DocumentCode :
2551731
Title :
Technical Method of Fault Diagnoses Based on Fractal Theory
Author :
Wang, Bingcheng ; Ren, Zhaohui
Author_Institution :
Shenzhen Univ., Shenzhen, China
fYear :
2012
fDate :
18-21 Oct. 2012
Firstpage :
285
Lastpage :
288
Abstract :
Four different fault signals are simulated and collected which are oil whip fault signal, rub-impact fault signal, rub-oil whip coupling fault signal and rub - oil whip - loose coupling fault signal In the experimental stage. Using fractal theory, fractal dimension of vibration signal of X direction and the Y direction are separately calculated and resultant dimension of these two directions is obtained. Then, resultant dimension and its dimension energy spectrum are used as characteristic quantity to achieve the feature extraction and recognition. Because resultant dimension is synthesis of the fractal dimension of two directions vibrates signal, it combines the vibration information of two directions. Therefore, it can reflect comprehensively and really mechanical system´s vibration characteristic, the reliability of failure diagnosis is greatly improved.
Keywords :
fault diagnosis; feature extraction; fractals; maintenance engineering; mechanical engineering computing; reliability; signal processing; vibrations; energy spectrum; failure diagnosis reliability; fault diagnoses; feature extraction; feature recognition; fractal theory; mechanical system; oil whip fault signal; rub-impact fault signal; technical method; two directions vibrates signal; vibration characteristic; whip-loose coupling fault signal; Couplings; Fault diagnosis; Feature extraction; Fractals; Rotors; Vibrations; Resultant dimension; fault diagnosis; fractal theory; general dimension energy spectrum;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Chaos-Fractals Theories and Applications (IWCFTA), 2012 Fifth International Workshop on
Conference_Location :
Dalian
Print_ISBN :
978-1-4673-2825-8
Type :
conf
DOI :
10.1109/IWCFTA.2012.66
Filename :
6383237
Link To Document :
بازگشت