Internal combustion engines [3d ed]

Citation preview

Internal Combustion Engines: Performance, Fuel Economy and Emissions

Combustion Engines and Fuels Group Organising Committee: Prof Paul Shayler (Chair)

University of Nottingham

Dr Frank Atzler

Continental Automotive

Prof Choongsik Bae

KAIST

Hugh Blaxill

Mahle Powertrain

Brian Cooper

Jaguar Land Rover

Prof Colin Garner

Loughborough University

Dr Roy Horrocks

Ford Motor Company

Dr Mike Richardson

Jaguar Land Rover

Dr Martin Twigg

Consultant

Dr Matthias Wellers

AVL Powertrain

Steve Whelan

Clean Air Power

Prof Hua Zhao

Brunel University

The Committee would like to thank the following supporters: Automobile Division

Internal Combustion Engines: Performance, Fuel Economy and Emissions

27–28 NOVEMBER 2013 IMECHE, LONDON

Oxford Cambridge Philadelphia New Delhi

Published by Woodhead Publishing Limited 80 High Street, Sawston, Cambridge CB22 3HJ, UK www.woodheadpublishing.com www.woodheadpublishingonline.com Woodhead Publishing, 1518 Walnut Street, Suite 1100, Philadelphia, PA 19102-3406, USA Woodhead Publishing India Private Limited, G-2, Vardaan House, 7/28 Ansari Road, Daryaganj, New Delhi – 110002, India www.woodheadpublishingindia.com First published 2013, Woodhead Publishing Limited © The author(s) and/or their employer(s) unless otherwise stated, 2013 The authors have asserted their moral rights. This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but the authors and the publisher cannot assume responsibility for the validity of all materials. Neither the authors nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from Woodhead Publishing Limited. The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited for such copying.

Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe.

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library.

Library of Congress Control Number: 2013954934

ISBN 978 1 78242 183 2 (print) ISBN 978 1 78242 184 9 (online)

Produced from electronic copy supplied by authors. Printed in the UK and USA. Printed in the UK by 4edge Ltd, Hockley, Essex.

CONTENTS

GASOLINE ENGINES C1370/018

Ultra boost for economy: realizing a 60% downsized engine

3

concept J W G Turner, A Popplewell, S Richardson, Jaguar Land Rover Ltd; A G J Lewis, S Akehurst, C J Brace, University of Bath; S W Bredda, GE Precision Engineering, UK C1370/023

Part-load performance and emissions analysis of SI combustion

19

with EIVC and throttled operation and CAI combustion M M Ojapah, Y Zhang, H Zhao, Brunel University, UK C1370/037

An optical investigation of a cold-start DISI engine startup

33

strategy P Efthymiou, M H Davy, C P Garner, G K Hargrave, J E T Rimmer, Loughborough University; D Richardson, J Harris, Jaguar Land Rover Ltd, UK

REAL WORLD EMISSIONS ASSESSMENT C1370/009

Emissions from Euro 3 to Euro 6 light-duty vehicles equipped

55

with a range of emissions control technologies J May, C Favre, D Bosteels, Association for Emissions Control by Catalyst, Belgium C1370/015

Engine emissions measurements from passenger cars at two different locations within the metropolitan area of Antwerp in Belgium and further statistical analysis D Savvidis, K Bounos, B Sochacki, University of Antwerp, Belgium; C Ioakimidis, University of Deusto, Spain

67

C1370/020

Analysis of the influence of vehicle usage pattern on the

79

optimum range extender driveline configuration for a compact-class passenger car M D Bassett, J Hall, M Warth, MAHLE Powertrain Ltd, UK

COMPRESSION IGNITION ENGINES C1370/039

Benefits of cylinder deactivation on a diesel engine and

95

restrictions due to low boost J P Zammit, M J McGhee, P J Shayler, University of Nottingham; I Pegg, Ford Motor Co., UK

FUELS C1370/036

Dual-fuel heavy duty engines drive the dash for natural

111

gas S Whelan, Clean Air Power Ltd, UK; H C Wong, Clean Air Power Ltd, USA C1370/040

High efficiency and low emission natural gas engines for

123

heavy duty vehicles M E Dunn, G P McTaggart-Cowan, J Saunders, Westport Innovations, Canada C1370/005

Meeting the challenges associated with low-carbon

137

alternative fuels through advanced CAE technologies A J Smallbone, K M Banton, A N Bhave, J Akroyd, M D Hillman, cmcl innovations; R C R Riehl, M Kraft, University of Cambridge; N M Morgan, Shell Global Solutions, UK C1370/011

Laminar burning velocity as a fuel characteristic: Impact on vehicle performance R Cracknell, B Head, S Remmert, Y Wu, Shell Global Solutions, UK; A Prakash, Shell Global Solutions, USA; M Luebbers, Shell Global Solutions, Germany

149

FUELS, COMBUSTION AND SPARK IGNITION ENGINES C1370/004

Expanding applications of tracer-based two-line PLIF

159

technique for combustion measurements M Anbari Attar, H Zhao, M R Herfatmanesh, Brunel University, UK C1370/027

Lean boost CAI combustion in a 2-stroke poppet valve GDI

169

engine Y Zhang, H Zhao, Brunel University, UK C1370/026

Combustion process and PM emission characteristics in a

179

stratified DISI engine under low load condition H Oh, J Jung, C Bae, Korea Advanced Institute of Science and Technology, Republic of Korea; B Johansson, Lund University, Sweden C1370/002

Particulate matter emissions from gasoline direct injection

193

spark ignition engines F Leach, R Stone, University of Oxford; D Fennell, D Hayden, D Richardson, N Wicks, Jaguar Land Rover Ltd, UK

AFTERTREATMENT OPTIONS, DESIGN AND CONTROL C1370/006

Advanced, combined exhaust aftertreatment systems for

205

light-duty diesel engines to meet next emission regulations Th Körfer, Th Schnorbus, B Holderbaum, Th Wittka, FEV GmbH, Germany C1370/038

Advanced integrated exhaust aftertreatment systems and

219

the mechanisms of NOx emissions control M V Twigg, TST Ltd, UK C1370/041

Gasoline direct injected particulate emissions control at stage 6 P Rounce, M Brogan, P Eastwood, Ford Motor Co., UK

AUTHOR INDEX

231

This page intentionally left blank

GASOLINE ENGINES

This page intentionally left blank

Ultra boost for economy: realizing a 60% downsized engine concept J W G Turner, A Popplewell, S Richardson Powertrain Research, Jaguar Land Rover Ltd, UK A G J Lewis, S Akehurst, C J Brace Department of Mechanical Engineering, University of Bath, UK S W Bredda GE Precision Engineering, UK

ABSTRACT The paper discusses Ultra Boost for Economy, a collaborative project part-funded by the Technology Strategy Board, the UK’s innovation agency. ‘Ultraboost’ combines industry- and academia-wide expertise to demonstrate that it is possible to reduce engine capacity by 60% and still achieve the torque curve of a large naturally-aspirated engine, while encompassing the attributes necessary to employ such a concept in premium vehicles. In addition to achieving the torque curve of the Jaguar Land Rover 5.0 litre V8 engine, the main project target was to show that such a downsized engine could in itself provide a viable route to a 35% reduction in vehicle tailpipe CO2, with the target drive cycle being the New European Drive Cycle. In order to do this vehicle modelling was employed to set part load operating points representative of a target vehicle and to provide weighting factors for these points. The engine was sized by using the fuel consumption improvement targets while a series of specification steps, designed to ensure that the required full-load performance and driveability could be achieved, was followed. The intake port in particular was the subject of much effort, and data is presented showing its performance versus a current stateof-the-art production design. The use of a test-cell-based charging system, while the engine-mounted charging system was being developed and characterized in parallel, is discussed. This approach allowed development of the base engine and combustion system without the complicating effects of the charging system performance coming into play. Finally, data is presented comparing the performance of the engine in this guise with that when the engine-driven turbocharger was used, showing that the peak torque and power targets have already been met.

