Research On Noise Sources Separation Technique Of Complex Mechanical System And Software Realization

The traditional and single measures of signal analysis and processing are difficult to obtain accurate quantitative results of the noise source separation, because there are many mechanical and electrical equipments inside the submarine so that the phenomena of coupling is very serious. Therefore, the method to quantify the noise source separation based on fuzzy reasoning, neural networks and other modern diagnostic techniques has become a focus. This paper will have a systematic research on quantifying the noise source separation using fuzzy chromatography diagnostic technique based on previous studies and practical research, including:(1) Feature ExtractionFeature extraction is an extremely important part of quantifying the noise source separation which will directly affects the result. Therefore, for the sake of forming effective knowledge of the steady-state characteristics and transient characteristics about submarine radiation noise, this article establishes a variety of information processing method in allusion to time-domain, frequency domain and time-frequency joint domain. It respectively contents the steady-state signal to curb environmental noise interference and transient signal to breakthrough short sequences high-resolution spectral analysis. It extracts submarine radiation noise and major machinery vibration noise and also establishes feature database which will do contribute for further analysis.(2) Establishment of Fuzzy Judgment MatrixThe establishment of fuzzy judgment matrix is the basic of fuzzy AHP diagnostic technique. Consequently, this paper forms a complete method to set up fuzzy judgment matrix by the formation of " feature extraction—establishing characteristic database—establishing membership function—establishing fuzzy judgment matrix"(3) The Study on Diagnostic Technique of Fuzzy AHPThis paper gives the concept of diagnostic technique of fuzzy AHP for submarine noise source through learning typical AHP, and not only establishes related diagnosis framework but also has in-depth research about the principle and process.(4) The Experimental Study on Separating Noise Source for Submarine ModelsFirstly, we do feature extraction to experimental data through time-domain, frequency domain and time-frequency joint domain based on the experimental results, establishing characteristic database and membership function. Secondly, We analyse the experimental results by the diagnostic technique of fuzzy AHP. Lastly, we do accurate assessment by procedure and operation method, discussing the effect of diagnostic technique of fuzzy hierarchy in practical application.(5) The Development of the Software about Separating the Noise SourceThe paper develops intelligent diagnosis system for vibration noise source of complex mechanical systems based on the research of quantifying the noise source separation in submarine. This system gives technical support for quantifying the noise source separation in submarine and lays a good foundation for noise source diagnosis in submarine based on the fuzzy theory as well.