【摘要】 变压器油中溶解气体分析(DGA)是电力变压器故障诊断的重要方法。针对模糊C均值聚类算法用于溶解气体成分分析时存在的问题,将加权模糊核聚类方法(WFKC)引入到电力变压器故障诊断中,建立了一个新的变压器故障诊断模型。该法首先考虑到样本中不同特征对聚类结果的不同影响,利用基于样本相似度的加权方法对样本特征进行加权,然后将样本从输入空间映射到高维特征空间,在特征空间实现加权模糊核聚类。形成的模型充分考虑了不同特征对聚类结果的不同影响,能有效改善复杂数据集的聚类性能,提高了故障诊断的正确率。案例分析表明,该法能快速有效地对样本进行聚类,从而验证了该法在变压器故障诊断中的有效性和可行性。更多还原

【Abstract】 Dissolved gas analysis(DGA)is an important method to diagnose the fault of power transformer.To solve the problems existed in fuzzy c-means clustering algorithm which is applied in DGA,the weighted fuzzy kernel clustering(WFKC) algorithm is introduced into the fault diagnosis of power transformers to build a new fault diagnosis model.In the algorithm,firstly considering that the different effects of the different attributes on cluster results,so the similarity based weighting method is used to assign weight to features of the transferred samples,and then weighted fuzzy kernel clustering in the feature space is realized when the transferred samples in the original space is mapped into high-dimensional feature space.The new fault diagnosis model can adequately consider that the different effects of the different attributes on cluster results and effectively improve the clustering capability for the complex dataset,and the correct rate is effectively improved.WFKC is applied in practice to analyze fault diagnosis of power transformers,The results demonstrate that this algorithm can cluster the samples fast and efficiently,and WFKC is feasible and valid.更多还原