Next Generation Electrified Vehicles Optimised for the Infrastructure (Automotive Engineering : Simulation and Validation Methods) 3031476824, 9783031476822

Citation preview

SpringerBriefs in Applied Sciences and Technology Automotive Engineering : Simulation and Validation Methods Bernhard Brandstätter · Alois Steiner   Editors

Next Generation Electrified Vehicles Optimised for the Infrastructure

SpringerBriefs in Applied Sciences and Technology

Automotive Engineering : Simulation and Validation Methods Series Editors Anton Fuchs, Virtual Vehicle Research GmbH, Graz, Austria Hermann Steffan, Graz University of Technology, Graz, Austria Jost Bernasch, Virtual Vehicle Research GmbH, Graz, Austria Daniel Watzenig, Virtual Vehicle Research GmbH, Graz University of Technology, Graz, Austria

The book series will cover current scientific issues concerning relevant simulation technology in vehicle development. Book chapters are reviewed contributions from science and industry and address approaches in the fields of system design and optimization, thermo- and fluid dynamics, noise and vibration aspects, vehicle dynamics and safety and vehicle electronics.

Bernhard Brandstätter · Alois Steiner Editors

Next Generation Electrified Vehicles Optimised for the Infrastructure

Editors Bernhard Brandstätter Virtual Vehicle Research GmbH Graz, Steiermark, Austria

Alois Steiner Virtual Vehicle Research GmbH Graz, Steiermark, Austria

ISSN 2191-530X ISSN 2191-5318 (electronic) SpringerBriefs in Applied Sciences and Technology ISSN 2570-4028 ISSN 2570-4036 (electronic) Automotive Engineering : Simulation and Validation Methods ISBN 978-3-031-47682-2 ISBN 978-3-031-47683-9 (eBook) https://doi.org/10.1007/978-3-031-47683-9 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 Chapters “Towards Digitalisation of the Charging Value Chain” and “Definition of a Set of Indicators for the EV Impact Assessment” are licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/). For further details see license information in the chapters. This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland Paper in this product is recyclable.

Preface

With the European Green Deal initiative, Europe is supposed to become the first climate-neutral continent by 2050. Within the Green Deal, the initiative “Fit for 55” aims at reducing the greenhouse gas emissions by at least 55% by 2030 compared to the levels of 1990. The electrification of vehicles plays a crucial role to meet these target values, as the transportation sector accounts for roughly one quarter of the global CO2 -emissions. The market penetration of electric vehicles is currently gaining momentum and global sales numbers are strongly increasing. In 2022, around 1.6 million EV were sold in the EU accounting for a market share of 12%. For the year 2030, the number of battery-powered electric vehicles is expected to grow to 30–40 million in the EU. On the research side for electric vehicles, the improvement of components as the battery, power electronics or brake systems are major topics. Further, the design and control of different systems within the vehicle is important to ensure energy efficiency and driving comfort. This book provides insights into current research topics of powertrains for the next generation of electric vehicles, related vehicle components (brake systems, e-axles) as well as their control. Selected Articles emerge from projects of the E-VOLVE cluster https://evolvecluster.eu/ (namely ACHILES and FITGEN). E-VOLVE stands for Electric Vehicle Optimized for Life, Value and Efficiency. The Cluster will produce greater impact acknowledging the importance of connecting parallel R&D activities funded on complementary areas, as stated by the European Commission. Further, as the interaction with the charging infrastructure becomes more and more important also the topic charging will be discussed in detail. Articles from the EUproject “XL-Connect” initiate the extension of system boundaries towards the grid, explain the way towards the digitalization of the charging value chain and describe measures for an impact assessment of EVs integration into the power grid. Finally,

v

vi

Preface

details about protocols and interfaces needed for advanced charging technologies as Vehicle-to-Grid will be given. Graz, Austria September 2023

Bernhard Brandstätter Alois Steiner

Contents

Powertrain Components and Control Future Brake Systems and Motion Control . . . . . . . . . . . . . . . . . . . . . . . . . . . Sebastian Müller and Thomas Raste

3

Advanced Methodologies for Brake Validation of EV Vehicles . . . . . . . . . . 17 Gerard Pérez Griso, Fabio Squadrani, and Jérémie Clément Torque Vectoring Testing in X-in-the-Loop Simulation Environment . . . . 27 Asier Alonso Tejeda and Pablo Prieto Arce Modelling and Control Solution of an E-axle for Third-Generation Electric Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Andres Sierra-Gonzalez, Paolo Pescetto, Elena Trancho, and Gianmario Pellegrino Smart Charging and Vehicle-to-Grid Towards Digitalisation of the Charging Value Chain . . . . . . . . . . . . . . . . . . . 51 Alois Steiner, Anna Eisner, Sandra Trösterer, Rainer Schruth, and Annika Hämmerle Definition of a Set of Indicators for the EV Impact Assessment . . . . . . . . . . 59 Martin Strelec, Pavel Hering, and Per Janecek Protocols and Interfaces for EV Charging . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Christopher Hecht, Jan Figgener, and Dirk Uwe Sauer

vii

Powertrain Components and Control

Future Brake Systems and Motion Control Sebastian Müller and Thomas Raste

Abstract The megatrends of mobility are clearly defined: Digitalization, electrification, autonomous mobility and sustainability. The brake of the future is therefore becoming increasingly intelligent to be able to meet requirements through automated driving and electrification. And this in newly conceived vehicles with modified architecture. This functional expansion requires a deep understanding of the system to combine uncompromising safety and sustainability in future braking and motion systems, which will also be modular and distributed in the long term. This is because digitization and connectivity are leading to a fundamental realignment of the electrical and electronic architecture (E/E architecture) of vehicles. Just like the electrification of the powertrain and the growing possibilities of automated driving (AD). And finally, brakes must also become sustainable and contribute to lower CO2 and dust emissions. Brake Systems of the future must continue to support safety, efficiency, comfort and, in the future, more and more additional functions. Modularization and scalability will allow the brake of the future to be integrated into cross-vehicle platforms. Keywords Dry Brake System · EMB · Electrification · Sustainable · Dust emissions · Efficiency · Safety

Abbreviations ABS AD ADAS AVP ECO EPB

Antilock Braking System Autonomous Driving Advanced Driver Assistance Systems Automated Valet Parking Ecological/Energy Efficient Driving Electronic Parking Brake

S. Müller (B) · T. Raste Continental Automotives Technologies GmbH, Guerickstraße 7, 60488 Frankfurt/Main, Germany e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_1

3

4

S. Müller and T. Raste

ESC MPC Euro NCAP TCS WLTP HPC

Electronic Stability Control Model Predictive Control European New Car Assessment Program Traction Control System Worldwide harmonized Light Vehicle Test Proceedure High Performance Computer

1 Introduction Vehicles are currently being rethought. This is due to global megatrends: With the electrification of the powertrain and the growing capability for automated driving (AD), vehicle architecture is changing. Digitalization and networking are bringing about a fundamental reorganization of electrical and electronic architecture (E/E architecture), which is increasingly geared towards software. Because it is the software that will determine the character of the car and the driving experience in the future! Bits are taking the place of horsepower. Apps and services are expanding the car into an immersive experience space that is becoming increasingly safe and comfortable. The brake of the future is therefore becoming increasingly intelligent in order to be able to meet requirements through automated driving and electrification. And this in newly conceived vehicles with modified architecture. This functional expansion requires a deep understanding of the system to combine uncompromising safety and sustainability in future braking systems, which will also be modular and distributed in the long term. This is because digitization and connectivity are leading to a fundamental realignment of the electrical and electronic architecture (E/E architecture) of vehicles. Just like the electrification of the powertrain and the growing possibilities of automated driving (AD). And finally, brakes must also become sustainable and contribute to lower CO2 and dust emissions. As a result, electrification is also bringing about a comeback of the drum brake, for example.

2 Braking System as a Vehicle Subsystem At a first glance, the vehicle as a system can be divided into two top level vehicle subsystems: vehicle body and vehicle chassis, which have different tasks.

Vehicle body:

Vehicle chassis:

top level vehicle subsystems

Host driver, passenger and goods

Move the vehicle

Future Brake Systems and Motion Control

5

But for a more detailed view on the vehicle, namely for modeling a passenger car, cruising on the road there are still more subsystems at this level: ● ● ● ● ● ● ●

Human driver Virtual driver Vehicle body Vehicle Chassis Assistance systems … … wheels, wheel suspension, axes, drivetrain, batteries, tank, auxiliary equipment,

For a clear distinction to the other vehicle subsystems we use the following formal Definition: “Chassis” is the Aggregation of vehicle elements for performing and controlling vehicle motion. Humans the chassis essentially consists vehicle control structures, allowing the driver for driving, steering and braking.

Chassis Drive pedal

Powertrain

Driven wheels

Steer wheel

Steer installation

Steered wheels

Brake pedal

Brake installation

Braked wheels

Vehicle Control Structures

In today’s passenger vehicles the powertrain uses by-wire technology for translating a driver’s acceleration pedal command into driving action whereas in historic vehicles a mechanical connection between pedal and engine was employed. The steering still uses mechanical coupling between steer wheel and steered wheels, but a steer actuator is used to support the driver. The braking system is presently in a phase of transition. The conventional direct mechanical-hydraulic connection between brake pedal and wheel brakes, which was interrupted only during antilock and stability control interventions, is increasingly being replaced by a by-wire transmission of the driver brake command. This by-wire braking technology has the advantage that no brake booster is needed anymore, that there are no irritating pedal effects in the event of stability control intervention and that the brake system is ready for automated driving. Obviously, there is a clear technology trend to a full by wire chassis. This trend is intensified by the fact that the use of the by-wire technology is a prerequisite for automated driving. Another aspect which calls for a switch to the use of by-wire

6

S. Müller and T. Raste

technology is that besides the vehicle functionality that is covered by the separate controls, enhanced functionality can be achieved by combined control of steering, drive, brake and suspension. Future chassis technology still provides the classic control lines drive, steer and brake but instead of being independent from each other, they will interact. An already existing example for this is regenerative braking. On the input side, a virtual driver is put aside to the human driver. This computer based virtual driver does not communicate with the control lines by steer wheel and pedals but communicates via electronic signals. A vehicle chassis for future mobility comprises sensors, interfaces, control installations and actuators for driving, braking, steering and wheel suspension.

3 Brake System Functions Main task of brake system is to provide specific braking of every vehicle wheel, being adequate to commands for vehicle deceleration or holding and to motion stabilization needs. This means that in a system science view, the brake system inputs are driver and virtual driver commands as well as sensor signals indicating the state of vehicle motion. Brake system outputs are braking torques, exerted on the vehicle wheels in case of a rotating wheel and in case of a stopped wheel, the output is a wheel torque threshold up to which the wheel can be held in stopped state for park braking purpose. Since the brake system outputs are torques that act between wheels and chassis, actuators are needed, to provide them. For frictional braking clamping force actuators are used to produce the needed torques by pressing friction linings against a rotor. For electric motor braking, torques are produced directly by electric traction motors working as generators. Since the resulting wheel torque is the sum of respective torque components, software-controlled coordination of wheel braking actions is mandatory.

Future Brake Systems and Motion Control

7 Brake System

HD

Vehicle Motion Interaction Control by Human (HD) and virtual driver (VD)) VD Interaction

Yaw Rate, Long, Lat

Pedal Actuation Unit Driver Brake Request Determination

Functions NBrake ABrake BSA (Brake System Application)

Clamping Force Controller

Wheel Brake FL

Clamping Force Controller

Wheel Brake FR

Clamping Force Controller

Wheel Brake RL

Clamping Force Controller

Wheel Brake RR

Wheel braking torque FL

Wheel braking torque FR

Wheel braking torque RL

Interface Pow ertrain

Interf ace Steering

Wheel braking torque RR

Steering Angle Wheel Speed Sensor

Interf ace Suspension

Brake System as system of interest

Brake system not only decelerates and holds the vehicle at standstill, but also includes assistance functions that are indispensable in today’s cars.

4 List of the Most Important Functions of the Brake System and Cross-Functional Functions Service Braking The service braking function allows for decelerating the vehicle and to halt it safely, quickly and effectively, whatever its speed and load, on any regular up or down slope of pathway. It must be possible to graduate this braking action. The driver must be able to achieve this braking action from his driving seat without removing his hands from the steering control. Usually, the service brake HMI control element is the brake pedal. Secondary Braking The secondary braking function takes over if the primary service braking function is affected by a fault. By applying a secondary braking control element, the driver must be able to halt the vehicle within a reasonable distance in the event of failure of an arbitrary component in the service braking chain. It must be possible to graduate this braking action. The driver must be able to obtain this braking action from his driving seat without removing his hands from the steering control. For implementation of these provisions, it is assumed that not more than one single failure inside the service braking system will occur at the same time. The secondary braking regulation, referred to above, holds for vehicles, being controlled by human drivers, for those with virtual drivers there exists not yet a regulation. It can be assumed that a comparable Secondary Braking Function for automated driving will be compulsory.

8

S. Müller and T. Raste

Regenerative Braking The Regenerative Braking Function splits the sum of brake torques, actually demanded by the human and virtual driver and driver assistance systems like ACC into one part to be done by regenerative braking and another part to be done by frictional braking. Wheel Rotation Stabilizing Functions (ABS, TCS Functionality) Main purpose of the ABS (Anti-Lock Brake System) function is to avoid locking of the wheels in case of overbraking, which happens if the tire-road frictional contact is not able to withstand the full amount of horizontal forces occurring during braking. This changing of tire-road frictional contact from stick to slip must be avoided for keeping the vehicle steerable and preventing the tire from being destroyed. ABS function reduces braking intensity wheel by wheel to an optimum value, related to the current wheel and vehicle state of motion. The Traction Control System function (TCS) is designed as a slip control system to prevent the driven wheels of a vehicle from excessive wheel slip when accelerating the vehicle. Vehicle Motion Stabilizing Functions (ESC Functionality) The ESC Brake system is an electrohydraulic brake system, performing the aforementioned ABS and TCS functionality, extended by the active yaw control AYC. The latter stabilizes vehicle motion by individually proper chosen wheel braking torques. By this, the vehicle stays driver-controllable even if skidding occurs when cornering. The underlying effect is that braking one single wheel will not only decelerate the vehicle but also effects vehicle yawing torque. The ESC system takes into account vehicle and wheel speeds, angular yaw speed, steer angle and a possible braking command to calculate and to perform adequate wheel specific braking actions. If needed, this is combined with an engine torque reduction command. Result is a most exact vehicle reaction on driver’s vehicle controlling commands, even in case of adverse driving conditions. In future vehicles, wheel individual electric drive motors can be controlled to support the yaw control function. And yaw control will be no longer an emergency intervention, but a permanently active feature. This will be addressed in the chapters about so called “motion control”. Vehicle Distance Control Functions (E.G. ACC) Vehicle distance control is another driver assistance function and thus a little step towards automated driving. If it is activated by the driver, it keeps the distance to the preceding vehicle and if necessary, it activates braking up to a predetermined deceleration level. Full braking is restricted to be commanded by the driver himself.

Future Brake Systems and Motion Control

9

Dynamic Motion Control Function (DMC) Defines and controls the vehicle motion on the road, with focus on efficient driving, stability and agility. Standstill Functionalities The item called SSM (Standstill Management) collects driver assistance functionalities dealing with holding the vehicle in standstill and driving off from standstill. It provides functionalities ranged from directly executing a driver’s command to automatically actuating and releasing the parking brakes automatically depending on the driving situation. A direct driver command is e.g. the manual pulling of a park brake switch lever. And an example feature for automated control of park brakes is to release them, when driving off.

5 Automotive Trends What does this mean for the brakes? Changes that are sometimes radical in the long run! A look back helps to understand this: Until now, braking systems have mainly been mechanical systems with vacuum brake boosters and hydraulic power transmission from the brake pedal to the wheel brake (pressure generation, valves, lines, brake calipers and drum brakes). Electronic safety systems such as ABS and ESC ensure that the brakes proactively contribute to driving safety in extreme situations – without the driver having to do anything. At the same time, the brake must now also contribute to vehicle efficiency – i.e. help avoid CO2 – and reduce particulate emissions during friction braking in the future. With digitalization and networking, electric drives and AD capability, braking systems must therefore perform a number of additional tasks. To this end, Continental, the long-standing and globally proven brake system specialist, is developing future brake system technologies: Future Brake Systems (FBS). A journey into the future of brakes. And an innovation roadmap with a far-reaching, step-by-step transformation.

10

S. Müller and T. Raste

6 Dry Brake System—The Brake Can Dissolve into Modules Next big step in brake system evolution is to get rid of hydraulics. This is done by using an electromechanic brake actuator (EMB actuator) on every single wheel brake. Hydraulic lines are no longer needed. Each EMB actuator is supplied with signals and electric energy by an electric cable. EMB actuator and wheel brake form a unit called wheel brake control unit (WCU). The according dry by wire brake system is also referred to as pure by wire system and will be described in detail in the following. In the very long term, hydraulics could be completely eliminated: To this end, all four wheel brakes could be actuated electromechanically—and thus braking could be completely “dry.” Today’s concentration of pressure generation and modulation with corresponding control intelligence would then no longer be necessary. A dry brake system would consist of the four dry wheel brakes (calipers or drums) and a set of software function blocks that could run on several of the existing high-performance computers (HPC) with integrated wheel controllers for safety and redundancy reasons to provide the redundancy required for safety. To make this long-term transformation to Dry Brake System at all possible, it is necessary to encapsulate the individual functions of a braking system as independent products in modular, validated and proven software blocks that can be integrated into different vehicles thanks to standardized interfaces in accordance with the principle of re-use. Functional decomposition of dry by wire brake system reveals system boundaries and main functions.

