MPC based Fan Control for Automotive Applications

Abstract

A major drawback of pure electrical vehicles is their low range compared to conventional cars. The present work focuses on increasing cruising range of electric vehicles by reducing the power consumption for e-motor and power electronics cooling.Standard implementations for cooling circuit control which are taken over from conventional cars with internal combustion engine usually involve a temperature based bang-bang control. This implementations are simple to implement but lead to oscillating temperatures. To meet hard temperature constraints usually safety offsets have to be introduced leading to increased cooling ventilation which results in significantly higher power consumption mainly caused by the cooling package fan.The proposed approach aims to enhance control performance in order to significantly reduce temperature safety margins in reference values resulting in a more efficient cooling and a lower power consumption. The approach involves a separation of the plant model into a linear and a nonlinear part, which is then advantageously used for control. The proposed control approach consists of a standard linear observer plus a linear model predictive control. Nonlinearities caused by the radiator are addressed separately.A possible real-time execution on standard automotive hardware is also addressed. Results are shown on simulation examples, as well as on first testbed experiments.