基于神经网络的大体积混凝土温度预测与控制

Forecast and control on temperature of massive concrete based on neural network

大体积混凝土的水化热若不能及时散发 ,会产生很大的温度应力 ,导致出现温度裂缝。为了避免温度裂缝的产生 ,人们必须预测和控制大体积混凝土的温度形成。文章引用“虚厚度”的概念 ,解决了不同边界条件的处理问题 ,通过对各种影响因素的分析并结合 BP神经网络模型的特点提出了大体积混凝土温度预测模型。通过工程实例 ,用 MATL AB语言实现了对大体积混凝土温度的预测 。

The thermal cracking will appear for the sizeable thermal stress if the hydration heat of massive concrete does not emanate in time,so the forming of temperature of massive concrete should be forecasted and controlled to avoid the appearance of thermal cracking. In the paper,the concept of 'virtual thickness'is introduced to resolve the matter of treating the boundary condition,and based on the analysis of the factors that influence the temperature of the massive concrete,a temperature forecasting model is put forward according to the characteristic of artificial neural network model. An example project is presented and the forecast of the temperature of massive concrete is realized with MATLAB. The artificial neural network (ANN) method has been proved to be feasible and practical.