基于主元分析算法的回转窑故障诊断

A Study on Fault Diagnosis of Rotary Kiln Based on Principal Component Analysis Algorithm

针对回转窑故障会引发一系列危害的问题,提出了基于主元分析结合贡献图的故障诊断方法,首次应用于回转窑系统。首先将数据进行预处理;然后从总体相关系数矩阵出发求解主成分,通过累计贡献率确定主元个数,建立主成分表达式;以回转窑系统为对象进行仿真分析,将实时运行数据输入建立的主元模型,计算相关统计量T2和SPE,并与一定置信度的控制限相比较,监测过程运行的状态是否发生异常;最后采用贡献图方法进行诊断分析,识别故障原因。仿真结果表明,该方法能够从准确度和高效性等方面很好地进行回转窑故障诊断。

In view of the rotary kiln failure triggered a series of hazard problems,this paper proposes the fault diagnosis method based on principal component analysis algorithm combined with the contribution plot.The data is preprocessed first;Then embarks from the overall correlation coefficient matrix to solve principal components,determining principal component number by the cumulative contribution rate,establishing the principal component expression.The simulation analysis of rotary kiln system is carried out.Take real-time running data to established PCA model,calculating related statistic T2 and SPE,and compare with a certain degree of confidence of control limit,monitoring process of running state whether an exception occurs.Finally,adopt the method of contribution plot,making the diagnostic analysis and identifing the cause of the problem.