Fault detection for discrete-time systems in a networked environment

Abstract

This article is concerned with the fault detection (FD) problem for a class of linear discrete-time systems in a networked environment. The plant and its controller are located at diverse places, which are connected over network media to form a networked control system (NCS). A set of Kronecker delta functions is employed to characterise the delay and missing phenomenon, simultaneously, in both the backward and forward channels. After properly augmenting the states of the plant and the FD filter, an H _∞ FD filter can be designed in solving a certain set of linear matrix inequalities (LMIs). The residual is generated based on the data that can be obtained exactly at the controller and FD filter side. To illustrate the proposed methodology, we apply it to the leak FD for a discrete model, which is generated from the linearisation and discretisation of a practical internet-based nonlinear three-tank system. The fault can be detected within 1 min after its occurrence.

Keywords:

• fault detection
• networked control systems
• random communication delay
• packet dropout
• three-tank system

Acknowledegments

This work was supported by the NSFC under Grants 60721003 & 60736026, and the 973 project under grant 2009CB320602.