基于决策分类方法的水泥回转窑故障诊断方法

Fault diagnosis method of cement rotary kiln based on decision classification method

提出了一种基于决策分类的水泥回转窑故障诊断方法。该方法利用水泥生产过程中产生的回转窑历史数据结合模糊决策树算法生成故障诊断模糊推理规则库,利用在线数据可对模糊规则库进行增量学习,不断完善故障诊断规则库。故障诊断阶段,将在线数据形成的特征向量与故障诊断规则库进行匹配,判断是否发生故障以及相应的故障类型。该方法对故障数据进行模糊化处理,提高了故障诊断结果的可解释性以及泛化性能,适用于实际的水泥生产过程。

A fault diagnosis method of cement rotary kiln based on decision classification is proposed.The method using the cement production process of rotary kiln historical data combining with fuzzy decision tree algorithm to generate fault diagnosis rule base of fuzzy reasoning,fuzzy rule base to make use of online data can be incremental learning,constantly improve fault diagnosis rule base.In the fault diagnosis stage,the feature vector formed by online data is matched with the fault diagnosis rule library to determine whether there is a fault and the corresponding fault type.This method is applied to the actual cement production process by fuzzy treatment of fault data,which improves the interpretability and generalization of fault diagnosis results.