基于人工神经网络的钢筋混凝土受弯构件正截面抗裂预测

钢筋混凝土构件抗烈度预测受多种条件因素的影响,现有的方法由试验实测数据建立的数学模型误差较大,因此有必要寻求一种精度较高的方法进行抗烈度预测。通过实测试验数据,训练形成一个三层BP网络,其中78组数据作为学习样本,另外9组数据则作为测试样本,建立了人工神经网络预测钢筋混凝土正截面抗裂性能的方法,还对其他模型抗裂性能的计算值与实测值进行了比较。该方法预测值与试验值吻合良好。结果表明,提出的人工神经网络预测钢筋混凝土正截面抗裂度预测方法具有对直接参与训练的数据仿真效果好,整体预测精度高,与理论分析得出的结论基本一致,可用于受弯构件抗烈度预测。

Predicting for Cracking Resistance of Normal Section of Reinforced Concrete Flexural Members Based on Artificial Neural Network

Based on test results,a three-1ayer back-propagation network is trained and formed by using the experimental data,among which 78 groups are used for training samples,while the remaining 9 groups are used for testing samples.The method for predicting the cracking resistance of normal section of reinforced concrete flexural members is set up with the method of artificial neural network(ANN).Good agreement is reached between the predicted results and the test data.It is concluded that the ANN method is feasibl...