中文摘要：

基于人工神经网络原理,对20个CFRP筋夹板式锚具拉拔试验数据建立BP神经网络模型,分析锚固长度、螺栓总预紧力、螺栓数量、CFRP筋直径等参数对锚具极限抗拔力的影响。分析结果显示,训练好的神经网络模型可以较好地预测CFRP筋夹板式锚具的极限抗拔力。利用神经网络模型进行参数分析,表明锚具极限抗拔力在样本空间内随锚固长度、螺栓总预紧力、螺栓数量的增加而增加,与试验研究结果相符。本文可为CFRP筋夹板式锚具的应用及改进提供参考。

英文摘要：

Based on experimental results of 20 pull-out tests on clamping anchors for CFRP rods,a prediction model using artificial neural network is established to anal yze the influence of various parameters on bearing capacity of the clamping anch or.The parameters include the bonded length,the total pre-tightening forces a pplied to the bolts,the bolt number,and the diameter of CFRP rods.The predict ed values by the trained neural network agree well with the experimental values. Parametric study is performed and the analytical results reveal the effects of parameters on the bearing capacity of the anchor,which is verified by the exper imental results.This paper can be referenced by the application and improvement of clamping anchors for CFRP rods.