60 this is downloaded from a PC to the HIL to run on the FPGA array. For example, this could be a model of a two-level inverter with six transistor switches and three-phase permanent magnet (PM) motor. Typhoon has built its own specialist processing system that is implemented as custom logic on the FPGA and runs the design, avoiding the need for designing or compiling a design for an FPGA, which can be challenging. “We built a processor architecture on the FPGA, and use different modules and translate to the architecture,” says Gartner. “This gives us flexibility and control over the whole model, and the users don’t need third-party tools as the FPGA is pre-compiled.” Developers can use the system to test the physical ECU. “The time-step of the simulation is important; for us this is sub-microsecond, typically 250 ns, for a drive system,” Gartner says. “The other thing is the latency from the control signal coming into the ECU on the outside. “This is in the microsecond range for us, through the analogue to the digital converter (ADC), apply to the model, and then the digital-to-analogue (DAC) conversion on the output. “For an inverter, we track the thermal effects, the losses in the motor, and we can track in real-time. For the motor, the more detailed the model you can run the better it is, so we model and simulate how flux saturates with the currents, and how the field is impacted by the geometry and the slots. “We can import that detailed data from the customer with their measurements. We cover a range of fidelity for the model, increasing the level of fidelity through the development process as the data improves.” Digital backbone “A few years ago, people came to us for a subjective assessment through a driving simulator, but the simulator is just one aspect with MIL and mechanical HIL,” says Safdar at Ansible Motion. “We get involved right at the beginning of any development, from the components to modifying existing base models of vehicles. We can quickly load the models on the system, and put a driver in the simulator to tell developers how a driver will feel with handling, safety and performance. “We are finding engineers are becoming much more bold in trying different concepts. Now they can push the boundaries with less risk.” To do this, Ansible has developed a databus technology that allows all the different HIL technologies, testbeds models and simulators to be combined. The distributed data bus (DDB) is driven by a network of Windows PCs with software developed by Ansible. A traditional simulator connected to a powertrain testbed can cover 80% of calibration and then take it on track for safety validation. This federated HIL system offers the ability to do 95% of calibration. The next step is a dynamic platform with six degrees of freedom (6DoF) for 100% validation, he adds. “Dampers are difficult to model in the virtual environment. We have done an integration with four active damper systems with individual attributes when driving on the road, monitoring the dampers, so we’ve taken it to that extreme.” “The third area that’s difficult to create a good model for is the steering. That’s a nightmare for integration, and we worked with MdynamicX for the physical hardware and integrated their technology into our simulators, so what we offer now is way beyond just the driving simulators. “In our distributed data bus we align the time stamps and synchronise the signals, and present them in real-time to the driver, so we have data-logging March/April 2025 | E-Mobility Engineering The smart cell emulator (Image courtesy of Typhoon HIL)
RkJQdWJsaXNoZXIy MjI2Mzk4