ABBREVIATIONS ATDC BDC BMEP BTDC CAHU

After top dead centre Bottom dead centre Brake mean effective pressure Before top dead centre Combustion air handling unit

___________________________________________ © The author(s) and/or their employer(s), 2013

3

CPS DCVCP DF DI EAT EGR IEM IMEP JLR MOP NA NEDC

pbrake pind PCP SI TDC UB VSwept WCEM

mech

1

Cam profile switching Dual continuously-variable camshaft phasing Downsizing factor Direct injection Exhaust after treatment Exhaust gas recirculation Integrated exhaust manifold Indicated mean effective pressure Jaguar Land Rover Maximum opening point Naturally-aspirated New European Drive Cycle Brake mean effective pressure Gross indicated mean effective pressure Peak cylinder pressure Spark-ignition Top dead centre Ultraboost (Ultra Boost for Economy) Swept volume Water-cooled exhaust manifold Mechanical efficiency

INTRODUCTION

1.1 Spark-ignition engine downsizing Spark-ignition (SI) engine downsizing is now established as a ‘megatrend’ in the automotive industry, providing as it does an affordable solution to the twin issues of reducing tailpipe CO2 emissions and improving fuel economy while providing improved driveability from gasoline engines. The ‘downsizing factor’ is here defined to be DF 

VSweptNA − VSweptDownsized VSweptNA

,

Eqn 1

where DF is the downsizing factor, VSweptNA is the swept volume of a naturallyaspirated engine of a given power output and VSweptDownsized is the swept volume of a similarly-powerful downsized alternative. To the OEM the attractions of a downsizing strategy include that gasoline engine technology is very cost-effective to produce versus diesel engines (especially when the costs of the exhaust after treatment (EAT) system are included), that there are still significant efficiency gains to be made due to the losses associated with the 4stroke Otto cycle, and that pursuing the technology does not entail investing in completely new production facilities (as would be required by a quantum shift to electric or fuel-cell vehicles, for example). The advantages of downsizing a 4-stroke spark-ignition (SI) engine stem chiefly from shifting the operating points used in the engine map for any given flywheel torque, so that the throttle is wider-open to the benefit of reduced pumping losses. At the same time, the mechanical efficiency increases, this being defined as

4

mech  where

pbrake , pind

Eqn 2

mech is the mechanical efficiency, pbrake is the brake mean effective pressure

(BMEP) and

pind is the gross indicated mean effective pressure (IMEP) [1].

Thermal losses also improve and, in the case of downsizing and ‘decylindering’ from a Vee-configuration engine to an in-line one, crevice volume losses can be markedly reduced and there are potentially significant bill of materials (BOM) and manufacturing cost savings, too. These savings can help to offset the additive technologies required to recover the power output, because some means of increasing specific output has to be provided to retain installed power in a vehicle. This is normally done by pressure charging the engine, with turbocharging generally being favoured because it allows some exhaust gas energy recovery. There are significant synergies with other commonplace technologies such as direct injection (DI) and camshaft phasing devices, too [2]. To date production downsized engines have generally been configured with a DF in the region of approximately 40%, with one research engine shown with this value at 50% [3]. Consequently the Ultraboost project was formed with the major tasks of specifying, designing, building and operating an engine with a minimum of 60% downsizing factor. Through the results obtained it was intended to establish whether 60% is a practical limit for the approach or whether there would be benefit in further downsizing, and that such a downsized engine could in itself provide a route to a 35% reduction in vehicle tailpipe CO2 (importantly, without the use of hybridization other than a Stop/Start system). Consequently a primary aim of the project was to achieve the power and torque curves of the Jaguar Land Rover 5.0 litre AJ133 naturally-aspirated V8 engine with a pressure-charged engine of approximately 2.0 litre capacity. These curves are reproduced in Figure 1, together with the associated BMEP values required from the downsized engine at peak torque, peak power and 1000 rpm. The CO2 emissions and fuel consumption improvement was to be demonstrated by using dynamometer measurements and vehicle modelling, with the target drive cycle being the New European Drive Cycle (NEDC). 515 Nm at 3500 rpm

283 kW / 380 bhp at 6500 rpm

415 Nm at 6500 rpm

400 Nm at 1000 rpm

26.1 bar

25.1 bar

Corrected Power / [kW]

Corrected Torque / [Nm]

32.4 bar

Engine Speed / [rpm]

Fig. 1: Target power and torque curves and selected associated BMEPs for a 2.0 litre engine

5

1.2 Ultraboost project partners The Ultraboost project comprised eight partners, Jaguar Land Rover (JLR), GE Precision Engineering, Lotus Engineering, CD-adapco, Shell, the University of Bath, Imperial College London and the University of Leeds. It started in September 2010 with a duration of three years. JLR is the lead partner, with responsibility for engine build, general procurement, engine-mounted charging system integration and project management. GE Precision provided engine design and machining capabilities as well as background knowledge on the design of high-specific-output racing engines. Lotus Engineering provided a dedicated engine management system (EMS), 1-D modelling and knowhow on pressure-charged engines, and support for engine testing. All engine testing was to be conducted at the University of Bath, where dedicated boosting and cooled exhaust gas recirculation (EGR) rigs were used for initial testing of the demonstrator engine. CD-adapco supported the design process with steady-state and transient CFD analysis primarily in order to support intake port design, which is discussed in detail below. Shell provided test fuels and autoignition know-how. Imperial College specified the charging system components, with support from both JLR and Lotus, and tested them in order accurately to characterize them so that the 1-D model was as robust as possible. Finally, the University of Leeds developed their autoignition model to assist with the 1-D modelling process. This project structure was reviewed in an earlier publication [4], where some of the background detail to the establishment of the projects targets was also discussed. 1.3 Phases of the Ultraboost project The project was split into several parts. In Phase 1, a production JLR 5.0 litre AJ133 V8 engine was commissioned on the test bed at the University of Bath using the Denso engine management system (EMS) then used for production. This was then replaced by the Lotus EMS, which was demonstrated to be capable of controlling the engine and giving exactly the same performance at full and part load, including matching the steady-state fuel consumption of the production engine and Denso EMS combination to ≤ 0.5%. This phase therefore set the fuel consumption benchmarks for the project’s downsized engine design and proved the capability of the Lotus EMS when controlling a direct-injection engine with many high-technology features, including multiple-injection strategies. In parallel with the Phase 1 engine test work, Phase 2 specified, designed and procured the core Ultraboost engine (known as UB100). To do this the pooled knowledge of all the parties was used, resulting in a current industry best-practice high-BMEP engine with some additional novel features. The Phase 2 test programme utilized a test bed combustion air handling unit (CAHU) and a speciallydesigned EGR pump rig. It was primarily intended to prove out the efficacy of the newly-developed combustion system. The testing portion of this phase also permitted fuel testing to be undertaken without the complicating effects of an engine-driven charging system, although this important subsystem would also be specified, modeled, procured and validated in a parallel work stream within this phase. Phase 3 was intended to comprise any necessary redesign of the UB100 engine coupled with mounting the engine-driven charging system. The engine was then to be known as UB200. The present paper discusses some of the engine-specific technologies configured and tested in Phase 2; the results of the fuels testing and of the Phase 3 engine will be reported separately in later publications.

6

Ultimately, the level of achievement of the project targets will be demonstrated by a combination of direct measurement (power, torque, driveability etc.) and modelling (by the application of gathered minimap fuel consumption data to a vehicle performance model, this being necessary since the baseline AJ133 engine is no longer fitted to the target vehicle).

