用神经网络方法预测新老混凝土的粘结强度

Prediction Model for Bonding Strength of New-Old Concrete Based on Neural Network(NN)

对新老混凝土粘结的微观机理进行了分析 ,认为新老混凝土粘结层呈叠层结构 .老混凝土的洇水、老混凝土粘结面上粗骨料周围所形成的“过渡层”及新混凝土收缩微裂缝是造成新老混凝土粘结强度降低的主要原因 .使用水泥净浆界面剂 ,可弥补上述的某些不足 ,从而使新老混凝土的粘结强度得到提高 ;采用BP神经网络方法对新老混凝土粘结劈拉强度及粘结抗折强度进行了预测 ,对预测模型的精度和预测值的合理性也给予了必要的分析 。

The micro-mechanism of new-old concrete bond is analyzed firstly, and the new-old concrete is considered as a kind of layered composite material. The water saturation degree in old concrete, the 'transition layer' of coarse aggregate on old concrete bond surface and the shrinkage micro-cracks of new concrete are the main factors which reduce new-old concrete bonding strength. Some shortcomings above can be overcome, the bonding strength can be increased by applying pure cement slurry on concrete bond face. At last, the split-tensile bonding strength and the flexure bonding strength are predicted with BP neural network. The precision of prediction model and the rationality of predicted results are analyzed. It is shown that the neural network is a perfect method to deal with the prediction of concrete bonding strength.