用神经网络理论进行纤维水泥砼的配合比设计

RESEARCH ON MIX RATIO DESIGN OF FIBER CEMENT CONCRETE IN BRIDGE DECK PAVEMENT WITH ARTIFICIAL NEURAL NETWORK THEORY

根据公路桥梁桥面铺装对纤维水泥砼的使用性能要求，将一批试验数据作为样本，通过神经网络对其进行学习和训练，得到了相应的神经网络权值矩阵和阈值向量．计算结果与实际试验结果的对比表明，采取本文方法可以提高纤维水泥砼配合比的设计效率．

A design method for the fiber cement concrete mix ratio in highway bridge deck pavements is presented by means of the artificial neural network theory. The weight matrices and bias vectors in accordance with input neurons and output ones can be obtained by learning and training a set of specimens, which come from the lab test. The calculating results in the paper have a good agreement with the test data and therefore the present method can be used efficiently in practice.