2

ENGINE DESIGN

2.1 Derivation of engine swept volume At the start of the project the actual swept volume was unconstrained. In order to establish this parameter, vehicle modelling was employed to set part load operating points representative of the target vehicle and to provide weighting factors for these points. The engine swept volume was then determined by using the fuel consumption improvement targets and a series of specification steps designed to ensure that the required full-load performance and driveability could be achieved; these were informed by previous work undertaken by JLR [5]. The engine was then designed in conjunction with 1-D modelling which helped to combine the various technology packages of the project. These included an advanced charging system (discussed in a previous paper [6]) and a valvetrain system with the necessary variability to deliver target performance. The modelling also helped to determine the flow characteristics required of the intake port. Ultimately this had stretch targets set for it to ensure the necessary charge motion for fuel mixing and to help suppress knock, and was subjected to a full transient CFD analysis. This is discussed later. In Phases 2 and 3 of the project the 1-D model was also used to guide testing, primarily to set intake and exhaust system boundary conditions to make them representative of what could be expected of the real charging system. It was also used to calculate the extra torque that the core engine would have to produce for the results to be representative of the combined engine and charging system. It was also used to help to explain trends in the results. 2.2 General engine specification From this preliminary work the engine was specified as shown in Table 1. The undersquare nature of the engine is readily apparent; this helps to shorten the flame travel to the benefit of knock and to reduce thermal losses. It also possibly benefits preignition, the causes of which are believed to include oil being ejected from the piston top land, and reducing the bore diameter directly reduces the top land area [7,8]. Effectively, the engine is one bank1 of a heavily-modified AJ133 V8, with a new bore and stroke, a flat-plane crankshaft and attendant firing order. This approach was taken because the bearings and scantlings of the AJ133 engine would easily be capable of handling the performance. A CAD image of the UB100 engine, fitted with the original log-type exhaust manifold, is shown in Figure 2. The engine management system was configured to be capable of controlling the many functions on the engine as detailed in Table 1 and ultimately also the selected charging system components, including the supercharger clutch and bypass system [6]. The engine has been designed to withstand a peak cylinder pressure (PCP) of 130 bar, with known further countermeasures should it be considered advantageous to increase this to a higher level (for instance, when investigating high-octane fuels).

1

The active bank is the A Bank (on the right-hand side of the engine).

7

The aluminium alloy piston itself is safe to a PCP of 145 bar for the sort of duty cycle a research engine is typically used for. Table 1: Ultraboost UB100 engine specification General architecture

4-cylinder in-line with 4 valves per cylinder and double overhead camshafts

Construction

All-aluminium AJ133 cylinder block converted to single-bank operation on the A Bank (right-hand side) Siamesed liner pack to facilitate reduced bore diameter Dedicated cylinder head

Bore

83 mm

Stroke

92 mm

Swept volume

1991 cc

Firing order

1-3-4-2

Combustion system

Pent-roof combustion chamber with asymmetric central direct injection and spark plug High-tumble intake ports Auxiliary port-fuel injection Possible second spark plug position in an under-intake-port location

Compression ratio

9.0:1

Valve gear

Chain-driven double overhead camshafts with fast-acting dual continuously-variable camshaft phasers (DCVCP) Cam profile switching (CPS) tappets on inlet and exhaust

Fig. 2: CAD images of assembled UB100 engine, as originally tested with a log-type exhaust manifold; note coolant bypass pipe for the absent B Bank cylinder head 2.3 Intake port design and flow-rig performance In order to achieve the necessary air motion and mixture preparation in DISI engines there has been a general evolution of high-tumble intake ports; this has only been made possible by the simultaneous adoption of pressure charging to overcome the flow loss generally associated with this move. It is worth noting that under-port placement of the injector had a symbiotic relationship with this evolution

8

of the general port configuration of DISI engines, but nevertheless the situation has arisen that flow rate is seen as a worthwhile trade for tumble (and hence improved mixture preparation and charge cooling). Obviously, any loss in flow capability can be expected to manifest itself in increased charge cycle (pumping) work, and so a prime desire for Ultraboost was to achieve a balance of flow and tumble considered to be significantly beyond the current state of the art. This was especially important given the high BMEP rates and specific power targeted by the project. While it is accepted that it is of primary importance to have high charge motion near to top dead centre (TDC) when the spark is initiated, high tumble has another function earlier in the cycle as a means to homogenize the air, fuel, residuals, oil droplets and temperature as fully as possible. Near to TDC piston geometry has an important effect with regard to the bulk flow breakdown and the generation of microturbulence, but during the intake stroke the importance of its geometry gradually lessens towards bottom dead centre (BDC). Thus intra-cycle CFD should be employed to determine the best overall engine geometry but the air flow rig can be used as a good differentiator early in the port development process. This section briefly discusses how this process was followed within the project and compares the performance of the adopted port with a current production turbocharged DISI engine benchmark. Initially, a target was agreed upon based upon the JLR engine database and the knowledge of the other partners. Several ports were then designed which fitted the cylinder head package. With these ports designed, CD-adapco then brought their capabilities to bear in two distinct stages of the process: a first calculation stage where the steady-state flow characteristics were determined, and a second one where full transient calculations were carried out. During the first part of this process many ports were schemed. From these, 20 were designed and analyzed under steady-state conditions. Filtering led to five being chosen and carried forward to the second transient analysis stage. Finally one port design was selected and machined into the first UB100 cylinder head, with the other available heads being held back from machining should it be found necessary to implement any changes as a result of engine performance testing. After the design had been created and the first head machined, the ports were flow tested on Lotus Engineering’s cylinder head air flow rig. These results were compared to data from the BMW N20 2.0 litre I4 engine which had also been measured on the same rig. Although the N20 engine is rated at a BMEP level significantly below that which Ultraboost was targeting, it was still considered to be the current state of the art in terms of specific power, BMEP and the fact that it had a central DI combustion system employing a multi-hole solenoid injector [9]. The results of this flow rig testing are shown in Figures 3 to 5 and are discussed below. Figure 3 presents the outright flow capability of the inlet port in comparison with the BMW engine. The Ultraboost flow at the maximum valve lift of 10.5 mm is 182 CFM, and that for the BMW at a similar lift is 139 CFM. Figure 4 shows the related flow coefficients, with Ultraboost having 0.633 and the BMW 0.520 at the same 10.5 mm valve lift condition. From this it can be seen that the port flow performance of Ultraboost in comparison to the N20 is extremely good, despite the Ultraboost engine having a 1 mm smaller bore diameter. Part of this increased flow will be due to the 5.9% larger throat area of Ultraboost, but this does not in itself account for the fact that the Ultraboost port flows nearly 30% more air than that of the N20 at 10.5 mm valve lift.

