PREDICTING STRESS CONCENTRATION ACTORS IN TENSIONLOADED SHAFTS USING ARTIFICIAL NEURAL NETWORKS

Abstract

This paper presents a novel approach to determining the stress concentration factor (Kt) for tension-loaded machine parts using artificial neural networks (ANNs). Analytical methods for calculating Kt rely heavily on empirical data and standardized charts, which are often limited to specific geometries and load conditions. To overcome these limitations, we trained an ANN model using a comprehensive dataset of empirical Kt values, covering a wide range of dimensions for tension-loaded shafts. The input parameters for the ANN model were the key geometric dimensions: the smaller diameter, the larger diameter, and the radius at the critical section of the shaft. By leveraging the ANN's capability to learn complex, non-linear relationships within the data, the model was able to accurately predict the stress concentration factor for any given set of input parameters. The results demonstrate that the ANN-based approach can serve as a reliable and efficient tool for engineers, reducing the reliance on timeconsuming finite element analyses or limited empirical charts and providing quick and accurate predictions of Kt across a wide range of applications.