神经网络工况识别的混合动力电动汽车模糊控制策略

Fuzzy control strategy for hybrid electric vehicle based on neural network identification of driving conditions

采用模糊控制可以改进混合动力电动汽车(HEV)的燃油经济性和排放性,但是对模糊控制器进行优化时通常只针对某一典型工况.不同的城市的行驶工况有一定差别,影响了模糊控制改善混合动力电动汽车性能的效果.研究中以广州和上海市主干道行驶工况为例,首先建立了一个模糊控制策略,并采用遗传算法,以汽车燃油经济性和排放性为优化目标,分别针对广州和上海主干道行驶工况对模糊控制器中隶属度函数进行优化.然后建立了一个基于模糊神经网络的行驶工况识别方法,通过识别广州和上海的主干道行驶工况,对控制策略中模糊控制器的隶属度参数进行相应调整,结果证明采用模糊神经网络识别行驶工况的HEV模糊控制策略可以进一步提高汽车的燃油经济性和排放性能.

The fuzzy control strategy can improve the fuel consumption and reduce the emission of hybrid electric vehicle（HEV）, but the parameters of control strategy are always optimized under a typical driving condition which is different from different cities. We study the fuzzy control strategy based on the urban driving conditions of Guangzhou and Shanghai. First, we propose a fuzzy control strategy and optimize the parameters of membership functions by applying the genetic algorithm to the urban driving conditions in Guangzhou and Shanghai. Second, we identify the urban driving conditions in these two cities based on the fuzzy neural network. The results of identification are applied to adjust the parameters of membership functions in the fuzzy control strategy for the HEV. The simulation results show that the HEV fuzzy control strategy based on the fuzzy neural network identification of driving conditions improves the fuel consumption and reduces the emission.