Fault Detection and Isolation Indices for Large-Scale Systems

Multivariate statistical process monitoring techniques have been developed to detect and isolate abnormal situations of modern industrial processes that became more complicated and are classified as large-scale systems. Several fault detection and isolation indices have been proposed for multivariate statistical process monitoring. This paper discusses these indices and compare their performances by applying for an industrial benchmark, the Tennessee Eastman chemical process. The efficiency of these indices is measured by four key performance indicators (KPIs), i.e., fault detection time delay, false alarm rate, missed detection rate, correct fault isolation.