Elsevier

Procedia CIRP

Volume 75, 2018, Pages 155-160
Procedia CIRP

Meta-Model Based on Artificial Neural Networks for Tooth Root Stress Analysis of Micro-Gears

https://doi.org/10.1016/j.procir.2018.04.031Get rights and content
Under a Creative Commons license
open access

Abstract

Micro-transmissions, consisting of micro-gears with a module <200µm, are used in manifold industrial applications, e.g. the medical industry. Due to the technological limits of their manufacturing processes, micro-gears show large shape deviations compared to their size, which significantly influence their lifetime. Thus, for micro-gears a model has been developed to enable a prognosis of their lifetime based on areal measurements of the gear geometry, finite elements simulations as well as lifetime experiments. To significantly reduce the amount of experiments, existing prior knowledge is additionally used as input to the lifetime model by means of Bayesian statistics.

To enable a time-efficient application of the model for industrial series production, in this article the application of a machine learning approach based on artificial neural networks is investigated.

The uncertainty of the model is evaluated according to the principles of the Guide to the Expression of Uncertainty in Measurement (GUM).

Keywords

metrology
machine learning
uncertainty

Cited by (0)