基于神经网络的钢筋混凝土碳化深度预测研究

Forecast Research on Carbonation Depth of Reinforced Concrete Based on Neural Network

碳化使混凝土的内部组成及结构发生变化,直接影响混凝土结构的性质及耐久性。钢筋混凝土的碳化,会破坏钢筋钝化膜,导致钢筋发生锈蚀,降低钢筋混凝土结构件的耐久性,从而影响到建筑物安全。碳化深度的影响因素很多,实验结果证明,采用BP神经网络,建立碳化时间序列与深度的关系模型,使用MATLAB进行仿真,预测值误差正常。神经网络能对钢筋混凝土的碳化深度进行预测。

The carbonation changes the internal composition and structure of concrete, which affects directly the prop- erties and durability of concrete structure. The carbonation of reinforced concrete will destroy the rebar passive film, result in the corrosion of the rebar,, and reduce the durability of reinforced concrete structure to affect the safety of buildings. There are many factors which affect the carbonation depth. The results of experiments demonstrate that the relational mod- els of the carbonation time series and depth can be set up based on BP neural network and the errors of prediction fall into the expected range by means of the MATLAB simulation, the carbonation depth of reinforced concrete can be predicted with the neural network.