Fault Diagnosis For Linear Time Delay System Based On Observer Approach

Fault detection and diagnosis(FDD) is receiving more and more attention due to the increasing demands for high safety and reliability of dynamic systems.The model-based FDD strategy is a hot research area. However,most of the achievements in the FDD field are for delay-free systems.Time delay is an inherent characteristic of many practical systems,such as chemical processes,paper manufacturing industry, traffic control systems and networked control systems.Furthermore,in many large lag systems,the large lag is replaced by small lag and pure time delay,which can be regarded as time delay systems.Time delay usually results in unsatisfactory performances and is frequently a source of instability.So the research of FDD theory for the time delay systems appears more important.The development of fault diagnosis techniques was introduced and diagnosis methods had been summarized.Some issues needed to be completed on the research of FDD theory for the time delay systems were pointed out,such as robust for the various uncertainties and the sensitivity to faults.A fault diagnosis method based on unknown input observer was proposed for the state-delay linear systems with fault.Under the conditions of time delay item decoupled,disturbance item decoupled and none decoupled,relative observer gain matrix and residual system were designed.A proper threshold was selected for faults diagnosis.For linear state-delay systems with modeling uncertainties and unknown input disturbance,the design of fault diagnosis robust observer was researched.The observer-based robust fault detection filter problem was translated into a model matching problem,so that the influence of uncertainty and disturbance was minimized to increase the robustness to unknown inputs,while the effect of fault was maximized to enhance the sensitivity to fault.Based on the H_∞optimization technology,a linear matrix inequality(LMI) approach to design FDD was proposed.The solvable condition of this optimization problem and the solutions were derived.A fault diagnosis approach based on an adaptive observer was proposed for state delay systems with both nonlinear perturbations and unknown inputs.The sufficient condition for the existence of observer and its design method were presented in terms of linear matrix inequality. Robust stability of the error system was proofed.Threshold judgment formula for fault detection was proposed with high sensitivity and zero false alarm rates.The proposed adaptive observer achieved two targets: first as a detection observer to detect fault,then as a diagnostic observer to estimate the timely change of fault curve.The method can diagnose both the abrupt fault and soft fault.An example was given to demonstrate the effectiveness of the proposed approach.Many simulation researches were done to illustrate the efficiency of the proposed methods.In the end,a brief conclusion of this thesis was given-and further research in this field was expected.