HIL simulation testing

As intelligence in vehicles rapidly grows, real-time simulation testing of advanced electronic control units is becoming an integral part of automotive software development.

February 4, 2014 - Montreal, Quebec

The explosion in the number of control units that have become standard equipment in automotive systems over the past 20 years has driven hardware-in-the-loop (HIL) testing to become a standard process in controller validation. In this process, electronic control unit (ECU) software is tested by running the actual controller with a real-time simulation of the system that it is designed to control, with actual signals exchanged between the ECU and the simulator (see Figure 1). The HIL testing process delivers substantial cost and time savings as the ECU can be tested even before the actual plant is built or available. In addition to this, the risk of equipment damage is minimized and fault or off-design operating conditions can be safely created. HIL testing has been successfully used for testing mechanical systems, which require system models to run at sample rates of around 1KHz, thereby greatly reducing ECU development and testing time. However, with the advent of several new technologies, ranging from electric drives to self-driving vehicles, HIL technology is now entering its second generation, where complex control algorithms spanning multiple timescales will need to be tested not only in isolation but also as part of an integrated system, with several controllers interacting with each other.

In the fiercely competitive world of automotive development, several new challenges have emerged as OEMs race to develop products to satisfy multiple requirements in terms of maximizing safety, fuel-efficiency and ride-quality while minimizing costs. In particular, the growing numbers of EV and HEV programs, in conjunction with advanced technology programs for reducing fuel consumption in IC engines, rely on increasingly complex control systems. The first level of analysis for these control algorithms involves basic functional testing in which each ECU is tested with a model of the primary system that it will control. For example, the motor-drive control of an EV or HEV will be tested using an HIL simulation of a permanent-magnet synchronous machine (PMSM) motor model with switching transitions that occur in microseconds. At the second level, the ECU must be tested in the context of a bigger vehicle system to understand its interactions with other subsystems. In this example, the electric drive's performance will now be simulated with real-time dynamic models of the transmission, vehicle dynamics and IC engine (for hybrid powertrains). The models of the other subsystems need not be as complex as the primary controlled subsystem model, which in this case is the electric drive. Furthermore, they have dynamics with time-constants in the order of milliseconds.

At this level of testing, communication integritywith other ECUs is also tested using a restbus simulation. In the final stage of testing, multiple ECUs that form an integrated system need to be tested together to ensure that control conflicts do not occur between them and that all systems revert to fail-safe modes under fault conditions. As an example, a hybrid-vehicle ECU integration test would involve testing of the high-level hybrid ECU that manages the activation of electric motor and/or IC engine, the engine management system, the motor-drive controller, the automatic braking system controller and the transmission controller.

Creating an HIL platform that addresses the various testing profiles described above requires that it be both flexible and open. In particular, high-fidelity models used for specific ECU functional testing are usually created in specialized software packages. In addition, custom hardware, unique to the particular subsystem, may need to be integrated, requiring specialized interface hardware and software drivers. Finally, the test environment needs to be flexible to cover the various modeling, test automation, software and hardware needs of a complete test program.

Bearing these requirements in mind, HIL platforms for ECU development should:

address the range of leading-edge automotive technologies, including EV/HEV solutions; provide high-power, cost-effective computing to enable real-time simulation of multiple systems with different timescales; integrate all required hardware for both single-ECU and multi-ECU testing; and be open and flexible, providing an environment in which the most effective modeling and testing software can be deployed for HIL testing.

For the EV/HEV market, OPAL-RT's HIL platform includes a suite of field-programmable gate array (FPGA)-based, high-fidelity, real-time motor simulation solutions that integrate leading FEA models from companies such as Jmag, Infolytica and Ansys, which include non-linear effects such as field saturation and harmonic effects. The entire real-time environment also enables the use of industry-standard modeling environments, such as Matlab/Simulink, AMESim and Dymola.

