Fretting Fatigue Life Prediction for Aluminum Alloy Based on Particle-Swarm-Optimized Back Propagation Neural Network

Abstract

Fretting fatigue is a specific fatigue phenomenon. Due to the complex mechanisms and multitude of influencing factors, it is still hard to predict fretting fatigue life accurately, despite there being many works on this topic. This paper developed a particle-swarm-optimized back propagation neural network to predict the fretting fatigue life of aluminum alloys using the test data gathered from the published literature. A commonly used critical plane model, the Smith, Watson, and Topper criterion, was used as a contrast. The analysis result shows that the proposed fretting fatigue life prediction neural network model achieves a higher prediction accuracy compared to the traditional SWT model. Experimental validation demonstrates the effectiveness of the model in improving the accuracy of fretting fatigue life prediction. This research provides a new data-driven methodology for fretting fatigue life prediction.