论文部分内容阅读
考虑模糊聚类的数据离散功能,粗糙集理论对决策系统的约简能力,以及模糊神经网络在模式识别方面具有的优势,提出了模糊c 均值(FCM)—粗糙集—自适应模糊神经网络推理系统(ANFIS)集成进行故障诊断的方案:首先,应用FCM聚类方法离散故障诊断数据中的连续属性值;然后,基于粗糙集理论计算诊断决策系统的约简,按照实际需要确定诊断条件;最后,根据系统约简设计ANFIS进行故障诊断。4135柴油机的实际诊断结果验证了文中提出集成故障诊断方案的可行性。在数据充分的条件下,该方案可以推广应用于其它机械设备。
Considering the data discrete function of fuzzy clustering, the reduction ability of rough set theory to decision system, and the advantages of fuzzy neural network in pattern recognition, this paper proposes fuzzy c-means (FCM) - rough set-adaptive fuzzy neural network inference System (ANFIS) to integrate fault diagnosis scheme: Firstly, the FCM clustering method is used to discretize the continuous attribute values in the fault diagnosis data; then, the reduction of the diagnostic decision system is calculated based on rough set theory, and the diagnosis conditions are determined according to actual needs; finally , According to the system design ANFIS reduction for fault diagnosis. The actual diagnosis results of 4135 diesel engine verify the feasibility of the integrated fault diagnosis scheme proposed in this paper. Under the condition of sufficient data, the scheme can be widely applied to other machinery and equipment.