Based On The Fractal Theory And Support Vector Machine (SVM) Of Gearbox Fault Diagnosis Research

As an essential part of the transmission system parts, gearbox，s running status qualityconcerns significantly. However, during operation of the gearbox will be disturbed byenvironmental factors, so that the data collected by the sensor contain unwanted noise. Thispaper based on the full analysis of the current situation of the significance of gearbox faultdiagnosis and diagnostic methods, put forward a short time-division-dimensional filter asshort-time fractal fuzzy control parameters on real data mining filtered, and then multiplefractal spectrum algorithm to obtain the characteristic parameters of the fault, and finally theuse of support vector machine fault diagnosis and identification methods. By building agearbox failure experimental platform, validated this approach can be very good to make theright diagnosis of the gearbox failure diagnosis.This paper details the fractal theory and several fractal dimensions， then proposedgrid dimension as the short-term adaptive fuzzy control parameters, to combine the twoapproaches for signal noise reduction. This article describes the fully multifractal spectrumand its algorithm and then proposed an idea that using the a (q)and f (a (q))maxand boxdimensions as fault diagnosis characteristic parameters. And support vector machine forverification, fault identification works well. Using particle swarm optimization algorithm tooptimize the kernel function of support vector machines (SVM) and support vector machine(SVM) penalty factor. To compare the performance of before and after optimization, provedefficiency of the optimized is better. Enhance support vector machines for the generalizationability, the introduction of multi-core support vector machine learning, using a linearcombination of different kernel functions will be integrated. By contrast multicore supportvector machine not only proved better able to complete the diagnostic tasks, and more than single-core cases have a higher ability to promote.