Fault Diagnosis For Nonlinear Systems Based On Observers

With the advent of large-scale and high-powered production system, how to improve the safety performance of the control system effectively has become an urgent problem. Fault diagnosis is a technique which could detect troubles in time by processing information during system running, and take corresponding measures to ensure the safety and reliability of the system. Combining the theory of state observer, adaptive control, fuzzy logic, neural network and genetic algorithm, this paper aims to find a way to diagnose the fault of the nonlinear system which has the uncertainty factors based on observer. The main research work is as follows:Firstly, this paper introduces the research status of fault diagnosis technology, analyses the main methods of fault diagnosis at present and claims the fault diagnosis technology which based on observer in detail.Secondly, based on the unknown input observer, this paper designs a way to diagnose faults of nonlinear system which has disturbances and actuator failures. This design solve the problem by offering a method which could decouple unknown factors by limiting the parameter of the observer, and could designing the unknown-input and declining-dimension observer to make this diagnosis.Then, the intelligent algorithm is introduced to the fault diagnosis of control system and a new fault diagnosis method based on adaptive fuzzy observer is proposed. This method uses the adaptive fuzzy system to compensate the nonlinear higher order term and design the state observer of the system. Meanwhile, threshold logic is used to detect the fault of the system and increasing the use of the approximation function of RBF neural network to realize the line tracking of the fault.Finally, the genetic algorithm which has the function of parallel search is introduced into the adaptive fuzzy system, improving the membership function and the connection weight of the Gaussian function and putting forward the scheme to diagnose the faults of the adaptive system based on genetic optimization. The design and its performance of the fault diagnosis scheme are examined by a numerical example and the simulation results show that the method of fault diagnosis for nonlinear systems with uncertain factors is of great effect.