神经网络在自密实混凝土流动性能中的应用

Application of Neural Network in Prediction for Flowing Property of Self-compacting Concrete

影响自密实混凝土流动性能的因素多而复杂。为此,本文在自密实混凝土配合比试验的基础上,通过建立神经网络中使用最广泛的BP网络模型,来预测自密实混凝土的流动性能。计算结果表明,由于神经网络方法综合考虑了影响自密实混凝土流动性能的各种因素(水胶比、水泥重量、粉煤灰、粗骨料和外加剂),因此具有较高的预测精度。

The factors that affect the flowing property of self-compacting concrete are much and complex. Based on the ratio tests of the self-compacting concrete, this paper establishes the BP neural network modal that is widely applied in all kinds of neural networks to predict the flowing property of the self-compacting concrete. The calculation results show that the neural network method has a good predicting accuracy because it considers the various influence factors（such as the water gel ratio, cement weight, fly ash, coarse aggregate and the admixture） of the flowing property of the self-compacting concrete.