混合动力汽车模糊神经参考控制策略优化

为提高并联式混合动力汽车控制策略精准性,建立了基于发动机效率的模糊逻辑控制器,进一步使用神经网络模型对模糊逻辑控制器的隶属度函数进行在线学习,引入变尺度优化方法的改进型学习算法,完成了隶属度函数的在线学习后的优化;通过模型参考构成闭环在线修正,降低输出转矩的误差。通过循环仿真实验,利用模糊神经参考模型控制策略,发动机工作效率点与实时工况的匹配率更高稳定性更好,发动机平均效率提高4. 16%,峰值电源荷电状态保持在稳定的容量范围内,整车燃油经济性得到改善。因此该控制策略具有较强的工程实用性。

Optimization of Fuzzy Neural Reference Control Strategy for Hybrid Electric vehicle

In order to improve the accuracy of control strategy of parallel hybrid electric vehicle. A fuzzy logic controller based on engine efficiency was established for the parallel hybrid electric vehicle. Further the neural network model was used to study the membership function online of fuzzy logic controller based on engine efficiency. By using an improved learning algorithm with a variable metric optimization method, the online learning optimization of membership function had been accomplished. Closed-loop online Correction based on Model reference is established, the output error was further corrected. The simulation results show that the control strategy of neural fuzzy reference model could match efficiency of engine working point and real time condition is higher and stability was better. The engine average efficiency is increased by 4.16%. Moreover , this strategy could ensure the peak power state of charge within a relatively stable capacity range, the hundred kilometers fuel consumption is reduced. It is concluded that control strategy has strong engineering practicability.