基于RBF神经网络对体外预应力梁抗剪强度的建模与预测

Modeling and prediction of externally prestressed concrete beam shear strength based on Radial Basis Function neutral network

采用基于径向基(RBF)神经网络对体外预应力混凝土梁抗剪强度进行建模,预测开裂荷载和极限荷载,并将预测值、试验值、规范公式值进行比较.结果表明,神经网络计算方法是体外预应力混凝土梁抗剪强度研究中的一种很有潜力的方法.

Presents the model based on radial basic function neutral network to predict shear capacity of the externally prestressed concrete beams. The values predicted with the model are compared with the experimental ones and the calculated ones based on codes. The results prove that neurocomputing is a practical and promising tool for the analysis of shear strength of the externally prestressed concrete beams.