Evaluation of the effect of surface roughness parameters on fatigue of TC17 titanium alloy impeller using machine learning algorithm and finite element analysis

Abstract

The exploration of the fatigue life of TC17 titanium alloy impellers using machine learning has become a key approach in fatigue studies. Here, finite element simulation analysis of a TC17 titanium alloy impeller is employed to quantitatively describe its mechanical status under actual working conditions. Fatigue tests considering various surface roughness parameters are performed to reveal the failure mechanism of the matrix material and analyze the effect of Ra and Rz on the fatigue life. Finally, an ANN algorithm with backpropagation is employed as the fundamental algorithm. Through the integrated application of the two surface roughness parameters and maximum stress, the fatigue life of the TC17 impeller could be evaluated. The theoretically predicted fatigue life is closer to the experimental value when both Ra and Rz are applied as the input data. This fatigue life evaluation approach using an ANN algorithm provides a novel method for investigating fatigue failure in mechanical components, and the fatigue failure mechanism investigation also enrich the study of TC17 titanium alloy property.