Search

In order to predict the long-term creep life of P91 steel, this work proposed a coupled model, combining the error-trained back-propagation artificial neural network (BP-ANN) and an improved θ model. Both the short-term rupture life obtained by creep experiments and the related creep data of National Institute for Materials Science (NIMS) database were used to validate the above model. In this model, creep parameters (temperature, stress) were used as the input and relative error of 5 θ model extrapolations as the output. Then, the prediction errors of BP-ANN method were introduced to correct the 5 θ model extrapolation results to achieve a more accurate prediction. Consequently, the long-term creep life of P91 steel was predicted. It demonstrated that with the proposed model, the prediction capability for long-term creep life of P91 steel with 200,000 h is much better than those of other θ methods, which has a prediction uncertainty less than 5%. • A model coupled the improved 5 θ model with BP-ANN was proposed. • The proposed model based on error-training predicted the long-term creep life of P91 steel successfully. • The prediction accuracy of proposed model was much better than those of the 4 θ and 5 θ models. [ABSTRACT FROM AUTHOR]

The reliable prediction of long-term creep life is of great importance in order to ensure the service safety of high-temperature components. In this work, the novel Wilshire-Cano-Stewart (WCS) relationships were introduced to model the creep behavior and predict the long-term creep life of P91 steel based on the short-term creep data. The accelerated creep experiments were firstly carried out at the temperature range of 580–620 °C, while the stresses varied from 135 to 200 MPa. According to the analyses of normalized-mean-squared error (NMSE) and mean logarithmic error (MLE), it demonstrated that the WCS model can express all the obtained creep data very well due to the average relative errors lower than 10%. Meanwhile, these values are much lower than those of traditional Time-Temperature-Parameter (TTP) models. In addition, the WCS model can provide the reasonable estimates of 200,000 h creep life with accelerated creep measurements lasting up to only 2000 h. • WCS model made full use of Wilshire and Sinh model respective advantages and eliminated known weaknesses. • Results of WCS model are better than the traditional L-M model in life prediction. • Results of WCS model are accurate enough in life prediction. [ABSTRACT FROM AUTHOR]