Predictive Modelling for the Role of Material Properties and Operating Conditions in Fatigue Life Prediction using RMS-ANNs Approach

Material fatigue is now an essential aspect of modern industrial engineering as it has a significant impact on the strength and reliability of mechanical components in industrial systems. The study of fatigue mechanisms under different operating conditions is critical to achieving durable, long-lasting designs. This experimental study investigates the fatigue behaviour of locally fabricated materials subjected to rotational bending. The main objectives are to determine the fatigue limit for a given cycle and the endurance limit. In order to characterise the fatigue strength of the material, a group of identical specimens were subjected to cycles of a specific shape. To analyse the effect of three operational variables - hardness (H), surface roughness (Ra) and applied stress (σ) - on the life (number of cycles) of the specimens, experiments were carried out using a Taguchi L18 mixed design. In conclusion, the results were statistically analyzed using Response Surface Methodology (RSM), Analysis of Variance and Artificial Neural Network (ANN). The findings show that the applied stress is the most significant factor affecting specimen fatigue (lifetime), contributing 38.53%, followed by hardness. Additionally, the high value of the coefficient of determination, derived from the Taguchi method and the developed RSM, clearly demonstrates a strong correlation between the predicted and experimental data.