Simulation wrings out EV motors in the test lab

March 20, 2013
Electromagnetic simulation has advanced to where it can reduce the amount of track time devoted to finding problems in electric-vehicle drivetrains.

Authored by:
Nick Keel
Dr. Ben Black
Advanced Control & Simulation
National Instruments Corp.
Austin, Tex.
Edited by Leland Teschler
[email protected]
Key points:
• Advanced test systems have gotten to the point where they can run finite-elementanalysis models while monitoring ongoing tests of real-world equipment.
• Special test gear able to handle high voltages now help wring out electric-vehicle motors.
National Instrument Corp.

Software takes aim at energ efficiency

For the first time ever, an electric car earned Motor Trend magazine’s Car of the Year award. Electrics and hybrids are in high demand with federal fueleconomy and carbon-emission standards dictating new vehicles must hit an average 54.5 mpg by 2025. So automakers are looking for ways to quickly develop efficient electric-motor systems at a lower cost.

However, the growing use of electric motors in vehicles has created challenges for embedded-control-system developers and test engineers. Control algorithms for electric-drive engine-control units (ECUs) must run much faster than those controlling internal-combustion engines. The traditional approach to hardware-in-theloop (HIL) testing is too slow for wringing out electric-motor ECUs.

The simulation of electric motors in real time lets engineers examine many types of transient and fault conditions that would be difficult or impractical to evaluate with real systems. Many of these conditions, such as faults on motor terminals or faults between dc and ac buses, have historically been impossible to implement during HIL testing. By combining high-fidelity, finite-element analysis with high-performance fieldprogrammable gate-array (FPGA) hardware, simulation models can include complex nonlinear behavior to represent motors accurately, and the signals from the model running in FPGA can then connect to other hardware at the high I/O rates needed for complete testing.

Electric-motor emulators can be used at early stages in developing and testing hybrid drives and other drives with electric motors. Any electricmotor type can be mapped using an integrated emulator which will behave like the real motor in dynamic turning-speed operation. Emulators also let developers model physical phenomena, such as harmonic vibrations and acoustic effects, so control schemes can account for them. The system responds like a real electric motor, but without any moving parts. This means engineers can run a wider range of tests and collect more accurate data, which reduces time spent field testing and produces better embedded software.

To perform anything more complex than a functional test of an ECU, the HIL system must represent complex nonlinear behavior and execute the simulation models approximately every microsecond to adequately represent the motor response to high-speed motor-control signals.

Power electronics introduce another challenge for developing and testing electric-motor-control systems: The test system must be able to handle voltages that range from 20 to 600 V and current that could be in excess of 500 A. These types of tests often run on a dynamometer, but dynamometers are limited in the amount of test coverage they can provide. They often fail to accurately represent actual vehicle dynamics. This forces developers into more field testing to fully validate an electric-vehicle powertrain.

Thus, it can be advantageous to have a test system capable of handling high-power signals while accurately simulating vehicle dynamics. This makes testing more efficient and cuts the cost of fielding electric and hybridelectric vehicles.

FEA and motor simulation
One of the most challenging problems facing engineers who simulate advanced-motor drives is in getting adequate model fidelity and simulation step times that are fast enough. A simple constant-parameter model may be enough for some HIL tests, but today’s motor drives often need something closer to real life. Ditto for optimizing the performance of control systems handling the electric motors that characterize the automotive and energy industry.

This is where high-fidelity finite-element-analysis (FEA) models come in. Engineers can use FEA to simulate complex nonideal behavior such as cogging torque, and design a controller to reduce torque ripple. Similarly, they can simulate the variation in motor inductance at high currents, which greatly affects the motor-torque production. This lets designers tune and test the motor control ECU more effectively.

Lower-fidelity models do not adequately represent cogging torque, motor inductances at high currents, and other nonlinearities. The inability to allow for these effects makes HIL testing less effective. The result is usually more field testing and development time for testing embedded control software.

As a simulation method, FEA provides highly accurate electric-motor models with enough fidelity to account for nonlinearities. However, the extreme mathematical complexity of FEA has historically limited it to use in softwareonly implementations. It can often take hours to simulate a few seconds worth of real-world operation. Simulation models must run in real time for HIL testing of electricmotor systems. High-fidelity models must be simplified so they run within the limits of processor-based systems. This makes HIL tests less effective. It takes a processor-independent, hardware-based simulation to get the closedloop update rates needed for electric-motor HIL testing.

