Aeroengine Gas Path Fault Diagnosis Based On Intelligent Technology

The fault diagnosis of aeroengine is an important research area in aeroengine industry. It means a great deal to aeroengine. It has become one research focus in home and abroad currently. This thesis is motivated by the problem of fault detection and diagnosis for aeroengine gas path components. The neural networks fault diagnosis system based on intelligent technology is developed by use of neural networks'characteristic of nonlinear mapping to impend over aeroengine which is a complicated nonlinear dynamic system.Firstly, this thesis has described the fault diagnosis of aeroengine and the technology of neural networks. Then, the neural networks fault diagnosis system has been constructed with BP networks. A method based on the BP networks for fault diagnosis of aeroengine over full the envelope, and a software model of fault diagnosis have been constructed using the Visual C++ language. It can detect the gas path fault on line.Secondly, rough set and particle swarm optimization have been introduced. A new method based on neighborhood rough set model and neural networks for fault diagnosis of aeroengine is presented in this thesis. And the method has tested and proved to be right by the experiment. At last, the BP neural network learning algorithm has been optimized using particle swarm optimizer. Simulation results show that the algorithm based on PSO is more efficient and accurate.