eDrive - Summer 2013 - (Page 10)

Feature Article Model-Based Development of Motor Drive Controller for Hybrid Electric Vehicles Kunal Patil, Ph.D., Applications Engineer, HIL Modeling Mahendra Muli, Director of Marketing & New Business Development dSPACE, Inc. NA A hybrid drivetrain combines well-known, state-of-the-art powertrain technology with powerful electric motors and energy storage device in various configurations. Hybrid electric vehicles (HEV) typically have complex interaction between different powertrain devices and incorporate a complex electronic control unit (ECU) network. The control functions are spread over a distributed network of electronic control units. Use of powerful motors and their control requirements further adds demands to complex vehicle drive technology. ECUs for controlling electric motors are often incorporated into complex and distributed vehicle functions, so it is essential to test their interaction with other ECUs. Sophisticated, high fidelity, real-time simulation technology enabling closed-loop testing is required for electric motor ECU testing. Model-based design (MBD) represented by the V-cycle is a widely recognized approach in the development of ECU software, spanning from offline controller development to the final implementation on production hardware. This article describes the necessary transformation steps for an electric motor controller model, from the early controller design phase to the final implementation on production hardware. The V-cycle for motor controller development involving phases of control design, rapid control prototyping and target implementation is explained. Controller Development Process Plant and Controller Model Development Use of systematic model-based control design has been widely accepted in the automotive industry. MBD helps in effectively managing complexity by validating performance requirements throughout development, impacting quality and at the same time shortening development time. In a MBD process, the control algorithm is designed together with the physical system (aka plant) for validation. Therefore, the first step in the controller development process is to create accurate plant models of appropriate fidelity for real-time computation. Integration of the controller algorithm models, together with the validated plant models, provides a complete simulation environment for model-based, closed-loop development and testing of the motor controller. In this article, dSPACE components of Automotive Simulation Models (ASM) libraries are used to model motor controller and plant models. ASM implemented in a MATLAB/ Simulink modeling environment can be used in real-time hardware systems such as dSPACE HIL simulators. ASM provide simulations of key automotive systems, such as engines, vehicle dynamics, brake hydraulics, turbocharger, electrical system, 10 Summer 2013 exhaust system, traffic and trai1er. ASM can be used both on a dSPACE simulator for hardware-in-the-loop testing of ECUs, and also during the design phase of controller algorithms for early validation using PC-based offline simulation. Use of such well-defined, documented, open and structured models allows the user to quickly change software architectures, controller implementation and parameters. Due to large implications of the overall vehicle operation on electric motors, this article references the design of a parallel hybrid electric vehicle (PHEV) model for developing a motor controller for an automotive application. In parallel hybrid vehicles, torque is generated by both the IC engine and the electric motor. The torque generated by the IC engine is directly supplied to the wheels, with assistance from an electric motor that is mechanically coupled to the driveline. The same structure needs to be reflected in the simulation model. Figure 1 shows the top level of a parallel hybrid vehicle ASM model. The model incorporates seven components: soft-ECU, gasoline engine, battery, motor, drivetrain, vehicle dynamics and the environment. The engine, motor and drivetrain Figure 1. ASM Parallel HEV are connected via shaft speeds and shaft torques. The drivetrain is coupled to the vehicle dynamics model with the tire torque and tire speed interface. Plant Model Computation The PHEV model is designed to be used in a real-time hardware system in such a way that the reading of actuator signals, the model calculations themselves and the output of sensor signals coming from ECU is performed in one simulation step. The step size of 1ms is used for the entire model. The focus of this study is to develop electric drive controller development, but the overall model requires other controllers (e.g. engine condSpace Article Continued on Page 12 www.e-DriveOnline.com http://www.e-DriveOnline.com

Table of Contents for the Digital Edition of eDrive - Summer 2013

eDrive - Summer 2013
Installing an Energy-Efficient Motor Helps Reduce Losses While Increasing Profits
Volatility and Oversupply: Why the Motion Control Market Underperformed in 2012
Model-Based Development of Motor Drive Controller for Hybrid Electric Vehicles
Motors & Drives
Motion Control
Components
Industry News
New Linear Motor Designs Improve Speed and Positioning
Marketplace

eDrive - Summer 2013