FPGAs provide the kind of processing speed and low latency that electric-motor simulation demands. Their latency from input to output, for example, is typically measured in nanoseconds. However, FPGAs have limited available resources, and electric-motor models must often be simplified to operate within these limitations. This simplification typically means reducing the model to a linear time-varying representation small enough fit in the FPGA.

Nevertheless, there’s a way to bridge the gap between FPGAs and FEA simulation demands. Tools such as the JMAG add-on for NI VeriStand, a real-time testing program, let FEA-generated parameters update the FPGAbased model as the simulation runs. Here, FEA-generated lookup tables hold the important motor parameters. These update the model at each time step based on the current state of the simulated system. This setup provides a good balance between speed and model complexity.

The motor simulation itself is computed as a straightforward set of linear equations that can be programmed into the FPGA to execute quickly and deterministically. The lookup tables for the motor parameters represent highly complex and nonlinear behavior captured by the FEA tool. They reside in memory on the FPGA board. The combination of these two approaches yields a simulation that can run quickly but that can also represent complex phenomena.

Electric-motor emulation
Signal-level testing is quite helpful for developing control algorithms and evaluating ECU performance. But developers must also validate the high-power, high-speed switches in the system. Dynamometers are common tools for power-level testing, but they cannot accurately represent high-frequency dynamic loads that are a part of validating system performance. So engineers must perform this kind of testing on a test track. The problem is that such field tests can be hit or miss — you are as likely to uncover an unforeseen problem as to miss it completely, depending on track conditions. So it can take a long time running on a variety of tracks to build up confidence in an electric-powertrain design. Thus, an HIL test system capable of integrating power electronics and simulating vehicle dynamics can potentially save a lot of track time if it provides test coverage not found in dynamometer testing.

In the case of an electric-drive dynamometer, the original electric motor in the drivetrain couples to a load motor used to simulate road conditions. This electric-motordynamometer configuration combines a real-life electric drive with a real-life controller that includes high-voltage power electronics. The problem with this procedure is that this drivetrain does not accurately simulate the movements of the vehicle that will eventually hold the electric motor. Thus, it is almost impossible to model mechanical feedback from the vehicle to any level of accuracy.

This is a serious challenge because electric-motor speed does not always correspond to vehicle speed, especially in hybrid drives. In many real-world scenarios there are intermediate states with gears disengaged and no drive to the wheels, where the electric motor turns without any load. None of the resulting turning-speed dynamics can be reproduced in a rotating test system. The inert masses involved are significantly different than those in the vehicle. So the ECU must be wrung out in a test vehicle on the track.

An inverter test system, or electric-motor emulator, bridges the gap between the HIL test system and vehicle field testing. It can interact with high-power signals associated with inverters and lets engineers reproduce load and ambient conditions in the lab exactly as they arise in a drive inverter during real-life driving. Consistent front loading of everything from road tests to component tests reduces costs and minimizes development cycles which ultimately cuts time to market and yields embedded software of a higher quality.

© 2013 Penton Media, Inc.

Sponsored Recommendations

From concept to consumption: Optimizing success in food and beverage

April 9, 2024
Identifying opportunities and solutions for plant floor optimization has never been easier. Download our visual guide to quickly and efficiently pinpoint areas for operational...

A closer look at modern design considerations for food and beverage

April 9, 2024
With new and changing safety and hygiene regulations at top of mind, its easy to understand how other crucial aspects of machine design can get pushed aside. Our whitepaper explores...

Cybersecurity and the Medical Manufacturing Industry

April 9, 2024
Learn about medical manufacturing cybersecurity risks, costs, and threats as well as effective cybersecurity strategies and essential solutions.

Condition Monitoring for Energy and Utilities Assets

April 9, 2024
Condition monitoring is an essential element of asset management in the energy and utilities industry. The American oil and gas, water and wastewater, and electrical grid sectors...

Voice your opinion!

To join the conversation, and become an exclusive member of Machine Design, create an account today!