基于人工神经网络(BP)的混凝土抗裂性能指标预测

Prediction of concrete crack-resistant performance index based on artificial neural network(BP)

通过对人工神经元及BP网络的简要介绍,结合三峡工程大坝混凝土试验实测数据,应用人工神经网络的理论和方法,利用MATLAB语言编制了基于神经网络理论的混凝土抗裂指标预测程序,实现了对混凝土抗裂指标值的预测,证明了人工神经网络在混凝土抗裂指标预测方面的可行性与可靠性。

The paper briefly describes the artificial neural elements and BP network. Based on the measured data from test on concrete of Three - Gorges Dam, and applying the theory and method of artificial neural network, the prediction program of concrete crack - re- sistant index is prepared by use of MATLAB language, which realizes the prediction of concrete crack - resistant index value and proves the availability and reliability of artificial neural network in the prediction of concrete crack - resistant index.