关键词: 数据挖掘, 关联规则算法, 神经网络, 风险维修, 故障诊断

Abstract: Aiming at hydraulic equipments characteristics in steel enterprises, a maintenance decision-making method with risk control was proposed. In this method, only one calculation was needed to judge whether the equipment was in the fault state, the potential fault type and the probability value. Compared to the traditional risk maintenance methods, the fault diagnosis accuracy was improved, as well as the diagnosis speed, which provided better reference for on-line decision-making. The experiment proved the effectiveness of this method.

Key words: data mining, association rules arithmetic, neural network, risk-based maintenance, fault diagnosis