Multi-parametric energy management system with reduced computational complexity for plug-in hybrid electric vehicles

Due to the limited computational capabilities of commercial control hardware, the implementation of model-based optimal control approaches remains a challenging problem. Among the model-based approaches, model predictive control (MPC) is infamous for its cumbersome computational cost especially for designing a hybrid vehicle powertrain energy management system (EMS). To resolve this issue, two multi-parametric model predictive EMSs for a plug-in hybrid electric vehicle (PHEV) are introduced, by considering the limited memory size of a control hardware. One of the EMSs is designed based on an improved control-oriented model that is derived by using the control-relevant parameter estimation (CRPE) approach. The results of simulation using Autonomie software shows significant fuel saving by using these EMSs compared to a baseline controller, while maintaining real-time capabilities.