Model-based fault diagnosis for safety-critical chemical reactors: An experimental study

This article presents an experimental application of fault detection, isolation, and estimation in a chemical reactor system, introducing a functional observer-based approach without the need for linear approximation. The residual signal generators, functioning as disturbance-decoupled functional observers, provide fault size estimates and enable fault isolation through multiple generators operating independently. The experimental study focuses on the 3-Picoline oxidation process, deriving a discrete-time model, and constructing specific residual generators for coolant inlet temperature and feed concentration faults. Fault diagnosis employs Fast Fourier Transform (FFT) filtering and Generalized Likelihood Ratio (GLR), facilitating on-the-fly detection during the experiment. The effectiveness of fault detection, disturbance decoupling, and estimation is experimentally validated.