| Vessel safety can be better guaranteed by the good condition of marine engine and the related equipments. Up to now, research as regarding to condition monitoring and fault diagnosis technique of marine engines has been paid widespread attention. New methods and equipments of diagnosis are coming forth. With regard to the methods based on monitor, process and analysis of vibration signals, its theory becomes more sophisticated, analysis and testing equipments are more perfect. The veracity and reliability of diagnosis are greatly improved. And these methods are applicable on board. The study on vibration monitoring and stimulant spectral identification of marine power equipment is part of the scientific and technological developing project supported by Shanghai Municipal Educational Commission. The main aspects contained in this thesis are as follows.1) The developing situation and tendency of machinery fault diagnosis based on vibration signal processing and related technologies are summarized and analyzed in this paper. Directions for the further investigation of marine machinery fault diagnosis are proposed upon this foundation.2) Some useful methods of vibration signal progessing applied to fault diagnosis and its realization in "MATLAB" Language are introduced systematically. On the basis of characteristics of vibration signals of diesel engines, theories and methods of time series analysis and modeling applied to small sample signal analysis are emphasized.3) A computer program named "System of vibration signals analysis and fault diagnosis of marine engines " was developed with "MATLAB" Language. On the basis of program construction, the operational processes of system are brought forward from the following aspects: normal methods of signal processing; time series analysis and modeling; neural network analysis and so on. Establishment of this program provides many facilities for vibration signal processing of diesel engines.4) The surface vibration signals of diesel engines were measured and sampled by modern advanced measuring technologies. Analysis and related fault diagnosis of operation conditions of diesel engines are carried out by means of differentmethods of technologies of vibration signal processing. Analysis indicates that vibration signals of marine engines can deftly reflect working conditions of diesel engines, the condition monitoring system of diesel engine based on vibration signal analysis can promote working efficiency of diesel engines, detect fault in time, and finally improve the safety of vessels.5) Time series analysis and neural network are integratedly applied to vibration diagnosis of diesel engines in this paper. Characteristic parameters of diesel engine faults are obtained from time series analysis. Different faults of diesel engines are recognized through related neural network, which was founded on the basis of the characteristic parameters. Veracity of diagnosis can be extremely improved by combining the advantages of the two methods. This paper takes the diagnosis of abnormal valves clearances and recognition of loading condition of diesel engines as examples, and discusses the realizing process of this method and the related conclusions.6) At last, the conclusions from this study are presented and emphasized, and the developing directions of vibration diagnosis technology are discussed. The main conclusions are reached as follows.(1)The responding vibration amplitude of engine cylinder covers, which is stimulated by the air pressure, will increase with the engine load. And the responding signal power is mainly concentrate in the band from 1.5kHz to 2.5kHz, and the power is more concentrated with the increasing of the diesel engine loads.(2)The responding vibration amplitude of engine cylinder cover, which is stimulated by the valve impact, will increase with the augment of the valve clearance. And the responding signal power is mainly concentrated in the band from 2kHz to 5kHz, and the power is more concentrated with the increasing of the vaive clearan... |
PDF Full Text Request