9

200 180 160

Flow / [CFM]

140 120 100 80 60 40 20 0 0

1

2

3

4

5

6

7

8

9

10

11

Valve Lift / [mm] Ultraboost

BMW

Fig. 3: Inlet port flow comparison for the Ultraboost Phase 2 and the BMW N20 cylinder heads A comparison of non-dimensional tumble number is made in Figure 5. The N20 offers significantly higher tumble at low lift; however it employs valve shrouding in order to increase tumble in that area of the curve, a specific requirement because of its use of Valvetronic mechanically-variable valve train [9]. The adoption of this form of valve train makes it especially important to generate high tumble at low valve lifts, where valve lift and duration are the primary means of controlling load while minimizing throttling loss. As a consequence Valvetronic only utilizes the high lift region during high load operation, and so a compromise at low load is presumably considered acceptable for the N20 engine. 0.7

Flow Coefficient / [Cf]

0.6

0.5

0.4

0.3

0.2

0.1

0 0

1

2

3

4

5

6

7

8

9

10

11

Valve Lift / [mm] Ultraboost

BMW

Fig. 4: Inlet port flow coefficient comparison for the Ultraboost Phase 2 and the BMW N20 cylinder heads

10

Non-Dimensional Tumble Number

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0 0

1

2

3

4

5

6

7

8

9

10

11

Valve Lift / [mm] Ultraboost

BMW

Fig. 5: Inlet port non-dimensional tumble number comparison for the Ultraboost Phase 2 cylinder head and the BMW N20 Conversely, Ultraboost was only ever to be fitted with two-step CPS tappets and so the achievement of high outright tumble rates was considered to be paramount, even for part-load operation (where greater in-cylinder air motion would still result albeit at the expense of relatively higher throttling loss). The use of valve shrouding in the N20 is reflected in the values for the tumble ratio for the two ports, Ultraboost giving 1.626 and the BMW 1.868. The fact that the Ultraboost port gives high tumble throughout the majority of the effective high-lift cam profile – from 7 mm to 10.5 mm – was considered a success, especially when paired with the high flow coefficient. This was also borne out by the fact that the port has not had to be changed since it was finalized; the engine is extremely knock tolerant and does not suffer from preignition, both of which would be expected to benefit from extremely good homogenization of the charge at the point of ignition, as discussed earlier. 2.4 Water-cooled exhaust manifold The integrated exhaust manifold (IEM) is becoming a common technology for production engines [10,11], and is particularly advantageous for turbocharged units since it allows the removal of a large degree of component protection over-fuelling at high load [2,12]. Unfortunately, because of the bore pitch and cylinder head bolt spacing necessarily inherited from the AJ133 engine, it was not feasible to design an IEM into the Ultraboost cylinder head (shown in the left-hand side of Figure 6). There was, however, an interest in investigating a water-cooled exhaust manifold (WCEM) from the point of view of assessing the full-load heat rejection. At the same time, the new WCEM permitted a more advantageous geometry than the original log manifold, and mitigated the fact that the original’s outlet geometry was restrictive. It also permitted the provision of a flow splitter which could separate all the cylinders completely, pulse divide numbers 1 and 4 from 2 and 3, or permit full mixing (all at the entry to the turbine). This is shown in situ in the right-hand side of Figure 6.

11

Fig. 6: Ultraboost cylinder head showing large bore pitch and cylinder head bolt spacing inherited from the AJ133 engine (left) and the water-cooled exhaust manifold with enlarged outlet area and flow splitter in situ (right)

3

ENGINE TESTING AND RESULTS

In the present work, the results quoted have been gathered mostly with the CAHU system, the GT-Power model being used to apply boundary conditions so that the brake results are representative of those to be expected when the engine and charging system are combined during Phase 3 of the project. Thus, in the area where the supercharger would operate, its drive torque as determined using the 1D model was added to the values shown in Figure 1 to give the target brake torque. Where the turbocharger would operate by itself, just the pressure and temperature boundary conditions were sufficient to establish whether the engine was capable of meeting the torque targets with the eventual engine-mounted charging system in place. All results reported here were gathered using commercially-available 95 RON fuel supplied by Shell; it complied with EN228 and had 5% ethanol content by volume. Other fuels will be tested as part of the project and reported in later publications. Testing to date has shown that the engine can generate the performance required to achieve the target torque curve. Furthermore, among other investigations, specific tests have been carried out in the areas of intake temperature (to show the combustion system’s sensitivity to this parameter) and PFI/DI split ratio. The engine showed no particular sensitivity to air intake temperature, being capable of delivering target performance at up to 80°C (the design target is 35°C), demonstrating a very robust combustion system and justifying the effort expended on the intake port design. This is further supported by the engine’s response to PFI/DI split ratio, shown in Figure 7, where 100% DI fuelling gave the most performance; in fact the performance of the engine is broadly constant down to and including 70% of the total fuel load being supplied by the DI system. This result is attributed to optimum air-fuel mixing and the maximum use of the latent heat of the fuel being ensured by the very high tumble flow, while PFI operation not only removes most (but not all) of this effect but also displaces more oxygen (13).

12

Fig. 7: Results of DI/PFI split loop at 2000 rpm and constant intake pressure of 2.2 bar Abs. Percentage of total fuelling supplied by direct injection shown in key. Conducted at constant intake/exhaust valve maximum opening points (MOPs) of 88° ATDC / 96° BTDC Results using the CAHU and with the engine in the configuration shown in Figure 2, i.e. with the log-type exhaust manifold, are shown in Figure 8. Here it can be seen that up to 4000 rpm the UB100 engine exceeded the target torque by the equivalent of the predicted supercharger drive torque of approximately 48 Nm, but that its performance started to dip thereafter. This was despite the intake manifold conditions supplied by the CAHU being exactly as called for by the 1-D model. Investigation revealed that this was due to the exit area of the log manifold being too small for the exhaust gas mass flow, causing it to choke. This situation was an artifact of not considering the waste gate flow in the original specification of the log manifold, and so for that reason the design and procurement of the WCEM described above was accelerated, since it had the correct sizing. In order to alleviate the problem of manifold restriction, the WCEM was fitted and the engine tested again with the intake pressure and temperature boundary conditions supplied by the CAHU as determined by the 1-D model. However, performance was again limited, this time by PCP in Cylinder 2. Examination of the individual cylinder pressure traces and those for the exhaust and intake manifolds showed that a wave dynamic effect was causing Cylinder 2 to generate more BMEP than the others, eventually reaching the PCP limit prematurely in that cylinder. To circumvent this issue, it was decided to conduct an early test with the selected Honeywell GT30 turbocharger [6] instead of using the CAHU. This test also allowed engine-based verification of the turbocharger run-up line as an input to the choice of supercharger pulley ratio for the next-phase UB200 engine. The result of this test is shown in Figure 9, where it can be seen that the engine achieves the fullload torque curve from 3000 rpm onwards and has thus has technically delivered both the maximum torque and power targets.

13

Fig. 8: UB100 performance versus AJ133 target torque curve. Intake and exhaust conditions were derived from the 1-D model. Required supercharger drive torque is approximately 48 Nm up to 3500 rpm. Original log-type exhaust manifold fitted; the dip in performance from 4000 rpm onwards is discussed in the text

524 Nm 600

382 bhp

Torque / [Nm]

500

400

=1

300

 = 0.9

200

100

10% EGR 0 0

1000

2000

3000

4000

5000

6000

7000

Engine Speed / [rpm] AJ133 Baseline

UB100 on Boost Rig

UB100 using Engine-Driven Turbocharger

Fig. 9: Engine performance with selected GT30 turbocharger (supercharger not fitted). Relative air-fuel ratio () and EGR rate for the curve with the engine-driven turbocharger also shown The results shown in Figure 9 were obtained with 10% EGR from 4000 rpm onwards, and slight enrichment to =0.9 above 5500 rpm. This was very much an exploratory test using the UB100 build specification and future work with the final UB200 configuration will concentrate on removing and hopefully eliminating enrichment throughout more of the speed-load range.

14

Even when operated at a gross BMEP of 28 bar at 1000 rpm the engine does not suffer from preignition. Even after subtracting supercharger drive torque, this is in excess of the project target output at that speed. The reasons for this absence of preignition are the subject of further investigation, since to the degree found it appears to be outside of the experience of the engine research and development community within the context of engine downsizing. A single surface preignition failure has been experienced to date, when the WCEM was first fitted and the engine operated using the CAHU and without the turbocharger; however, this situation was complicated by the fact that the second side-mounted spark plug was fitted to the cylinder head in use (but inoperable at the time). In this case it is surmised that the removal of enrichment fuelling and EGR enabled by the high heat removal of the WCEM drove the exhaust port temperature to a point where the electrodes of the second spark plug overheated, followed by the exhaust valves. As such, this is not considered to have been a lowspeed preignition (LSPI)-type event leading to superknock. More extensive cooled EGR testing has been also carried out and this is the subject of ongoing research. As part of this, pre- and post-catalyst EGR take-offs have been investigated, because there are conflicting views in the literature as to which is the more beneficial [14,15]. From the work to date, it is thought that such conflicting evidence may be confused by the use of high- or low-pressure EGR loops, which are in themselves known to affect turbocharged engine performance in different ways, regardless of whether the EGR is catalyzed or not [2,16]. Results of this investigation will be published at a later date. A full optimization of part-load fuel economy, based on a 15-point speed-load minimap, will be performed using UB200, but tests to date (again using the 1-D model to set boundary conditions and to guide control parameters such as camshaft timing) show every indication that the 35% fuel economy target can be met invehicle.