Future Brake Systems and Motion Control Sense

11

Plan

Act Emulate pedal feel

Brake pedal Capture pedal command Encoder wheel

Capture wheel speeds

Compute

Vehicle motion Sensors

Capture vehicle status data

braking

Actuate wheel brakes

Generate friction force

Wheel brake control unit WCU

action

Vehicle communication bus

Communicate braking torque request to drive system

Vehicle drive system

Key:

Brake system boundary

Frictional braking torques at wheels

Function Component

Electric braking torques at wheels

Physical interaction Signal flow

Functional Decomposition of Dry Brake-by-Wire System

The most characteristic feature of the pure by wire brake system under consideration here is the use of electric brake actuators, driven by a motor at each wheel instead of hydraulic fluid. Motivation for this step is the common trend to electrification (replacement of hydraulics by electromechanics) of the vehicles and of course, the automation of future vehicles. To achieve this, all four-wheel brakes could be actuated electromechanically—and thus provide completely “dry” braking. Today’s concentration of pressure generation and modulation with corresponding control intelligence would then no longer be necessary. An dry braking system would then consist of the four dry wheel brakes (calipers or drums) and one Wheel Control Unit (WCU) per wheel, which thanks to sensors and clever algorithms would communicate with the existing high-performance computers (HPC).

Motion Control Motion control systems and their building blocks are described from various viewpoints. A reference architecture based on model predictive control allows to systematically assess future demands on composability of the defined services. Vehicle functions and components that increase the safety are essential to modern vehicles since they have been introduced with Antilock Braking System (ABS) in the 70 s. Since then, the number of safety systems grew from Traction Control

12

S. Müller and T. Raste

System (TCS) and Electronic Stability Control (ESC) up to Advanced Driver Assistance Systems (ADAS) that improve not only the safety but also the comfort of the driver. The goal is to reach a level where those functions can take over the driver’s responsibility to control the vehicle autonomously and safely. The vehicle motion is affected by several constraints: the performance limits of the actuators, tire-road friction circles constraining the admissible forces at the wheels, fault tolerant control reconfiguration and safety limits as well as energy-related dynamic control allocation [3]. These constraints are systematically considered by equality and inequality constraints of a Model Predictive Controller (MPC). Future motion systems for autonomous and manual driving should meet increased requirements for energy efficiency, safety and driving dynamics. A modern motion and stability system has several operating modes that it can transition to [1]. The transition is either activated, and turned off by the driver or the transition is built-in, with no turning off option, e.g., ABS. The following figure summarizes the functional requirements relevant for Achiles.

Needs Automated driving/ remotable

Functions

Modes

The system shall have closed loop control and control allocation capabilities

Key

Motion and Stability System

Mode

Robust to all conditions

The chassis shall be controlled based on a target vehicle

Personalization

If one system fail, the other systems available shall compensate

Self-Safe Braking

Wheel control shall accurately blend between friction and electric motor brake torque

Standstill active

EPB open

EPB closed

Use Case

Scenario

Priority

Standstill idle

Hill hold

Stability active

Splitmue

Rollover

Stability idle Chassis Limit

Chassis & Wheel

AD idle

AD active

Sport OFF

Lane change

Sport ON

ECO ON

ECO OFF

Sport Steer

Sport Drive

AD safe

AD normal

AD Cruise

AD Park

AD Safety

AD Fault

AVP Type2

Euro NCAP

Limp aside

Self-Safe Steering

Energy saving

Normal Drive

Normal Brake

ECO Drive

ECO Brake

Throttle response

Decel. perception

City cycle

WLTP

Race track

Connected

Sinus sweep

Over take

Toll gate

Achiles Goal: Cover them all with a Chassis and Wheel module

Motion control functional requirements

The concept of motion control is based on the principle of inverse dynamics. That is quasi working back from the desired kinematic motion considering the vehicle’s inertial properties to establish the corresponding control variables, namely the dynamic tire forces and the corresponding wheel torques and wheel steering angles [2]. To compensate for the inevitable disturbances, measures such as appropriate controls and disturbance compensation must be used. An important aspect is the allocation of motion commands to the individual actuators. In so-called overactuated systems, the number of control variables exceeds the number of degrees of freedom of vehicle motion. To facilitate coordinated actuator operation, rule-based or optimization-based allocation algorithms can be used. For all this tasks appropriate services are provided. It is beneficial to split up the control system into a few cascading levels and distinguish between managing and observing services. A first extension stage covers the three horizontal degrees

Future Brake Systems and Motion Control

13

of freedom, i.e., longitudinal, lateral and yaw motion. In a maximum extension stage the dynamic motion control system covers all three translational and rotational dimensions of the body motion as well as the wheel dynamics. The following figure summarizes the main services packages at the cascading levels.

Vehicle Manager Services

Vehicle Observer Services

ï

Trans./Rot. Position Tracking Control (ADAS/AD)

ï

Trans./Rot. Position Estimation (ADAS/AD)

ï

Trans./Rot. Velocity Request Determination

ï

Trans./Rot. Position Limits Determination (ADAS/AD)

ï

Trans./Rot. Disturbance Estimation

Chassis Manager Services

Chassis Observer Services

ï

Trans./Rot. CoG Velocity Tracking Control

ï

ï

Wheel Force/Slip Request Determination (Force Allocation)

ï

ï

Wheel Speed/Torque/Steer Angle/Vertical Force Request Determination (Direct Allocation, optional)

ï

Trans./Rot. Velocity Limits Determination

ï

Vertical Force Estimation

Trans./Rot. CoG Velocity Estimation (Odometry) Wheel Force Limits Coordination

Wheel Manager Services

Wheel Observer Services

ï

Wheel Slip and Torque Tracking Control

ï

ï

Wheel Steer Angle Tracking Control

ï

Wheel Force Limits Determination

ï

Actuator Request Determination

ï

Tire-Road Friction Limits Determination

ï

Actuator Limits Coordination

Wheel Slip and Torque Estimation

Motion control services

The motion control logical architecture illustrated in the following figure shows the interconnections of the services from a signal-based control perspective rather than a service-based communication. Service-oriented communication is in general well established since years, but because of hard safety and real-time constraints not yet common in the motion and stability domain.

Motion and Stability Control

Vehicle Manager

xd yd ψd

Chassis Observer

Vehicle Observer

x,y,ψ

Chassis Manager

vd,AD βd,AD ωd,AD

Tracking Control

v β ω

Wheel Observer

Wheel Manager

Fxd,i,MPC Fyd,i,MPC

Tm,j Tb,i sx,i

MPCCha

AD Planning Target Dynamics

vd Fxd βd ωd

βd ωd

Tmd,j,MPC Tbd,i,MPC

MPCWhl

sxd,i Twd,i

VC

δfd

δrd,VC STM

Driver

axd δfd

Arb

vd,Drv βd,Drv ωd,Drv

CACha

iCar

vd βd ωd

Fxd Fyd Mzd

Arb

Fxd,i,CA Fyd,i,CA

CAWhl

iTire

Fxd,i Fyd,i

Twd,i

Arb

Tmd,j,CA Tbd,i,CA

Tmd,j Tbd,i

Smart Actuator

i=1,.,4 j=1,2

Achiles dynamic motion control logical architecture with single track model (STM), virtual rear steer angle control (VC), static control allocation (CA), inverse vehicle model (iCar), inverse tire model (iTire), model predictive control (MPC) and arbitration (Arb) services

14

S. Müller and T. Raste

The logical architecture is composed for the Achiles configuration with six individually controllable smart actuators: Friction brakes at each wheel and electric motors at each front wheel. The dynamic motion control includes feedforward and feedback control parts. Input to the system is driver requested acceleration axd and front steer angle deltafd and from an AD planning module a targeted vehicle pose with longitudinal and lateral position xd , yd and yaw angle psid . Outputs are the commanded motor torques Tmd and brake torques Tbd . Not shown are the sensor inputs and the capability feedback from the actuators. The feedforward control includes customizable target vehicle dynamics which provides a desired vehicle state with speed vd , side slip angle betad and yaw rate omegad . The target dynamics is translated into total longitudinal and lateral forces Fxd , Fyd respectively and yaw moment Mzd . The chassis MPC tracks the desired vehicle state and ensures agility and stability by torque vectoring with dynamic control allocation of the total forces and moment to eight horizontal wheel forces Fxd,i , Fyd,i , i = 1,…,4. Inverse tire models determine for each wheel the desired longitudinal wheel slip sxd,i and wheel torque Twd,i . Wheel MPCs track the torques and slips and dynamically allocate the torques at the front wheels to the electric motors and friction brakes to maximize energy efficiency by recuperation. Conclusion The result of the Achiles project is a framework on how to develop and operate modern motion control and future brake systems. Moreover, the flexible services will gradually substitute the monolithic electronic stability control software. The current market trend shows that such a composition of services is essential for autonomous driving as well as vehicle architectures that are built around high-performance vehicle computers and zone control units. Thus, the proposed motion control architecture will be key to derive new requirements and assess future demands on composability of the defined services. For safety reasons, a vehicle which is suitable for automated driving has two independent electrical on-board power networks. Future braking systems can and must take advantage of this to maximize the availability of the braking function even in the event of a possible failure. This holds for braking commands of both driver and automated driving system. At present, it is not yet possible to predict the layout of future standard braking systems for passenger cars with automation levels 3, 4 or 5 in detail. It is therefore expected that there will be competition between different future brake system concepts. For automated driving, the braking action shall be controllable by signals and the braking system shall have, in addition to normal operation, reserve modes of operation which ensure this function even in the event of failures of any kind. Depending on the nature of the failure, the corresponding reserve mode may be accompanied by degradation of some functions, for example a lower deceleration capability. Reserve mode of operation in today’s passenger car, whereby the driver takes over operation of the wheel brakes by means of a “direct hydraulic access from pedal to wheel brakes”, can no longer be used.

Future Brake Systems and Motion Control

15

In addition to these imperative requirements for automated driving, further development objectives must be pursued in the design of new braking systems: To improve energy balance of battery-electric vehicles, usage of recuperative braking should be maximized in driving mode and quiescent energy consumption of the braking system should be low. Besides these functional development goals, production engineering goals are also important for the vehicle manufacturer. In particular, the time required for the installation and commissioning of the braking system during the production of cars is to be reduced. In this context, there is a discussion of a waiver of brake fluid, i.e. the replacement of the electrohydraulic brake system by an electromechanic brake system. For vehicles with automation level 4, stowable or retractable pedals have already been shown at automobile fairs. They can only be realized with reasonable effort as so-called by-wire pedals. Such pedal module only delivers signals and does not provide any mechanically usable output when applied. The conventional brake pedal with a pressure rod that presses a hydraulic cylinder, which is installed outside the cabin will no longer exist. new brake system architectures will be modular and scalable. Previously centralized components can thus be positioned more freely in the vehicle. Brake systems of the future must continue to support safety, efficiency, comfort and, in the future, more and more additional functions. Modularization and scalability will allow the brake of the future to be integrated into cross-vehicle platforms.

References 1. Raste T, Kuhlmann A, Jokic M (2022) Characteristics of modern motion and stability systems. Aachen Colloquium Sustainable Mobility, Aachen 2. Raste T, Hohm A, Eckert A (2021) Holistic motion control for personalized and efficient vehicle dynamics. FISITA World Congress, Prague 3. Raste T (2021) The role of constraints in future motion systems. ACIMobility Summit, Braunschweig

Advanced Methodologies for Brake Validation of EV Vehicles Gerard Pérez Griso, Fabio Squadrani, and Jérémie Clément

Abstract Electric vehicles are the present and the future of the automotive industry, and the testing and validation sector has to adapt and prepare for the new requirements and challenges they may bring. This document goes through some methodologies that can be used in electric vehicle validation. Although they are based on testing standards for traditional combustion engine vehicles, they take into account new parameters and evaluate different behaviors. Keywords Brake · Testing · Electric · Energy · Sustainability · Regenerative · Recovery · Battery · Driving · Automotive · Efficiency · Friction · Development · Standards · Methodologies · Validation · Coast down · Blending · Pedal feel · Thermal · Durability · NVH · Instrumentation · Brake booster · Electronic

Abbreviations NVH EV ICE SOC AMS ABS

Noise, vibration and harshness Electric vehicle Internal combustion engine State of Charge Auto Motor und Sport (German magazine) Anti-lock Braking System

G. P. Griso (B) · F. Squadrani · J. Clément Applus IDIADA, Tarragona, Spain e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_2

17

18

G. P. Griso et al.

1 Introduction Electric vehicles are quickly taking over the automotive industry. They have brought with them many innovative and quick-evolving technologies to the braking system standards, such as regenerative braking, electronic brake boosters, or the still evolving brake-by-wire. Furthermore, their systems, morphology and dimensions are not the same as conventional combustion engine vehicles, which have been ruling the market in the past years. Electric vehicles trend to be heavier and bigger, and require less friction brake usage in normal driving. To add more, user expectations can be really demanding in some aspects in comparison to conventional vehicles. Energy efficiency, NVH and braking feeling refinement present new challenges to the manufacturers and suppliers, that must be constantly evolving their technologies in order to keep up with such a fast-growing market, with new and powerful competitors. With all these new challenges and requirements, the testing sector is also evolving. Standard procedures and test methods are being adapted or even new ones are being implemented. New instrumentation is being used in order to achieve valuable data to analyse the performance of the systems involved in the braking of the vehicle. This document will go through all these new technologies and their implications on testing methodologies and development.

2 Chapter 1—EV Brake Testing Methodologies Following the implementation of new types of components and technologies, electric vehicles demand new methodologies for validation testing procedures. In this section, some of the new procedure standards will be explained.

2.1 Regenerative Brake on Coast Down Some vehicles use the regenerative brake while the accelerator pedal is released. This feature provides a certain level of deceleration that will simulate the engine brake of an ICE vehicle, but at the same time charges the battery. In many vehicles, the level of deceleration can be adjusted by the driver or programmed in different driving modes, ranging from pure rolling coast down to the so-called one-pedal modes, that allow the vehicle to be driven almost entirely using only the accelerator pedal. These modes can be usually managed using buttons, menus or even levers. In order to evaluate the function of coast-down regenerative braking, a coast-down maneuver can be performed in a proving ground. The test needs to be conducted in a flat and straight proving ground track with enough distance to cover the whole coast down.

Advanced Methodologies for Brake Validation of EV Vehicles

19

Fig. 1 Coast down results for different modes

The exact test conditions will change depending on the scope of the validation, but the basics are to start with the vehicle at a certain speed, and the throttle must be released until the vehicle stops. During this time, vehicle speed and longitudinal acceleration will be recorded and later analyzed. To fully characterize the behavior of this feature, the test can be performed in a range of: – Different SOC. Battery status and SOC can affect the regeneration capability. – Different modes. Each mode will provide a different deceleration level and it is necessary to characterize it. – Different initial speeds. The speed at which the pedal is released may affect the behavior of the regenerative braking logic. The results obtained can be processed and presented in the form of graphs or tables that allow to analyze the behavior of the vehicle in each condition. The main parameter to take into account is the evolution of deceleration vs time (Fig. 1). Also, other parameters may be electric current to the motors or braking torque, if available. Once the results are represented or compiled, the main points to be analyzed are: – Absolute deceleration. To determine the maximum or average deceleration that each condition provides. Specially interesting to analyze the differences that different SOC can provide. – Deceleration continuity. To evaluate if there is an oscillation in the deceleration level during the whole maneuver. – Initial deceleration increase. To evaluate the time it takes to reach the maximum deceleration after the accelerator pedal has been released – Regen disconnection speed. To determine the speed value where regen braking starts to decrease the deceleration, if it does.

2.2 Regenerative Brake Blending As two separate systems (friction and regen brake) are being used for braking the vehicle, there is a need to merge them. The system has to be able to use both systems

20

G. P. Griso et al.

to their best capabilities while keeping the differences in behavior unnoticeable for the driver. The objective of brake blending tests is to evaluate how the system manages to merge both systems, how the driver can feel this behavior, and the contribution of the regenerative braking in the whole stop energy. To characterize these parameters, a constant application maneuver can be performed. Starting from a certain speed, the brake pedal is applied until the target deceleration is reached. Then, the pedal travel is kept constant until the end of the stop. This type of maneuver can be performed in different parameters to evaluate the differences or the logic that the vehicle uses in each of the conditions. Depending on the scope of the testing, varying parameters could be: – – – –

Different start speeds Different modes Different target decelerations Different SOC.

After the tests are performed, the results can be presented in the form of graphs or tables, and the following parameters can be analyzed: – Deceleration continuity. To check if deceleration has fluctuations that the driver may feel uncomfortable or difficult to predict. Specially interesting to check the points where the system changes from using regenerative brakes to friction brakes. – Pedal force fluctuation. To evaluate if the driver can feel variations in the needed pedal force while pedal travel is kept constant. – Brake pressure behavior. To identify if the system uses brake pressure and how it applies it during the stop (Fig. 2). – Regenerative brake contribution. To analyze the magnitude of brake pressure, regenerative torque or electric current during the stop, in order to quantify the contribution of the regenerative brake in the whole stop.

