Cascade System Active Fault Tolerant Control Design

The bringing forward and application of fault tolerant control technology has obvious operation significance to increase reliability and security of industry system. This paper introduces basic idea and method of active fault tolerant control, which is based on malfunction diagnosis, neural network, system identification and control law reconstruction, after discussion about application of process system fault tolerant control technology, and completes an active fault tolerant control system design of a cascaded control system aiming at sensor malfunction. First, the malfunction diagnosis unit of sensor signal is built based on wavelet analysis method. Then, model of the controlled object is identified using neural network and reconstruction rules of the fault tolerant control law of the cascaded system are deducted according to state feedback control law reconstruction theory. While operating, fault tolerant control system reconstructs system control law according to instructions from malfunction diagnosis unit so that the controlled system runs stably. Simulation shows the sufficient advantage of wavelet in analyzing unstable signal. The task of sensor malfunction diagnosis is well completed using this method which provides a good base for succedent work.. The stable running of the controlled system under the reconstructed control law is also indicated in the simulation curves. The feasibility of the design is proved by the similarity of the operating characteristic between the reconstructed system and the original one.