当前位置: 首页 > >

混合动力电动汽车动力系统的通信系统的设计

发布时间:

Powertrain Communication System Design of Hybrid Electric Vehicle
Lifang Wang, Fang Li, Chenglin Liao
Institute of Electrical Engineering Chinese Academy of Sciences Beijing, China wlf,lifang,liaocl@mail.iee.ac.cn

Abstract— In the automotive system design flow, protocol design is an essential step which mainly includes message allocation, message priority assignment and message period assignment. This work provides an approach in the message period assignment through experiment and validation. The experiment and validation takes the powertrain system of hybrid electric vehicle as the platform. Under the baud rate of 500Kbps, and take the messages from EMS(Engine Management System) as the example, simulations are done under different message periods. The results show that when the message EEC1 sending by EMS becomes to 100ms, the engine speed can shake dramatically. It restricts the period of message EEC1 must be smaller than 100ms. Messages from other ECUs in the powertrain system of hybrid electric vehicle can be selected under the experiments similar to the process of the EMS. The approach in this work can be taken as a reference for designing CAN bus protocol of hybrid electric vehicle. Keywords- CAN bus, powertrain system, hybrid electric vehicle, message period

concentrate on the protocol itself such as system bus load, message transfer delay, etc. In this paper, Powertrain system model of hybrid electric vehicle and CAN bus communication are all built using the tool of Matlab/Simulink. The system can offer an efficient and visible environment for not only analyzing CAN bus real-time performance, but also hybrid electric vehicle performance, and finally support the system protocol design. The paper first describes automotive system design flow, in which period assignment is the emphasized in the protocol design. Then, the powertrain system architecture of hybrid electric vehicle is given, which embracing the powertrain module and CAN communication module. CAN bus communication module established by SimEvents of Matlab/Simulink is introduced in detail. Finally, experiments and simulations are done to obtain the period optimization result through the system design. II. AUTOMOTIVE SYSTEM DESIGN FLOW

I.

INTRODUCTION

Due to the crisis of energy resources and increasingly pollution of the environments, research of electric vehicle is emphasized more and more. The vehicle system of hybrid electric vehicle, pure electric vehicle and full cell vehicle are the key research point, while the protocol and communication system for powertrain system of electric vehicle is also one of the most critical technology, which guarantees the real time performance and control dependability of the powertrain system. Controller Area Network (CAN) [1][2][3] is a communication protocol that most widely used in the automotive area. In the automotive system design flow, the protocol design is one of the most important stage, while experiment and validation is a direct approach to support the protocol design results. In the stage of experiment and validation, simulation modeling is easy to be established to estimate the behavior of communication systems in worst-case situations. There are already several scholars established different CAN bus models [4][5]using tools such as Sate flow and OPNET, etc. However, in these works, they do not emphasize the control performance of the application, but
This work was supported in part by National High Technology Research and Development Program of China (863 Program), ‘Research of Network, Bus and Protocol of Electric Vehicle’ (2007AA11A134).

The automotive system design flow is shown in figure 1. After the system architecture is established, task allocation assigns tasks to ECUs according to the functionality of the system. Correspondingly, signals that exchanged in or between ECUs are all designed, and then it goes to the mission of the protocol design.

System Architecture

Task Allocation

Message Assignment Priority Assignment Period Assignment

Protocol Design

Experiment and Validation

Fig

1. The design flow of automotive system

978-1-4244-2487-0/09/$25.00 ?2009 IEEE

The protocol design is divided into three aspects, message assignment, priority assignment and the message period assignment. Message assignment is the first step and it assigns the signals from ECUs to different CAN bus messages. The second step assigns priority to messages, and the last step assigns the message periods to messages. In this work, the period assignment of the protocol design process is focused on. To validate the protocol design, experiment and validation is done. The results of the period assignment may trigger the design iterations in order to obtain the final feasible design. Experiment and validation can be done in simulation environments or the hardware-in-the-loop platform. This approach is done experiment and validation through simulation in Matlab/Simulink. In the next section, the hybrid electric vehicle is chosen as the application platform, and the detailed explanation are shown. III. THE PLATFORM OF HYBRID ELECTRIC VEHICLE

inertia module is also employed to emulate the driver’s real operation.

