As vehicles get more complicated, vendors face the challenge of simulating the increasingly complex systems inside of them.
With more hybrid and electric vehicles on the market, for example, the powertrain options have expanded far beyond gasoline and diesel. While internal combustion engines are still a popular choice among drivers, alternative powertrains may include propulsion systems that rely on components like compressed air or full battery electrics.
So, how can simulation providers ensure that they're modeling features like the powertrain with high accuracy? In a live presentation this month titled Trends in Passenger-Vehicle Component Design Simulation, a reader had the following question for speaker Sam Akehurst, Deputy Academic Director for the Institute for Advanced Automotive Propulsion Systems at University of Bath:
"What are the main challenges in simulating powertrain systems? What do we struggle to model now?"
Read Akehurst's edited response below.
Sam: That's quite a broad question. Fundamentally, if we have enough time, we can model most things that we understand in physics and chemistry, but it's the trade-off of "Have we got enough time to do that simulation, or is it actually faster to go and do the experiments?"
There are some specific phenomena that are hard to model, particularly NVH (noise, vibration, and harshness), and some of the acoustics around turbocharging. As we move to electrification: NVH, acoustics around e-Drives, and [scenarios] where you've got lower background noise (but higher frequencies) are particularly challenging.
And then there are things that we really still don't quite understand at a fundamental level.
So, we work with people looking at fundamental battery chemistries and degradation mechanisms. There are some fairly stochastic kinds of random events that trigger, at an atomic scale, battery degradation. Because we don't fully understand [these events], they are very hard to model within powertrains.
Beyond that, actually representing drivers in a real-world [situation], and interacting with traffic, is quite difficult to do. If we can have the real human in those kind of simulations, it's probably easier to model. And also, measuring things that happen at lots of different time-bases, so highly dynamic things happening alongside gradual thermal events is also a challenge.
Have more questions? Share them below in our comments section.