铝电解槽阴极的电压信号能够反应铝电解槽的运行状况。通过对采集到的铝电解槽的阴极电压进行小波去噪,对重构后的包含主要信息的信号进行无量纲特征参数提取。选出对信号故障最敏感的两个特征,运用K均值聚类分析的方法,将信号的特征分为正常的一类和有故障的一类,从而确定信号是否存在故障和故障发生的时间,这对于铝电解槽的故障在线诊断提供了借鉴。
The voltage signals of the cathode of aluminum reduction pot can reflect the operation status of the aluminum reduction pot. By wavelet de - noising of collected cathode voltage for aluminum reduction the dimensionless characteristic parameters of the reconstructed signal containing the main information are extracted, and two characteristics which is the most sensitive characteristic to the fault signal are chose. By using the method of K - means clustering analysis, the signal features are divided into normal class and a kind of fault. It can determine whether there is fault signal and the happening time of the fault. The research results can provide reference for the fault diagnosis of aluminum reduction pot on line.