Model predictive controller based optimal energy management strategy for a fuel cell hybrid vehicle

In today’s world, the environmental pollution and climate issues caused by fossil-fuelvehicles have made it a necessity to develop alternative technologies capable of beingmore environmentally friendly. Electric and hybrid vehicles have emerged as alternativetechnologies to fossil-fuel vehicles within this context, as they offer energysources with less environmental impact. Despite the importance of electric vehiclesdue to zero tailpipe emissions, challenges such as limited range, long charging times,lower prevalence of charging stations compared to gas stations, and shorter batterylife have led to increased research on hybrid vehicles based on the principle ofutilizing multiple energy sources simultaneously. Internal combustion engine-basedhybrid vehicles are not environmentally friendly due to their use of fossil fuels. Inthis regard, fuel cell hybrid electric vehicles draw attention due to several advantagessuch as zero carbon emissions, longer lifespan of the fuel cell compared to thebattery, and shorter charging times. However, the distinct dynamics of batteries andfuel cells reveal the necessity of an energy management strategy that can efficientlyand harmoniously utilize these sources in fuel cell hybrid electric vehicles.In this thesis, the energy management problem in a fuel cell hybrid electric vehicle isaddressed. Firstly, in the study, an electrical powertrain model of a fuel cell hybridelectric vehicle is developed in the Cruise-M simulation environment, providing arealistic platform for testing energy management strategies using its comprehensivecomponents. The specifications of the first-generation Toyota Mirai are consideredfor the designed plant model. Subsequently, a rule-based energy management strategy (RB-EMS) is developed not only to be a benchmark method, but also to assessthe representability of the presented plant model. Furthermore, a model predictivecontrol-based energy management strategy (MPC-based EMS) is designed and employedto solve the energy management problem in an optimization-based manner.The developed strategy can adjust the weight factors of the cost function in realtimeand improves the the controller performance by enhancing the effectiveness ofthe optimization. It is observed that the designed plant model can produce realisticresults considering the experimental outputs of the power sources presented in thedynamometer test results conducted on a Toyota Mirai. According to the simulationresults, it is revealed that the power allocation performance of the MPC-basedEMS is more satisfactory compared to the RB-EMS. The strategy has improved fueleconomy by reducing fuel consumption overall by 3% compared to the RB-EMS andhas been able to keep the battery’s state of charge within a certain range. Ultimately,it is understood that the presented powertrain model has the potential tobe a realistic virtual platform for studying different energy management strategiesextensively. Additionally, the real-time adjustability of controller parameters basedon the system’s instantaneous conditions via a feedback mechanism in the modelpredictive control approach demonstrates its potential for furthe