基于KFCM-MNN并联式混合动力汽车能量管理策略  

Energy management strategy for parallel hybrid electric vehicles based on KFCM-MNN

作　　者：孔慧芳[1] 朱翔[1] KONG Huifang;ZHU Xiang(School of Electric Engineering and Automation,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2018年第4期485-489,共5页Journal of Hefei University of Technology：Natural Science

基　　金：国家科技支撑计划资助项目(2014BAG06B02);中央高校基本科研业务费专项资金资助项目(2014HGCH0003)

摘　　要：为了提高并联式混合动力汽车的燃油经济性,文章提出了一种基于核模糊c聚类(kernel fuzzy cmeans clustering,KFCM)的多神经网络(multi-neural network,MNN)能量管理设计方法。采用动态规划全局优化离线仿真得到全局最优解,使用KFCM算法对全局最优解数据集合按照车辆运行模式作聚类划分,针对每一个聚类建立局部神经网络。训练后的MNN模型结构根据实时工况,将多个局部神经网络的输出联结作为能量管理策略的输出,以实现发动机和电机转矩的实时优化分配。仿真结果表明,基于KFCM-MNN的能量管理策略,具有对动态规划能量管理策略很好的学习模拟能力,是一种准最优的控制策略。In this paper,considering multi-mode modeling,an energy management strategy based on kernel fuzzy c-means clustering(KFCM)multi-neural network(MNN)mode is proposed,in order to improve the fuel economy of parallel hybrid electric vehicles.The global optimal solution set from offline simulations is optimized globally with dynamic programming(DP),and KFCM algorithm is adopted to make a partition for the global optimal solution set,then local neural network is established for each cluster.According to real-time conditions,the trained MNN mode joins the outputs of several neural networks as the final output of the proposed energy management strategy,to realize the real-time optimal distribution of the engine and motor torque.Simulations results show that the designed energy management strategy has the excellent learning simulation capability to DP and is a quasi-optimal control strategy.

关 键 词：并联式混合动力汽车 动态规划 多神经网络(MNN) 核模糊c聚类(KFCM) 能量管理策略 

分 类 号：TP273.3[自动化与计算机技术—检测技术与自动化装置]