基于模糊免疫识别方法的电力变压器故障诊断

Power Transformer Fault Diagnosis Based on Fuzzy Immune Recognition Approach

在人工免疫识别系统的基础上,结合模糊推理,提出了基于模糊免疫识别的充油变压器故障诊断方法。代表相同B细胞的人工识别球经克隆选择和资源竞争后产生具有分类性能的记忆抗体,但其资源分配数目和人工识别球的刺激水平成线性,易使少数优秀人工识别球占据过多资源,引起人工识别球种群早熟收敛。将模糊推理应用于资源分配机制中,可使分配兼顾刺激水平和抑制水平,改善收敛性能。实例分析表明,以H2、CH4、C2H4、C2H6、C2H2等5种气体含量为特征参数,将模糊推理与人工免疫识别系统相结合的方法,能有效诊断变压器内部潜伏性故障。

On the basis of artificial immune recognition system and combined with fuzzy reasoning, a fault diagnosis method for oil-filled transformer based on fuzzy immune recognition is proposed. After the clone selection and resource competition of artificial recognition ball （ARB） representing the same B cells the memory antibodies possessing classification performance are produced, however because of the linear relation between the recourse allocation of memory antibodies and the stimulation level of ARB, a few of excellent ARBs are easy to hold too much resource and it causes the premature convergence of ARB population. The Application of fuzzy reasoning in resource allocation can make the allocation considering both stimulation level and suppression level, thus the convergence performance can be improved. Case analysis results show that the method integrating fuzzy reasoning with artificial immune recognition system in which the content of five gases, i.e., H2, CH4, C2H4, C2H6 and C2H2 are taken as characteristic parameters can diagnose the internal latent faults in oil-filled transformer effectively.