摘要
为确定钢筋混凝土梁的受火损伤程度,采用布谷鸟搜索(CS)算法优化支持向量机(SVM),提出一种以受火时间为指标的火灾损伤识别方法.首先,建立适用于T型简支梁的火灾损伤识别方法,用T型简支梁数值模拟验证了该方法的有效性,通过与SVM识别结果对比发现,CS-SVM识别结果更加接近真实受火时间.然后,在简支梁火灾损伤识别算法的基础上,提出了适用于钢筋混凝土连续梁火灾损伤识别的逐级递减识别方法.对5跨连续梁进行了实例计算分析,验证了其准确性,该方法大大降低了识别样本量,更适用于实际工程.
In order to determine the degree of fire damage of reinforced concrete beams,cuckoo search(CS)algorithm was used to optimize support vector machine(SVM),and a fire damage identification method using fire time as an indicator was proposed.Firstly,a fire damage identification method suitable for T-shaped simply supported beams was established.The numerical simulation of the simply supported beams verified the effectiveness of the method.By comparing with the recognition results of the SVM algorithm,it was found that the CS-SVM recognition results were closer to the real fire time.Then,based on the fire damage identification algorithm of simply supported beam,a stepwise decreasing identification method suitable for fire damage identification of reinforced concrete continuous beam was proposed.The accuracy of 5-span continuous beam was verified by an example calculation and analysis.This method greatly reduces the identification sample size and is more suitable for practical engineering.
作者
宋苏萌
刘才玮
苗吉军
肖建庄
顾振健
SONG Sumeng;LIU Caiwei;MIAO Jijun;XIAO Jianzhuang;GU Zhenjian(School of Civil Engineering, Qingdao University of Technology, Qingdao 266033,China;Department of Structural Engineering, Tongji University, Shanghai 200092, China)
出处
《青岛理工大学学报》
CAS
2021年第2期19-26,85,共9页
Journal of Qingdao University of Technology
基金
国家自然科学基金资助项目(51608289)
中国博士后基金资助项目(2018M632640)
山东省博士后创新项目(2019057)
青岛市博士后应用研究资助项目(2018103)。
关键词
钢筋混凝土梁
布谷鸟搜索算法
支持向量机
动力测试
损伤识别
reinforced concrete beam
cuckoo search(CS)algorithm
support vector machine(SVM)
dynamic test
damage identification
