基于神经网络的碳纤维布与混凝土极限粘结强度分析

PREDICTING THE ULTIMATE BOND STRENGTH OF CFRP BONDED TO CONCRETE USING ARTIFICIAL NEURAL NETWORKS

碳纤维布与混凝土的极限粘结强度问题属于高度非线性问题,难以建立精确的数学表达式进行分析。对基于拉出试验的极限粘结强度数据进行分析,建立了人工神经网络,对极限粘结强度进行仿真预测。神经网络的建立考虑了碳纤维布的厚度、宽度、粘结长度、弹性模量、抗拉强度和混凝土试块抗压强度、抗拉强度、宽度这8个参数,运用了118组试验数据对网络进行训练,对15组数据进行了预测分析。将神经网络计算结果同4种经验公式计算结果进行比较,其精度明显高于其他4种模型。结果表明,运用人工神经网络对碳纤维布与混凝土的极限粘结强度进行预测是可行的。

The application of artificial neural networks （ANNs） to predict the ultimate bond strengths of CFRP bonded to concrete is investigated. An ANN model is built, trained and tested using 118 sets of test data collected from the technical literatures. The data used in the ANN model are arranged in a format of 8 input parameters that cover the compressive strength, tensile strength and width of cube concrete; the thickness, the width, the bond length and the elastic modulus of the CFRP plate. 15 sets of experimental data are used for predicting. The results predicted by the ANN model are compared with those obtained by 4 empirical equations. The results show that ANN is a feasible tool for predicting the ultimate bond strengths of CFRP bonded to concrete.