insulation of the bags from top has neither a significant effect from October. A lot of anaerobic tests with various potentially treatable substrates with solids above and below 10 % were performed. Through a gradual development of the tests, the most appropriate conditions for anaerobic digestion of biomass/biowaste with content of low to very high moisture were identified (Chen et al. 2008; Kumar et al. 2010; Duan et al. 2011). Biowaste from the manufacture of confectionery (raw materials unsuitable for consumption by humans) is possible, in most cases, to classify as highly concentrated and valuable raw material for the production of biogas. The authors (Xiaohu et al. 2013) appeared that during high-solid anaerobic digestion of dewatered sludge, the addition of food waste not only increased the system stability but also greatly increased the volume production of biogas in comparison with mono-digestions. In the literature it is possible to trace a series of works dealing with mesophilic single-stage anaerobic digestion of different mixtures of fruit and vegetable waste or biological waste generated during their processing. For example, authors (Bouallagui et al. 2005) point out that the main problem of single-stage digestion is too fast acidification or noticeable accumulation of lower fatty acids.

Materials and Methods Laboratory Model Semicontinuous mesophilic high-solid anaerobic codigestion of the silage of winter triticale with mixed waste of fruit and vegetables and with biogas biscuit meal of EKPO-EB type was carried out. The partially mixed horizontal bag digester was used. The total volume of the bag was 0.5 m3 and working volume was 0.3 m3. The laboratory model of fermenter consisted of a circular steel front equipped with a short cylindrical nozzle of diameter 0.5 m to fasten the bag and, further from the internal bag of fermentation and external (tempering) bag, a heater unit for recirculation of hot air between the bag and a thermal insulation (mineral wool, 30 mm). The heating unit located at the rear end of the model consisted of two electric heaters (2  400 W), air duct fan (type Vents VKMz 100 with a controller of rotations), aluminum pipes for air recirculation (2  4 m), and a thermostat with a meter of electricity consumption. Vaulted ends of fermentation bag were blown by warm air (75  C) from the fan. Further, warm air flew in the space between internal and external bag toward the front of fermenter (heating of the batch to 40  C). The cooled air (30–35  C) was sucked from the front to an aluminum pipeline back to the fan and the heating coils. Air heating was chosen with regard to the limitation of piping in contact with the bag which might eventually physically interfere. Typical power consumption of the heater was 333 W. The model fermenter was placed horizontally on the laboratory desk. A container with volume 0.03 m3 was used for dosing the feed mixture. The container was connected to the self-priming pump with a rubber impeller (with power 1,500 W). The stirring of the fermenter content was provided in two ways. The recirculation was switched-on manually several times a day using the selfpriming pump. The dose was sucked-off by a hose with a diameter DN50 (located at the bottom to the rear of the bag) and returned to the opposite end of the bag (to the front) or vice versa. This stirring by the pump was performed daily at 9.00 for 15 min before dosing and after each dose of substrate for 5 min.

Page 2 of 11

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_30-1 # Springer-Verlag Berlin Heidelberg 2014

A constant volume of dose was ensured by overflow of digestate over the hose DN50 bended at a constant height (slightly above the axis of stirrer). The overpressure of biogas (10–100 Pa) was checked by a manometer (type U) and belayed by a liquid lid. Biogas flew to a cube-shaped sink for biogas (volume 0.02 m3) placed above the steel front of fermenter. Biogas production was measured by a drumtype gasometer. Biogas composition was measured by a portable IR/electrochemical analyzer. The laboratory model fermenter was developed in a research project (TA01020959) supported by the Technology Agency of the Czech Republic in cooperation with the Centre for Environmental Technology, VSB – Technical University of Ostrava and the company CERNIN Ltd. The aim of the project is to develop and launch a biogas station, nondemanding for design, low-cost, and potentially mobile, suitable for small farms, food production, etc. The company CERNIN Ltd. is now ready for producing the mobile container biogas plants working in accordance with the model described above. The cost of the typical 10 kWe unit will be around 185,000 EUR. The view on the model at laboratory digestion is shown in Fig. 1.

Feedstock The starting batch was created from mixing materials imported from the biogas plant Pustějov II. The composition of batch (300 kg) was as follows: the silage of triticale of 20 wt% and the starting digestate – inoculum of 80 wt%. The digestate with the content of dry matter of 5.7 wt% came from the first fermenter of biogas plant Pustějov. The dry matter content of the starting batch was 13 wt%. Because of the swelling of silage, this batch had the character of “rigid” or “non-pumpable” mass even at this low dry matter. During the model test, the silage of winter triticale Agostino was fermented with several other biowastes from the food industry. Among these wastes belonged the biogas biscuit meal of EKPO-EB type, the biological waste generated from sorting fruits and vegetables, and meat and bone meal. Further the small doses of Ca(OH)2, NaOH, KOH, and NaHCO3 in order to increase the pH were added. The silage of winter triticale Agostino is an agricultural substrate for the production of biogas, but the biogas plants in the Czech Republic are only used minimally. The substrate is characterized by medium content of total dry matter (approximately 45 wt%) with organic matter content in the dry matter (loss on ignition of the dry matter) of approximately 96 wt% and the high percentage of starch and fiber in the dry matter. The substrate was put into the digester and further wasn’t stored. The biscuit meal of EKPO-EB type is a substrate which is consisted of a number of biowastes from confectionary production. The main components of the biscuit meal include waste wafer mass, waste fat dough, chocolate filling, fat mass filling, and starch from jelly production. The substrate is characterized

Fig. 1 The model horizontal fermenter Page 3 of 11

Handbook of Renewable Energy DOI 10.1007/978-3-642-39487-4_30-1 # Springer-Verlag Berlin Heidelberg 2014

Table 1 Substrate and starting batch parameters

Winter triticale silage Starting digestate (inoculum) Biscuit meal EKPO-EB Meat and bone meal Mixed waste of fruit and vegetables Starting batch

pH – 4.1 8.2 5.3 5.2 3.8 7.0

TS wt % 45.1 5.7 91.7 96.4 10.7 13.0

VS wt % 43.6 4.6 90.3 72.3 9.9 11.3

VSTS wtTS % 96.6 80.9 98.5 75.0 92.7 87.0

by the high content of total dry matter (approximately 92 wt%) and loss on ignition of the dry matter of approximately 99 wt%. The total dry matter contains high percentage of starch, carbohydrates, and fat by low percentage fiber. The biscuit meal was supplied by the company CERVUS, Ltd. Olomouc. The substrate was stored in the original packaging – paper/PE bag about volume of 0.05 m3 with limiting access of fresh air. Meat and bone meal is a substrate with a high content of nitrogen and phosphorus, which can coferment only in limited additions. Apart from the mineral component, the meal is characterized by very good anaerobic degradability of proteins and lipids. For the test, the sample of meat and bone meal from the third category imported from a rendering plant REC Ltd. Mankovice was used. The content of total dry matter is high (about 96 wt%) by loss on ignition of the dry matter of 75 wt%. The ration C:N does not exceed 5:1. The substrate was stored in a barrel with a volume of 0.05 m3 with an airtight seal. The biological waste generated from sorting fruits and vegetables before the distribution to retail networks (supermarkets, hypermarkets) was tested. The biological waste was delivered by Hortim International, Ltd. Co. It is a mixture of commonly grown fruits and vegetables from the European Union (e.g., apples, carrots, onions, cabbage, etc.) with citrus fruits and bananas. Biowaste (HORTIM) with content of solids average of 10 wt% contains mainly saccharides and a minimum of anaerobic indecomposable substances. The materials was shredded of meat grinder (snail) to particles