A data-driven fault detection approach for static processes with deterministic disturbances

Based on the well established model-based fault detection techniques, in this paper, a data-driven fault detection approach for static processes with deterministic disturbances is proposed. The basic idea behind this approach is, first identify the maximum influence of the unknown input on the measurement using the fault-free recorded data, and then apply the existing model-based schemes to solve the fault detection problem. The performance and effectiveness of the proposed scheme are demonstrated through a laboratory continuous stirred tank heater (CSTH) setup.