4

CONCLUSIONS

The Ultraboost project is a major collaborative engine research project part-funded by the Technology Strategy Board, the UK’s innovation agency. It is led by Jaguar Land Rover with industry- and academia-wide support. It seeks to realize a concept engine downsized by 60% (from a naturally-aspirated 5.0 litre V8 engine baseline) utilizing production technologies and with attributes suitable for deployment in premium saloons and SUVs. To that end it employs direct injection, independent cam profile switching and camshaft phasing for the intake and exhaust sides of the engine and an advanced charging system including two stages of charge air cooling. It also utilizes low-pressure cooled EGR and is fitted with a water-cooled exhaust manifold. Overall, it is designed to withstand peak cylinder pressures of 130 bar. Thus far it has been demonstrated that the power and torque targets can be met with the selected turbocharger and, when operated using a facilitated charging system, that it can deliver the gross BMEP necessary to meet the low-end torque target. With this level of downsizing preliminary fuel economy tests show that 35% fuel economy improvement in the target vehicle should be achievable.

15

5

ACKNOWLEDGEMENTS

The authors of this paper would like to thank all of the other members of the Ultraboost consortium for their involvement and the Technology Strategy Board for their continued support, without any of whom this project would not have been possible.

REFERENCES 1. 2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

16

Heywood, J.B., “Internal Combustion Engine Fundamentals”, McGraw-Hill Book Company, New York, USA, 1988, ISBN 0-07-100499-8. Turner, J.W.G., “Interactions Between Charge Conditioning, Knock and SparkIgnition Engine Architecture”, Ph.D. Thesis, Loughborough University, April 2011. Hancock, D., Fraser, N., Jeremy, M., Sykes, R. and Blaxill, H., “A New 3 Cylinder 1.2l Advanced Downsizing Technology Demonstrator Engine”, SAE paper number 2008-01-0611, SAE 2008 World Congress, 14th-17th April, 2008. Salamon, C., McAllister, M., Robinson, R., Richardson, S., Martinez-Botas, R., Romagnoli, A., Copeland, C. and Turner, J.W.G., “Improving Fuel Economy by 35% through combined Turbo and Supercharging on a Spark Ignition Engine”, 21st Aachen Colloquium, pp. 1317-1346, Aachen, Germany, 8th-10th October, 2012. McAllister, M.J. and Buckley D.J., “Future gasoline engine downsizing technologies - CO2 improvements and engine design considerations”, paper number C684/018, I.Mech.E. Internal Combustion Engines Conference, pp. 19-26, London, UK, 8th-9th December, 2009. Copeland, C., Martinez-Botas, F., Turner, J., Pearson, R., Luard, N., Carey, C,. Richardson, S., di Martino, P. and Chobola, P., “Boost System Selection for a Heavily Downsized Spark Ignition Prototype Engine”, 10th International Conference on Turbochargers and Turbocharging, London, UK, 15th-16th May, 2012. Zahdeh, A., Rothenberger. P., Nguyen, A., Anbarasu, M., Schmuck-Soldan, S., Schaefer, J. and Goebel, T., “Fundamental Approach to Investigate PreIgnition in Boosted SI Engines”, SAE paper number 2011-01-0340 and SAE Int. J. Engines 4(1):246-273, 2011, doi:10.4271/2011-01-0340. Palaveev, S., Spicher, U., Magar, M., Mass, U., Schießl, R. and Kubach, H., “Premature Flame Initiation in a Turbocharged DISI Engine - Numerical and Experimental Investigations”, SAE paper number 2013-01-0252 and SAE Int. J. Engines 6(1):2013, doi:10.4271/2013-01-0252. Steinparzer, F., Unger, H., Brüner, T. and Kannenberg, D., “The new BMW 2.0 litre 4-cylinder S.I. engine with Twin Power Turbo Technology”, 32nd Vienna Motor Symposium, Vienna, Austria, 5th-6th May, 2011. Helduk, T., Dornhöfer, R., Eiser, A., Grigo, M., Pelzer, A. and Wurms, R., “The new generation of the R4 TFSI engine from Audi”, 32nd Vienna Motor Symposium, Vienna, Austria, 5th-6th May, 2011. Ernst, R., Friedfeldt, R., Lamb, S., Lloyd-Thomas, D., Phlips, P., Russell, R. and Zenner, T., “The New 3 Cylinder 1.0L Gasoline Direct Injection Turbo Engine from Ford”, 20th Aachen Colloquium, Aachen, Germany, 11th-12th October, 2011, pp. 53-72. Turner, J.W.G., Pearson, R.J., Curtis, R. and Holland B., “Improving Fuel Economy in a Turbocharged DISI Engine Already Employing Integrated Exhaust Manifold Technology and Variable Valve Timing”, SAE paper number 2008-01-2449, SAE International Powertrain Fuels and Lubricants Meeting, Rosemont, Illinois, USA, 7th-9th October, 2008. Anderson, W., Yang, J., Brehob, D. D., Vallance, J.K. and Whiteaker, R.M., “Understanding the Thermodynamics of Direct Injection Spark Ignition (DISI)

14.

15.

16.

Combustion Systems: An Analytical and Experimental Investigation”, SAE paper number 962018, SAE International Fall Fuels & Lubricants Meeting & Exposition, San Antonio, Texas, USA, 14th-17th October, 1996. Hoffmeyer, H., Montefrancesco, E., Beck, L., Willand, J., Ziebart, F. and Mauss, F., “CARE - CAtalytic Reformated Exhaust Gases in Turbocharged DISIEngines”, SAE paper number 2009-01-0503 and SAE Int. J. Fuels Lubr. 2(1): 139-148, 2009. Roth, D.B., Keller, P. and Becker, M., “Requirements of External EGR Systems for Dual Cam Phaser Turbo GDI Engines”, SAE paper number 2010-01-0588, SAE 2010 World Congress, Detroit, Michigan, USA, 13th-15th April, 2010. Cairns, A., Fraser, N. and Blaxill, H., “Pre Versus Post Compressor Supply of Cooled EGR for Full Load Fuel Economy in Turbocharged Gasoline Engines”, SAE paper number 2008-01-0425, SAE 2008 World Congress, Detroit, Michigan, USA, 14th-17th April, 2008.

