基于克隆选择分类算法的电力变压器故障诊断

变压器油中溶解气体分析是进行电力变压器故障诊断的一种有效方法,将克隆选择分类算法引入电力变压器油中溶解气体分析,利用免疫克隆选择原理学习并提取表征故障样本特征的记忆抗体集,然后用最邻近分类法对故障样本进行分类。人工免疫系统具有良好的自学习和自记忆能力, 使得克隆选择分类算法具有很强的非线性分类和泛化能力。经大量实例分析,并将其结果与IEC三比值法和BP神经网络等方法的结果相比较表明,该算法能有效对电力变压器单故障和多故障样本进行分类,并具有较高的诊断精度。

Power Transformer Fault Diagnosis Based on Clonal Selection Classification Algorithm

Dissolved gas analysis is an effective and important method for power transformer fault diagnosis. Here, the clonal selection classification algorithm is led into the analysis of gas dissolved in transformer oil, using immune clonal selection principle the memory antibody set characterizing the characteristics of fault samples is studied and extracted, then the nearest neighbor method is used to classify the fault samples. Because artificial immunization system possesses good self-learning capacity and self-memory capacity, so the clonal selection classification algorithm also possesses strong capacity of nonlinear classification and generalization. A lot of fault samples are analyzed by this algorithm, and the results are compared with those obtained by IEC three-ratio method and BPNN. Comparison results show that the proposed algorithm can effectively classify the single fault samples and multi-fault samples of power transformer, and the precision of fault diagnosis can be evidently improved.