The HIL system provides a combination of CPU and FPGA-based simulation, as well as FPGA controlled input/output signal management, enabling execution of full-vehicle models in time-steps less than 100µs, with high-fidelity motor models running at time-steps of around 1µs.

To enable hardware integration, customized signal routing and conditioning boxes must be provided as part of the solution, and there should be the ability to integrate a range of third-party hardware, such as fault-insertion units and load-boxes. Applications such as battery-management systems require specialized signal conditioning to source and synch currents with high-precision voltage outputs, while motor-drive emulation may require the measurement of high current levels in the order of tens to hundreds of amperes.

For the purposes of test development, management and automation, the software environment must enable the development of GUIs for interaction with the real-time model; development of test scripts that can be executed in real-time and synchronized with model execution; collation of models, scripts and test results in a database; and statistical analysis of test results. To meet these requirements, OPAL-RT has integrated MBtech's PROVEtech:TA test-automation software. PROVEtech:TA offers a user-friendly interface as well as database support for the execution of test-management tasks, such as the administration of test libraries, test suites and test results.

With a wide variety of control elements, PROVEtech:TA facilitates access to the signals needed for test scenarios. For example, the user can set the gas pedal position then measure the resulting vehicle speed. Comprehensive diagnosis and fault simulation modules complete this software.

PROVEtech:TA was developed in close cooperation with many end users, taking into account their requirements in accessing signal information inside a simulation environment. It is much more convenient to run test software independently of the simulation software, so PROVEtech:TA was created to be executed on a host PC instead of the simulator itself. Only a small part of PROVEtech:TA is integrated on the simulator, taking care of the communication between the simulation model and PROVEtech:TA. Test engineers write test scripts that obtain signal information and validate it against expected references. Alternatively, they can use the real-time-automation-engine, where portions of the test script get transferred to the simulator to be executed in real-time with the speed of the test model, thereby having access to the exact timing of signals and making it possible to verify multiple signals in parallel.

With its team-work support features, PROVEtech:TA makes it possible to host various projects in one database, providing teams distributed across the world with access management for multiple users and library creation for software elements that can be commonly used over different projects. By utilizing the features already mentioned, the ntegration of PROVEtech:TA with the RT-LAB distributed real-time platform gives a flexible solution for a diverse range of test activities, easing the workload of engineers while ncreasing test coverage and reducing execution time through the high level of automation.

While the pace of development of computing technology continually increases and computing resources become increasingly powerful, it is important to remember that certain core requirements will always drive the foundation of a strong HIL solution. First, as system complexity grows and high-fidelity models demand more computational power, the HIL platform should be able to rapidly exploit the latest processing technology. Second, as ECUs are sensitive to jitter in I/O signals, I/O signal generation should be done on FPGAs independent of, but synchronized with, the multi-core CPUs handling the models. Third, the platform should be open in terms of hardware and software integration so that it can cope with new devices, protocols and tools. Fourth, the solution needs to be modular and expandable but delivered as a comprehensive solution to ensure that testing can begin with minimal integration effort on the part of the tester. Finally, efficient and cost-effective customization and maintenance services need o be readily available as part of the solution.

OPAL-RT and MBtech are working together to ensure that the HIL solutions of today and the uture meet the needs of the industry as it gears up for one of the most intensely competitive and innovative periods in its history.

N.B.: This article was published in Engine Technology International magazine // January 2014

About OPAL-RT TECHNOLOGIES:

OPAL-RT is a world leading developer of open, Real-Time Digital Simulators and Hardware-in-the-Loop testing equipment for electrical, electro-mechanical and power electronics systems. Our simulators are used by engineers and researchers at leading manufacturers, utilities, universities and research centers around the world.

Our unique technological approach integrates parallel, distributed computing with commercial-off-the-shelf technologies. Customers perform Rapid Control Prototyping, System Integration, and Hardware-in-the-Loop testing of electric drives, electronic controllers and power distribution networks in a variety of industries including automotive, aerospace, electric ships, power generation, rail, and industrial manufacturing.