摘要
针对利用单一判据对隔爆开关永磁机构储能电容进行故障诊断时存在较高的误判率的问题,提出了一种基于信息融合理论的电容失效诊断方法。通过分析储能电容的失效机理,得到其有效的故障特征参数,从而建立了储能电容状态评价的证据体系;为避免经验赋值的主观性,提出了基于模糊熵的最小模糊度优化模型,以此模型求取储能电容诊断模糊系统隶属度函数中的未知参数,并构建基于隶属度的电容状态基本概率分配函数。利用D-S证据合成规则实现多种故障信息的融合,并根据基本概率分配(BPA)决策准则来判断储能电容的状态。根据现场采集的数据,利用这种方法进行了实例分析,得到了其储能电容的当前状态评估结果,分析结果验证了该方法的有效性。
When the faults in energy-storage capacitor in flameproof switchgears are diagnosed with single evidence, the diagnostic system has a higher misdiagnosis. To overcome this problem, this paper proposes a capacitor failure diagno- sis method based on information fusion theory. Firstly, through analyzing the failure mechanism of storage -capacitor, the feature parameters of capacitor were extracted, and the evidence source for state assessment was then obtained. Second- ly, the basic probability assignment for capacitor state was determined according to fuzzy membership functions to avoid the subjectivity of membership assigned with expert experiences, and the membership function parameters were estab- lished by a minimum fuzzy optimization model. Finally, the different evidences were integrated by D-S combination rule, and according to BPA decision rule, the condition of the energy storage capacitor in practical operation can be as- sessed. By using the method proposed and the operation data, the current state of the energy-storage capacitors were obtained, and the evaluation study results demonstrate its effectiveness.
出处
《煤炭学报》
EI
CAS
CSCD
北大核心
2014年第10期2121-2127,共7页
Journal of China Coal Society
基金
国家自然科学基金资助项目(51207110)
中国博士后科学基金面上资助项目(2013M531210)
关键词
储能电容
失效诊断
D-S证据理论
模糊隶属度
模糊熵
storage-capacitor
failure analysis
D-S evidence theory
fuzzy membership
fuzzy entropy
