基于人工神经网络的P91钢管硬度预测其剩余寿命的方法

Creep Rupture Life Estimation of P91 Steel Pipes Through Artificial Neural Network Based Hardness Prediction

采用人工神经网络对不同硬度区间的P91钢管进行了持久寿命评估,建立了相应的持久寿命评估模型,并通过不同硬度下的试验值对该模型进行了验证。结果表明:基于人工神经网络的持久寿命评估模型预测值与试验值较为吻合,误差大多控制在±5%以内;所建持久寿命评估模型可以快速无损评估P91钢管在某硬度下的持久寿命,不仅可以节约试验成本,还可以通过人工神经网络的自学习提高持久寿命预测的可靠性。

The creep rupture life of P91 steel pipes with different hardness was evaluated using artificial neural network(ANN) technology, while an assessment model was established for the creep rupture life, which was verified with the test data obtained under different hardness conditions. Results show that the predictive data of the creep rupture life assessment model based on artificial neural network are in good agreement with the experimental data, and most of the output errors can be controlled within ±5%. Based on this model, the creep rupture life of P91 steel pipes could be evaluated quickly and nondestructively under any hardness conditions. The model helps to not only save the test cost, but also improve the reliability of the creep rupture life prediction through self-learning of artificial neural network.