Structural Fault Detection Using Weighted Principal Component Analysis ( WPCA )

Fault that occurred in a system is actually affecting the quality of the products produced and as a result, the process monitoring is required to eliminate the fault in the system and eventually increase and met the performance specification. Principal Component Analysis(PCA) is a method that have been introduced in process monitoring to detect the fault in the system and it has been categorized as one of the method of Multivariable statistical process monitoring (MSPM) as its ability to monitor multivariable system. The extension of PCA is proposed which is Weighted Principal Component Analysis (WPCA) to deal with the situation of useful information being submerged and reduced missed detection rate of T2 statistic. The main idea of WPCA is building conventional PCA model and then using change rate of T2 statistic along every principal component (PC) to capture the most useful information in process, and setting different weighting values for PCs to highlight useful information.WPCA method will be focusing on how to detect structural fault since most of the literatures only focusing on the variable change. In this paper, structural fault will be simulated using that CSTR model which will be developed using MATHLAB software. Lastly, the process data will be collected and tested with WPCA.