模糊遗传算法的神经网络方法在变压器故障诊断中的研究

Study of the FGA-BP Method for Transformer Fault Diagnosis

基于油中溶解气体数据,避开了传统的比值方法,采用模糊遗传算法的神经网络(FGA-BP)方法来诊断变压器故障。该方法是在遗传寻优过程中,用模糊控制的方式对杂叉率与变异率进行动态调节,并结合神经网络来对故障实现模式识别。为了优化模式特征量,文章用灰关联分析方法对样本集进行了筛选,得到了一组模式特征向量,以此DGA数据作为FGA-BP的输入值,经过FGA-BP运算后,优化出一组用于模式识别的权重与阈值,在此基础上,结合实例对该诊断方法进行了分析与探讨。

Based on the Dissolved Gas Analysis (DGA) data, ratio method is the nor mal method for the fault diagnosis of oil-immersed electric power transformer. I n this paper, the novel method of Fuzzy Genetic Back-propagation neural network (FGA-BP) is applied to transformer fault diagnosis instead of the ratio method. The novel method combines GA and BP,during genetic algorithm's optimize,crossov er rate and mutation rate are adjusted dynamically by fuzzy control.To optimize model samples,in the first part samples set are filtered through the method of d egree of grey relational analysis, then the data of the model samples are operat ed by FGA-BP and a group of weighs and biases are found. Finally examples are gi ven.Compared to the other traditional method, the results have demonstrated the robustness of the method. Some conclusions are provided.