基于人工免疫算法的变压器故障诊断方法

Artificial immune algorithm for fault diagnosis of power transformer

人工免疫系统是一种由生物免疫系统启发的学习外界物质的自然防御机理的学习技术。基于人工免疫识别原理提出一种电力变压器故障诊断方法,利用增加抗原、记忆抗体类别信息等方法对变压器故障样本进行学习,获取更多表征故障样本特征的记忆抗体集并进行分类。经过对故障数据的Matlab实验,将结果与IEC三比值法进行比较,表明该算法有较高的诊断准确率。

Artificial immune system is a kind of learning technology which is stimulated by a biological immune system and studies the natural defense mechanism of outside material. Based on the principle of immune recognition, this paper proposes a power transformer fault diagnosis method,which can get more characterization fault memory antibody characteristics of the sample collection and classification through the study of the fault samples by increasing the use of antigens, antibodies and memory types of information and so on.From the Matlab experimental data and the comparison of the results and the IEC of three rations, we can get the conclusion that this algorithm can get higher accuracy of diagnosis.