17

This page intentionally left blank

Part-load performance and emissions analysis of SI combustion with EIVC and throttled operation and CAI combustion M M Ojapah, Y Zhang, H Zhao Centre for Advanced Powertrain and Fuels, School of Engineering and Design, Brunel University, UK

ABSTRACT Direct Injection (DI) gasoline engines are staging a come-back because of its potential for improved fuel economy through principally the engine down-sizing by boosting, stratified charge combustion or possibly Controlled Auto Ignition (CAI) at part load operations. The problem with the Spark Ignition (SI) engine is its inherent low part-load efficiency. This problem arises due to the pumping loses that occur when the throttle closes or partially opens. One way of decreasing the pumping losses is to operate the engine lean or by adding residual gases. It is not possible to operate the engine unthrottled at very low loads due to misfire. However, the load can also be controlled by changing the valve closing timing – either early or late intake valve closing. Both strategies reduce the pumping loses and hence increase the efficiency. However the early intake valve closure (EIVC) can be used as mode transition from SI to CAI combustion. In order to investigate and develop a more efficient DI gasoline engine to overcome the stated challenges, an advanced research single cylinder engine with electro hydraulic valve actuation has been developed and operated under different combustion modes. In this paper, the performance, combustion and emission were measured and compared between throttled SI, EIVC and CAI combustion with negative valve overlap (NVO). At part-load condition, it is found that the CAI combustion produced the lowest fuel consumption and NOx emissions. The EIVC operation led to a moderate improvement in the fuel conversion efficiency over the throttled SI operation but it was characterised with the slowest combustion and worst HC emissions. The particulate emission results showed that soot is the dominant particles in the exhaust, which could be reduced by leaner mixture combustion.

1 INTRODUCTION Internal combustion engines have been widely used in automotive vehicles for more than a century and are still the dominant powerplant for automotive applications. Due to limited sources of fossil fuel, growing energy demand and stringent emissions regulations, automotive industries are forced to develop practical technologies to improve the fuel economy and reduce the emissions. The CO2 emissions from vehicles and Particulate Matter have attracted public attention due to its significant effects on global warming and health implications. Automotive industries have the ultimate aim of developing electric vehicles or hydrogen fuelled vehicles which emit zero tail pipe emissions, internal combustion engines will likely

___________________________________________ © The author(s) and/or their employer(s), 2013

19

continue to dominate the power source of vehicle for decades, this is certain with increasing development and production of renewable fuels and advanced engines for flexible fuel vehicles. To improve fuel economy and reducing CO2 emissions of the gasoline engine, the most robust and promising technology is through engine downsizing which reduces a vehicle’s fuel consumption. In addition, Early Intake Valve Closing (EIVC) can be effective in reducing the pumping loss at part-load operation of a SI gasoline engine. The EIVC timing is a concept where the intake valve is closed before BDC when the required quantity of air is admitted into the cylinder instead of closing the intake throttle. The EIVC can be achieved by both reduced valve lift and duration with a low lift cam lobe using a mechanical variable valve actuation system or a shorter duration with a camless valve actuation system. In the recent years, different mechanical variable valve lift systems have been implemented in production engines which range from the 2-step cam profile switching [1, 2] to continuous mechanical variable lift mechanisms, such as BMW Valveronics and Fiat’s multi-air valve actuation system. Such mechanical systems allow the valve lift and duration to be changed though not independently. Fully Flexible Variable Valve Actuation (FFVVA) systems are capable of independent control over the valve lift, opening and closing timings, and duration. The FFVVA camless system can be either electromagnetic, electro-pneumatic or electrohydraulic and drive individual engine valves directly [3-13]. They offer more control freedom and improved engine performances and emissions. They have been used to operate the 2/4-stroke switchable gasoline engine [14] and air-hybrid engine [15]. Several studies have been performed on the effectiveness of EIVC to reduce the fuel consumption. An early study by Kreuter et al [16] suggested that the EIVC with a maximum valve lift of 2mm or less could result in higher flow-in velocities even at low engine load and speed. The resulting micro turbulences in the area between valve and valve seat effectively supports a mechanical mixture preparation and helps to compensate the reduction of in-cylinder turbulence which is principally associated with unthrottled load control by means of early intake valve closing. Vogel et al [17] investigated the effect of EIVC on fuel consumption and pumping losses using a Ford Zetec, 4 cylinder, 2.0l engine fitted with secondary valve assembly between the cylinder head and the original intake manifold of the engine. The aim was to implement EIVC with the secondary valve. Such strategy enabled 70% reduction in pumping loss and 4% improvement in fuel economy. Kreuter et al [18] recorded noticeable fuel consumption improvement at low speed and low load conditions. It was also reported that the EIVC resulted in poor combustion as a result of reduced in-cylinder turbulence, temperature, and reduced mixing through the inlet valve. Soderberg and Johansson [18] investigated the load control by the EIVC and LIVC control using symmetric and asymmetric valve timings. Their results indicated longer flame development period with EIVC than the throttled SI, while varied combustion duration results were found for LIVC. The pumping mean effective pressure (PMEP) was reduced for both EIVC and LIVC compared to throttled SI. Urata et al [19] developed Hydraulic Variable-valve Train (HVT) that can be closed at arbitrary timings and capable of withstanding engine speeds up to 6000rpm, they reported reduced pumping loss of 80% and about 7% reduced fuel consumption, they also observed lower in-cylinder compression temperature and increased combustion fluctuation under low load. More recently, Patel et al [20] reported that the EIVC could be used to improve part-load fuel efficiency in a direct injection gasoline engine.

20

Another effective way to reduce part-load fuel consumption is the controlled autoignition (CAI) combustion operation [21]. One of the most effective and practical ways to achieve CAI combustion is through the negative valve overlap method, which involves the earlier closure of the exhaust valve to trap burned gases and retarded intake valve opening to control the amount of air into the cylinder. In this paper, results from a single cylinder camless engine will be presented and analysed for the EIVC and throttled SI combustion operations and the CAI combustion mode at a typical part-load condition. In addition to the combustion and engine efficiency analyses, both gaseous emissions and particulate emissions are shown and compared between the three operational modes.

2 EXPERIMENTAL SET UP All experiments were conducted in a newly commissioned single cylinder engine research facility, as shown in Figure 1a. It comprises of a unique single cylinder direct injection gasoline camless engine, the high pressure hydraulic pack for the electro-hydraulic actuators, an AC dynamometer, a supercharger unit, emission measurement and analysers, data logging and analysis system. A Denso double slit injector was used and fuel injection pressure was set to 100bar. An AVL supercharger system is connected to the engine’s intake system to supply the compressed air at a preset boost pressure and temperature through the closed loop control of heaters and heat exchangers. A laminar air flow meter is installed in the engine’s intake system to measure the intake mass flow rate. Both the intake and exhaust pressures were recorded by two piezo resistive pressure transducers. The single cylinder camless Engine is coupled to an AC dynamometer for motored and firing operations. The engine is capable of 2 and 4 stroke cycle operations through the flexible electro-hydraulic actuated intake and exhaust valves. The engine has a bore of 81.6mm and a stroke of 66.94mm respectively. The engine details are given in Table 1. The engine speed range is 0-6500 rev/min in the fourstroke mode and up to 4000rpm in the two-stroke mode.

(a)

(b)

Figure 1: (a) Overview of the single cylinder engine testing facility; (b) schematic diagram of the EMS VIE system for PM measurements

21

Table 1 Engine Specification Bore×Stroke

81.6mm×66.94mm

Compression ratio

11.78:1

Combustion Chamber

Pent roof / 4 valves

Valve train

Electro-hydraulic actuation

Fuel Injection

Direct Injection

Fuel

Standard Gasoline (RON 95), E15, and E85

Air/Fuel Ratio

Lambda=1, 1.1, and 1.2

Intake Temperature

25 C

Spark timing

MBT for EIVC and throttled SI operation

o

The instantaneous relative air to fuel ratio, Lambda, was measured by a lambda sensor connected to a MOTEC lambda meter. In parallel, exhaust gas was sampled from the exhaust pipe for gaseous and particulate emission measurements. A Horiba 7170DEGR gas analyser system was used to measure the levels of carbon monoxide (CO), carbon dioxide (CO2), total unburnt hydrocarbons (THC), and nitrogen oxides (NOx) from the exhaust. The exhaust particle measurement in this study was done with an Electrostatic Mobility Spectrometer EMS VIE, which measures and outputs the size classification and size analysis of airborne particles within the size range of 5 to 700 nm, as shown in Figure 1b. The size separation of particles is based on the principle of electrical mobility. The electrical mobility of a particle determines the drift velocity of a charged particle under the influence of an electrical field. The Differential Mobility Analyser, DMA, within the EMS is used for this purpose. Within the DMA the particles are classified according to their electrical mobility, gas flow velocity, geometry of the DMA, and the strength of the electrical field. The strength of the electric field is varied by changing the voltage between the electrodes of the DMA. The sample from the exhaust of the engine was allowed to pass through a charger or neutraliser to establish a well defined distribution of electrical charges on the particles before it is fed into the DMA. After classification, the concentration of the particles in the output sample flow is determined by the Faraday Cup Electrometer FCE. Typically it took up to four minutes for a single measurement of particle sample reading. The data can be displayed on-line and then converted into Excel format for post-processing and analysis of the particle size distribution as a function of the engine operating conditions and variables being investigated.

