Commuting-pattern-oriented optimal sizing of electric vehicle powertrain based on stochastic optimization

Owing to the dynamic randomness of traffic flow, driving patterns stochastically change from time to time and person to person, which inevitably triggers significant variability in the energy efficiency of electric vehicle (EV) powertrains. For optimizing the expected energy efficiency of the EV powertrain in probable driving operations while reducing the variability in energy efficiency, this paper engineers an integrated stochastic optimization (SO) method for the design scheme and control strategy of EV powertrains. To evaluate the candidate design scheme and identify the superior design scheme in random driving operations, an instantaneous optimal control and a Monte Carlo simulation-aided iterative searching process are developed and utilized as critical components of the SO. According to simulation validation, by optimizing the energy consumption in extreme operations, the SO improves the expectation of the energy efficiency of the EV powertrain by 26.4% and reduces the variability in energy efficiency by 90.4%. Moreover, the proposed SO has no side-effect on the energy consumption in typical/frequent driving operations. Even operated in the driving cycle from which the deterministic optimization (DO) result is obtained, the increase in energy consumption of the SO result is less than 5% compared with the energy consumption of the DO result.