Research On Machinery Faults Diagnosis Technology Of Nuclear Power Equipment Based On HMM And PSO

According to nuclear power equipment is more redundancy and a complicated system, which only accumulated less data and fault samples in this paper. There is large amount of information embedded in the machinery equipments, and the non-stationary signals of this duration are always high instability and poor reappearance. In view of these characteristics, Particle Swarm Optimization (PSO) optimizing Hidden Markov Model (HMM), as a basic modeling and recognizing tool in this paper, introduces to implement online monitoring of the nuclear power equipment.The system is capable of adopted ensure and enhance the nuclear power equipment safety and reliability in a nuclear power station.The paper firstly studies the principium,common models and some related arithmetic of HMM,PSOand QPSO.Then, with the non-stationary signal traits of nuclear power equipment ,constructes models of the system fault diagnosis in nuclear power equipment based on HMM and QPSO (Quantum behaved Particle Swarm Optimization). Last, a series of collecting fault data and fault diagnosis methods verifies the feasibility and accuracy. The main contents in this paper are studied as follows:Chapter one introduces the current status in home and abroad and research significance ofmachinery fault diagnosis technology for nuclear power equipment. Analysises the feasibility of fault diagnosis in nuclear power equipment based on QPSO optimizing HMM. Finally, the research main contents, the total frame and innovations of this paper are presented.Chapter two discusses several commom feature extaction methods of non-stationary signals .According to vibration signal characteristics of nuclear power equipment. The study on a new Time-frequency analysis method for machinery fault diagnosis based on Optimization Lcoal Mean Decomposition (LMD) technology is applicated.Chapter three researchs the theory of Hidden Markov Model and Particle Swarm Optimization,introuces to improve the basic algorithm for Particle Swarm Optimization. Introduces practicle application of PSO algorithm.puts forward machinery fault diagnosis method for nuclear power equipment based on Particle Swarm Optimization algorithm optimizing Hidden Markov Model.Chapter four studies deeply Quantum-behaved Particle Swarm Optimization algorithm and characteristics, QPSO optimizing Continuous Gaussian mixture Hidden Markov Model (CGHMM), establishes corresponding fault dagnosis model for nuclear power equipment.Chapter five verifies various theories and methods are presented in front chapter, given concrete experimental plans and results, then, analysises it.Chapter six sums up the main achievements in this paper,and presented prospect.This research can promote the progress of on-line monitoring and technology of fault diagnosis, have important theoretical and practical value in guaranteeing the nuclear power system's safe and reliable operations and increasing the combat effectiveness of the national defence.