纯电动汽车烟花算法自学习模糊控制策略研究

摘要

针对纯电动汽车复合电源模糊控制策略靠人工经验的主观性与局限性问题,基于advisor软件建立超级电容与蓄电池复合电源纯电动汽车模型,利用新的爆炸式搜索算法—烟花算法,调用advisor软件对复合电源模糊控制策略隶属度函数及模糊规则进行闭环学习,并与遗传算法、交叉蜂群算法的学习效果进行比较.结果表明:烟花算法学习的模糊控制策略与经验模糊控制策略相比,在FTP、NEDC、UDDS和CHINACITY工况下能耗仍能再降低5.47%、5.23%、4.25%和4.15%,且烟花算法与其他算法相比在复合电源纯电动汽车模糊控制学习中的求解精度更高.%Aiming at the problem of subjectivity and limitation of pure electric vehicle couple synergic electric power supply system fuzzy control relying on human experience,building the super capacitor and battery electric vehicle based on advisor software,studying the FM and rules of fuzzy control system calling advisor software uses a new explosive search method firework algorithm,and compares the effect with GA and cross ABC algorithm.The result shows:the energy consumption of FTP,NEDC, UDDS and CHINACITY still can reduce 5.47%,5.23%,4.25% and 4.15% under the fuzzy control strategy improved by firework algorithm comparing the fuzzy control strategy based human experience.Also,firework algorithm has better accuracy compared with other algorithm under the question of couple power supply electric vehicle fuzzy control studying strategy.