Diesel Engine Fault Diagnosis Based On Particle Swarm Optimization And Fuzzy Neural Network

The diesel engine is a widely used reciprocating motive power machine, which has lots ofparts and a complicated structure. The diesel engine is often used in complex working conditions,the possibility of failures is big. Fault diagnosis of diesel engine will bring great economic valueand social value. Because of the diesel engine fault features and limitition of various conditions,traditional method of diesel engine fault diagnosis is difficult to meet the requirements. In recentyears, with the development of computers and artificial intelligence, through the use of neuralnetworks, fuzzy logic and artificial intelligence expert system technology such as computerdiagnostics is an important direction of diesel engine fault diagnosis research. Therefore thisessay carries out profound study and research on feasibility of diesel engine fault diagnosis basedon PSO & FNN optimized neural network system.First of all, this essay describes the present state and various methods of diesel oil enginefault diagnosis, and learns the basic structure and working process of the diesel engine.According to the vibration mechanism of diesel oil engine, this essay determines the vibrationdata acquisition test program, and then, preprocesses the collected data. Processing the signalwith wavelet packet, the essay extracts the wavelet packet energy of the band as the eigenvalue ofthe fault diagnosis, which is used to determine the work status of the diesel engine.Secondly, Particle Swarm Optimization algorithm easily gets stuck at local optimal solutionand low convergence accuracy, Two Subpopulation Swarm PSO Algorithm with AdaptiveMutation(ATPSO) was proposes. The search range of the algorithm was extended through thetwo subpopulation swarm, which have different inertia weight. It also adopts the crossbreedingmechanism in genetic algorithm, and adaptive global optimum mutation to accelerate theoptimization convergence and improve the search capabilities of particles. According to the testof classic functions, this new algorithm performs very well. At he same time, Taking into account the complementary nature of the particle swarm optimization algorithm and fuzzy neural network,the improved particle swarm optimization algorithm is applied to the optimization of fuzzy neuralnetwork parameters.Finally, the essay uses the algorithm based on particle swarm optimization of fuzzy neuralnetwork to determine the fault of diesel engine. Bringing eigenvalue into the combined network,and determining the statues of diesel engine. The results show that the combined algorithmachieved a better performance.