Control Oriented Modeling and Nonlinear Model Predictive Temperature Control of a Vehicle Cabin

The topic of modeling the temperature dynamics of a vehicle cabin has been tackled frequently. The orientation towards control however has rarely been explicitly researched. In this work, a zonal model is presented that solves many issues of currently used dynamic thermal vehicle cabin models. Firstly, all equations are based on physical modeling and do not rely on experimental data. Secondly, air flow dynamics in the vehicle are modeled on a top level, allowing for a physically motivated calculation of zonal masses. The resulting high order model is then thirdly optimized in terms of calculation efficiency by using steady state relations for high frequency dynamics, yet keeping a close resemblance between model and system. The conformity for short term predictions is shown in simulation. Applicability of the model for control purposes is then demonstrated using a nonlinear model predictive control algorithm. The deviation of the reduced order model from the higher order model is shown to be small in simulation. Furthermore a controller based on the reduced order model is capable of driving the higher order system to steady state in the presence of disturbances.