2.3 Pedal Feel Pedal feel testing is broadly used in the automotive industry for all kinds of vehicles. With the introduction of new generation boosters and regenerative braking, pedal feel and vehicle braking behavior testing require new points of evaluation and analysis. The key points that will have to be considered in electric vehicle pedal feel testing are: – Differences between driving/braking modes. Initial regenerative-induced deceleration, and regenerative brake contribution can be different throughout different modes. Its influence in driver perception needs to be evaluated. – Regenerative brake contribution. To evaluate how much the regenerative brake can contribute to stopping the vehicle, because as more regenerative brake can be used, greater is the efficiency of the vehicle.

Advanced Methodologies for Brake Validation of EV Vehicles

21

Fig. 2 Brake blending test results example

– Pedal feel. Depending on the type of booster system there can be physical separation between the pedal and the braking system. Also, braking output and pedal stiffness can be electronically controlled. This means that if the system is not mature enough or tuning is not good, the driver can feel the brake application as strange or not natural. Parameters as the lack of linearity or dispersion between different stops can give a hint of this naturality. – Hysteresis and delay. Depending on the type of system and its maturity, a slight delay can exist between the moment of brake application and when the vehicle starts to decelerate, especially in fast applications, as the systems needs some response time. The same happens in the difference between the application and release moments, where a bad system can give a bad feeling or it can give the sensation of the pedal being stuck.

22

G. P. Griso et al.

In order to evaluate all these parameters, testing can be performed as in conventional vehicles. Starting from a certain speed, a brake application is performed in a way that deceleration linearly increases until maximum deceleration is achieved. Application speed is defined as deceleration versus time, but it can also be pedal travel versus time or force versus time. Another test that can be done which can help to evaluate hysteresis is a controlled application and release. Also, it is interesting to test many different conditions, that can be: – – – – –

Different modes (Fig. 3) Different SOC Different application speeds Different start speeds Different brake temperatures

Fig. 3 Pedal feel test results example

Advanced Methodologies for Brake Validation of EV Vehicles

23

2.4 Low µ Tests As it happens in the acceleration phase, in the deceleration phase there must be a control of the wheel slip in order to maintain the stability of the vehicle. In conventional vehicles, this task is performed by the ABS system only, which actuates in the friction braking system to limit braking torque when wheel slip increases. In the case of vehicles with regenerative braking, torque has to be modulated when there is a danger of losing wheel speed control. This is especially critical in low adherence surfaces, where only a light coast-down regenerative torque can produce wheel slip. This can be evaluated in many ways, for example: – Braking in different adherence surfaces. To evaluate if the regenerative brake is able to modulate its torque or just disconnects to give the ABS full control. – Coasting down in low adherence surfaces. As releasing the accelerator pedal can provide some deceleration from the regenerative brake, doing so in a very low adherence surface can induce wheel slip. The key point is to see if the system can modulate the torque in such situations to keep the vehicle stability. – Adhesion changes and µ split. To evaluate if the regenerative brake control adapts fast enough to µ changes and keep stability in µ split braking, as a vehicle with only ABS would.

2.5 Thermal Tests Electric vehicles trend to be heavier due to the weight of the battery, and sometimes more powerful than their ICE competitors. This presents a challenge to thermal testing, as it will be more demanding for the braking system, due to the great amounts of energy it has to dissipate. On the other side, a well-designed regenerative brake can help in this task by absorbing part of such energy. Brake testing then, must evaluate the capability of the vehicle to cope with this kind of situations while keeping safety and comfort. Thermal tests that can be done to evaluate the performance of the braking system are: – High speed fade: Consisting of 10 consecutive stops from high speed to 60 km/h with a deceleration of 0.6 g for the first 9 stops, and the last stop with maximum deceleration. The goal of this test is to dissipate a large amount of energy through the brakes, and to evaluate if the vehicle is able to stop at maximum deceleration in the last stop. If the regenerative brake contributes to absorb part of the energy, the thermal impact on the brakes will be reduced and the performance will be greater. – AMS test: Consisting of 10 consecutive stops from 130 to 0 km/h with a panic application. This test will evaluate the evolution of the stopping distance and performance for each stop, as temperature will keep increasing. A good result would be one where stopping distance does not increase, while pedal feel and vehicle stability is maintained during the test. As in the high speed fade, if the

24

G. P. Griso et al.

Fig. 4 AMS test results example

regenerative brake can contribute to slowing down the vehicle, performance will be greater (Fig. 4).

2.6 Brake Durability Brake durability testing has been used throughout the years to validate vehicles and braking systems in terms of NVH phenomena, brake comfort and material durability. With the introduction of EVs, previously mentioned factors are still in study, but new parameters can be evaluated using the same type of test. This means, that by driving a certain amount of kilometers and recording the necessary data, statistics of the following parameters can be studied: – Amount of energy recovered from the regenerative system, by measuring battery current – Regenerative vs friction contribution for different levels of deceleration stops (Fig. 5) – SOC evolution along the durability route – Regeneration capability depending on SOC

Advanced Methodologies for Brake Validation of EV Vehicles

25

Fig. 5 Durability test results example

3 Conclusions Most of the procedures mentioned in this article were already used in traditional ICE vehicles, but with new technologies and requirements brought by electric vehicles, will mean that some procedures will be used in a different way, some of them will see their importance reduced, or new ones will take place. New parameters have to be evaluated and key points have changed from the original requirements. The introduction of further innovative technologies will mean that new client requirements and points to validate will appear, and more adaptations or completely new testing methodologies will have to be created. Because of this, manufacturers, engineering companies, and even regulatory administrations will have to work in combination to advance with new technologies, follow the growth of the industry, and satisfy the needs of costumers and society. Acknowledgements This chapter was based on work conducted in the EU project ACHILES by Applus IDIADA and the rest of the partners involved.

Torque Vectoring Testing in X-in-the-Loop Simulation Environment Asier Alonso Tejeda and Pablo Prieto Arce

Abstract Yaw moment control approaches such as Torque Vectoring (TV) can improve cornering response and vehicle handling. Electric vehicles with multimotor powertrains are particularly suitable for implementing these strategies since they allow to vary the torque applied to each wheel. For simulating and validating these algorithms, a modular X-In-the-Loop (XIL) for Electronic Control Unit (ECU) testing for Electric Vehicles (EV) has been developed. With this setup, the time and cost of the feature developments is reduced, covering from modelling level to hardware and communications level. Keywords Torque vectoring · Electric vehicle · X in the Loop · Vehicle dynamics · Optimal control

Abbreviations XIL ECU EV TV NN MPC LUT FWD ABS ESP ESS

X-In-the-Loop Electronic Control Unit Electric Vehicles Torque Vectoring Neural Network Model Predictive Control Lookup tables Front Wheel Drive Antilock Brake System Electronic Stability Program Energy Storage System

A. A. Tejeda (B) · P. P. Arce Tecnalia, Basque Research and Technology Alliance (BRTA), Donostia-San Sebastian, Spain e-mail: [email protected] P. P. Arce e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_3

27

28

EPS CCU CAN MIL KPI HIL

A. A. Tejeda and P. P. Arce

Electric Power Steering Central Control Unit Controller Area Network Model in the loop Key Performance Indicators Hardware in the loop

1 Introduction Torque Vectoring strategies are state of the art algorithms that enhance handling and stability of Electric Vehicles (EV). To validate these algorithms and fulfil safety requirements, a vast number of simulations and tests need to be performed. A highly customizable X-in-the-Loop (XIL) environment helps to follow the V-cycle development process defined by the ISO 26262 [1] and carry out a testing process from simulation to vehicle. The objective of this paper is to present some of the results of an optimal control-based TV algorithm developed during the ACHILES project [2] and tested at different levels of validation in a XIL environment.

2 Torque Vectoring Torque vectoring algorithms, in its simplest consideration for electric vehicles with a motor per wheel, are composed of at least 3 different layers: 1. Layer 1: It is a Yaw Rate Reference generator that stablishes the desired yaw rate of the vehicle which defines the target dynamic behavior. It is usually calculated knowing the driver inputs like the steering wheel angle (δ), and the vehicle dynamic states, such as velocity (v) and longitudinal acceleration (ax ). 2. Layer 2: This layer implements a High-Level Controller considering the previously calculated yaw rate reference (ψ˙ r e f ), driver inputs (steering wheel angle (δ) and throttle and braking pedals positions) and vehicle dynamic states (actual ˙ velocity (v), lateral acceleration (a y ) and longitudinal accelerayaw rate (ψ), tion (ax )). The usual outputs of this controller are the overall longitudinal force demand (Fx ) and the overall yaw moment demand (Mz ). 3. Layer 3: It is a control allocator which calculates the torque demanded by each wheel (τi j ). State estimators and observers, which can be common to all 3 layers, are frequently implemented when a needed variable is difficult to measure. Most common estimators are related to tyre forces due to the complexity that tyres present when trying to measure supported forces. During the ACHILES project several torque vectoring

Torque Vectoring Testing in X-in-the-Loop Simulation Environment

29

strategies have been developed, exploring their capabilities and suitability. Among the most remarkable algorithms two separate groups can be distinguished. On the one hand, intelligent control-based algorithms have been developed. In [3], an energy efficient intelligent torque vectoring approach based on fuzzy logic controller and neural network tyre forces estimator is presented. This algorithm follows the aforementioned 3-layer-structure with a yaw rate reference generator based on a bicycle model, a second layer composed by a fuzzy yaw moment controller and a vertical tyre forces estimator for a longitudinal torque distribution algorithm based on a novel artificial neural network (NN) and a third layer for the motor torques calculation. Results showed a significant improvement in terms of lateral dynamics and energy consumption. On the other hand, optimal control-based approaches have been developed. In [4], a comparative study on optimal control-based torque vectoring algorithms is carried out. The study covers the performance of a Linear Quadratic Regulator and a lineal and non-lineal Model Predictive Controller in some dynamic manoeuvres. Results show that the non-lineal MPC provides the best results in terms of dynamics, but also requires a high computation effort compared to the lineal approaches. Hence, linear optimal approaches are yet cost-effective. These latter approaches neglect nonlinear dynamics so having an accurate model is key to achieve the best performance. Linear bicycle models rely upon different hard to measure parameters and state variables like tyre cornering stiffnesses or sideslip angle. The performance of the linear approaches was considerably improved when estimating the sideslip angle with an observer and the cornering stiffnesses with a linear Kalman filter. In [5], a Nonlinear Model Predictive Control for Energy-Efficient Torque-Vectoring is presented. In this case, the model considers the non-linear dynamics of each tyre, which implies that the output torque of each wheel is the outcome of the optimization. This way, the third layer is removed. During the article, the effect of the vehicle understeer characteristic on energy consumption is discussed and the set of energy-efficient understeer characteristics is used for the generation of the reference yaw rate of the TV controller. Additionally, the powertrain losses, longitudinal tyre losses and lateral tyre losses are included in the cost function of the controller. The weights of the cost function can be adapted to an energy-oriented controller for energy saving or yaw rate tracking-oriented controller for dynamic performance.

3 Simulation Models 3.1 Powertrain Model and Virtual ECUs The developed powertrain is a front wheel drive (FWD) architecture, with two inwheel motors supplying torque to the front wheels. The motor and inverter, the battery and a new brake concept were modelled by third party members of the consortium, shared as s-functions and integrated into the proposed environment.

30

A. A. Tejeda and P. P. Arce

Other components are represented by low-fidelity models. Power consumption of auxiliar devices is also considered. Virtual ECUs aim to control the actuators in the virtual vehicle, generating the communications and computing the control signals. A brake ECU runs the ABS and ESP functions and distributes the mechanical and electrical braking torques. An Energy Storage System (ESS) estimates the state of the battery and computes the power limitations. The motor torque is controlled by the inverters ECU. The Traction ECU computes the demanded torque and runs the torque vectoring algorithm. Finally, the Electric Power Steering (EPS) and Central Control Unit (CCU) emulate the required communications to the rest of ECUs. CAN messages and heartbeat signals to detect communication failures are proprietary and specifically programmed for the ACHILES project.

3.2 Driver Models and Dynacar Vehicle Model To repeat some recorded manoeuvres with the real vehicle, a vehicle dynamics manoeuvre generator has been developed. It consists of a PID, that computes timevelocity and time-steering series to calculate the steering commands and throttle and braking pedals positions. With this controller, the TV algorithms can be adjusted and validated, comparing the improvement in terms of dynamics with the baseline vehicle data. Once the algorithm is adjusted, circuit tests need to be performed to mimic real driving conditions. To do so, an automated driver has been developed. It is a more sophisticated MPC approach that emulates the behaviour of a human driver, allowing to parametrize this behaviour to ensure objective testing and repeatability. For the final validation of ECUs and algorithms a driving cockpit for a human driver is included in the setup. The position of the vehicle is displayed in a 3D visor and the inputs from the driver are received from a Logitech G920 driving station. A high-fidelity vehicle dynamics model is essential for validating the developed torque vectoring algorithm. The proposed vehicle model is Dynacar, a vehicle dynamics simulation model developed by Tecnalia that represents a “rolling chassis” using a multibody formulation. Multibody models compared to analytical models, provide higher accuracy, at the cost of higher computational load. A trade-off is always required, especially when hard real-time performance is required, as it is the case [6]. Dynacar model’s suspensions are modelled as macro-joints that capture the kinematics in the form of Lookup tables (LUT) and tyres forces are modelled using the Pacejka 2006 Magic Formula. Dynacar is a forward-looking vehicle model that reacts to the traction and braking torques and the steering wheel position. Input torques are transmitted by the powertrain. The validity of Dynacar’s formulation has been demonstrated in [7–9]. The virtual vehicle has been tuned and assessed with experimental data obtained in tests performed by the real vehicle. More than 130 tests were performed, collecting data from Coast down, Steering Frequency Response, Double Lane Change, Step Steering and Ramp Steering tests at multiple speeds. Unavailable parameters were initially tuned using expert knowledge and

Torque Vectoring Testing in X-in-the-Loop Simulation Environment

31

Fig. 1 Comparison of real and virtual data during a step steer test

experience, to mimic the response measured with the real baseline vehicle. A finer calibration has been performed using optimization techniques based on surrogate models. In Fig. 1 a step steer manoeuvre (ISO 7401:2003 and ISO TR 8726:1988) that compares the measured real yaw rate and lateral acceleration to the virtual model values is depicted.

4 Model in the Loop With the simulation setup assembled, simulations at Model-in-the-Loop level can be performed. These simulations are carried out in Matlab/Simulink and signals between sub-models are interchanged by means of virtual wires and virtual CAN messages. Simulations at MiL level usually present issues to perform at Real Time. In [6] the MiL model components execution times when running in Simulink are analyzed. After preliminary tests, the torque vectoring strategy can be assessed at MiL level. Among all the algorithms presented in Torque Vectoring, a linearized yaw rate tracking-oriented optimal control-based strategy with the formulation presented in [5] will be studied from now on, as optimal controllers show promising results even at their simplest considerations. Since the tests to be performed are used to assess the vehicle dynamics and to reduce computation effort, the energy terms of the optimization have been simplified. This way, more than 130 manoeuvres have been performed. Two indicators have been chosen for the quantitative assessment of the results: the lateral acceleration and the yaw rate. The improvement of any of these indicators implies increasing the adherence range of the vehicle. This fact entails an improvement in terms of dynamics, stability and safety.

32

A. A. Tejeda and P. P. Arce

4.1 VDA Lane Change ISO 3888-2 ISO 3888-2 defines the double-lane change maneuver to test the obstacle avoidance performance of a vehicle. In the test, the driver: ● ● ● ●

Accelerates until vehicle hits a target velocity. Releases the accelerator pedal. Turns steering wheel to follow path into the left lane. Turns steering wheel to follow path back into the right lane.

Typically, cones mark the lane boundaries. If the vehicle and driver can negotiate the maneuver without hitting a cone, the vehicle passes the test. The virtual driver presented before aims to follow the target trajectory defined by the standard. The effect of the driver is not negligible since the driver tries to continuously correct the maneuver in closed loop. The same driver parameterization and tuning has been used to obtain comparable results. The next tests present the critical speed of the tandem driver and vehicle, which will be lower than the results obtained with the physical vehicle. This is the accepted way to proceed for running this test virtually. Four Key Performance Indicators (KPIs) have been selected for this test: 1. vcrit : is the critical speed, which is the maximum entry speed measured at the first cone—the higher the better—at which the considered vehicle configuration manages to successfully complete the obstacle avoidance. 2. Rear slip angle: the peak absolute value of the rear axle slip angle, which assesses vehicle stability. 3. Sideslip angle: the peak absolute value of the sideslip angle, which assesses vehicle stability. 4. Maximum lateral acceleration. The results in Table 1 have been obtained simulating the maximum critical velocity for both vehicles, 50% split and TV. That is, both vehicles perform the maneuver at the maximum velocity without hitting a cone. As expected, the vehicle governed by the Torque Vectoring overperforms the results from the baseline, permitting a 5.16% higher speed for the same maneuver and obtaining higher lateral acceleration dynamics. The behavior has also been analyzed performing the maneuver at the same velocity (at the critical velocity of the baseline vehicle). Table 1 Results at each maximum critical velocity for 50% split and TV

Baseline

Torque vectoring

61.04

64.19

Rear slip angle (°)

5.65

6.00

Sideslip angle (°)

2.15

2.37

9.16

9.67

vcrit (km/h)

Max ay

(m/s2 )

Torque Vectoring Testing in X-in-the-Loop Simulation Environment

33

Fig. 2 Results at the critical velocity of the baseline and TV vehicle

Figure 2 depicts the trajectory of the baseline vehicle at the critical speed, which is always slightly lower than the launch speed, as seen below. It also depicts a detail from the same maneuver performed by the vehicle governed by the TV on the right top. In this detail, it can be seen how the cone is not hit, and also how the yaw rate is restrained. It can also be observed that although the entrance velocity is the same for both vehicles, the end velocity of the TV vehicle is higher, which means that the vehicle performs the maneuver more efficiently. This can be better observed looking at the stability indicators. The rear slip angle, sideslip angle and maximum lateral acceleration are reduced 3.81%, 13.46% and 1.95%, respectively.

