Distributed Robust Monitoring for Three-Tank Systems Based on Dynamic Average Consensus

The three-tank system (TTS) is indeed an extremely important benchmark system in the field of process control. The study of distributed monitoring of TTS is of great significance for the monitoring paradigm of large-scale, complex, and networked industrial processes. This paper proposes a robust distributed monitoring method based on an average consensus algorithm, aiming to address the challenges of accurate monitoring in interconnected systems caused by state coupling and unknown disturbances. Existing data-driven methods often face the difficulty of effectively decoupling state coupling when designing distributed monitoring systems; simultaneously, unavoidable local disturbances within the system can propagate through coupling relationships, further affecting monitoring accuracy. To address these issues, this paper first designs a distributed adaptive observer capable of identifying subsystem parameters and states online, thereby effectively handling state coupling. Based on this, a distributed adaptive residual generator based on a projection method is proposed to decouple and suppress disturbances, improving monitoring robustness. Finally, simulation experiments on a TTS verify the feasibility and effectiveness of the proposed method in complex coupling and disturbance environments.