61 HIL testing | Deep insight coming from different areas, the vehicle physics, and then we take control of their software, which goes into the DDB and then into the motion-control PC. Then we have a growing network of HIL applications, from an IR sensor model to a full powertrain testbed, and even connecting four or five testbeds to have multiple vehicles for sensor interactions. The DDB bus is based on user datagram protocol (UDP) Ethernet packets that offer more deterministic links than other Ethernet protocols. “UDP is accessible by most software, so we’ve made it open and flexible to allow developers to write their own drivers to blend into the platforms,” says Safdar. “If you look at the structure of the control system we have deliberately created partitions so the audio has a separate stream, graphics, motion for the motors, HIL – separate for data, separate for physics – and we evaluate how much processing power you need, depending on the refresh rates and size of the packets. We can connect up to 18 PCs, but the maximum so far has been 11, so there is room to expand.” “The other thing is the software. There is an office-based version and the real-time version. Depending on which physics engine you are using, it is Windows-based, so if the software demands Linux for real time we can bring in a third party such as dSpace, NI or Speedboat into our system, and we provide an interface. This is where we are spending a lot of time designing the system as it’s very diverse.” For example, OEMs can run real-time models of different tyres at the DDB level to test vehicle performance under varying conditions. “So there are very different ways to bring those models and data back into the system to keep flexibility, with multiple physics engines and multiple tyre engines to test out the options.” The same applies to testing battery technologies in different virtual vehicles and driving them in a simulator to see how their performance changes. “We are seeing a lot of people doing this in tools such as Matlab where developers write their models. The thing with our system is you can change the chemical properties of the batteries on the go, so that can give you different weather configurations, changes to the motors, and change the weight, the aerodynamics. Currently, it’s the weight versus the range of the battery that’s the concern – that is the biggest topic we see right now,” says Safdar. “We can integrate the battery models into a car of their liking, and that changes the properties of the vehicle, changing the pack density, handling, road grip, and what that means for the range. We are seeing a lot of complaints from end-customers about the ride and comfort of EVs with the low centre of gravity and heavy battery packs, and this helps focus on the balance of comfort and range,” he adds. Custom rigs ZF is using HIL and model engineering in its r&d department for servo drives for noise, vibration and harshness (NVH) testing of vehicles. The company has developed a custom system that combines two HIL systems. “We have a DSpace controller with two consoles – a traditional dynamometer arrangement with a drive motor and with a torque cell. With something like that we can get the dyno to replicate a load, and at the same time the model that is in the HIL can get the unit under test away. “What we have done is added a second console that we can use for NVH, and in particular structural noise. In EVs, everything is so much quieter and we want to look at this,” he says. “We designed this ourselves, and it’s similar to the dyno with a drive motor and a large metal rig on rubber E-Mobility Engineering | March/April 2025 A HIL rig for testing noise, vibration and harshness (Image courtesy of ZF)
RkJQdWJsaXNoZXIy MjI2Mzk4