Analysis of Heat Treat Growth on Carburized Ring Gear and Multivariate Regression Model Development

Accurate assessment of heat treat (HT) growth on carburized ring gears is of critical importance when developing new gears or implementing various design/process changes on current production gears. Traditional approaches include conducting expensive and time consuming HT trials with green and after-HT measurements. An extensive database was created in order to develop a predictive model. Various statistical analyses were performed in Minitab. Ring gear HT growth on measurements between pins expressed in % growth was found to give better predicting power than delta (mm) growth. The best subset model with green hardness data utilizes 8 factors (material, key geometrical features) and yields 98.8% R². The model developed from a larger dataset without green hardness yields 92.4% R². On-going work includes continuous updating the database and refining the model. Application of these models will minimize the number of trials needed for new product launches and shorten the development cycle.