基于CS-SVR优化算法的结构健康监测数据修复研究  

Research on Structural Health Monitoring Data Restoration Based on CS-SVR Optimization Algorithm

作　　者：满建浩 刘才玮[1] 刘峰 苗吉军[1] 李翔 王甫来[2] MAN Jianhao;LIU Caiwei;LIU Feng;MIAO Jijun;LI Xiang;WANG Fulai(School of Civil Engineering,Qingdao University of Technological,Qingdao 266033,China;Qingdao International Airport Group Co.,Ltd.,Qingdao 266300,China)

机构地区：[1]青岛理工大学土木工程学院,青岛266033 [2]青岛国际机场集团有限公司,青岛266300

出　　处：《结构工程师》2022年第4期75-81,共7页Structural Engineers

摘　　要：健康监测对大跨度空间结构的安全运行具有重要意义,而对监测结果进行评估与预警时,数据的部分缺失将会直接影响到结果的可靠性。为了解决准确重建缺失数据的问题,本文建立了基于布谷鸟搜索算法优化支持向量机的数据插补模型。首先对监测数据进行预处理,获取监测数据的时间序列;然后利用支持向量机对监测数据进行建模,并用布谷鸟搜索算法优化支持向量机参数;最后,以深圳大运会主场馆的监测数据对模型进行工程验证且通过与不同插补方法的比较证明CS-SVR模型在健康监测数据修复方面的优越性。Structural health monitoring is of great significance in the safe operation of large-span spatial structures.When the monitoring results are evaluated and early warning,some missing data will directly affect the reliability of the results.In order to solve the problem of accurately reconstructing the missing data,this paper established a data interpolation model based on the cuckoo search algorithm to optimize the support vector machine.First,the monitoring data is preprocessed to obtain the time series of the monitoring data;then the support vector machine is used to model the monitoring data,and the cuckoo search algorithm is used to optimize the support vector machine parameters;finally,the model is modeled with the monitoring data of the main stadium of the Shenzhen Universiade Perform and compared with different interpolation methods to prove the superiority of the CS-SVR model in health monitoring data interpolation.

关 键 词：结构健康监测 数据缺失 布谷鸟搜索算法 支持向量机 

分 类 号：TU317[建筑科学—结构工程]