基于Matlab神经网络工具箱 ,设计了一个用于结构钢焊条熔敷金属化学成分预测的神经网络模型。该模型采用前向神经网络 ,利用BP算法 ,建立了熔敷金属化学成分与力学性能之间的关系。用此神经网络 ,既可预测结构钢焊条熔敷金属化学成分 ,也为深入研究结构钢焊条化学成分提供了有利的工具。 Based on the Matlab neural network toolbox, a neural network model for the chemical composition prediction of deposited metal in structural steel electrodes was designed. The model uses forward neural network and BP algorithm to establish the relationship between chemical composition and mechanical properties of deposited metal. Using this neural network, it can not only predict the chemical composition of the deposited metal of the structural steel electrode, but also provide a useful tool for further studying the chemical composition of the structural steel electrode.