Optimization Of Component Sizes And Control Parameters Of Plug-in Hybrid Electric Bus

Plug-in Hybrid Electric Bus (PHEB) is one of the most potential publictransportations to alleviate the energy shortages and urban environment pollution. Thecomponent sizes and control parameters of PHEB impact the vehicle cost, energyeconomy and dynamic performance significantly, which leads to two categories ofoptimization requirement: component sizes should be optimized by PHEB manufacturersaccording to a given city’s geographic characteristics; control parameters should beoptimized by engineers for the settled bus route based on its regular traffic information.From these two categories of requirement aforementioned, the optimization ofcomponent sizes and control parameters of PHEB are researched in this paper.Firstly, the Chongqing Hengtong CKZ6116PHEV quick-charge plug-in gas/electrichybrid bus with single-shift parallel configuration is treated as the prototype. Thecharacteristics of single-shift parallel hybrid system are analyzed and the model of PHEBand its components are built.Secondly, a multi-objective optimization model of PHEB component sizes andcontrol parameters aiming at the minimization of energy consumption, component costs,and acceleration time is built. Then, with the goal-attainment method, the multi-objectiveproblem is converted into a single-objective problem. For the multidimensional,nonlinear, multimodal optimization problem of PHEB parameters, simple geneticalgorithm (SGA) might not powerful to solve the optimization problem. Therefore ahybrid genetic algorithm (HGA) which combines an enhanced genetic algorithm (EGA)with simulated annealing (SA) is proposed in this paper. Several techniques are adoptedin HGA, such as real-coded, orthogonal design, SA, adaptive mechanisms, et al. Thesimulation results show that the proposed HGA has a superior performance than both SGAand EGA, and might provide a feasible method for the optimization of PHEB parametersin the engineering applications.Finally, in view of the widely used rule-based (RB) control strategy optimized by GAis still lack of the global optimization property, while global optimization algorithms havedifficulties in real-time applications, this paper proposed a dynamic programming-rule based (DPRB) control method, which consists of a real-time controller and an off-linedynamic programming (DP). Meanwhile, GA is adopted to replace the quantizationprocess of DP permissible control set to reduce the computation burden. Simulations areconducted based on the historical running information of a test route. Simulation resultsdemonstrate that the proposed DPRB might distribute electric energy more reasonablythroughout the bus route and improve the vehicle economy. DPRB might give apracticable solution, which is a tradeoff between the applicability of RB and globaloptimization property of DP.