基于多向主元分析的多传感器故障诊断

研究一种基于MPCA的多传感器故障诊断方法．这种方法把过程测量空间分为主元子空间和残差子空间．在残差子空间，首先用Q统计指标检测出传感器是否存在故障，如果Q统计指标超限，在主元子空间应用T^2统计量和相应的T^2统计量的贡献率，识别出引起过程异常的主要传感器变量并剔除．然后用同样的方法继续判断其它的传感器故障．仿真实例验证了该方法的有效性．

Multiple Sensor Fault Diagnosis Based on Multiway Principal Component Analysis

Monitoring batch processes to ensure their safe operation and to produce consistently high-quality products is needed. Multiway principal component analysis(MPCA) is a nonlinear modeling methodology for batch process. Previous publications have focused upon the application of statistical analysis for sensor fault identification through data reconstruction. These reconstruction based methods do not address the problem of fault propagation to other sensor measurements and as a consequence misleading fault identification can result. Based on MPCA, this paper use a multiple sensor faults diagnosis method. By using the T~2-statistic in conjunction with the associated contribution plot, multiple sensor faults can be identified in a sequential manner. Simulations verify the effectiveness of the method.