摘要
为了快速得到受火后RC(Reinforced Concrete)梁中钢筋的温度,减少中间环节,提出一种利用BP(Back Propagation)神经网络预测受火后RC梁钢筋温度的方法。考虑截面高度、截面宽度、混凝土保护层厚度和受火时间等参数的影响,应用有限元软件ABAQUS对钢筋进行受火后温度场模拟,应用数学分析软件MATLAB编写BP神经网络程序进行钢筋温度预测。程序结果表明,BP神经网络得到的预测值与有限元软件ABAQUS得到的期望值误差较小,结果较为准确,说明BP神经网络可作为预测钢筋温度的一种工具。
In order to quickly obtain the reinforcement temperature of Reinforced Concrete(RC) beam and reduce the intermediate link,a BP(Back Propagation) neural network is proposed to predict the reinforcement temperature of RC beam after fire.Consider the influence of parameters such as beam height,beam breadth,concrete cover thickness and fire exposure time.The finite element software ABAQUS was used to simulate the temperature field of reinforcement,and the MATLAB math software was used to write the BP neural network program to predict the reinforcement temperature.The program results show that the predicted values obtained by BP neural network and the expected value obtained by the finite element software ABAQUS are less error,and the results are more accurate,indicating that BP neural network can be used as a tool to predict the reinforcement temperature.
作者
蔡斌
许龙飞
郝丽妍
CAI Bin;XU Longfei;HAO Liyan(School of Civil Engineering,Jilin Jianzhu University,Changchun Jilin 130118,China)
出处
《北方建筑》
2019年第4期23-27,共5页
Northern Architecture
关键词
钢筋混凝土
火
BP神经网络
reinforced concrete
fire
BP neural networks