3 THREE ENGINE OPERATING MODES Because of the flexibility of the camless system and engine control software, many engine operation modes were achieved through different combinations of valve timings and durations. Figure 2 shows the valve timings and injection timings used in the three operation modes in this study.

22

BDC

TDC 1)

BDC

BDC

4-stroke Throttle-controlled SI

BDC

TDC

2)

BDC

4-stroke Intake valve throttled SI

TDC 3)

BDC

4-stroke Negative Valve Overlap CAI

Exhaust valve 1 Exhaust valve 2 Intake valve 1 Intake valve 2 Injection Timing Figure 2: Valve timings and Injection timings for the 3 operation modes Mode 1: 4-stroke throttle-controlled SI mode This is the conventional spark ignition mode used in the production gasoline engine. Engine load is controlled by the throttle opening, and its combustion process is initialized by the spark discharge followed by flame propagation. The engine was operated in this mode to obtain the baseline data. At part load, the partially closed throttle results in significant increase in the pumping loss, the main cause for the poor fuel economy of current SI gasoline engines. In this engine operation mode, fuel was injected earlier in the intake stroke to obtain homogeneous mixture. To prevent wetting the piston top, injection timing used in this paper was set 300 CA BTDC. The spark timing was set to the minimum advance for the best torque (MBT). Mode 2: 4-stroke intake valve throttled SI mode In order to reduce the pumping loss caused by the partially closed intake throttle at part load, intake valve opening duration can be used to regulate the amount of air into the cylinder with WOT. In this work, the intake valve opening (IVO) was fixed at the normal timing and the intake valve closing (IVC) was varied to throttle the intake air flow, injection timing was set at 300 CA BTDC. The spark timing was set to MBT. The valve lift was reduce to 2.0mm in order to simulate what could be achieved with a mechanical camshaft.

23

Mode 3: 4-stroke negative valve overlap CAI mode The CAI combustion was achieved by trapping a large amount of residual gas in the cylinder through the negative valve overlap between the earlier closed exhaust valve and retarded intake valve opening. The maximum valve lifts were 2.0mm. In this case, the intake air flow rate was determined by the amount of the trapped residuals, which was controlled by varying exhaust valve closing. Therefore, the engine could be operated with wide opened throttle. As the exhaust valve closed earlier, the fuel injection timing could be advanced into the exhaust stroke for better evaporation and mixing. In this work, the fuel was injected 440 CA BTDC.

4 RESULTS AND DISCUSSIONS Comparisons of Throttle SI, EIVC and CAI combustion In order to compare the results of the three different operation modes, the engine was operated at a typical operation condition of 1500rpm and 3.20bar IMEP with gasoline. This net IMEP used was equivalent to that generated in a suitable multicylinder production engine at a reference brake load of 2.62bar BMEP (a popular mapping point used by many vehicle OEMs).The oil and coolant temperature were held at 80oC. In order to study the effect of engine operation modes on combustion, performance and emissions, experiments were performed at different lambda values. The results are divided into four groups. Firstly, the results in Figure 3 show the combustion characteristics with lambda of 1.0 to 1.25. Engine gaseous emissions are plotted in Figure 4 for the three modes of operations at the three lambda values. Figures 5a to 5d show engine efficiencies as a function of lambda for each of the engine operation mode. Finally, the particle size distribution and particle number (PN) from the 3 combustion modes are shown in Figure 6, in which the xaxis displays the particle diameters, equally spaced on a log scale from 2.5nm up to 500nm. The PN concentration is displayed on the y-axis as dN/dlnD(#/cm3). Combustion Analysis The CAI and EIVC combustion modes were operated with wide open throttle (WOT) and throttled SI at part opened throttle. As shown in Figure 3a, EIVC results in 3.3 % reduction in ISFC compared to the throttled SI with the near stoichiometric mixture. The difference in ISFC between EIVC and the throttled SI modes disappears at λ=1.2. In comparison, the CAI combustion brings about 10.3% improvement in the fuel consumption over the EIVC and throttled SI operations. The minimum ISFC was obtained in the lean CAI combustion at λ=1.2. All three modes are characterised with very low and nearly constant cycle-to-cycle variability as shown by the COVimep results in Figure 3b. Figure 3c shows that the CAI combustion takes place earlier than the other two modes and it is slightly retarded with the leaner mixture. Though, the CAI operation starts later than the other two modes (Figure 3d) but it is much faster than the EIVC and throttled SI combustion (Figure 3e), due to the simultaneous burning of in-cylinder mixture during the CAI combustion process. Of all three modes, the EIVC mode was characterized with the longest combustion duration. This can be explained as follows. The low intake valve lift (2mm) and early closing of the intake valves result in loss of flow momentum inside the cylinder and the natural tumble movement is not supported long enough to produce turbulence required to enhance the combustion initiation and the rate of heat release. In addition, early closing of the intake valve reduces cylinder gas temperature and pressure at the end of the compression stroke, at which the flame propagation takes place.

24

(a)

(b)

(c)

(d)

(e)

(g)

(f)

(h)

Figure 3: Fuel consumption and heat release analysis of Throttled SI, CAI NVO, and EIVC

25

As shown in Figure 3f, the CAI combustion has the lowest pumping loss and hence the lowest ISFC seen in Figure 3a. In addition, the pumping loss of the CAI combustion is the least with the leanest mixture. This can be explained by the reduced heat loss to the cylinder wall from the lower burned gas temperature of the leaner mixture during the recompression stroke of the negative valve overlap period. Furthermore, it can be seen that the difference in PMEP between EIVC and throttled SI operations is greater than their difference in ISFC. This implies that some other factors have reduced the benefit of lower PMEP of EIVC than the throttled SI. One is the slightly longer combustion duration seen in Figure 3e, but the slower flame propagation in the EIVC mode alone is not sufficient to explain the diminishing difference in ISFC and the large difference in PMEP between the two engine operation modes. Figure 3g shows maximum pressure rise rate as a function of lambda for both CAI and SI combustion, which affects combustion noise. It can be seen that the maximum pressure rise rate for CAI was about 5 times higher than SI in the knocking region at lambda 1.0. To enable stable engine operation without knocking that may lead to engine damage the spark ignition timing was retarded. the retarded spark assisted CAI combustion helped to bring the peak cylinder pressure down by about 40% to safe level, though the maximum in-cylinder pressure is still some 75% higher than the throttled SI and EIVC engine operation at same load Figure 3h. In addition, the maximum cylinder pressure falls from 42 bar at lambda 1.0 to 36 bar at lambda 1.2. Gaseous Emissions All gaseous emissions were analyzed and converted into the indicated specific values (g/KWh). Figure 4a shows the CO emission from the throttled SI, EIVC and CAI combustion at 3 lambda values. As seen in Figure 4a, the ISCO emission is higher in the EIVC mode than the throttled SI and CAI combustion and it decreases with increasing lambda reaching a minimum at lambda=1.15. For throttled SI and CAI combustion the CO emission is about 50% lower than for EIVC at lambda 1 and reaches the same level at that of EIVC at lambda 1.1. As the lambda value is further increased to 1.25, there is about 16% improvement on CO emissions of EIVC over CAI combustion. Figure 4c shows the results for ISHC emission for the three modes of operations. The EIVC operation produces about 25% increase in ISHC emissions compared to throttled SI and CAI combustions. The higher CO and HC emissions results suggest that there are more locally rich mixtures formed in the EIVC mode. In the case of CAI combustion, fuel was injected into hot residual gas during the negative valve overlap period and hence could evaporate quickly and had more time to mix with air. With EIVC, intake valves are closed early in the induction stroke before BDC and the in-cylinder temperature is reduced. In addition, early closing of the intake valve and a relatively long period from valve closing to combustion resulted in weaker incylinder gas motion with reduced turbulence would result in poorer fuel evaporation and incomplete mixing. However, the interaction of fuel injection and intake air flow during the EIVC mode could have caused more fuel impingement on the piston when the fuel spray was deflected by the high speed air flow. Figure 4c shows the NOx emissions for the three combustion modes. It can be seen that the NOx emission for EIVC and throttled SI are around 9-13g/KWh and the CAI combustion is about 10 times less at about 1g/KWh. The NOx formation depends strongly on the combustion temperature. As the combustion temperature increases above 1500oC, NOx formation rates increase rapidly with increasing combustion temperature. The formation of NOx is also dependant on the combustion duration at elevated temperature. In CAI combustion, the peak combustion temperature is much reduced due to the high dilution ratio and the simultaneous burning of

