Study On Fault Diagnosis By Mechanical Noises Based On Blind Source Separation

Based on the "Study on Faulty Diagnosis over Mechanical Noises Based on Blind Source Separation"(Nature Science Fund Project of Zhejiang Province, No. 5001004), this paper researches on techniques of mechanical noises monitoring and fault diagnosis. Theory of Blind Source Separation and its application in the field of faulty diagnosis are analyzed, and some simulations are carried through. This paper also studies the structure and its important modules of the fault Monitoring and Diagnosis System for large rotating machinery, which was developed by our research group.In the first chapter, the developmental actuality and its trends in the field of fault monitoring and diagnosis on mechanicals are expressed. Principles and actuality inside and outside country of noises diagnosis on mechanicals are analyzed. Some problems of this method are issued. Then, the concept and mathematic model of BSS are introduced aimed at these problems. Some applications and research works inside and outside China are set forth.In the second chapter, the characteristics of noises are introduced. Some basic methods of supervision and monitoring are issued. On the basis of some guideline about measurement, the method of construction of faulty monitoring system on mechanical noises is brought forward aiming at requirement of BSS.In the third chapter, the article emphasized on the one of two kinds of mature BSS arithmetic, that is, BBS arithmetic based on Minimum Mutual Information. The flow chart of this arithmetic are given, and some simulations on imitate signals are worked out. Then, the results of BSS are expressed.In the fourth chapter, the other kind of BSS arithmetic, which is based on Maximum Entropy, are discussed. These two kinds of arithmetic are compared. After then, several simulations and results are demonstrated.In the fifth chapter, some research works are done on the system of faulty diagnosis. Precepts of monitoring and faulty diagnosis system on large rotary equipments are figure out. Its hardware configuration and software modules are discussed. The sub-system of BSS on faulty diagnosis over noises is designed briefly.In the last chapter, all of the works in this dissertation are summed up. Some limits and improvement in application of BSS are figured out. The future research of BSS on faulty diagnosis over mechanical noises is prospected.