A disassembly-free method for evaluation of spiral bevel gear assembly

摘要: Abstract The paper presents a novel method for evaluation of assembly spiral bevel gears. examination the approaches to problem gear control diagnostics without disassembly has revealed that residual processes in form vibrations (or noise) are currently most suitable this end. According literature, contact pattern is complex parameter describing position. Therefore, task determine correlation between and vibrations. Although vibration signal contains great deal information, it also spectral structure interferences. For reason, proposed three variants which effect preliminary processing on results. In Variant 2, stage 1, subjected multichannel denoising using wavelet transform (WT), 3 – combination WT principal component analysis (PCA). This procedure does not occur 1. Next, we features order focus information crucial regarding objective study. Given lack unequivocal premises enabling selection optimum features, calculate twenty rank them finally select appropriate ones an algorithm. Diagnostic rules were created artificial neural networks. We investigated suitability network types: multilayer perceptron (MLP), radial basis function (RBF) support vector machine (SVM).