A comparative study of extremum seeking methods applied to online energy management strategy of fuel cell hybrid electric vehicles

As an online adaptive optimization algorithm, the extremum seeking method (ESM) can be effectively employed to find an optimal operating point of a static nonlinear system in real-time. This paper presents a comparative study of different ESM schemes for online energy management strategy of fuel cell hybrid electric vehicles (FCHEVs). By applying the extremum seeking controller, the fuel cell system operating points can be maintained in a high efficiency region and thus saving the hydrogen consumption. In addition, battery state of charge (SOC) is considered as the input of penalty function for extremum seeking controller, in order to prevent over-discharging/over-charging of the lithium-ion battery during the FCHEVs operation. Different schemes of ESM presented in this comparative study consist of first-order ESM, high-pass filter based ESM, and band-pass filter based ESC. The main evaluation criteria in this comparative study include the utilization of lithium-ion battery, the fluctuation of fuel cell system output power, the fuel cell system efficiency and the hydrogen consumption. A Hardware-In-the-Loop (HIL) platform is used to experimentally validate the presented comparative study. Experimental comparison results show that, the performance of all the presented ES controllers is close to that of offline benchmark dynamic programming. The band-pass filter based ES controller is preferred to improve both the performance and durability of energy storage system (ESS) in FCHEVs, since this controller is found to have a good ability to limit the fuel cell power dynamics.