基于多变量统计过程监控的盾构机故障诊断

由于盾构机安装有众多监测仪表,且依靠单变量过程监控和人工诊断方式已经无法满足监测的要求,因此引入多变量统计过程监控方法(MSPM).主元分析(PCA)是应用最广泛的MSPM技术,PCA根据盾构运行监测的过程变量和历史数据建立数学模型,并计算统计监控量T2和平方预测误差δ,以及主元空间和残差空间的控制限,分析过程变量是否发生异常.最后以盾构刀盘驱动系统和螺旋输送液压系统为例说明MSPM的详细应用.

Fault diagnosis on shield machines based on multivariable statistical process monitoring

Due that numerous monitoring instruments are equipped on the shield machines,the single-variable process monitoring and manual diagnosis cannot meet the demands.In this study,the multivariable statistical process monitoring(MSPM),amongst which the primary component analysis(PCA) is the most frequently-applied technology,is first introduced.Based on the process variables and historical data of shield operational monitoring from PCA,a mathematical model for abnormal process variable analysis is then established to calculate the statistical monitoring variable,e.g.T2,the squared prediction error(δ) and the control limit between the primary component and residual spaces.Finally,shield cutter-plate driving system and screw conveying hydraulic system are used as examples for the detailed MSPM applications.