Cloud computing-based real-time global optimization of battery aging and energy consumption for plug-in hybrid electric vehicles

Abstract

This paper addresses two conflicts in the energy management strategy (EMS) for plug-in hybrid electric vehicles (PHEVs). One is the conflict between fuel economy optimization and battery state of health preservation, and the other is the conflict between global optimality and real-time capability. Inspired by the hierarchy structure of a computer, a two-layer internet-distributed EMS (ID-EMS) is developed using cloud computing and the internet of vehicles. The top layer in the cloud, which possesses powerful calculating capability, focuses on global optimality by utilizing machine learning technology and stochastic dynamic programming. The bottom layer on board, with limited computing power, employs a fuzzy controller to respond to real-time conditions while trying not to deviate too far from the global solution. Thus, a real-time global optimal EMS can be achieved. The ID-EMS is implemented on an internet-distributed vehicle-in-the-loop simulation platform whose in-loop vehicle makes it possible to test the ID-EMS in an on-road driving experiment. The ID-EMS outperforms a rule-based EMS in terms of overall cost by 6.8%, but it is surpassed by an acausal EMS with dynamic programming by 7%. These results suggest directions for the future development of EMS for PHEVs using cloud computing.