4.2 Steady-State Cornering, ISO 4138 The purpose of this test is to determine the steady-state circular driving characteristics of the test vehicle by continuously increasing lateral acceleration. The test is conducted at constant speed and shall be carried out with continuously incrementing lateral acceleration through a steering ramp input. It is performed until the limit of adherence is reached. Two indicators have been assessed to compare the response of both vehicle models: the jerk and the maximum lateral acceleration. The jerk represents the slope of the lateral acceleration during the ramping period of the test, while the maximum lateral acceleration corresponds to the constant section of the test when the limit of adherence is reached. The graphs in Fig. 3 depict the results of 12 different manoeuvres at speeds from 60 to 100 km/h. Tests included right and left manoeuvres. The results include the dynamical enhancement of the TV simulation model compared to the baseline model. It can be concluded that the TV overcomes the results of all indicators, especially for the jerk section, reaching up to 22% improvement. The yaw rate is enhanced when the TV is active, as demonstrated in the graph above. Over 65 km/h, the enhancement

34

A. A. Tejeda and P. P. Arce

Fig. 3 Yaw rate, jerk and lat. acceleration in 12 ramp steer manoeuvres

is over 5%, reaching a maximum 8.7% improvement at 87 km/h. Hence, the dynamic performance of the vehicle is improved.

4.3 Großglockner High Alpine Road Großglockner is a demanding mountain descend route which includes some challenging cornering situations with low curvature radius that are useful to assess stability and safety. As expected, the dynamical benefits of the TV are maximized for higher velocities. Since the automated driver self-adjusts the velocity of the vehicle for each curve based on the path itself, the velocity command acts like a maximum. For this reason, although the mean velocity is higher, the velocity during the curvy sections (especially in the beginning) does not change significantly from one test to another. Even though the driver tries to follow the same path with both models, the TV simulation model is more effective. Negotiating the same route with a TV strategy redounds in higher comfort. Figure 4 illustrates how the lateral slip angle of each tyre is minimized with the TV simulation model for the ~40 km/h simulation. For every curvature radius in Großglockner the TV approach presents low values of the lateral slip angle, assessing a safer and more stable performance. When the curvature radius is low, the baseline model struggles when cornering and needs to work in an unstable operating point for the tyres. This latter case presents peak values of the lateral slip angles of 18º, which are too high for any road car and out of bounds of the stability threshold.

Torque Vectoring Testing in X-in-the-Loop Simulation Environment

35

Fig. 4 Comparative study of the lateral slip angle depending on the curvature radius during Grossglockner

5 ECU in the Loop After validating the ECUs and TV algorithm at MiL level, all the models are deployed in a Real Time system. This setup is based on NI Veristand, which is a suitable software for developing Hardware in the Loop (HiL) systems. The developed models in Simulink are automatically converted into real time capable modules using NI Veristand compiler and the communications that were simulated virtually are replaced by physical communication networks. For this application, a NI PXIe-1071 chassis combined with a NI PXIe-8880 controller has been selected. There are also two PXI-8512 modules for the CAN interfaces and a PCI-6229 for the data acquisition (DAQ). The Logitech G920 driving station lets the user simulate real driving conditions. The same PXIe-8880 controller runs the virtualized ECUs and the Dynacar vehicle model at different cycle times. The CAN modules consist of a total of four CAN networks running at rates between 1 and 10 ms. The loop is closed with the RazorMotion from TTTech that has been implemented in the real vehicle. For the torque vectoring validation in Real Time, the same manoeuvre tests were performed. Furthermore, to assess the validity of the results, a correlation analysis is presented next. The correlation presented below consists of exciting both MIL and HIL frameworks with the same inputs and comparing the outputs. Inputs have been recorded during Driving-In-the-Loop maneuvers at HIL level. Inputs consisted

36

A. A. Tejeda and P. P. Arce

Fig. 5 Torques and dynamic variables during steering response test

of steering wheel, brake pedal and throttle positions, gear and parking brake and ignition key status. Outputs consisted of vehicle velocity, yaw rate and lateral acceleration (identically to previous tests) and motor torque requests as intermediate states. The following results are extracted from a representative steering response test for the baseline vehicle (50% torque split). This maneuver consists of accelerating the vehicle to 65 km/h while slowly varying the steering position, then the vehicle is slowly decelerated to 50 km/h while slowly varying the steering position and finally, sharp steering turns are reproduced. It is demonstrated that the correlation is remarkable, although a little error is observed at higher frequencies (Fig. 5).

6 Conclusions In this article, the process for testing and implementing a torque vectoring algorithm from simulation to vehicle level has been presented. Under the umbrella of the ACHILES project [2], different torque vectoring algorithms have been developed, as explained in Torque Vectoring. The theoretical results obtained at simulation level for all the proposals are outstanding but implementing computationally highdemanding algorithms in current embedded systems is a challenge to overcome. A non-linear model predictive control-based approach provides more chances to implement energy-related considerations and the best results in terms of dynamics. However, the more complex the formulation of the problem, the more the computational effort increases for ensuring a convergence to a minimum. Thus, linearizing the problem may slightly reduce the accuracy of the results but it guarantees that the solver will find a solution in an achievable task time. The overall results of the proposed algorithm showed that not only the cornering performance of the vehicle has been improved, but also the vehicle is safer, and its stability is enhanced.

Torque Vectoring Testing in X-in-the-Loop Simulation Environment

37

The proposed testing setup follows the ISO 26262V-cycle development process, which is an efficient workflow that covers from offline design to real-time testing. This way, the proposed torque vectoring algorithm was initially tuned in simulation environment just by considering a simplified powertrain and vehicle model. Once the algorithm was properly designed, it was introduced in the torque path of the Traction ECU, which is included in the simulation environment presented in Simulation models. This whole traction control that includes TV and other traction systems was then tested at Model in the loop level, where the rest of the ECUs and communications of the vehicle were virtualized. At this level, the proper integration of the algorithm in the code and overall performance was studied. Even though a preliminary study of the CPU time that is needed for solving the TV algorithm is mandatory before performing the MiL level tests, the real-time performance of the traction control must be thoroughly reviewed. The MiL simulation setup was developed in Matlab/Simulink, so compiling the code to a real-time target is straightforward using Simulink Compiler. With the models compiled and functional for the targets presented in ECU in the loop, the traction code was flashed into the physical ECU and the same tests were successfully reproduced. Ensuring a reliable performance at both simulation and RT levels entailed needing lower time for achieving the expected performance at real vehicle level. It is also relevant to consider that at vehicle level the torque vectoring algorithm may need a finer calibration related to tuning parameters, noisy signals filtering or variables estimators which are usually disregarded at simulation level. Acknowledgements The present work is supported by the ACHILES project, funded by the European Union’s Horizon 2020 Research and Innovation program, under grant agreement Nº 824311.

References 1. Joshi A (2017) Powertrain and chassis hardware-in-the-loop (HIL) simulation of autonomous vehicle platform. SAE Technical Paper 2017-01-1991. https://doi.org/10.4271/2017-01-1991 2. H2020 Achiles (h2020-achiles.eu) 3. Parra A, Zubizarreta A, Pérez J (2021) An energy efficient intelligent torque vectoring approach based on fuzzy logic controller and neural network tire forces estimator. Neural Comput Appl. https://doi.org/10.1007/s00521-020-05680-2 4. Alonso A, Parra A, Zubizarreta A, Sainz I (2021)A comparative study on optimal control based torque vectoring systems. In: 2021 IEEE vehicle power and propulsion conference (VPPC), pp 1–6. https://doi.org/10.1109/VPPC53923.2021.9699302 5. Parra A, Tavernini D, Gruber P, Sorniotti A, Zubizarreta A, Pérez J (2021) On nonlinear model predictive control for energy-efficient torque-vectoring. IEEE Trans Veh Technol 70(1):173–188. https://doi.org/10.1109/TVT.2020.3022022 6. Sainz I, Arteta B, Coupeau A, Prieto P (2021) X-in-the-loop simulation environment for electric vehicles ECUs. In: 2021 IEEE vehicle power and propulsion conference (VPPC), pp 1–6. https:// doi.org/10.1109/VPPC53923.2021.9699126 7. Parra A, Rodríguez AJ, Zubizarreta A, Pérez J (2020) Validation of a real-time capable multibody vehicle dynamics formulation for automotive testing frameworks based on simulation. IEEE Access 8:213253–213265. https://doi.org/10.1109/ACCESS.2020.3040232

38

A. A. Tejeda and P. P. Arce

8. Pena A, Iglesias I, Valera JJ, Martin A (2012) Development and validation of Dynacar RT software, a new integrated solution for design of electric and hybrid vehicles. EVS26 9. Parra A, Cagigas D, Zubizarreta A, Rodriguez AJ, Prieto P (2019) Modelling and validation of full vehicle model based on a novel multibody formulation. In: IECON 2019—45th annual conference of the IEEE industrial electronics society

Modelling and Control Solution of an E-axle for Third-Generation Electric Vehicles Andres Sierra-Gonzalez, Paolo Pescetto, Elena Trancho, and Gianmario Pellegrino

Abstract This work studies the use of an e-axle based on a six-phase IPMSM. In addition, it has a dc bus with a cascade configuration. Moreover, a dc/dc converter is incorporated between the battery module and the six-phase inverter to provide the vehicle with fast charging capabilities, while avoiding the use of power semiconductors with high nominal voltages. In this scenario, the control algorithm must cope with the non-linearities of the machine by providing an accurate setpoint command for the entire torque and speed range of the inverter. Therefore, cross-coupling effects between the windings must be considered, and the voltage of the cascade link capacitors must be actively controlled and balanced. Given this, the authors propose a novel control approach that provides all these functionalities. The proposal has been experimentally validated on a full-scale prototype 70 kW electric drive, tested in a laboratory and an electric vehicle under real driving conditions. Keywords E-axle modelling and control · Six-phase IPMSM · Cascaded dc-link · On-road validation · Electric vehicles

Abbreviations IPMSM LUT PI-FOC SiC VCT

Interior Permanent Magnet Synchronous Machine Look Up Table Proportional Integral Field Oriented Controller Silicon-Carbide Voltage Control Tracking

A. Sierra-Gonzalez (B) · E. Trancho Tecnalia, Basque Research and Technology Alliance (BRTA), 48160 Derio, Spain e-mail: [email protected] P. Pescetto · G. Pellegrino Politecnico Di Torino, 10129 Torino, Italy © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_4

39

40

A. Sierra-Gonzalez et al.

1 Introduction The transition to a circular economy and the profound decarbonization of our economies and lifestyles has never been more urgent. In December 2019, the EU authorized the European Green Deal Action Plan in response to these challenges. This plan aims to transform the Union into a modern, resource-efficient, competitive, and inclusive economy by 2050, decoupling economic growth from resource consumption and attaining carbon neutrality across all economic sectors. In this context, the decarbonization of transportation is crucial. To achieve carbon neutrality by 2050, this sector, which accounts for roughly one-fourth of the Union’s greenhouse gas emissions, must undergo a tremendous transition to electrification. In this scenario, the EU-funded FITGEN project aimed to develop a functionally integrated e-axle (motor-inverter-transmission) ready for implementation in thirdgeneration electric vehicles. The FITGEN e-axle is delivered at TRL-7 by the end of the project, and it is demonstrated on the FIAT 500e electric vehicle platform. The FITGEN e-axle aimed to increase the power density of the e-motor by 40% and the power density of the inverter by 50% compared to the best-in-class market-available technology at the project start. The daily electric driving range of the demonstrator is required to go from 740 to 1,050 km with three battery recharges, by using the e-axle in-built super-fast charging capability. To meet the FITGEN project’s goals, the e-axle depicted in Fig. 1 was designed and developed. This e-axle is equipped with a high-speed symmetrical six-phase IPMSM. The machine development takes advantage of a patented design with three layers of permanent magnets and flux barriers [1] and state-of-the-art patented Formed Litz wire winding technology [2]. Thanks to this, the machine prototype achieved a gravimetric power density of 5.3 kW/kg, stand-alone efficiency of 97.4%, low torque ripple, low back electro-motive force voltage, and a high reluctance torque. The electrical machine is driven by a six-phase inverter, with a switching frequency of up to 24 kHz. This inverter was designed using wide bandgap SiC power metal–oxide– semiconductor field-effect transistors. Thanks to the SiC technology’s higher efficiency, smaller size and better thermal performance, the inverter prototype exhibits a volumetric power density of 35 kW/l. In addition, a bidirectional dc/dc converter is affixed between the battery pack and power inverter on the e-axle. This converter increases the dc-link voltage to the

Fig. 1 General diagram of the FITGEN e-axle including a dc/dc converter, a six-phase SiC-based inverter with a cascaded dc-link capacitor, and a dual three-phase IPMSM

Modelling and Control Solution of an E-axle for Third-Generation …

41

800 V range, enabling ultra-fast charging compatibility. Furthermore, the 800 V dclink allows high dynamic performance of the IPMSM at high speed. Nevertheless, a cascaded dc-link configuration has been adopted to avoid the use of power semiconductors with high voltage ratings, which have greater conduction losses and are more expensive. This configuration consists of two series-connected three-phase inverter units, with each unit driving one of the machine windings sets [3]. Furthermore, the high-speed machine is coupled to a single-gear high-speed transmission. Such a gearbox was designed to operate at its top speed of more than 20,000 rpm while maintaining great efficiency. Reduced weight and cost were achieved due to the lack of a pump and corresponding control system thanks to a passive lubrication/cooling system. This work presents the design and validation results of the control solution proposed for FITGEN’s e-axle. First, the dynamic models used for the design of the controller are presented. Then, the design of the control solution is described, and the experimental validation results are shown. Finally, the conclusions are presented.

