In order to deal with the time series data from multiple sensors

a fault detection approach based on dynamic principal component analysis（DPCA）is proposed.With this approach

normal samples are used as training data to develop a DPCA model.T 2 statistic and SPE statistic are used to detect fault.SVI statistic performed as indexes of fault sensor diagnosis.Several simulations in sewage disposal system are used to validate the DPCA.The results show that the DPCA model can effectively extract the dynamic relations among process variables and sensor fault detection.