The architecture of the powertrain system of hybrid electric vehicle referred to is depicted in figure.2. The front axes of powertrain include the engine, ISG motor, clutch and transmission, while the rear axes of powertrain are drived by two wheel-side motors. The controllers of the powertrain system are communicated with CAN bus, and the HCU (Hybrid Control Unit) module plays the role of holistic management. Using CAN bus messages, HCU manages controllers in the powertrain system, and furthermore, according to the feedback information from controllers, HCU coordinates the energy allocation among the controllers in order to reach the optimum control of the powertrain system. For example, making use of CAN message, the HCU sends speed and torque command to the motor ECU, while motor ECU feedbacks information through CAN message to the vehicle controller, which may includes motor power/off requests, motor state and speed of the motor, etc. According to the design flow shown in figure 1, the experiment and validation is done to validate the protocol design, and this work selects the platform of hybrid electric vehicle designed in Matlab/Simulink, it is divided into two modules: the powertrain module and CAN communication module. A. The powertrain module 1 HCU Module The HCU model emulates the internal logic and algorithm of the HEV powertrain controller. It embraces the driver intention identification module, the energy management strategy module and the coordinated control module. According to the vehicle speed requirement of different operating conditions, the driver intention identification module calculates the control action of the driver, mainly about the opening degree of the accelerator pedal and brake pedal. In the inner of the module, a simple PID control is used, and an 2

Figure 2. The architecture of the hybrid electric vehicle

Engine Module

The engine module embraces the engine component and the engine management system (EMS) which manage the running mode of the engine and communicate with other ECU through CAN bus. Further, the engine component includes the indicator torque calculation module and the dynamic calculation module. The indicator torque module calculates the indicator torque of the engine based on the current engine speed, the throttle opening and the oil cut signals. It uses the steady data of the engine and also inserts dynamic correction, which enhancing the calculation efficiency. As for the dynamic calculation module, the formula used shows as follow:

T = I ×w+d ×w
Where:

w represents the engine speed.
T represents the engine torque.
d represents the damping. I represents the inertia moment.
3 Other modules

Motor/ISG module consists of the motor/ISG component and the Motor/ISG ECU. Four modes of the motor/ISG are realized: racing, driving, generating and speeding, In the battery module, the R-C circuit model of the battery and the battery management system are designed. B. CAN Communication Module [6] Since CAN bus is a communication protocol that uses event-based mechanism, to implement the mechanism efficiently, it is necessary to use the module - SimEvents, which extends Simulink with a discrete-event simulation model of computation. With SimEvents you can develop activity-based models of systems to evaluate system parameters such as congestion, resource contention, and processing delays. You can configure entities with user-defined attributes, and then aggregate entities and attributes to model data hierarchy and transport in applications such as packet-based networks, real-time operating systems, and computer architecture.

module, monitors the messages being transferred. For example, when a message sending error happens, the node will retransmits this message, while there is an error in the receiving process, the receiving node will discard the message. 3 CAN Bus Arbitration Module

CAN bus arbitration module emulates the arbitration mechanism of CAN bus. The module embraces CSMA/CD, Message arbitration, and Replicate module. In order to validate the correctness of CAN bus communication model established in Matlab/Simulink, necessary experiments have been done and the results was compared with CANoe experiment sufficiently. It concludes that CAN bus communication model established in Matlab/Simulink, which are used for simulation of hybrid electric vehicle powertrain system, can represent the message transfer sequences on the CAN bus correctly, and achieve an exact result of the performance analysis of the system. The model can be used as an efficient tool for ECU (Electric Control Unit) testing and CAN bus testing in hybrid electric vehicle. IV. EXPERIMENTS AND VALIDATION

Experiments and validation are done based on the platform of hybrid electric vehicle that has established above. The experiments help CAN bus protocol design of powertrain system of hybrid electric vehicle, which is done under the hypothesis that the baud rate of CAN bus is 500Kbps and there are no bus errors through the experiments. Although, in CAN bus protocol design, message priority and message period are two important aspects. Message period is emphasized in the experiments. In the automotive application area, considering the real-time requirement and the software programming, the message periods can be ranged from 10ms to 1000ms. For the hard real-time messages, normally the 10ms to 50ms is selected, while for the soft realtime messages, 50ms to 1000ms are selected. To decide the message periods in the CAN protocol, the experiments are implemented to study how different message periods can influence the control performance. Since engine is the most important component in the powertrain system of hybrid electric vehicle, and the message between the EMS and HCU can influence the vehicle performance dramatically, take the EMS system as the example, while the messages from other nodes are used as the bus load of the CAN bus. A. Message and Signal Design Part of the messages transferred on CAN bus is listed in table 1. In the table, the message name and its sending and receiving node are all listed. The message exchanged between HCU and EMS is VCU1 and EEC1. Since the message VCU1 embraces the control parameters of engine from HCU, it is hard real-time message, the message period should between 10ms and 50ms. While the message EEC1 takes the feedback information of the engine, the real time performance requirement is lower, thus the message period can be chosen from the 50ms to 100ms.