2 Dynamic Model of the Multiphase Electrical Machine Control systems rely on accurate representations of the electrical machine dynamics to provide stable and precise control of motor torque. Furthermore, dynamic models allow a fast and low-cost assessment of the performance of the e-axle in different operating scenarios and load conditions. By simulating the system behaviour, parameters like response time, efficiency, power consumption, and transient behaviour can be analyzed. This section describes the modelling approaches that can be used for six-phase machines. First, the mathematical equations modelling the dynamics in the natural per-phase variables are developed. Then, two vector model approaches, multiphase and double three-phase, are described. For a 6-phase machine, the stator voltage equation in the natural variable reference frame is: V = RI +

dI d PM dL I+L + , dt dt dt

(1)

where V = [v1 , v2 , v3 , v4 , v5 , v6 ]T and I = [i1 , i2 , i3 , i4 , i5 , i6 ]T are the per-phase voltages and currents, respectively. R is a 6 × 6 diagonal matrix, where each diagonal element represents the phase resistance. On the other hand, L is the 6 × 6 stator inductance matrix. Each element Lij (i, j ∈ {1, 2, . . . , 6}) represents the self (i = j) and mutual inductances (i = j) between phases i and j. Considering the matrix. The term  PM is the spatial distribution of the windings, L is a symmetric  6-dimensional flux linkage vector ( PM = ψ1 , ψ2 , . . . , ψ6 ) generated due to the permanent magnets. For IPMSMs, the elements of L and  PM vary according to the

42

A. Sierra-Gonzalez et al.

rotor electrical angle (θ e ). This is produced because of the variable magnetic reluctance of such rotor configuration [4]. Moreover, L and  PM varies with the stator currents due to magnetic saturation. The next step is to model the electromagnetic torque produced by the machine. This model can be derived from the conventional electrical input power equation (P = I T V ). First, the non-torque-producing terms are removed (copper and magnetic losses). Finally, the expression that relates output power and torque is used (P = ωmec Tem ) to get the electromagnetic torque equation: T em =

 PM 1 T dL I I + IT , 2 dθm dθm

(2)

Thus, (2) completes the mathematical representation of the electric machine. However, such a model is complex and highly coupled, not suitable for control system design. Therefore, vector transformations are applied to simplify the mathematical representation of the model by decoupling and eliminating dependence on rotor position. These transformations allow the design and implementation of the well-known FOC technique [5]. The first vector transformation T1 is obtained by multiplying the decoupling transformation matrix C1 in (3) and the rotating transformation matrix P1 in (4), i.e., T1 = P1 C1 . ⎡ ⎤ 1 cos(α) cos(2α) cos(3α) cos(4α) cos(5α) ⎢ 0 sin(α) sin(2α) sin(3α) sin(4α) sin(5α) ⎥ ⎢ ⎥ ⎥ 1⎢ ⎢ 1 cos(2α) cos(4α) cos(6α) cos(8α) cos(10α) ⎥ (3) C1 = ⎢ ⎥ 3 ⎢ 0 sin(2α) sin(4α) sin(6α) sin(8α) sin(10α) ⎥ ⎢ ⎥ ⎣1 ⎦ 0 1 0 1 0 0 1 0 1 0 1 ⎡ ⎤ cos(θe ) sin(θe ) 0 0 00 ⎢ − sin(θ ) cos(θ ) 0 0 0 0⎥ e e ⎢ ⎥ ⎢ ⎥ 0 0 cos(5θe ) sin(5θe ) 0 0 ⎥ ⎢ (4) P1 = ⎢ ⎥ ⎢ 0 0 − sin(5θe ) cos(5θe ) 0 0 ⎥ ⎢ ⎥ ⎣ 0 0 0 0 1 0⎦ 0 0 0 0 01 where α = π/3, is the spatial separation between adjacent phases in symmetrical six-phase machines. Applying T 1 to (1) and (2), the multiphase vector model is obtained: vD1 = Rs iD1 + LD1 vQ1 = Rs iQ1 + LQ1

diD1 − ωe LQ1 iQ1 dt

 diQ1 + ωe LD1 iD1 + ψPM1 dt

(5)

Modelling and Control Solution of an E-axle for Third-Generation …

vD2 = Rs iD2 + LD2

diD2 − 5ωe LQ2 iQ2 dt

diD2 − 5ωe LQ2 iQ2 dt   

= 3N p ψPM1 iQ1 + LD1 − LQ1 iD1 iQ1

vD2 = Rs iD2 + LD2 T em

43

(6) (7)

This vector model shows that both planes D1 − Q1 and D2 − Q2 are decoupled, where the frame D1 − Q1 contains the fundamental components, while the subspace D2 − Q2 comprises the 5th and 7th harmonics [5]. From (7), it is deduced that torque can be controlled by only regulating the D1 −Q1 plane currents. Therefore, the optimal current references are easier to calculate, with the advantage that cross-coupling between the two three-phase sets is considered. The second vector transformation T 2 is obtained by multiplying matrices C2 and P2 . ⎤ ⎡ 1 cos(2α) cos(4α) 0 0 0 ⎢ 0 sin(2α) sin(4α) 0 0 0 ⎥ ⎥ ⎢ ⎥ ⎢ 2 ⎢ 0.5 0.5 0.5 0 0 0 ⎥ (8) C2 = ⎢ ⎥ 0 0 1 cos(2α) cos(4α) ⎥ 3⎢ 0 ⎥ ⎢ ⎣ 0 0 0 0 sin(2α) sin(4α) ⎦ 0 0 0 0.5 0.5 0.5 ⎡ ⎤ cos(θe ) sin(θe ) 0 0 0 0 ⎢ −sin(θ ) cos(θ ) 0 0 0 0⎥ e e ⎢ ⎥ ⎢ ⎥ 0 0 1 0 0 0⎥ ⎢ (9) P2 = ⎢ ⎥ ⎢ 0 0 0 cos(θe − α) sin(θe − α) 0 ⎥ ⎢ ⎥ ⎣ 0 0 0 −sin(θe − α) cos(θe − α) 0 ⎦ 0 0 0 0 0 1 As previously, applying T 2 to (1) and (2), the double three-phase vector model is obtained: vd 1 = Rs id 1 + Ld vq1 = Rs iq1 + Lq

diq1 diq2 + ωe Ld id1 + M q + ωe M d id2 + ωe ψPM dt dt

vd 2 = Rs id 2 + Ld vq2 = Rs iq2 + Lq

did 2 did 1 − ωe Lq iq1 + Md − ωe Mq iq2 dt dt (10)

did 2 did 1 − ωe Lq iq2 + Md − ωe Mq iq1 dt dt

diq2 diq1 + ωe Ld id2 + M q + ωe M d id1 + ωe ψPM dt dt

(11)

44

A. Sierra-Gonzalez et al.





 3  N p ψPM iq1 + iq2 + Ld − Lq id1 iq1 + id2 iq2 2



 + M d − M q id1 iq2 + id2 iq1

T em =

(12)

In this model, the d1 − q1 and d2 − q2 planes are coupled by the terms Md and Mq . Therefore, the equations of voltages, currents, and torque are more complex than those of a three-phase machine. Another feature of applying T2 is that both planes rotate at the same speed, but there is a π/3 offset between them. Both planes can produce torque, as shown by (12). The proposed control solution benefits from both vector models.

3 Design and Validation of the Control Solution Figure 2 shows the block diagram of the control solution designed for the e-axle [6], which can be viewed in three stages. The first stage considers the required torque, dc-link and stator voltages, and rotor speed to calculate the optimal electric current setpoints. These are computed by LUTs, compiled using a high-fidelity mathematical model in the D-Q synchronous frame to simplify the representation of the complex non-linear and coupled behaviour of the IPMSM. To reduce the LUTS’s required dimensions and increase the controller’s robustness against system parameter variation, a VCT loop is incorporated. To achieve this, the VCT varies the speed fed to the LUTs and keeps the stator voltages in a safe range. The second stage comprises the dc voltage active balancing algorithm, necessitated by the cascaded topology of the 6-phase SiC-inverter, which balances the input voltages of the sub-inverters. To do so, first, the current references are transformed

Fig. 2 General block diagram of the proposed control solution

Modelling and Control Solution of an E-axle for Third-Generation …

45

(T 1→2 ) to the d–q synchronous frame to facilitate the independent regulation of each three-phase set. Then, a PI controller modifies the current setpoints to balance the DC voltage of each three-phase set. Note that only q-axis setpoints are modified to avoid disturbing the operation at high speeds or field weakening. As required in an automotive application, the algorithm operates correctly in motoring and regenerative braking. The third stage contains synchronous current regulation via PI-FOC featuring feed-forward decoupling and anti-windup schemes. Additionally, carrier-based Pulse-Width Modulation blocks synthesize the firing pulses for each sub-inverter. In addition to regulating torque, the system optimizes e-axle efficiency by adopting a variable dc-link strategy to dynamically adapt the inverter bus voltage based on the e-motor operating point. The dc-link variation is enabled by the high-performance dc/dc converter located between the battery and the 6-phase inverter. A custom control strategy minimizes the dc-link voltage without affecting control dynamics and minimizes switching losses in the drive and the dc/dc converter [7]. The proposed control solution was experimentally validated. First, the control system is validated at TRL-6 over a laboratory test bench, incorporating a fullscale prototype with a peak power of 135 ~ kW, developed within the context of the FITGEN project. These initial tests show the need to implement an interleaved PWM modulation scheme due to the presence of overlapping windings in the machine’s stator. After this modification, the controller is extensively tested in the laboratory under various operation profiles, including several driving cycles. Next, the drive prototype is integrated into a real EV, and the novel controller is validated at TRL-7, both in dynamometer and on-road tests. For example, Fig. 3 shows the experimental results obtained under the US06 driving cycle. The right-side charts present a stretch of the US06 cycle at maximum speed, demonstrating accurate torque tracking during sharp torque commands (Fig. 3a), proper voltage balancing (Fig. 3b), and regulation of the stator voltages below the maximum admissible value (Fig. 3b).

4 Conclusions The control approach conveniently combines two vector representations to meet automotive requirements such as torque control accuracy, safe operation, and low computational burden. This approach successfully addressed the highly coupled and non-linear nature of the six-phase e-axle with cascaded configuration. Thanks to the TRL-7 validation, the industrial applicability of dual three-phase IPMSM drives with cascaded dc-link configuration is assessed, and the proposed control solution is fully validated for their utilization in real automotive applications.

46

A. Sierra-Gonzalez et al.

(a)

(b) Fig. 3 Experimental results obtained under the US06 driving cycle: a mechanical speed (ωmec ) meas , current-estimated T est and commanded T ∗ electromagnetic and torque response (Measured Tem em em torque); b dc-link voltages (vdc1 , vdc2 , and vdc ) and set 1, set 2, and maximum stator voltages (vs1 , vs2 , vdcmax ). The right-side charts show a stretch of the US06 cycle at maximum speed

Acknowledgements This chapter was based on work conducted in the EU H2020 project FITGEN, (Grant agreement 824335).

References 1. Zhou T, Oeschger D, Teufel R, Muntean A (2018) Rotor for a synchronous drive motor. EU Patent EP3651316A1

Modelling and Control Solution of an E-axle for Third-Generation …

47

2. Dimier T, Cossale M, Wellerdieck T (2020) Comparison of stator winding technologies for highspeed motors in electric propulsion systems. In: International conference on electrical machines (ICEM) Gothenburg. https://doi.org/10.1109/ICEM49940.2020.9270943 3. Martino M, Pescetto P, Pellegrino G (2020) Advanced functionally integrated E-axle for Asegment electric vehicles. In: AEIT international conference of electrical and electronic technologies for automotive Turin. https://doi.org/10.23919/AEITAUTOMOTIVE50086.2020.930 7382 4. Rahman M (2016) Power electronics and motor drives. CRC Press, 2016, Ch. Permanent Magnet Machines, ISBN 9781138077478 5. Karttunen J, Kallio S, Peltoniemi P, Silventoinen P, Phyronen O (2014) Decoupled vector control scheme for dual three-phase permanent magnet synchronous machines. IEEE Trans Indus Electron 61. ISSN 1557-9948 6. Sierra-Gonzalez A, Pescetto P, Alvarez-Gonzalez F, Heriz B, Trancho E, Lacher H, Ibarra E, Pellegrino G (2023) Full-speed range control of a symmetrical six-phase automotive IPMSM drive with a cascaded DC-link configuration. IEEE Trans Indus Appl 59. ISSN 1939-9367 7. Pescetto P, Sierra-Gonzalez A, Alvarez-Gonzalez F, Kapeller H, Trancho E, Lacher H, Pellegrino G (2023) Active control of variable DC-link for maximum efficiency of traction motor drives. IEEE Trans Indus Appl. ISSN 1939-9367

Smart Charging and Vehicle-to-Grid

Towards Digitalisation of the Charging Value Chain Alois Steiner, Anna Eisner, Sandra Trösterer, Rainer Schruth, and Annika Hämmerle

Abstract The number of battery-powered electric vehicles in the EU is expected to increase to 30–40 million by 2030. This strong increase of electric vehicles is a major challenge for the energy system in Europe, but at the same time an opportunity to use new technologies such as smart charging or vehicle-to-grid. The EU project “XL-Connect” is investigating these new technologies, as the overall project goal is to optimise the entire charging chain—from energy provision to the end user—in order to create clear benefits for all stakeholders. This is to be achieved through the implementation of demonstration actions in combination with a digital twin of the charging chain. With the help of the digital twin, the so-called “upscaling” can be carried out to simulate the impact of large fleets of electric vehicles and their impact on the grid. Keywords Digital twin · Smart charging · Vehicle-to-grid · User behaviour

1 Introduction The expected strong increase in electric vehicles—30–40 million are expected in the EU by 2030—poses a major challenge for the energy system in Europe. Especially the local distribution grids may be pushed to their limits if the future owners of electric cars will charge their vehicles at the same time (e.g. mainly in the evening). Expensive reinforcements and extensions of the grid would be necessary to avoid partial shutdowns. An alternative to these reinforcement measures is the use of advanced charging solutions that can bring benefits to all stakeholders. Within the EU-project “XL-Connect”, the following advanced charging concepts will be investigated in detail: A. Steiner (B) · A. Eisner · S. Trösterer · R. Schruth Virtual Vehicle Research GmbH, Inffeldgasse 21a, 8010 Graz, Austria e-mail: [email protected] A. Hämmerle Neuman Aluminium Industries, Werkstrasse 1, 3182 Marktl, Austria © The Author(s) 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_5

51

52

A. Steiner et al.

● V1G: Smart charging, also called V1G, refers to the ability to modify the charging power and the charging time. This can help to reduce the grid load during peak hours and to decrease the cost of charging. ● V2B/V2H: Vehicle-to-building (V2B) refers to the transfer of energy from an electric vehicle to a non-residential building and vehicle-to-home (V2H) to a residential building. ● V2G: Vehicle-to-grid (V2G) refers to bidirectional energy flow between an electric vehicle and the grid. Thus, the owner of an electric vehicle becomes a “Prosumer” (consumer who also produces) of energy and can help to stabilize the grid during peak hours. To enable the usage of these technologies on a large scale, all involved stakeholders including the EV owners need to see a clear benefit. Thus, the overall objective of the XL-Connect project is to optimize the entire charging chain—from energy provision to the end user—and to create a convincing benefit for all stakeholders.

2 Investigation of the User Behaviour One key element to successfully apply advanced charging technologies is to understand the user behaviour related to these technologies. Figure 1 shows an overview how the models for the user behaviour shall be set up, divided in charging (Gridto-Vehicle (G2V)) and discharging (Vehicle-to-Grid (V2G)). For both elements the users will have different interests to consider. E.g. for charging the procedures factors like time/duration, costs or related CO2 -emissions might be interesting for the users. For discharging also the time/duration will be important, but probably also factors related to battery ageing (state of health (SoH)), the remaining battery state of charge (SoC) and some incentive (remuneration, free parking etc.) to attract EV owners to allow V2G will be important. In order to investigate these aspects a questionnaire has been set up and will be distributed through the project partners of XL-Connect. The structure of the questionnaire is depicted in Fig. 2. Starting with a collection of demographic data the survey splits into persons who already drive an electric vehicle and those who plan to buy or lease one. For the first group questions regarding details of the electric car, their usage in daily life and their charging behaviour are stated. The second group gets questions regarding the time frame, reasons for an electric car as well as the planned charging behaviour. If they are unsure if they want to buy or lease an electric vehicle, the reasons for that will be asked. Finally, all persons are questioned regarding their attitude towards advanced charging technologies as smart charging, vehicle-to-home and vehicle-to-grid.

Towards Digitalisation of the Charging Value Chain

53

Fig. 1 Investigation of the user behavior

3 Digital Twins and Virtual Demonstration Actions With the help of digital twins of the whole charging chain from the grid to the electric vehicles, different use cases using different advanced charging technologies can be investigated. Figure 3 shows an overview of different actors that will be represented in the digital twin and how the relations in terms of information and energy flow are expected to be. With the digital twin the so-called “upscaling” can be carried out—i.e. the number of electric vehicles can be easily changed/increased—to simulate the impact of large fleets of electric vehicles and their impact on the grid. Virtual demonstration action—Neuman Aluminium use case: In XL-Connect real-word demonstration actions are complemented by virtual demonstration actions, which allow the investigation of different parameters (e.g. number of electric vehicles, control strategies etc.). One of the virtual demonstration actions is the so-called “Neuman Aluminium use case”, which deals with a production site for aluminium parts and a possible use of vehicle-to-building to optimize the on-site energy management. In general, the company Neuman Aluminium, located in Lower Austria, produces aluminium parts and has an overall yearly energy demand of ~111,000 MWh according to their energy intensive production processes. Overall, the total energy demand in 2022/23 of this use case can be divided in ~36% electricity demand and ~64% natural gas demand. According to this high energy demand, Neuman has employed three small hydroelectric power plants with an overall size of 0.95 MWp and a photovoltaic system of size 1.1 MWp. Currently, these power plants produce

54

A. Steiner et al.

Fig. 2 Questionnaire for the investigation of the user behavior

~4,100 MWh/year. As the production could cover only ~10% of the needed electricity, Neuman wants to increase their renewable energy production by employing additional PV systems (up to 4 MWp) and two wind turbines (overall 10 MWp). In XL-Connect the target is to analyse how the expansion of renewable energy power plants in combination with different storage possibilities impacts the overall energy consumption and the resulting financial benefits of Neuman Aluminium. Therefore, three future scenarios will be investigated (Table 1). To analyse these scenarios, a three-step approach is applied. In the first step, the energy production and consumption data of Neuman Aluminium will be analysed in detail in order to determine the periods and amount of surplus of energy

Towards Digitalisation of the Charging Value Chain

55

Fig. 3 Overview of different actors in the charging chain

Table 1 Overview scenario setups Status Quo (MWp)

Scenario 1 (MWp)

Scenario 2 (MWp)

Scenario 3 (MWp)

Hydroelectric power plant

0.95

0.95

0.95

0.95

Photovoltaic power system

1.1

1.3

4.0

4.0

–

–

10

Wind power station –

production for the different scenarios. With these findings the impact of different sizes of battery storages can be investigated in a second step in order to achieve a self-consumption optimization or peak shaving. These results will also contain the possible economic benefit for Neuman Aluminium. In the third and last step, a vehicle-to-building solution for a parking area with 300 vehicles is investigated as an alternative storage solution to the battery storage by means of a simulation model. Therefore, the parking situation at Neuman Aluminium will be simulated considering the behavior of the EV owners in order to determine the available energy storage capacity during the day considering the different work shifts. When analysing this alternative solution, it is also important to find out under which conditions the employees would agree to (temporarily) discharge their vehicles. A schematic overview of the Neuman Aluminium use case can be seen in Fig. 4.

