基于神经网络的金属材料疲劳裂纹扩展规律的预测

人工神经网络是新型的复杂系统预测方法。本文针对金属疲劳裂纹扩展速率建立了 BP 神经网络,并以部分应力比 LD2锻造铝合金疲劳裂纹试验数据作为训练样本,训练建立好的 BP 神经网络；以另一部分应力比条件下的试验数据作为预测样本,验证训练好的 BP 神经网络的预测能力。仿真结果表明,BP 神经网络能够方便地获得不同应力比下的疲劳裂纹扩展速率,对训练样本和测试样本都具有良好的泛化能力。该方法充分利用了已有数据,减少了疲劳试验次数,具有工程应用价值。

Research on Prediction of Fatigue Crack Growth Rate of Metallic Materials based on Neural Network

Artificial neural network is a newly-developed method for predicting complex system.In this paper,BP neu-ral network was established to predict the fatigue crack growth rate of metal.Part of the fatigue crack growth experimental data under specific stress ratios from LD2 were used to train the BP neural network,while other data were used to test and verify the prediction validity of the established BP neural network.Results indicated that the neural network had strong fit-ting and generalization capability.So the neural network can be used for predicting the crack growth rate of different stress ratios based on the existing data,and cut down the experimental work,which is very promising for engineering application.