摘要
由于CCBII制动机具有多种工作模式,工况复杂,工作部件繁多且耦合性强等特点,经常有多故障并发。传统的故障诊断技术难以对其进行多故障诊断。本文在基于解析冗余关系的诊断技术上,引入差分进化算法进行多故障检测与定位。首先构建制动机各工作模式模型,并推导系统的扩展全局解析冗余关系以获得模式转变特征表和系统故障特征表,再引入一致性向量比较特征表,实现系统的模式追踪。当模式追踪失败后,构建候选故障假设集,对每个候选故障设计基于差分进化算法参数估计器,通过比较候选故障参数的平均适应函数值,定位同时发生的多个故障,实现制动机系统多故障实时在线检测。最后通过仿真实验和实际应用验证本文方法的有效性。
Due to the fact that the CCBII brake system is characterized by multiple operation modes,complex working environments and mutiple number of components with strong coupling,multiple faults may often hap-pen simultaneously.It is difficult to carry out multiple faults diagnosis for the CCBII brake system using the traditional fault diagnosis techniques.In this paper,the differential evolution algorithm was introduced into the diagnosis technique based on the analytical redundancy relations to implement multiple faults detection and lo-cation.Firstly,the models of working modes were established and the augmented global analytical redundancy relations（AGARRs）for the system were deduced to obtain the mode-change signature matrix and fault signa-ture matrix.Mode tracking was done by comparing the two signature matrixes with coherence vectors.Once mode tracking failed,the candidate fault assumption sets were established correspondingly.Furthermore,fault parameter estimators were designed on the basis of the differential evolution algorithm.Multiple faults were i-dentified among the hypothesis sets by calculating the values of the mean fitness function.The real-time online faults detection of the CCBII brake system was realized.Finally,simulation tests and practical applications ver-ify the effectiveness of the proposed scheme.
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2014年第3期67-74,共8页
Journal of the China Railway Society
基金
国家自然科学基金(61073103
61071096
61003233)
2010年度高等学校博士学科点专项科研基金(20100162110012
20110162110042)
关键词
多故障诊断
差分进化算法
模式追踪
CCBII制动机
multiple faults diagnosis
differential evolution algorithm
mode tracking
CCBII brake system
