摘要
针对混凝土内部钢筋腐蚀程度判别难、精确度低等问题,提出了将改进粒子群算法(PSO)与BP神经网络结合起来,通过对钢筋锈蚀机理及其影响因素的分析,建立了以混凝土内部温度、湿度、pH值、Cl-浓度和腐蚀电位为输入,钢筋腐蚀率为输出的改进PSO-BP监测模型,并将实测输入数据与仿真结果进行了对比。结果表明,改进PSO-BP算法的收敛性与准确性均优于PSO-BP算法和BP算法。
In view of the difficulty in judging the corrosion degree of steel bars in concrete and the low accuracy,an improved particle swarm optimization algorithm(PSO)was combined with BP neural network.Based on the analysis of the corrosion mechanism of steel bars and its influencing factors,an improved PSO-BP monitoring model with the internal temperature,humidity,pH,chloride ion concentration and corrosion potential of the concrete as input and the corrosion rate of steel bar as the output was established,and the measured input data were compared with the simulation results.The experimental simulation comparison shows that the convergence and accuracy of improved PSO-BP algorithm are better than those of PSO-BP algorithm and BP algorithm.
作者
匡浩
俞阿龙
徐新元
KUANG Hao;YU A-long;XU Xin-yuan(College of Electrical Engineering and Control Science,Nanjing Tech University,211800,China;School of Physics and Electronic Electrical Engineering,Huaiyin Normal University,Huai′an 223300,China;School of International Education and Exchange,Changzhou University,213164,China)
出处
《混凝土与水泥制品》
北大核心
2020年第6期77-81,共5页
China Concrete and Cement Products
基金
江苏省高校自然科学研究重大项目(16KJA460003)。
作者简介
匡浩(1993-),男,硕士研究生。联系电话:18360920634;通讯作者:俞阿龙(1964-),男,教授、博士。
