基于扩展Type-2模糊集的故障检测方法

Fault detection based on expanded Type-2 fuzzy sets

随着系统复杂程度的不断提高,对系统故障的诊断也越来越困难。提出一种扩展的Type-2模糊集合,利用扩展隶属函数的"肯定证据,排除假象"的特性,建立基于Type-2的模糊诊断方法,并给出了相应的设计算法。增加了负值的扩展隶属函数,将系统输出的残差清晰地划分为三个状态:正常、异常和故障。增加了一个状态,能够更好地得出清晰的诊断结果,有效地避免误诊和漏诊。针对三水箱系统,仿真结果表明了该方法的有效性。

As the raising of the complexity of systems, it is more and more difficult for the fault diagnosis. An expanded Type-2 fuzzy set is proposed, whose membership function has the character of ＂confirming evidence, eliminating feint＂. Then a fuzzy diagnosis system is developed based on Type-2, and the corresponding designing algorithm is also given. As the negative values are added to the membership function, the residual error outputs could be clearly divided into three statuses： normal, abnormal and fault. The diagnostic conclusion can be clearly obtained by the added status, and miss-diagnosis and misdiagnosis could be prevented. Aiming at a kind of nonlinear system, simulation result verifies the efficiency of this method.