双掺粉煤灰和矿渣混凝土强度的BP网络预测模型

双掺粉煤灰和矿渣混凝土的强度发展机理复杂,不能用传统的水灰比线性函数来预测,利用BP神经网络模型来预测其3、28和56d的抗压强度.结果表明:BP神经网络具有较强的非线性映射能力,预测结果比较理想,可以指导实际工程;早龄期的混凝土强度预测值与实测值之间的误差较大,随着粉煤灰和矿渣的二次水化反应逐渐充分,强度发展趋于规律化,预测误差相应变小.

Prediction model for compressive strength of concrete with binary fly ash and slag by BP neural network

Mechanism of strength development for concrete with binary fly ash and slag is complex;And the strength could not be calculated by traditional linear function of water/cement ratio.BP neural network was used to forecast the compressive strength for 3 days,28 days and 56 days. The results show that BP neural network had strong non-linear mapping ability,and its prediction was relatively ideal enough to guide practical construction.For the early age concrete with binary fly ash and slag,the difference of comp...