Real-time optimization strategy by using sequence quadratic programming with multivariate nonlinear regression for a fuel cell electric vehicle

Abstract

The energy management of a plug-in FCEV (Fuel Cell Electric Vehicle) strictly depends on the control of SOC (State of Charge) over a given trip distance. The SOC may be varied with the trip distance by updating an EF (Equivalent Factor), which is derived from ECMS (Equivalent Consumption Minimization Strategy). However, the EF is too complicated to estimate accurately in real-time with traditional method. A real-time optimization strategy by using SQP (Sequence Quadratic Programming) with MNLR (Multivariate Nonlinear Regression) is proposed for a plug-in FCEV. First, the real-time hydrogen consumption optimization problem for SOC trip distance adaptive is formulated by using ECMS. The EF is adjusted according to the trip distances and predefined SOC. Then, in order to improve the accuracy of EF, SQP method is utilized to optimize the fuel cell and battery efficiency. Thus, the MNLR is applied to construct the fuel cell and battery efficiency response surface models for real-time optimization application. Finally, numerical verification and hardware in loop experiments are conducted to validate the proposed strategy. The results indicate that the combination of SQP with MNLR made it possible to develop the proposed strategy capable of significantly improving the hydrogen economic performance of this FCEV.