Markov chain-based approach of the driving cycle development for electric vehicle application

Abstract

Driving cycle development is crucial in electric vehicles (EVs) power management design and evaluation. In this study, a Markov chain-based method is proposed to develop the typical driving cycle for electric vehicle application in Shenyang, China. The presented study is composed by three steps: (1) the driving data is divided by states according to two dimensional factors, namely the speed and the acceleration; (2) the state transition matrix and Markov transfer probability are calculated; (3) the driving cycle is established based on the Markov state using Monte Carlo simulation. Based on the above effort, ten candidate cycles satisfying the time length between 500 s and 3000 s are constructed, among which the typical driving cycle is determined by the performance value (PV) and sum square difference (SSD). A blended power management of hybrid electric vehicle has been applied on the established driving cycle, the test results indicate that the presented method can produce a very representative driving cycle.