基于灰色神经网络的疲劳裂纹预测方法研究

Research on Prediction Method of Fatigue Crack Growth Based on the Gray Neural Network

摘要:

为对构件疲劳损伤进行预测,提出了基于灰色神经网络模型的疲劳裂纹扩展预测方法。将灰色GM(1,1)模型向BP网络映射,建立了一维灰色神经网络GNNM(1,1)模型。基于灰色GM(1,1)模型的发展系数和灰作用量给出了GNNM(1,1)模型初始权值。应用建立的GNNM(1,1)模型预测了某不锈钢构件腐蚀疲劳裂纹的扩展,并与GM(1,1)模型的预测结果进行了对比,表明GNNM(1,1)模型具有更高的预测精度和模型精度。

Abstract:

To predict the growth of fatigue crack of the structural component, the prediction method of the fatigue crack’s growth based on the gray neural network model was presented. The gray GM(1,1) model was mapped to the BP network, and the one-dimension gray neural network GNNM(1,1) model was established. Based on the grow coefficient and the gray actuating quantity of the gray GM(1,1) model, the initial weights of the one-dimension gray neural network GNNM(1,1) model were acquired. The GNNM(1,1) model was applied to predict the growth of the corrosion fatigue crack of the stainless steel test piece, and the prediction result was compared with the prediction result of gray GM(1,1) model. The comparative analysis shows that the one-dimension gray neural network GNNM(1,1) model is more accuracy for the growth prediction of the fatigue crack.