基于信息熵模糊聚类和粗糙集理论故障的模糊判据研究

摘要

复杂电子系统的缓变故障是故障预报的难点之一,针对这一问题提出了基于信息熵模糊聚类和粗糙集理论的故障决策判据方法;该方法主要分两个步骤:以信息熵作为聚类标准,采用谱系的方法确定聚类数目,然后通过FCM模糊聚类构造故障决策表;利用粗糙集理论对故障决策表进行简化和最小化,最终形成带评价的简约故障决策判据;该方法改进了一般模糊聚类算法的不足,克服了先验信息和知识不准确、不完整、不一致情况下故障决策表获取与更新的困难;实际算例表明,信息熵模糊聚类方法比一般模糊聚类方法的聚类质量更高,更客观真实.%Soft fault of Complex electronic systems is one of the difficulties in fault prediction. Aiming at this problem, the paper proposes a fault decision criterion method based on information entropy, fuzzy clustering and rough set theory. In this study, algorithm combines information entropy and FCM (fuzzy c-means clustering algorithm) are analyzed, which improves the deficiencies of traditional FCM and overcomes the difficulties of decision table' s acquisition and update in the case of inaccurate, incomplete and inconsistent prior information or knowledge. A real calculated example shows that, compared to traditional FCM, information entropy-fuzzy clustering method has higher clustering quality and objectivity.