Is a design or architecture decision worth the effort? To answer this question, we will show how to use simulation data to train systems models. Our case study will be a battery electric vehicle, BEV, with a full drive cycle simulation. A typical motor drive simulation will generate data that will be used to train a reduced fidelity system model for use in the BEV drive cycle simulation. This system model will be trained over a broad range of design scenarios including parallel devices, adaptive gate drive, battery voltages, operating temperatures, etc. The objective is to enable designers to easily evaluate the impact of a design decision on system efficiency and the resulting impact on cost or range to determine: is the engineering effort worth it? The system simulation runs faster than realtime and can be used to quickly and easily compare the inverter efficiency impact and cost. Inverter efficiency improvements can have major impacts on the overall system cost with reduced battery pack needs and resulting structural changes to the vehicle.
