Machine learning based nominal root stress calculation model for gears with a progressive curved path of contact

The study aims to investigate the possibility of employing machine learning models in design non-involute gears. Such a model would be useful for calculations non-standard gears, where there are no available guidelines. aim is create decision-support accompanying Finite Element Method (FEM) simulations, from which data training was collected. Multiple numerical prediction were tested, i.e. linear regression, Support Vector Machine, K-nearest neighbour, neural network, AdaBoost, and random forest. firstly validated with N-fold cross-validation. Further validation done new FEM simulations. results simulations good agreement. best-performing ones forest AdaBoost. Based on results, constructed calculating nominal root stress gears progressive curved path contact proposed. can used as an alternative determining real-time, able calculate different number teeth, widths, modules, paths contact, materials, loads. Therefore, many combinations gear geometries analysed most suitable chosen.