Prediction of the fatigue life of natural rubber composites by artificial neural network approaches

• An artificial neural network model was established to predict tensile fatigue life of natural rubber composites.
• A sensitive analytical model was founded to deduce the most critical factor on fatigue property of NR composites.
• The average prediction accuracy of artificial neural network is 97.3%.
• Stress at 100% is the most important factor affecting fatigue life.

A back-propagation artificial neural network (BP-ANN) model was established to predict fatigue property of natural rubber (NR) composites. The mechanical properties (stress at 100%, tensile strength, elongation at break) and viscoelasticity property (tan δ at 7% strain) of natural rubber composites were utilized as the input vectors while fatigue property (tensile fatigue life) as the output vector of the BP-ANN. The average prediction accuracy of the established ANN was 97.3%. Moreover, the sensitivity matrixes of the input vectors were calculated to analyze the varied affecting degrees of mechanical properties and viscoelasticity on fatigue property. Sensitivity analysis indicated that stress at 100% is the most important factor, and tan δ at 7% strain, elongation at break almost the same affecting degree on fatigue life, while tensile strength contributes least.