正在加载图片...
第12卷第4期 智能系统学报 Vol.12 No.4 2017年8月 CAAI Transactions on Intelligent Systems Aug.2017 D0I:10.11992/is.201605001 改进D-S证据理论在 电动汽车锂电池故障诊断中的应用 夏飞23,马茜2,张浩23,彭道刚2,孙朋2,罗志疆2 (1.上海电力学院自动化工程学院,上海200090:2.上海发电过程智能管控工程技术研究中心,上海200090:3.同 济大学电子与信息工程学院,上海201804) 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DS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$证据理论处理高度冲突证据的缺陷,本文提出了一种基于神经网络改 进的D-$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DS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。 关键词:故障诊断:电动汽车:锂电池:改进证据理论:信息融合 中图分类号:TP301文献标志码:A文章编号:1673-4785(2017)04-0526-12 中文引用格式:马茜,夏飞,张浩,等.改进D-S证据理论在电动汽车锂电池故障诊断中的应用[J].智能系统学报,2017,12(4): 526-537. 英文引用格式:MAXi,XIA Fei,.ZHANG Hao,etal.Application of improved D-S evidence theory in fault diagnosis of lithium batteries in electric vehicles[J].CAAI transactions on intelligent systems,2017,12(4):526-537. Application of improved D-S evidence theory in fault diagnosis of lithium batteries in electric vehicles XIA Fei3,MA Xi2,ZHANG Hao23,PENG Daogang'2,SUN Peng'2,LUO Zhijiang'2 (1.College of Automation Engineering,Shanghai University of Electric Power,Shanghai 200090,China;2.Shanghai Engineering Research Center of Intelligent Management and Control for Power Process,Shanghai 200090,China;3.College of Electronics and Information Engineering,Tongji University,Shanghai 201804,China) Abstract:In this study,we used the improved Dempster-Shafer (D-S)evidence theory combination rules based on the neural network to construct a fault diagnosis process for an electric vehicle battery system.To avoid misdiagnoses and missed diagnoses caused by a single fault diagnosis method,we applied the D-S evidence theory combination principle to determine the result based on the back-propagation (BP)network and radial basis function (RBF) network fault diagnosis algorithm.However,to overcome the defects in the D-S evidence theory in dealing with highly conflicting evidence,we propose a D-S evidence theory combination principle based on an improved neural network.First,we apply a neural network to perform a preliminary diagnosis regarding battery failure and the accuracy of the network diagnosis.Then,we distribute indefinite information and construct a body of evidence.We also introduce a support matrix of this evidence to determine a new weighted body of evidence.We then integrate the credibility of every focal element into the D-S evidence theory combination rules to fuse the neural network body of evidence with the new weighted body of evidence.Lastly,based on the decision criterion,we determine the failure state of the lithium battery system.Our simulation results show that our proposed improved D-S evidence theory fusion diagnosis method is effective in the fault diagnosis of electric vehicles with lithium batteries. Keywords:fault diagnosis;electric vehicle;lithium battery;improved evidence theory;information fusion 电动汽车的动力锂电池系统对于整个电动汽 车而言,是保证汽车正常行驶和准确预估续驶里程 的基础。当前制约电动汽车发展的核心技术就是 收稿日期:2016-05-03 在行车过程中电池系统能否准确切实提供动力,确 基金项目: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 保安全出行。由于目前国内动力电池技术并非完 心项目(14DZ2251100):上海市电站自动化技术重点实验室 全成熟,电池故障在初期征兆不易察觉,因此对电 开放课题(13D72273800). 通信作者:张浩.E-mail:hzhangk@163.com第 12 卷第 4 期 智 能 系 统 学 报 Vol.12 №.4 2017 年 8 月 CAAI Transactions on Intelligent Systems Aug. 2017 DOI:10.11992 / tis.201605001 改进 D⁃S 证据理论在 电动汽车锂电池故障诊断中的应用 夏飞1,2,3 ,马茜1,2 ,张浩1,2,3 ,彭道刚1,2 ,孙朋1,2 ,罗志疆1,2 (1.上海电力学院 自动化工程学院 , 上海 200090; 2.上海发电过程智能管控工程技术研究中心, 上海 200090; 3.同 济大学 电子与信息工程学院,上海 201804) 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。 关键词:故障诊断;电动汽车;锂电池;改进证据理论;信息融合 中图分类号:TP301 文献标志码:A 文章编号:1673-4785(2017)04-0526-12 中文引用格式:马茜,夏飞,张浩,等.改进 D⁃S 证据理论在电动汽车锂电池故障诊断中的应用[ J]. 智能系统学报, 2017, 12( 4): 526-537. 英文引用格式:MA Xi, XIA Fei, ZHANG Hao,et al. Application of improved D⁃S evidence theory in fault diagnosis of lithium batteries in electric vehicles[J]. CAAI transactions on intelligent systems, 2017, 12(4): 526-537. Application of improved D⁃S evidence theory in fault diagnosis of lithium batteries in electric vehicles XIA Fei 1,2,3 , MA Xi 1,2 ,ZHANG Hao 1,2,3 , PENG Daogang 1,2 , SUN Peng 1,2 , LUO Zhijiang 1,2 (1. College of Automation Engineering, Shanghai University of Electric Power, Shanghai 200090, China; 2. Shanghai Engineering Research Center of Intelligent Management and Control for Power Process, Shanghai 200090, China; 3. College of Electronics and Information Engineering, Tongji University, Shanghai 201804, China) Abstract:In this study, we used the improved Dempster⁃Shafer (D⁃S) evidence theory combination rules based on the neural network to construct a fault diagnosis process for an electric vehicle battery system. To avoid misdiagnoses and missed diagnoses caused by a single fault diagnosis method, we applied the D⁃S evidence theory combination principle to determine the result based on the back⁃propagation (BP) network and radial basis function (RBF) network fault diagnosis algorithm. However, to overcome the defects in the D⁃S evidence theory in dealing with highly conflicting evidence, we propose a D⁃S evidence theory combination principle based on an improved neural network. First, we apply a neural network to perform a preliminary diagnosis regarding battery failure and the accuracy of the network diagnosis. Then, we distribute indefinite information and construct a body of evidence. We also introduce a support matrix of this evidence to determine a new weighted body of evidence. We then integrate the credibility of every focal element into the D⁃S evidence theory combination rules to fuse the neural network body of evidence with the new weighted body of evidence. Lastly, based on the decision criterion, we determine the failure state of the lithium battery system. Our simulation results show that our proposed improved D⁃S evidence theory fusion diagnosis method is effective in the fault diagnosis of electric vehicles with lithium batteries. Keywords: fault diagnosis; electric vehicle; lithium battery; improved evidence theory; information fusion 收稿日期:2016-05-03. 基金项目: 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). 通信作者:张浩.E⁃mail: hzhangk@ 163.com. 电动汽车的动力锂电池系统对于整个电动汽 车而言,是保证汽车正常行驶和准确预估续驶里程 的基础。 当前制约电动汽车发展的核心技术就是 在行车过程中电池系统能否准确切实提供动力,确 保安全出行。 由于目前国内动力电池技术并非完 全成熟,电池故障在初期征兆不易察觉,因此对电
向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有