改进D-S证据理论在电动汽车锂电池故障诊断中的应用

发布时间：2018-04-24 10:49

本文选题：故障诊断 + 电动汽车　；参考：《智能系统学报》2017年04期

【摘要】：针对电动汽车电池系统的故障采用基于神经网络的改进D-S证据理论组合规则完成诊断过程。为了避免单一途径的诊断可能造成故障漏检误检的状况,决策层采用D-S证据理论组合规则来确定基于BP网络和RBF网络两种故障诊断算法结果。然而为了克服D-S证据理论处理高度冲突证据的缺陷,本文提出了一种基于神经网络改进的D-S证据理论组合规则。首先,采用神经网络对电池故障进行初步诊断,结合网络诊断准确率来分配不确定信息并构造证据体,又引入了证据间的支持矩阵来确定新的加权证据体。然后,把各个焦元的信任度融入D-S证据理论组合规则,从而融合神经网络证据体及新加权证据体。最后,依据决策准则确定锂电池系统的故障状态。通过仿真实验验证了本文提出的改进D-S证据理论融合诊断方法在电动汽车锂电池故障诊断中的有效性。
[Abstract]:The improved D-S evidence theory combination rule based on neural network is used to complete the diagnosis process for the battery system of electric vehicle. In order to avoid the situation that single way diagnosis may lead to false detection, the decision layer adopts the combination rule of D-S evidence theory to determine the results of two fault diagnosis algorithms based on BP network and RBF network. However, in order to overcome the defects of D-S evidence theory in dealing with highly conflicting evidence, this paper presents an improved D-S evidence theory combination rule based on neural network. Firstly, the neural network is used to diagnose the battery fault, the uncertain information is allocated and the evidence body is constructed by combining the accuracy of the network diagnosis, and the support matrix between the evidences is introduced to determine the new weighted evidence body. Then, the trust degree of each focal element is incorporated into the D-S evidence theory combination rule, thus the neural network evidence body and the new weighted evidence body are fused. Finally, the fault state of lithium battery system is determined according to the decision criterion. The effectiveness of the improved D-S evidence theory fusion diagnosis method for lithium battery fault diagnosis of electric vehicles is verified by simulation experiments.
【作者单位】：上海电力学院自动化工程学院;上海发电过程智能管控工程技术研究中心;同济大学电子与信息工程学院;
【基金】：上海市“科技创新行动计划”高新技术领域科研项目(15111106800) 上海市发电过程智能管控工程技术研究中心项目(14DZ2251100) 上海市电站自动化技术重点实验室开放课题(13DZ2273800)
【分类号】：TP183;U469.72;U472.9

【相似文献】