摘要
本文针对传统故障字典法不能对软故障实施诊断的缺陷,证明了运用节点电压增量向量对软故障实施诊断的基本理论并建立了相应的软故障字典。使用该字典不仅能够诊断出每一元件所有的软故障,且具有模拟次数少、诊断过程简单的特点。同时,为了解决因元件容差而导致故障诊断率低的问题,运用模糊理论设计了一个新的故障判别模型。新的模糊模型中,在高斯型隶属函数的参数估计方法以及判别算法上均与以往的类似方法有很大不同。诊断实例证明了本文所给出理论的正确性,实验结果也表明了新的模糊模型能最大程度地克服因人为经验而造成的误差,有效地提高了故障诊断的精度。
This paper address itself to linear analog soft fault diagnosis by means of fault dictionary and fuzzy technique. To overcome the limitations of the classical fault dictionaries, a new dictionary, in which the node voltage increment vectors are used as signatures, is proposed in this paper. If the parameter tolerances are considered, a refined fuzzy solution is designed to isolate each fault. The dictionary construction process is simple, and the simulation time is few. In the fuzzy solution, the parameter estimation method of the Gaussian membership function and the faults isola- tion algorithm are different from those old ones＇. The results of the given examples prove the correctness and effective- ness of the new approach in fault diagnosis.
出处
《电子测量与仪器学报》
CSCD
2009年第8期14-19,共6页
Journal of Electronic Measurement and Instrumentation