56

A. Steiner et al.

Fig. 4 Schematic overview of the “Neuman Aluminium” use case

4 Conclusion and Outlook The EU project “XL-Connect” is investigating advanced charging technologies as smart charging, vehicle-to-building or vehicle-to-home and vehicle-to-grid. Therefore, demonstration actions are carried out to collect the necessary data to build up a digital twin of the “charging chain”—containing all elements and actors from the vehicles to the grid. One important element related to charging technologies is the investigation of the user behaviour. So, in order to calculate the possible benefits of technologies as vehicle-to-building or vehicle-to-grid, knowing when the vehicles are connected to the charging station as well as the willingness of the users to use these technologies (i.e. allowing to discharge their vehicles) is crucial. In order to investigate the charging behaviour of the EV users as well as their attitude towards advanced charging technologies a questionnaire has been set up. The results will be assessed also related to the differences between European countries. Complementary to real-word demonstration action virtual demonstration actions are performed to analyse the possible benefits of advanced charging technologies. A virtual demonstration action assesses the benefits of a battery storage or a parking area with 300 vehicles and vehicle-to-building as alternative solution are assessed. Therefore, the production and consumption data are analysed to determine the periods and amount of surplus of energy production. With these findings the impact of different sizes of battery storages can be investigated in order to achieve a self-consumption optimization or peak shaving for three future scenarios. Finally, a vehicle-to-building solution

Towards Digitalisation of the Charging Value Chain

57

for a parking area with 300 vehicles is investigated by means of a simulation model as an alternative storage solution to the battery storage. Acknowledgements

Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor the granting authority can be held responsible for them. The publication was written at Virtual Vehicle Research GmbH in Graz and partially funded within the COMET K2 Competence Centers for Excellent Technologies from the Austrian Federal Ministry for Climate Action (BMK), the Austrian Federal Ministry for Labour and Economy (BMAW), the Province of Styria (Dept. 12) and the Styrian Business Promotion Agency (SFG). The Austrian Research Promotion Agency (FFG) has been authorised for the programme management.

Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. The images or other third party material in this chapter are included in the chapter’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

Definition of a Set of Indicators for the EV Impact Assessment Martin Strelec, Pavel Hering, and Per Janecek

Abstract Technological, market, regulatory and societal drivers in the energy domain have significant impacts on power systems. These impacts are related to individual technical processes (e.g. operation, planning), which may be considered in different time frames. Power system impacts are mostly determined by measurable indicators, which can be defined in various ways. This chapter focuses on the presentation of commonly used measurable indicators for grid impact assessment in the field of EVs integration. Because technical processes may run in different time horizons, suitability of collected measurable indicators to three different time domains (real time, short term operation planning and long term planning) are discussed and assessed. Keywords Power system · Key performance indicators · Grid assessment

1 Introduction Reflecting world-wide decarbonization strategies, electromobility is increasing the penetration to the common life, while an EV momentum is foreseen in upcoming years in the ten-year horizon.1 This phenomenon will significantly affect many technological, social, behavioral, and economical domains, where the energy domain is strongly facing novel challenges such as massive EV integration into power systems, efficient utilization of renewable energy production and others. For the efficient and sustainable integration of novel technologies into power systems, core paradigms of distribution system operators (DSO) related to planning and operation activities need to be updated, which implies to consider a holistic view respecting different

1

https://www.iea.org/reports/global-ev-outlook-2021/prospects-for-electric-vehicle-deployment.

M. Strelec (B) · P. Hering · P. Janecek NTIS – Research centre, University of West Bohemia, Technická 8, 301 00 Plzen, Czech Republic e-mail: [email protected] © The Author(s) 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_6

59

60

M. Strelec et al.

Fig. 1 Considered time horizons

time horizons. In this contribution, various DSO’s procedures are considered, which are depicted in Fig. 1 along time axis. The dynamic range stands for the closest time range to target time, where dynamic electrical phenomena can occur (e.g. voltage transients, sub-synchronous response etc.). However, dynamic range will not be treated in this article. On time horizon longer than minutes, electrical dynamics can be neglected, and only electromechanical phenomena remain. According to various DSO’s procedures and targets, overall time horizon can be split in several categories: • Long-term grid planning—relates to regional or national development, which reflects needs driven by demographical, societal, technological and economical changes. Long-term power systems planning tasks consider a large time horizon (years to decades), focus on generation as well as transmission expansion planning, policy development, and investment decisions.2 In the context of massive EV integration, power systems need to be designed for robust and reliable operation under highly volatile spatiotemporal consumption and generation patterns driven by EV charging, and renewable production respectively. • Short-term operational planning—embraces decisions in power systems (e.g., distribution system) valid for near future of operational time (week-ahead, dayahead).3 On this horizon, optimal settings of controllable network elements are calculated under consideration of expected power injections, power network constraints (power system congestions, voltage constraints) and frequently under N-k security conditions. Controllable network elements can change network topology (e.g., switching equipment), power injections (e.g., controllable loads) or power network equipment settings (e.g., tap changer positions of transformers). In this time range, technical calculations are based on short-term power injections forecasts (e.g., renewable productions, power consumptions). In consideration of high presence of electromobility, short-term forecasts for technical calculation shall cover spatiotemporal EV mobility patterns causing volatile consumption demand across distribution networks. • Real-time operation—stands for time frame closely preceding target time (e.g., minute or seconds). Here, fast operational decisions are triggered to solve network problems which could not be anticipated in a short-term operational planning or 2

Ankita Singh Gaur et al., Long-term energy system planning considering short-term operational constraints, Energy Strategy Reviews, Volume 26, 2019. 3 S. Grenard and O. Carre, “Optimal short term operational planning for distribution networks,” CIRED 2012 Workshop: Integration of Renewables into the Distribution Grid, Lisbon, 2012, pp. 1–4. https://doi.org/10.1049/cp.2012.0718.

Definition of a Set of Indicators for the EV Impact Assessment

61

planning inefficiencies caused by inaccurate forecast of volatile injections. Mostly, these decisions are made by network automation or by operating centers of a DSO. For example, available power capacity for EV charging stations can be optimized according to actual state of localized part of power network. Although various abovementioned time horizons with different procedures and decisions are presented in power system domain, most of processes rely on technical calculations (e.g., load flow/state estimation) having similar calculation fundaments and functionalities from an abstracted point of view, which can be shared across various time domains. Technical calculations such as load flow or state estimation evaluate power system state and rely on several inputs (network topology and properties, network elements settings and power injections), which can be obtained from various sources as depicted in Fig. 2. Typically, input datasets comes from: (i) development scenarios in the case of network planning, (ii) expected power injections and network settings in the case of the short-term operational planning or (iii) actual measurement of power injections and current network settings in the case of real time operation. The EV activities in the real time affect mainly power injections on network nodes, where charging infrastructure is deployed, and are included in power injection measurements. In the case of long-term network planning or short-term operational planning, behavior models capturing spatio-temporal character of EV charging need to be developed to appropriately assign EV charging or discharging power profiles to relevant power network nodes. Two main approaches to calculate power system state are frequently applied: (i) load flow and (ii) state estimation. Generally, power network state is determined by four different variables (e.g., nodal voltage amplitudes, nodal voltage angles, active and reactive power injection). In the case of load flow, only two variables are known across power network, while the others are calculated by numerical algorithms such as Newton–Raphson, FastDecoupled and others. On the other hand, more than two state variables are known across a power network. In this case, the numerical solution of a power system state (i.e., load flow calculation) cannot be applied. Therefore, optimization algorithms (e.g., least squares, weighted least squares) are used for the determination of a solution in the case of state estimation. Both calculation approaches lead to full determination of a power system state. Based on the calculated state, measurable indicators can be evaluated regardless of the applied calculation algorithm. The power system state can be used for an assessment of a power system condition defined by various measurable indicators, which are described in following section in more details.

2 Measurable Indicator for Network State Assessment In this section, measurable indicators for network state assessment at various technological levels are collected. Following levels are considered: • Power network level—represents power system-wide level, where set of relevant network elements are considered.

62

M. Strelec et al.

Fig. 2 Segment of technical calculations

• Connection point level—stands for point in a power system where consumers (generally prosumers) are connected. • Technology level—is a level behind connection point, where given technologies are located. Power Network Level Peak Power Maximum Peak Power Injection This indicator represents maximum active power injection Pinmax j (t) over whole set of B buses in a power network for given timestamp t. The measurable indicator is defined as ( ) Pinmax j (t) = max Pin j,i (t) i∈{1,2,...,B} ,

(1)

Definition of a Set of Indicators for the EV Impact Assessment

63

where Pin j,i (t) is the active power injection at the i-th bus at time t. Minimum Peak Power Injection This index represents minimum active power injection Pinmin j (t) over the set of B buses for the given timestamp t. Hence, ( ) Pinmin j (t) = min Pin j,i (t) i∈{1,2,...,B} ,

(2)

where Pin j,i (t) is the active power injection at the i-th bus at time t. Average Power Injection avg This indicator represents the average active power injection Pin j (t) over the set of B buses for the given timestamp t, which can be defined as avg

Pin j (t) =

B 1 Σ Pin j,i (t). B i=1

(3)

Peak Power Injection Reduction This indicator tracks the relative reduction of the value of the maximum peak load on a given period (day, month, year) [1], which is expressed as peak

Pload (t) =

Pinmax j (t) ) max Pinmax (k) j (

,

(4)

k∈{k1 ,k2 ,...,k T }

where the set {k1 , k2 , . . . , k T } represents indices of T time instants of the considered time horizon. Power Injection Quantile This indicator in a probabilistic manner tracks the power injections over all the network buses. Selected quantile Q p of power injections (e.g., Q 90 − 90% quantile) provides the upper bound of power injections at time t with the given p-percentual probability. Is can be mathematically described as ( ) Q Pin jp (t) = Q p Pin j,i (t) i∈{1,2,...,B} .

(5)

Peak Power Load Factor This indicator gives the proportion of the maximum peak load power to the average load power at a given time t [1], which is calculated as pf

Pload (t) =

Pinmax j (t) avg

Pin j (t)

.

(6)

64

M. Strelec et al.

Peak Power Generation Factor This indicator gives the proportion of the maximum peak generation power to average power injection at a given time t [1], which is calculated as pf Pgen (t) =

Pinmin j (t) avg

Pin j (t)

.

(7)

Peaks Above Limits This indicator is designed to track the percentage of time where the net exported energy exceeds a certain limit [2]. Energy Losses This indicator is designed to track the volume of electrical energy losses in a distribution grid (e.g., conductors, transformers, etc.). Eloss (t) is provided as a result of load flow calculation [1]. Voltage Magnitudes Voltage Variation This indicator represents the difference between the actual voltage V act (t) supplied to MV/LV users and the nominal value V base (t) [1]. The variable is calculated as v˜ = V act (t) − V base (t).

(8)

Voltage Mean Variation This indicator represents average bus voltage magnitude variation [1], which is expressed as ΣB μv˜ (t) =

˜i (t) i=1 v B

.

(9)

Voltage Standard Deviation This indicator represents the standard deviation of the voltage magnitude variation over all the considered B buses [1], which is defined as / σv˜ =

ΣB

˜i (t) i=1 (v B

− μv˜ (t))2

.

(10)

Fulfilment of Voltage Limits This indicator is designed to evaluate the fulfilment of regulatory voltage limits in distribution networks [4].

Definition of a Set of Indicators for the EV Impact Assessment

65

Voltage Quality (Voltage Factor) This indicator is used to evaluate the number and duration of voltage dips involving a given user or group of users [4]. Branches Loading Relative Branch Loading This indicator is designed to track relative loading of a branch (e.g., power line, power transformer etc.), which is defined as r el Pbr,b (t) =

act Pbr,b (t) limit Pbr,b

, b ∈ {1, 2, . . . , M},

(11)

limit act where Pbr,b is nominal loading and Pbr,b (t) is actual loading of the branch b and M is the total number of selected branches.

Maximum Relative Branch Loading This indicator represents maximum value of the Relative branch loading index over the selected branches at a given timestamp t, which can be written as ( r el ) Pbrmax (t) = max Pbr,b (t) b∈{1,2,...,M} .

(12)

Quantile Relative Branch Loading This indicator provides p-percentual quantile of the relative branch loading over the selected set of branches at time instant t, which is defined as ( r el ) Q Pbr p (t) = Q p Pbr,b (t) b∈{1,2,...,M} .

(13)

Others Total Harmonics Distortions (THD) This indication represents the of the equivalent root mean square (RMS) voltage of all the harmonic frequencies over the RMS voltage of the fundamental frequency [1] and can be calculated as /Σ ∞ 2 n=2 Vn (t) , (14) T H D(t) = V1 (t) where Vi is the root mean square voltage of i-th harmonics. CO2 Emissions This indicator is designed to provide the amount of the direct C O2 emissions (in kg) for the consumption of a specific (electrical) energy amount over the time horizon T [1]. The calculation can be expressed as

66

M. Strelec et al.

C O2 (t) =

Σ

Pgen,i (t) · C O2mi x,i ,

(15)

i

where Pgen,i stands for power generated by i-th generating unit or set of generating units and C O2mi x,i is the corresponding CO2 mix of the i-th unit. Note, the actual value of C O2mi x,i strongly depends on the assessment methodology, which can embrace fossil as well as renewable energy generation units in some evaluation approaches. Reliability SAIFI System Average Interruption Frequency Index stands for an average number of power interruption for each consumption point (i.e. customer) [1]. S AI F I =

Ninterr , total Ncust

(16)

total where Ninterr is the total number of interruptions and Ncust stands for total number of customers.

SAIDI System Average Interruption Duration Index stands for an average outage duration for each consumption point [1]. S AI D I =

interr Tcust , total Ncust

(17)

interr total where Tcust is the customer interruption durations and Ncust stands for total number of customers

CEMI n Customers Experiencing Multiple Interruptions index indicates percentage of consumption points (i.e. customers) experiencing at least n interruptions (e.g., n = 4) [5]. C E M In =

interr,n Ncust , Ncust

(18)

interr,n where Ncust is the total number of customers experienced to more than n interruptions.

CAIDI Customer Average Interruption Duration Index, the total interruption duration for all consumption points (i.e. customers) divided by the total number of interruptions for all consumption point (i.e. customers) [6].

Definition of a Set of Indicators for the EV Impact Assessment

C AI D I =

67

interr Tcust . interr Ncust

(19)

ASAI Average Service Availability Index, the customers’ access to electricity as the percentage of hours during which the customers have access to electricity out of the total hours during the period of study [6]. Σ AS AI = 1 −

i Tr est,i

interr · Ncust,i

Ncust,i · T

,

(20)

interr where Tr est,i is the restoration of time of i-th interruption, Ncust,i represents total number of the customers interrupted by i-th outage, Ncust,i stands for the total number of customers during i-th outage and T is the considered period.

ASUI Average Service Unavailability Index, the unavailability of energy, or the percentage of interruption hours out of the total hours during the period of study [6]. ASU I = 1 − AS AI,

(21)

Consumption Point Level Average Power Demand This index is designed to track the average demand at the i-th bus over the time horizon T , which is calculated as avg Pload,i

T 1 Σ = Pload,i (t + k), i ∈ {1, 2, . . . , B}. T k=1

(22)

Maximum Power Demand This indicator represents the maximum demand at the i-th bus over the time horizon T . The indicator can be computed as ( ) max Pload,i = max Pload,i (t + k) k∈{1,2,...,T } , i ∈ {1, 2, . . . , B}.

(23)

Minimum Power Demand This indicator represents the minimum demand at the i-th bus over the time horizon T , which is defined as ( ) min Pload,i = min Pload,i (t + k) k∈{1,2,...,T } , i ∈ {1, 2, . . . , B}.

(24)

68

M. Strelec et al.

Peak to Average Load Ratio The peak-to-average demand ratio is a measure expressing ratio of the maximal avg max to annual average demand Pload [1]. It can be mathematannual peak demand Pload ically expressed as ratio Pload =

max Pload avg . Pload

(25)

Average Power Generation The indicator is a measure of average generation at the given bus over the time horizon T . The calculation of the indicator is defined as avg

Pgen,i =

T 1 Σ Pgen,i (k), i ∈ {1, 2, . . . , B}. T k=1

(26)

Maximum Power Generation This indicator represents the maximum generation at the given bus over the time horizon T , where the mathematical expression can be defined as ( ) max Pgen,i = max Pgen,i (t + k) k∈{1,2,...,T } , i ∈ {1, 2, . . . , B}.

(27)

Minimum Power Generation This indicator represents the minimum generation at the given bus over the time horizon T , where the indicator can be defined as ( ) min Pgen = min Pgen (t + k) k∈{1,2,...,T } .

(28)

Peak to Average Generation Ratio The peak-to-average generation ratio stands for a ratio between annual peak avg max and annual average generation Pgen , which can be calculated as generation Pgen ratio Pgen =

max Pgen avg

Pgen

.

