摘要
钢筋混凝土(RC)柱构件是钢筋混凝土结构中的关键构件,柱构件发生破坏很可能会导致结构发生严重破坏.因此,对RC柱破坏形态判别方法的研究尤为重要.为了研究RC柱破坏形态的判别方法,需对柱的破坏形态进行分类.利用太平洋地震工程研究中心(PEER)的数据库,对RC柱构件的破坏形态进行分类.根据人工神经网络(ANN)的分析模型并结合Levenberg Marquardt反向传播(LMBP)训练算法,得出一种RC柱构件破坏形态基于人工智能的判别方法.将111组数据中的89组数据作为训练数据,22组数据作为测试数据,并建立了RC柱构件破坏形态的ANN分析模型.结果表明,所提出的基于人工智能方法在判别RC柱构件破坏形态方面比传统方法表现得更好.
The reinforced concrete(RC)column member was the key member in the reinforced concrete structure,and the damage of the column member was likely to lead to serious damage to the structure.Therefore,the study of the damage pattern discrimination method of the reinforced concrete column became significant.To research the discrimination method of damage patterns of reinforced concrete columns,it was necessary to classify the damage patterns of columns.Utilizing the database of the Pacific Earthquake Engineering Research Center(PEER),the damage patterns of column members were classified.An artificial intelligence-based discrimination method for damage patterns of reinforced concrete column members was derived based on the analytical model of artificial neural network(ANN)and combined with the Levenberg Marquardt backpropagation(LMBP)training algorithm.Eighty-nine data sets out of 111 data sets were used as training data and 22 data sets as test data,and the ANN analytical model for damage patterns of column members was established.The results showed that the proposed artificial intelligence-based method performed better than the traditional method in discriminating the damage patterns of column members.
作者
万海涛
张梦雅
赵铁军
任文娟
付小燕
WAN Haitao;ZHANG Mengya;ZHAO Tiejun;REN Wenjuan;FU Xiaoyan(School of Civil Engineering and Architecture,Henan University,Henan Kaifeng 475004,China)
出处
《河南大学学报(自然科学版)》
北大核心
2025年第2期226-235,共10页
Journal of Henan University:Natural Science
基金
河南省科技攻关计划项目(242102320044)
河南省自然科学基金(232300420341)。
关键词
钢筋混凝土柱
破坏形态
判别方法
人工智能
准确率
reinforced concrete column
failure pattern
discriminant method
artificial intelligence
accuracy rate
作者简介
万海涛(1979-),男,江西抚州人,副教授,博士.研究方向:工程结构抗震;通信作者:赵铁军,E-mail:tiejun_zhao@henu.edu.cn。
