ARTIFICIAL NEURAL NETWORK APPROACH TO LOW CYCLE FATIGUE AND CREEP-FATIGUE INTERACTION LIFE PREDICTION OF MODIFIED 9Cr-1Mo FERRITIC STEEL

摘要

Low cycle fatigue (LCF) behaviour of normalized and tempered modified 9Cr-1Mo steel has been studied at various temperatures, strain amplitudes, strain rates and hold-times. The alloy in general showed a reduction in fatigue life with increase in temperature, increase in strain amplitude, decrease in strain rate and with an increase in the duration of hold time in tension. The capability of artificial neural network (ANN) approach to life prediction under LCF and creep-fatigue interaction conditions has been assessed by using the data from National Institute for Materials Science, Japan and that generated in our laboratory. It is demonstrated that the predictions are well within a factor of two.