(29)

Flexible Energy Demand This indicator represents consumed energy at a given bus over the time horizon T , where the consumption is of a flexible nature. E f lex_load,i =

T Σ k=1

P f lex_load,i (t + k) · h, i ∈ {1, 2, . . . , B},

(30)

Definition of a Set of Indicators for the EV Impact Assessment

69

where P f lex_load,i (t) is the flexible load at the i-th bus at time instant t and h is the length of the measurement sampling period. Inflexible Energy Demand This indicator represents consumed energy at the given bus over the time horizon T , where the consumption is of an inflexible nature (i.e. critical). Thus, overall measure can be determined as E crit_load,i =

T Σ

Pcrit_load,i (t + k) · h, i ∈ {1, 2, . . . , B},

(31)

k=1

where Pcrit_load,i (t) is the inflexible consumption at the at the i-th bus at time instant t. Energy Demand This indicator provides the consumed energy at a given bus i, i ∈ {1, 2, . . . , B}, over the time horizon T . The mathematical expression can be formulated as follows Eload,i =

T Σ

Pload,i (t + k) · h = E f lex_load,i − E crit_load,i .

(32)

k=1

Renewable Energy Generation This indicator provides energy generated by renewables at a given bus over the time horizon T [1], which can be evaluated by Er es,i =

T Σ

Pr es,i (t + k) · h, i ∈ {1, 2, . . . , B}.

(33)

k=1

Conventional Energy Generation This indicator represents energy generated by conventional generating units at a given bus over the time horizon T . The mathematical expression is E conv =

T Σ

Pconv (t + k) · h, i ∈ {1, 2, . . . , B}.

(34)

k=1

Energy Generation This indicator is a measure of energy generated at a given bus over the certain time period T, which is calculated as E gen,i = Er es,i + E conv,i , i ∈ {1, 2, . . . , B}.

(35)

70

M. Strelec et al.

On Site Energy Ratio This indicator represents a share of generation in the energy mixture, which is computed as the ratio of energy generation E gen to the energy demand E cons [2]. Thus, ratio E gen =

E gen . E cons

(36)

Non-consumed On-Site Energy The indicator provides the share of generated energy over the time horizon T [2], which is not consumed on-site I ΣT I I I k=1 Pload (t + k) − Pgen (t + k) + Pstorage (t + k) match E site = . (37) Eload On Site Renewable Energy Ratio On site renewable energy ratio Erratio es provides share of RES in the energy mixture. It is the ratio of energy RES generation Er es to energy demand E cons calculated as Erratio es =

Er es . E cons

(38)

Demand Flexibility Ratio Demand flexibility ratio E ratio f lex defines the ratio of the flexible to total energy consumption [1], which is defined as E ratio f lex =

E f lex_load . E cons

(39)

Demand Flexibility Request The indicator determines the amount of the requested energy flexibility demand over the time horizon T [1], where the mathematical expression is sp

E f lex_load =

T 1 Σ sp P (t + k) · h, T k=1 f lex_load

(40)

sp

where P f lex_load (t) is requested flexible load at time t. Demand Request Participation This indicator represents the ratio of the End User’s actual participation to Demand Flexibility over the requested Demand Flexibility over the time horizon T [1], which is calculated as

Definition of a Set of Indicators for the EV Impact Assessment

sp_act

E f lex_load = 1 −

I ΣT II sp I k=1 I P f lex_load (t + k) − P f lex_load (t + k)I · h sp

E f lex_load

71

.

(41)

Power Imported from the Grid This indicator is designed to track the power imported from a connected distribution grid, which is defined as { import Pgrid (t)

=

Pgrid (t), i f Pgrid (t) ≤ 0 , 0, other wise

(42)

where Pgrid (t) is the power flow between the connected distribution grid. Power Exported to the Grid This indicator is designed to track the power exported to a connected distribution grid, which can be formulated as { ex por t Pgrid (t)

=

Pgrid (t), i f Pgrid (t) > 0 . 0, other wise

(43)

Energy Exported to the Grid This indicator is a measure of the amount of energy exported to a connected distribution grid on a given time horizon T [1], which can be calculated as ex por t

E grid

=

T 1 Σ ex por t P (t + k) · h. T k=1 grid

(44)

Energy Imported from the Grid This indicator is a measure of the amount of energy imported to a connected distribution grid on the given time horizon T [1], which is specified as import

E grid

=

T 1 Σ import P (t + k) · h. T k=1 grid

(45)

Energy Mismatch Ratio This indicator provides the ratio of exported energy to imported energy on the given time horizon T [1] defined as ex por t

ratio E grid

=

E grid

import

E grid

.

(46)

72

M. Strelec et al.

Capacity Factor The indicator is designed to provide the ratio between the energy exchanged between the consumption point and the grid and the energy exchanged that would have occurred at nominal connection capacity [2], which is expressed as CF =

I I I Pgrid (t)I max Psite (t)

.

(47)

Maximum Hourly Surplus Difference This indicator evaluates over the time horizon T the maximum value of how much greater the hourly local renewable supply Pr es is than the demand Pload during that hour [2]. Thus, the mathematical expression can be defined as follows max Psite = max(Pr es (t + k) − Pload (t + k))k∈{1,2,...,T } .

(48)

Volume Shifted Flexibility Factor This indicator reflects the ability to shift electric load from peak to off peak hours over the time horizon T , in terms of energy shift compared to a reference profile Pbase_load [3]. Hence, following formula can be introduced shi f ted

E f lex_load =

I I I T II Pbase_load (t + k) − P act Σ f lex_load (t + k)I k=1

Pbase_load (t + k)

.

(49)

Technology Level Electromobility (EV) EV Penetration Level This indicator is designed to track the percentage of EVs registered in a specific country over the total number of registered vehicles [1]. The indicator can be calculated as E V pen =

EV Ncar , Ncar

(50)

EV where Ncar is the overall number of cars and Ncar is the count of electric cars.

EV Peak Demand This indicator tracks the peak electricity demand for the charging of EVs over the time horizon T [1] defined as peak

PE V

= max(PE V (t + k))k∈{1,2,...,T } ,

(51)

Definition of a Set of Indicators for the EV Impact Assessment

73

where PE V (t) is the electricity demand for the EV charging at time t. Battery State of Charge (SoC) SoC is defined as the remaining capacity of a battery, it is affected by its operating conditions such as load current and temperature [1], which is defined as E SoC (t) =

E act (t) , E max (t)

(52)

where E act (t) is the actual battery energy and E max (t) is the maximum battery capacity at time t. Depth of Discharge (DoD) DoD is used to indicate the percentage of the total battery capacity that has been discharged [1]. For deep-cycle batteries, they can be discharged to 80% or higher of DoD. The calculation can be expressed as E DoD = min(E SoC (t + k))k∈{1,2,...,T } .

(53)

Average SoC This indicator is designed to measure average SoC of a battery over the time horizon T . If the battery spends a significant amount of time at a high SoC, it will degrade faster than if it is left and maintained at a mid-level state of charge [1]. Some batteries are more sensitive to this than others, but generally it is known that the higher the average SoC over the battery life, the faster it will degrade. Similarly, if a battery is kept at very low average SoC, it will also degrade quickly. Hence, avg

E SoC =

T 1 Σ E SoC (k). T k=1

(54)

3 Suitability of Performance Indicators for Different Time Horizons In the previous sections, the collection of measurable indicators for a grid assessment has been presented. These indicators cover a wide range of grid aspects, where every indicator cannot be fully reasonable in some time horizons. This section focuses on the assessment of the suitability of proposed indicators to considered time horizons. A suitability assessment is performed for each indicator and time horizon- where appropriate - is represented by filled squares (Table 1).

74 Table 1 Suitability of discusses measurable indicators for various time horizons

M. Strelec et al.

Definition of a Set of Indicators for the EV Impact Assessment

75

Resulting from the table above, different measurable indicators are suitable for different time horizons, always depending on the context. The assessment was made based on preferences, possibilities and focus of the EU-project “XL-Connect” (Project title: Large scale system approach for advanced charging solutions), where grid impact of a massive EVs integration will be evaluated in proposed time horizons.

4 Conclusion An evaluation of grid impacts always depends on assessment purpose, which can be considered from various stakeholder’s point of view (e.g. distribution system operators (DSOs), municipalities, charge point operators (CPOs)) and procedures are mostly connected with different time horizons. However, the majority of available measurable indicators for reflecting grid properties are not universal from the time horizon point of view. Therefore, a suitability of indicators needs to be considered in various grid assessment. In the project XL-Connect, tools and methods enabling a massive penetration of EVs into power systems are focus of the research and development activities, where an evaluation of the EV impacts on power system is one of the key components of the project. In the project, the presented collection of measurable indicators will be used to impact assessment of EV penetration from various perspectives (i.e. consumption point, power network, EVs or battery). The suitability of specified indicators for various time horizons were evaluated, which will be considered in the EV penetration assessment under different process context (e.g. short-term operational planning, planning processes). The evaluation of indicators represents preferences, requirements and outcomes from research activities in the XL-Connect project. Acknowledgements

Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor the granting authority can be held responsible for them.

References 1. inteGRIDy project (2017) D 1.4. - inteGRIDy global evaluation metrics and KPIs. Integridy project website. http://www.integridy.eu/sites/default/files/integridy/public/content-files/delive rables/inteGRIDy_D1.4_Evaluation_Metrics_KPIs_v1.0.pdf

76

M. Strelec et al.

2. Airò Farulla G, Tumminia G, Sergi F, Aloisio D, Cellura M, Antonucci V, Ferraro M (2021) A review of key performance indicators for building flexibility quantification to support the clean energy transition. Energies 3. Pramangioulis D, Atsonios K, Nikolopoulos N, Rakopoulos D, Grammelis P, Kakaras E (2019) A methodology for determination and definition of key performance indicators for smart grids development in island energy systems. Energies. 4. UpGrid Project (2018) D8.5: summary of results obtained in WP8 and recommendation 5. Harder WJ (2018) Key performance indicators for smart grids master thesis on performance measurement for smart grids 6. Mansouri MR, Simab M, Bahmani Firouzi B (2021) Impact of demand response on reliability enhancement in distribution networks. Sustainability

Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. The images or other third party material in this chapter are included in the chapter’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

Protocols and Interfaces for EV Charging Christopher Hecht, Jan Figgener, and Dirk Uwe Sauer

Abstract This chapter introduces the most widely used communication protocols, physical connectors and involved parties in public and private electric vehicle charging. The listed communication protocols relate to vehicle-charging station communication (IEC 61851-1, ISO 15118, CHAdeMO, GB/T) and other communication paths (OCPP, openADR, SPINE, OCHP, OICP, eMIP, OCPI, and OSCP). The introduced physical connectors are those that can be found in Europe and the United States (Type 2, CCS, Tesla, MCS, CHAdeMO) and East Asia (Type 1, GB/ T, CHAdeMO, ChaoJi). Virtually all electric vehicles today are equipped with one or two of the listed standards. The third component of this chapter is the introduction of the involved parties that organize the charging market such as CPOs, EMPs, customers, roaming platform operators, energy suppliers, grid operators, and aggregators. The roles and the relationship between the roles are outlined. Keywords Interfaces · Protocols · Communication · Charging stations

C. Hecht (B) · J. Figgener · D. U. Sauer Chair for Electrochemical Energy Conversion and Storage Systems, Institute for Power Electronics and Electrical Drives (ISEA), RWTH Aachen University, Jägerstrasse 17-19, 52066 Aachen, Germany e-mail: [email protected] Helmholtz Institute Münster (HI MS), IEK 12, Forschungszentrum Jülich, 52425 Jülich, Germany Institute for Power Generation and Storage Systems (PGS), E.ON ERC, RWTH Aachen University, Mathieustrasse 10, 52074 Aachen, Germany D. U. Sauer Jülich Aachen Research Alliance, JARA-Energy, Templergraben 55, 52056 Aachen, Germany © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024 B. Brandstätter and A. Steiner (eds.), Next Generation Electrified Vehicles Optimised for the Infrastructure, Automotive Engineering : Simulation and Validation Methods, https://doi.org/10.1007/978-3-031-47683-9_7

77

78

C. Hecht et al.

1 Introduction The world of electromobility is a fast-changing ecosystem of various actors that need to interact with each other in a variety of ways, while at the same time reducing complexity towards the user as much as possible. This chapter defines the most relevant parties, physical connectors, software communication protocols. We hope that in doing so, we offer readers a better understanding over which part of the system may be of relevance for their particular challenge. Outside the scope of this chapter is a more general market description such as how the public and private charging infrastructure market develops or what different vehicle manufacturers focus on.

2 Involved Parties In the world of charging infrastructure, several parties are defined. In practice, these roles are sometimes overlapping, if for instance an energy supplier also operates their own charging stations. The defined roles in the context of this work are: – – – – – – –

Charge point operator (CPO) E-Mobility service provider (EMP) Customer Roaming platform operator Energy supplier Grid operator Aggregator

The parties communicate with each other and exchange energy and payments via the protocols and interfaces that are introduced in this chapter. Each role is introduced in the following in brief. Figure 1 shows a simplified block diagram of these parties and their interactions. Note that there may be other parties of relevance such as navigation system operators or public administrators, but interfaces are less unified towards these actors and they are omitted in this work. Charge Point Operator (CPO) A CPO is responsible for the correct operation of a charging station. This includes construction, maintenance, and operation. For smaller CPOs, much or all of the technical processes can be outsourced to backend operators who act as white label companies and ensure that all IT processes run smoothly. Hardware maintenance and planning can also be outsourced to local maintenance firms. E-Mobility Service Provider (EMP) The EMP is the party which has a contractual relationship with the customer. As such, they offer RFID-cards, phone apps, contracts embedded via Plug and Charge, and alike. These tokens allow customers to identify themselves with CPOs. The EMP

Same entity or direct contract

Customer

Uses grid access

Single contract for access and payment

Grid operator

Purchases electricity

CPO Offers flexibility

Charging via a contract

EMP

Roaming platform operator

79

Ad-hoc charging

Single contract for access and payment

Protocols and Interfaces for EV Charging

Energy supplier

Aggregator

Fig. 1 Simplified block diagram of the involved parties and their interactions. Note that the diagram is restricted to the most relevant parties and their most relevant interactions

is billed by the CPO for the customers’ usage of the charging station. In turn, they send a bill to the customer. Oftentimes, EMP and CPO are the same entity. If this is not the case, then they can either have bilateral contracts or, more typical, interact through a roaming platform operator. Customer The customer is the person or organisation operating the electric vehicle that should be charged. The customer typically has a contractual relationship with the EMP or uses ad-hoc charging where the energy is paid on the spot with a credit card or via an online payment service. Roaming Platform Operator Since there are thousands of CPOs in Germany alone,1 it is not feasible that each CPO should have a direct contractual relationship with each EMP. Instead, several roaming platforms have emerged with the three largest ones in Europe being Hubject,

1

https://www.electrive.net/2021/12/23/bundesnetzagentur-meldet-50-901-oeffentliche-ladepu nkte/.

80

C. Hecht et al.

Gireve, and e-clearing.2 Among these, Hubject is currently the market leader. CPOs and EMPs enter a contract with the roaming platform to allow the usage of charging stations for customers while they are at CPOs other than their home CPO. The roaming platform operator is responsible for the operation of the IT infrastructure of the roaming platform. Energy Supplier The energy supplier is the organisation that provides energy to the charging stations. As such, the energy supplier has a contractual relationship with the CPO. Grid Operator The grid operator is the entity operating the electricity grid that supplies the charging station. In most electricity system, a difference is made between the distribution system operator (DSO) and transmission system operator (TSO). The DSO operates the low to medium voltage grid that is directly connected to the charging station. The main responsibility of the DSO is to ensure that electricity can be supplied to the charging station at all times and to ensure that the local network is not overloaded. The TSO in turn operates higher voltage levels. In the context of charging infrastructure, two key tasks of the TSO are relevant: They need to ensure that no overloading on the higher grid levels occurs and they have to maintain the grid frequency at the desired frequency, which is 50 Hz in Germany. Aggregator If the vehicle and charging station are used to perform flexibility services, an aggregator is usually required. The reason for this is that minimum bid sizes for control power start at one megawatt. Arbitrage trading would theoretically be possible with powers starting from 25 kW at the wholesale electricity market, but the administrative processes are usually only done by large aggregators with vehicle pools. For a single vehicle, 43 kW are theoretically possible via the Type 2 AC interface, but practically no vehicles with 43 kW and few with 22 kW are on the market. Instead, most vehicles charge with a power of approximately 10 kW3 and consequently, 100 vehicles or other assets are required to reach a pool of one megawatt. The role of the aggregator is to organize the pool and ensure that all marketed assets actually deliver the advertised service. Examples of such companies are The Mobility House, Jedlix, and Clever, although they typically are not limited to the role of an aggregator, but perform other market activities as well. Given the still young stage of the market, no market shares can be determined at the time of writing.

2

https://medium.com/extrawest/eroaming-hubs-for-cpo-emp-which-one-to-choose-2d975c 49ead4. 3 https://www.sciencedirect.com/science/article/pii/S258900422201906X.

