Many industrial processes, such as heat transfer and chemical diffusion reactions, are typical distributed parameter systems (DPSs) characterized by strong spatiotemporal (S-T) coupling. Any component within these systems may malfunction and result in significant safety risks. This article proposes a model-based framework for the collaborative diagnosis of S-T faults and sensor anomalies in DPSs. First, based on the reduced-order model obtained through the spectral method, two sets of observers are established for process faults and sensor anomalies, respectively. Fault detection and isolation (FDI) algorithms are developed by leveraging the characteristics of these two fault types. Next, using an unknown input observer (UIO), a cooperative fault estimation algorithm capable of handling the coexistence of both fault types is designed. The stability and convergence of the proposed method are ensured through the Lyapunov direct method. Finally, numerical simulations are conducted on a heat-transfer rod. The results demonstrate that the FDI algorithm can detect and isolate S-T faults and sensor anomalies effectively. Moreover, the root-mean-square error (RMSE) of the intensity estimation remains below 0.31, further verifying the effectiveness of the proposed collaborative diagnosis algorithm.
Wang et al. (Thu,) studied this question.