26

autoignited mixture produces much shorter combustion duration, resulting in the ultra-low NOx emission seen in Figure 4c. The earlier completion of combustion leads to lower exhaust temperature in Figure 4d due to longer effective expansion process.

(a)

(b)

(c)

(d)

Figure 4: Gaseous Emissions Analysis of Throttle SI, CAI NVO, and EIVC Efficiency Analysis The combustion efficiency is the ratio of heat liberated (QhrMEP) to the theoretical heat in the fuel (FuelMEP). The amount of heat liberated is less than the theotretical value because of incomplete combustion of UHC and CO emissions. Therefore, the combustion efficiency can be calculated with good accuracy from the exhaust gas analysis as:

1

.

1

Where G_CO is the CO emission mass flow rate, G_HC is the HC emission mass flow rate, and LHV the Low heating value of the fuel. Figure 5a shows the combustion efficiency for throttled SI, EIVC and CAI combustion modes. It can be seen that the combustion efficiency of the EIVC combustion is slightly lower than the other two modes, which is caused by the higher CO and HC emissions as a result of reduced valve lift and duration as discussed previously. In general, combustion efficiencies are lower than that from well designed modern SI engines, because of the use of a side-mounted fuel injector and un-optimised combustion chamber in the single cylinder engine.

27

The gas exchange efficiency in Figure 5b shows that the EIVC has an average of 91% efficiency brought down by the reduced valve lift and duration. CAI combustion has the highest gas exchange efficiency that increases linealy with the leaner mixture. The throttle SI has the lowest gas exchange efficiency 86.6%, the reason for the largest pumping loss seen in Figure 3f. The theoretical thermodynamic efficiency of the engine is shown in Figure 5c, which defines the upper limit of the efficiency obtainable from the complete combustion and is calculated from the ratio of IMEPgross to the heat liberated (QhrMEP) as follows: 2 Where 3 4 Where Vs is the displacement of the engine. The EIVC mode exhibits the lowest theoretical thermodynamic efficiency even if the combustion process were assumed complete. This is a result of its longer combustion durations than the other two modes. The net indicated engine efficiency (ηind) represents how efficient the fuel energy is converted into the net indicated work and takes into account of the effects of combustion duration, combustion phasing, combustion efficiency, and thermodynamic efficiency, and the pumping work of the cycle. It is calculated from the ratio of the net work of the cycle to the theoretical heat in the fuel (FuelMEP). as follows: 5 Where 6 mf = Fuel Flow rate QLHV = Lower Heating Value of Fuel Figure 5d compares the net indicated efficiency of the EIVC, throttled SI and CAI modes for the lambda sweep of 1 to 1.25. The figure shows average indicated efficiency of 33.71% for EIVC, 33.75% for throttled SI and 36.675 for CAI combustion. The indicated efficiency is similar for EIVC and throttled SI, however compared to the CAI the EIVC is some 2.9% lower in indicated efficiency on average. However, the ISFC of the EIVC mode could be further improved by optimising the valve lift, timing, and duration.

28

(a)

(c)

(b)

(d)

Figure 5: Performance Analysis of Throttle SI, CAI NVO, and EIVC Particulate emission The particulate emissions were measured and analyzed in number and their size distributions. It can be seen from Figure 6a, Figure 6b, and Figure 6c that for lambda = 1, 1.1, and 1.2 the CAI mode is characterized with greatest number of particles in total while the EIVC showed the least number of particles. For all cases examined, the particles are dominated by the small particles in the range of 20nm, which are typically associated with soot particles in the nucleation mode. There are two plausible causes to the production of soot particles at greater numbers in the CAI combustion mode. In the case of CAI operation, fuel was injected into the hot burned gas at high temperature but with little oxygen. This would cause hydrocarbons in gasoline to undergo thermochemical decomposition at elevated temperatures without the participation of oxygen, known as pyrolysis, in which soot particles could be produced. Although most of such soot particles would have been oxidised during the combustion process, the lower combustion temperature in CAI mode rendered the soot oxidation process less effective. In the case of EIVC and throttled SI operation, fuel was injected into air and residual gas at much lower temperature. As a result, there was not soot particles production from pyrolysis. The majority of particles detected in the exhaust was liquid condensates, as indicated by their much smaller particle sizes. Therefore, the higher production of particulates in nucleation mode due to pyrolysis and less oxidation at lower combustion temperature and post combustion oxidation in the

29

exhaust due to low exhaust temperature resulted in the higher particulate emissions from the CAI operation. In addition to the soot particles, there are also significant number of very small particles down to 5nm in diameter. Such particles are typically liquid fuel condensates in the exhaust. Furthermore, Figures 6d to 6f show that due to the reduced fuel rich combustion the leaner mixture helps reducing the particle emissions of all sizes in all three combustion modes, except a small range of particle sizes in the throttled SI engine where lambda 1.1 mixture caused a rise in particle emissions.

(a)

(b)

(c)

(d)

(e)

(f)

Figure 6: Particulate Emission Analysis of Throttled SI, CAI NVO, and EIVC

5 CONCLUSIONS A single cylinder direct injection engine has been developed and commissioned on a transient engine test bed, which is capable of 2/4 stroke operation as well as different 4-stroke operations by means of an electro-hydraulic camless system. In this research, experiements were carried out to study and compare the combustion

30

characteristics, performance, gaseous and particulate emissions with the EIVC, throttled SI, and CAI combustion with NVO. Based on the experimental results and analyses, the following observations were made: 1.

2.

3.

4.

5.

6.

Of all three operational modes, at 3.2bar IMEP load condition, CAI shows the lowest ISFC and highest indicated engine efficiency because of its lowest pumping loss and fastest combustion. A 3.3% reduction in ISFC of EIVC over throttled SI was obtained due to reduced pumping work. The engine out ISHC emissions for EIVC was on the average 25% higher compared to the throttled SI and CAI combustion, which is attributed to the poor mixture formation process due to the reduced valve lift and duration in the EIVC mode. The particle size distribution is characterised with a peak between 15nm to 20nm in diameter, indicating the presence of nucleation mode particles. The leaner mixture helps to reduce the total particle emissions for all combustion modes. The fuel injection during the NVO period was assumed to lead to the production of soot particles (