Distribution of Correlation Coefficients in the Helical Gear Deviation Network

Abstract

This study aims to develop a system for measuring and analyzing gear tooth surface deviations in hobbed gears and identifying manufacturing errors in hobbed gears that were the cause of the deviations. In our previous study, we proposed a method to create a gear network focusing on the correlations among each gear tooth helix data obtained by measuring a single gear. In this study, we analyzed the adjacency matrices composed of correlation coefficients generated by the tooth profile deviation data from hobbing simulation. Besides, we examined the effectiveness of this method to represent the tooth profile deviation using networks by clarifying the effect of mounting errors of hob cutters on matrix elements, which can be expressed as the weight distribution, i.e., correlation coefficient distribution. This approach makes it possible to estimate the hob mounting errors (eccentricity) from the correlation coefficient distribution of groups determined by the tooth number difference.