A benchmark of Canonical Variate Analysis for fault detection and diagnosis

The early detection and diagnosis of faults can improve the energy efficiency of industrial processes by avoiding the inefficient operation of faulty equipment as well as minimizing unplanned shutdowns, downtime and extensive damage to other parts of the system. In addition, industrial needs are evolving fast towards more flexible schemes. As a consequence, it is often required to adapt the plant production to the demand, which can be volatile depending on the application. This is why it is important to develop tools that can monitor the condition of the process working under varying operational conditions. Canonical Variate Analysis (CVA) is a multivariate data driven methodology that can be applied to detect and diagnose faults in industrial systems. This method has the ability to capture the process dynamics more efficiently than other similar data driven algorithms. The aim of this work is to provide a benchmark case to demonstrate the ability of CVA to detect and diagnose artificially seeded faults in a large scale test rig and measure the impact of those faults on the system performance, in particular its energy efficiency. The results obtained suggest that CVA can be effectively used for fault detection using real process data. The faults introduced were successfully detected in the early stages of degradation, and the source of the faults was identified using contribution plots.