Distributed fault detection for a class of large-scale systems with multiple incomplete measurements

This paper is concerned with the problem of distributed fault detection for a class of large-scale systems with multiple uncertainties in measurements and communications. As a divide et impera approach is used to overcome the scalability issues of a centralized implementation, the large-scale system being monitored is modelled as the interconnection of several subsystems. A local fault detector is formed for each subsystem based on the measured local state of the subsystem as well as the transmitted variables of neighboring measurements. Phenomena such as the sensor saturation, the signal quantization, and the packet dropouts are addressed, where a unified model is proposed to capture these issues. The goal is to design a set of consensus based fault detectors such that, for all unknown disturbance and uncertain information, the estimation errors between the global residuals and the faults are minimized. By using the Lyapunov stability theory and some stochastic system analysis, a sufficient condition for the existence of desired fault detectors is established and the fault detector gains are computed by solving an optimization problem. A case study on the interconnected continuous stirred-tank reactor (CSTR) systems is finally given to show the effectiveness of the proposed design.