基于数据挖掘技术的旋转机械故障诊断研究

Study on Rotary Machinery Fault Diagnosis Based on Data Mining

针对旋转机械故障诊断需人工干预、效率低、故障样本难获取等问题,提出基于数据挖掘技术的旋转机械故障诊断方法。提取不同故障的时域及频域特征频率,构成典型的故障样本数据库;采用粗糙集和决策树融合的诊断算法生成故障诊断的决策。用轴承故障诊断案例证明模型的有效性。

Considering the disadvantages existing in conventional fault diagnosis methods for rotating machinery, such as necessity of manual intervention, low efficiency and difficulty to obtain fault samples, a fault diagnosis method was proposed based on data mining technology. The characteristic frequency of time domain and frequency domain in different fault was extracted to build a typical fault samples database. Syncretic diagnosis algorithm of rough set and decision tree was used generate the decision-making of fault diagnosis. A diagnostic case of a bearing proves the effectiveness of the model.