Protocols and Interfaces for EV Charging

81

Table 1 Dominant connector types used across the world. Connector types in brackets are those that either have a lower market share or have not yet been rolled out AC charging

DC charging

Europe, USA, South Korea

Schuko, Type 2, Type 1, Tesla

CCS, (CHAdeMO, Tesla/NACS, MCS)

China

GB/T AC, (ChaoJi)

GB/T DC, (ChaoJi)

Japan

CHAdeMO, (ChaoJi)

CHAdeMO, (ChaoJi)

Rest of the world

Largely undecided

3 Physical Connectors This section introduces the physical connector pieces that are used in electromobility worldwide. After an initial phase of uncertainty, several dominant connectors have evolved depending on geographical area and type of power transfer. Table 1 provides an overview over which connectors are used in which mature markets. Table 2 expands on this by showing the most dominant applications of each connector type as well as the rated voltage, current, power, and voltage mode. During the early stages of electromobility, many connectors were tried and in use. This uncertainty was a hindrance especially in the very early years since the risk of purchasing a charger or vehicle that would soon become obsolete was high. For AC charging, the debate was settled relatively soon and Europe, the USA, and Japan decided to use Type 1 or Type 2 connectors. Since it is simple to build adapters between the two systems, few standardization issues in terms of the physical layout existed. The question was rather what power level would be adequate. The Type 2 connector used in Europe, for instance, is specified to allow up to 43 kW charging power. Very few vehicles, however, support such a high power rating, as it is not required for everyday mobility. Additionally, a large rated power of the onboard converter for transforming AC to DC typically leads to larger losses. The vast majority of vehicles nowadays therefore is able to charge with 11 kW via the Type 2 charger4 with a few exceptions also supporting up to 22 kW of charging power. China is the only major electric vehicle market that does not use either of the two options. The Chinese GB/T AC charger is very similar to the Type 2 charger though with the key difference being that male and female layout are swapped between plug and socket. On the DC side, the competition was and is much fiercer though. For a long time, CCS and CHAdeMO were both installed on vehicles and charging stations in Europe and the US. With the decision of the European Union to require all charging stations to host a CCS connector, CCS was de facto declared as the standard charger in Europe. Since the decision, a clear trend towards CCS can be observed and CHAdeMO is losing ground in these markets. South Korea followed the move in 20185 leaving only Japan as the only country using predominantly CHAdeMO. China introduced another 4 5

Mobility charts. https://chargedevs.com/newswire/south-korea-to-officially-adopt-ccs-fast-charging-standard/.

82

C. Hecht et al.

Table 2 Main purpose and technical ratings of EV connectors Connector

Vehicles

AC or DC

Max ratings

Type 2

Cars

3-phase AC

240 V 64 A 43 kW

Type 1

Cars

1-phase AC

120/240 V 70 A 16.8 kW

NACS by Tesla

Mostly Tesla cars, but in North America, several manufacturers announced a switch to NACS.

1-phase AC or DC

240 V (AC), 500 V (DC) 48 A (AC), 400 A (DC) 11.5 kW (AC), 250 kW (DC)

CCS

Cars (fast-charging), trucks

DC

1000 V 500 A 350 kW

MCS

Trucks, Airplanes, Ships

DC

1250 V 3000 A 3.75 MW

CHAdeMO

Cars (fast-charging)

DC

1000 V 400 A 400 kW

GB/T

Cars, trucks

3-phase AC or DC

440 V (AC), 1000 V (DC) 63A (AC), 250 A (DC) 28 kW (AC), 250 kW (DC)

ChaoJi

Cars, trucks

DC

1500 V 600 A 900 kW

DC fast-charger in 2015 under the standard GB/T 20234.3-2015 commonly referred to as GB/T DC. A further option in the market is the standard introduced by the company Tesla. It is the only standard successfully introduced by a single company and is used at Tesla superchargers globally. In 2022, Tesla published the previously proprietary standard under the acronym NACS and several large car manufacturers have since announced their intent to adopt the standard. For future developments, higher charging powers are required for trucks and heavy duty vehicles, which the EU and the USA aims to achieve through the MCS charger while China and Japan push the ChaoJi connector. Both are in a test phase at the moment. Overall, the DC charging market is much more contested and has undergone significant transformation following technical developments such as larger batteries with higher power

Protocols and Interfaces for EV Charging

83

capabilities. It is also uncertain how the remaining regions of the world will decide and how challenging a retrofit from one connector to the other would be if vehicles are sold secondhand from one area to another. Type 2 The type 2 connector (also called “IEC 62196 type 2” or “Mennekes”) is the standard AC charger in Europe. It supplies 220–240 V across 3 phases with amperages of 16 A, 32 A, or 64 A leading to charging powers of 11 kW, 22 kW, and 43 kW respectively. For private wallboxes, 16 A/11 kW is mostly used while public charging stations mostly offer 32 A/22 kW. Type 1 The type 1 connector (also called “SAE J1772 type 1” or “J plug”) is a connector type frequently used in the USA and in Japan for AC charging. It delivers 120 or 240 V depending on area of application on a single phase. Current ratings between 6 and 70 A are theoretically possible, but particularly the higher current ratings have not seen a wide adoption. The cable is locked in place with a manually operated hook. If the button holding this hook in place is pushed, power flow is interrupted. While convenient and safe, this aspect makes it possible for unauthorized third parties to stop the charging process and even to steal the cable. Tesla/NACS Tesla is the only carmaker that developed its own charging standards (since 11.2022 also known as North American Charging Standard or NACS) that is still in widespread use. The connector can be used both in AC and DC mode. It is rated at up to 48 A at 240 V in AC mode leading to a charging power of 11.5 kW. In many markets, however, Tesla vehicles will be equipped the local connector types. In the DC variant, the connector is rated for 500 V and 200 A with maximum allowed current reaching 400 A. This would allow up to 100 kW of charging power at rated current and 200 kW at peak current, but 250 kW are typically stated as Tesla does not specify a current rating outside maximum temperature boundaries.6 Given that vehicle batteries cannot set their voltage at will, lower values will have to be expected in practice. Tesla’s V3 charging stations support the Tesla connector and the CCS connector in Europe. General Motors, Ford and others have announced support for NACS in new car models in North America. CCS The Combined Charging System (CCS) is the fast-charger standard that is widely used in Europe and the USA. The connector extends the local AC variants with an extra pair of pins added below the connector for DC charging. CCS is designed to 6

https://tesla-cdn.thron.com/static/HXVNIC_North_American_Charging_Standard_Technical_ Specification_TS-0023666_HFTPKZ.pdf?xseo=&response-content-disposition=inline%3Bf ilename%3D%22North-American-Charging-Standard-Technical-Specification-TS-0023666. pdf%22.

84

C. Hecht et al.

provide between 200 and 1000 V with the goal of supporting the two typical voltage levels 400 and 800 V. The maximum rated current is 500 A. Since the system is not specified to simultaneously supply maximum voltage and current, the maximum power output is 350 kW reached with 920 V and 380 A.7 In practice, there are three factors limiting the realized power level: Since the voltage is the battery pole voltage, it cannot be chosen freely and will be lower in most applications. Additionally, only some vehicles are able to charge at ~800 V and especially vehicles with a smaller battery are only able to charge at ~400 V. This limits real-world charging speeds for many vehicles to approximately 150 kW. Lastly, several components such as the connector, cable, or battery may suffer from overheating. Liquid cooling in cables and connectors as well as air or liquid cooling for the battery may remedy this issue, but make the device more complex. MCS The Megawatt Charging System (MCS) is a new charging standard proposed for electric trucks and other heavy-duty vehicles. The connector only contains DC contacts and no AC contacts. The system is rated at 1250 V and 3000 A leading to a power rating of 3.75 MW. Except for the physical connector, the system is largely based upon the CCS standards. Vehicles with an MCS connector will very likely also host a CCS connector to enable charging at existing infrastructure. Although the system is not yet in use in practice, many governments especially in Europe are actively pushing for a fast deployment of the standard to support the expected growth in electric trucking. CHAdeMO CHAdeMO (“CHArge de MOve”) is the connector developed by Japanese manufacturers. The connector solely hosts DC contacts. For AC charging, a type 1 connector is typically installed additionally. An early key feature of CHAdeMO is the built-in bidirectional capability, which has only been introduced in 2022 for CCS connectors. In the 2.0 specification, CHAdeMO is able to charge at 400 kW with 1000 V and 400 A. For several years, CHAdeMO and CCS were sold with vehicles in Europe and the USA, but in recent years, CCS became the dominant standard at least in Europe. GB/T GB/T (name derived from the National Standards of the People’s Republic of China labelled “GB”) is a set of standards that describe charging connectors used in China. The AC variant is similar to the Type 2 charger, but male–female are switched between plug and socket. The connector is rated at up to 440 V and 63 A leading to a maximum power output of 27.7 kW. Similar to Type 2 connectors, however, most vehicles are equipped with an onboard charger that is rated at about 11 kW, which

7

https://www.researchgate.net/publication/336146725_Fast_and_Ultra-Fast_Charging_for_Bat tery_Electric_Vehicles_-A_Review.

Protocols and Interfaces for EV Charging

85

limits the practical power level. The DC variant of the GB/T charger is currently rated at up to 1000 V and 250 A leading to a maximum power output of 250 kW.8 ChaoJi ChaoJi (also called CHAdeMO 3.0) is a joint development from Chinese and Japanese actors and is intended as a next-generation DC charger of CHAdeMO and GB/T. Although not yet used and not yet accepted by standardization agencies, the standard is specified to provide between 500 and 900 kW at a voltage of up to 1500 V and 600 A.9 Similar to CCS, one concept is to co-locate the ChaoJi connector with the AC versions used nowadays in combination with GB/T in China and Type 1 in Japan and globally.

4 Communication Protocols Communication protocols enable the interaction between the various parties introduced earlier. This section introduces the key ones used in electromobility today. An overview is provided in Table 3 and further details can be found in the remainder of this chapter. We do not claim that the list is complete and some standards are certainly missing. The focus lies on the most relevant standards to avoid overloading this book chapter. Missing standards that may also be of relevance are10 : – IEC 61850-90-8: Object model for EVSE to CPO communication – ISO/IEC 27000: Data security – IEC 61140, IEC 62040, IEC 60529, IEC 60364-7-722, ISO 6469_3, ISO 17409: (Electric) safety – IEC 61439-7: Power switches for electric devices. IEC 61851-1 The IEC 61851-1 is a protocol supported by virtually every electric vehicle in Europe and many other areas. The protocol enables only very simple communication through the setting of resistance and pulse width values. Resistance values indicate the current carrying capacity of the cable and pulse width modulated signals are used to communicate the current carrying capacity of the charging station. The IEC 61851-1 is used as a first-contact communication protocol for AC charging stations as well as being used in very simple charging devices such as emergency chargers. ISO 15118 The ISO 15118 is a set of standards and protocols that define more extensive vehiclestation communication. It is superimposed on IEC 61851-1. The latest version ISO 8

https://www.mdpi.com/2624-6511/4/1/22#B76-smartcities-04-00022. https://www.chademo.com/technology/high-power. 10 After https://www.evs27.org/download.php?f=defpresentations/EVS27-4C-2840401.pdf. 9

86

C. Hecht et al.

Table 3 Overview of protocols showing communicating parties and the deployed communication channel Protocol

Parties

Communication channel

IEC 61851-1

Vehicle—Charging station

Resistance and pulse width modulation between ground and control pilot or ground and proximity pilot

ISO 15118

Vehicle—Charging station

XML files are exchanged through OSI layers defined in the protocol, mostly via control pilot and ground

CHAdeMO and GB/T

Vehicle—Charging station

CAN Bus or analogue signals can be used for signal transport

OCPP

Charging station—Backend

JSON files send via a secure HTTPS tunnel

OpenADR

Flexibility aggregator—Assets Flexible transportation layer

SPINE

Energy management Flexible transportation layer system— Smart home devices with TCP/IP natively supported through SHIP protocol

OCHP, OICP, eMIP, and OCPI Roaming platform operator—Backend OSCP

Distribution system operator or others—Backend

TCP/IP based transportation layer using API calls that deliver JSON files TCP/IP based transportation layer using API calls that deliver JSON files

15118-20 can be used for communication required to perform services such as bidirectional charging, Plug and Charge, scheduling of charging, and wireless charging if it becomes available. ISO 15118 can be used by Type 1, Type 2, CCS, and MCS connectors. CHAdeMO and GB/T CHAdeMO and GB/T define both a physical connector as well as a communication protocols. The CHAdeMO-protocol has been published in six major versions starting at 0.9, the most popular version worldwide,11 up to the 3.0 version. GB/T defined the protocol GB/T 27930 that is built atop SAE J1939. The ChaoJi-connector is a co-development between China and Japan and both protocols will be allowed on the ChaoJi-connector.

11

https://www.chademo.com/technology/protocol-development.

Protocols and Interfaces for EV Charging

87

OCPP The Open Charge Point Protocol (OCPP) is a protocol developed by the manufacturer-neutral organisation Open Charge Alliance. It enables the communication between the charging station and the IT backend of the charge point operator. Many charging stations currently support either version 1.5 or 1.6. These versions can be used for all required tasks when operating a charging station such as authentication and communication of meter readings. The newer versions starting from 2.0 are designed to support more advanced charging applications such as smart charging, bidirectional charging, and Plug and Charge. OpenADR OpenADR (Open Automated Demand Response) is a protocol published by the OpenADR alliance. The protocol is designed to be used between an aggregator of flexibility assets and the assets in its pool. Most charging stations do not support the protocol leading to the necessity of translation hardware. Oftentimes, this hardware uses OCPP to communicate with the charging station and OpenADR for communication with the aggregator. SPINE by EEBus SPINE (Smart Premises Interoperable Neutral-Message Exchange) is a protocol developed and supported by the EEBus initiative and built for communication between devices in a smart home environment. Some wallboxes and charging stations support the protocol. Others require a translator that uses OCPP to communicate with the charging station or wallbox. OCHP, OICP, eMIP, and OCPI The Open Clearing House Protocol (OCHP) is an open source protocol primarily developed by the operator e-clearing. The Open Intercharge Protocol (OICP) is a protocol developed by the roaming platform operator Hubject. The eMobility Interoperation Protocol is managed by Gireve, an organisation formed by various actors primarily in France. Lastly, the Open Charge Point Interface protocol (OCPI) is a protocol developed by eViolin, which serves stakeholders mostly in the Netherlands. The protocols serve almost identical purposes and allow communication between a CPO or EMP and a clearinghouse. The purpose of the protocols is to allow roaming charging if EMP and CPO do not have a direct contractual relationship with each other or are the same company. In Germany, Hubject and e-clearing are the dominant roaming platform operators and as such, their respective protocols are the most used ones in the market. OSCP The Open Smart Charging Protocol (OSCP) is a protocol used to exchange predictions of available power between a CPO and a distribution system operator or other parties. OSCP is supported by the Open Charge Alliance and currently available in version 2.0.

88

C. Hecht et al.

5 Trends The ecosystem developing around charging can at times be complicated to understand, partially due to the ongoing developments in the field. In the early days, many manufacturers developed own standards and industry-wide standards have only become established over time. This process is far from finished, however, and changing requirements mean that involved parties, physical connectors and communication protocols will have to continue to evolve. Currently, the biggest drivers for this change are higher power levels for larger vehicles and smart or bidirectional charging. Higher power levels have driven the development of the 800 V standard in CCS and the 1250+ V for MCS and ChaoJi connectors. These higher voltages allow for a higher power transfer without increasing current and thereby cooling requirements on plugs and cables. Smart and bidirectional charging increased the number of involved parties as vehicles became a controllable device instead of a simple load. Energy suppliers have to adapt their business model in response, grid operators have the possibility to actively control charging to avoid grid overloads, and aggregators bring the flexibility to the market. Similarly, most expansions of the involved communication protocols (ISO 15118-20,12 OCPP 2.X,13 OpenADR, SPINE) over the last few years have happened in order to accommodate intelligent and bidirectional charging. With vehicles becoming ever smarter and the entire energy industry becoming ever more interconnected, the boundaries between parties and assets are becoming blurry. If a vehicle battery is used to power an electric heat pump, communication and control of the heat pump will have to be connected to the interfaces provided by the vehicle. For these interconnected systems, the communication and control of other smart devices consequently have to be considered as well, but listing all possible standards would be outside the scope of this chapter. With passing time, new developments such as electrified flying,14 shipping15 or battery-electric trains16 will continue to enter the market and continue to change how charging interfaces and protocols look like. To allow for the innovative ideas of all actors in the market, it is important that interfaces and connectors remain open and accessible to established and new actors in the market. A successful model for these developments would be the USB connector that nowadays enables communication and power transfer between a countless variety of devices using only a small set of physical connectors.

12

https://www.iso.org/standard/77845.html. https://www.openchargealliance.org/protocols/ocpp-201/. 14 https://www.sciencedirect.com/science/article/pii/S2666691X2200032X. 15 https://www.nature.com/articles/s41560-022-01065-Y. 16 https://www.nature.com/articles/s41560-021-00915-5. 13

Protocols and Interfaces for EV Charging

89

6 Conclusion This chapter provides a snapshot of the current state of the market by outlining the key parties, connectors, and protocols in use today. While it is impossible to name every possible partner and all standards in existence, this chapter can serve as an overview for an interested reader aiming to understand the technology ecosystem in place. We further highlight some of the most relevant trends that the charging industry currently follows, namely an increase in power and intelligence for all components and parties involved. Acknowledgements

The content presented in this chapter is based on work done in the project BeNutz LaSA (FKZ: 01MV20001A) funded by the German Federal Ministry for Economic Affairs and Climate Action according to a decision of the German Federal Parliament.