Figure 3. CAN Bus Communication System

Figure 3 presents a CAN bus communication system including two nodes. Each node contains message sending module, message receiving module, controller module and CAN bus arbitration module. 1 Message Sending and Receiving Modules

essage sending and receiving module are designed to realize the function of sending and receiving messages in the nodes. Message sending module sends the packaged message entities to the controller module. Different from the message sending module, the receiving module receives messages from the controller module, and unpacked the signals from messages. 2 Controller Module

Controller module makes use of the queuing theory to implement CAN bus data link. In the controller module, message sending and receiving queue are realized by “Queue module” in SimEvents library, they simulate the sending and receiving buffer of the CAN controller in the node. Error handling, which including sending and receiving monitor

TABLE I Message Name VCU1 Sending Node HCU

MESSAGE MODEL Signal Definition Target throttle control Engine oil cut Engine speed Message Period 10ms50ms 50ms100ms

Receiving Node EMS

figure 4 and table II. The results show that the control performance of the hybrid electric vehicle can be influenced by the message period. For example, when the period of message EEC1 becomes to 100ms, the engine speed can shake dramatically. So the message EEC1 must be smaller than 100ms. The recommended message period should be 50ms. Messages from other ECUs in the powertrain system of hybrid electric vehicle can be selected under the experiments similar to the process of the EMS. The approach in this work can be a reference for designing CAN bus protocol. V. CONCLUSIONS

EEC1 TC1 ETC1 ETC1 VCU6 ETSC1 ETC2 BC1

EMS HCU AMT AMT HCU AMT AMT BMS

HCU AMT HCU HCU BMS HCU HCU HCU

Throttle opening

B. Experiment Results Under different message period of VCU1 and EEC1, experiments are done and the curve of engine speed is recorded.

Protocol design is an important step in the automotive design flow. It mainly embraces message allocation, message priority assignment and message period assignment. This work put an emphasis on the message period assignment through simulation. The simulation takes the hybrid electric vehicle as the platform. Powertrain system model of hybrid electric vehicle and CAN bus communication model are all built in Matlab/Simulink. Under the baud rate of 500Kbps, experiments are done under different periods of message VCU1 and EEC1, which sends by HCU and EMS separately. The results show that when the message EEC1 becomes to 100ms, the engine speed can shake dramatically. It restricts the message EEC1 smaller than 100ms. The approach in this work can be taken as a reference for designing CAN bus protocol of hybrid electric vehicle. ACKNOWLEDGMENT The authors would like to thank the anonymous reviewers for their helpful comments REFERENCES
[1] International Standard Organization (ISO). “Road vehicles – Interchange of digital information – Controller area network (CAN) for high-speed communication”, ISO 11898, 1993. G. Cena, A. Valenzano. “An Overview of Controller Area Network”. Computing & Control Engineering Journal, vol.10, Issue 3, June 1999, pp.113-120. Wolfhard Lawrenz. CAN System Engineering from Theory to Practical Application. New York: Spring-Verlag, 1997, ch.1 C. Bayilm, I.Ertürk, C. Ceken. “Modeling Controller Area Networks Using Discrete Event Simulation Technique”, in Complex Computing Networks. vol.104. Springer Berlin Heidelberg: Germany, 2006, pp.353358. Carvalho, A., Portugal, P. “On Dependability Evaluation of Field Bus Networks: a Permanent Fault Analysis”, in Proc. 27th Annual Conference of the IEEE Industrial Electronics Society, Denver, Colombia, USA, 2001, pp.299–-304. Fang Li, Lifang Wang, Chenglin Liao. “CAN(Controller Area Network) Bus Communication System Based on Matlab/Simulink” , in Proc. 4th IEEE International Conference on Wireless Communications, Networking and Mobile Computing, Dalian, China, 2008, in press.

[2] Figure 4. The engine speed under different message periods [3] TABLE II Experiment EXPERIMENTS UNDER DIFFERENT MESSAGE PERIODS [4] Message period
VCU1 EEC1

Engine Speed [5] Small fluctuation in the engine speed Large fluctuation in the engine speed; Stabilize to a constant value. Large fluctuation in the engine speed; Can not stabilize to a constant value

1 2 3

10ms 50ms 10ms

50ms 50ms 100ms

[6]

As mentioned in table I, period of message VCU1 can ranged from 10ms to 50ms, and period of message EEC1 can be 50ms to 100ms. Three experiments results are shown in




友情链接: