Model Predictive Control of the Air-conditioning System for Electric Bus

Abstract

The varying amount of passengers that were carried on the electric bus is an obsessional disturbance for the cabin temperature control and energy management. This paper targets at evaluating the role of passenger number forecasting in the energy management of an electric bus based on model predictive control (MPC). Markov-chain model is used to forecast the future number of passengers, and thus provides references for the controller to achieve better energy economy. Dynamic Programming (DP) is used to optimize the power behaviors of the AC system within a MPC structure. Simulation results show that MPC used in the AC system can achieve a feasible and reliable temperature performance compared with the DP-solved benchmark. The energy consumption is improved by 8% compared with that of DP with a fixed number of passengers.