混凝土碳化深度的人工神经网络分析与预测

Analysis and Forecast for Carbonation Depth of Concrete by Artificial Neural Network

基于水灰比、碳化龄期、相对湿度以及抗压强度对混凝土碳化深度影响的试验研究成果,利用BP网络和RBF网络按内推与外推两种方式,对混凝土碳化深度进行分析与预测.通过两种网络的分析与预测性能比较,建议优先采用RBF网络进行内推预测,对于外推预测则只能采用BP网络.

Based on the experimental of the carbonation depth of concrete influenced by results the water to cement ratio, the age of carbonation, the relative humidity and the compressive strength of concrete, the analysis and forecast were carried out for the carbonation depth of concrete by BP network and RBF network with the models of inside forecast and outside forecast. After comparing the properties of the analysis and the forecast of the two kinds of networks, it is suggested that the RBF network is preferentially selected for inside forecast, and the BP network is only used for outside forecast.