A systematic approach of vehicle plant model development for vehicle virtual testing & calibration

With continuous pressure for reducing vehicle development time & cost, without compromising reliability and quality from system, it is imperative to move from Road to Lab development. Every OEM is currently focusing on using virtual environment to complete the virtual testing and functional check for the system before hitting the road. This method of developing software and calibration in virtual environment basically brings repeatability and reproducibility to the test case which is tough to achieve in highly dynamic real world. Also this process is cost efficient and reduces the time for function development and brings high level of functional maturity before testing it in the vehicle. Current paper deals with defining a systematic and novel approach for vehicle plant model development which can be used for virtual testing & calibration. Vehicle plant model will be developed in multiple software platforms where in most of the vehicle modeling will be done in AVL CRUISE and will be brought into MATLAB as S-function. The control logic will be tested in MATLAB and will be linked to AVL CRUISE mode in MATLAB. This paper focus on methods to verify the accuracy of plant model with real vehicle data. Focus will also be on application of plant model for various development activities. With required level of vehicle plant model accuracy, this paper will discuss on the closed loop simulation results of vehicle (including the functional logic developed) for fuel economy and drivability estimation for real world and legislative cycles.