基于粒子群优化SVR-ARMA组合模型频率预测  

Frequency Prediction Based on SVR-ARMA Combination Model Improved by Particle Swarm Optimization

作　　者：刘哲[1] 丁阳[1,2] 严加宝[1,2] LIU Zhe;DING Yang;YAN Jiabao(School of Civil Engineering,Tianjin University Tianjin,300072,China;Key Laboratory of Coast Civil Structure Safety of Ministry of Education,Tianjin University Tianjin,300072,China)

机构地区：[1]天津大学建筑工程学院,天津300072 [2]天津大学滨海土木工程结构与安全教育部重点实验室,天津300072

出　　处：《振动．测试与诊断》2020年第2期374-380,423,共8页Journal of Vibration,Measurement & Diagnosis

基　　金：国家重点基础研究发展计划(“九七三”计划)资助项目(2011CB013606)。

摘　　要：为实现环境激励下复杂钢结构的损伤预警,提出一种基于粒子群优化(particle swarm optimization,简称PSO)的支持向量回归(support vector regression,简称SVR)-时间序列(auto-regressive and moving average model,简称ARMA)组合模型用于频率预测,并结合均值控制图法将其用于复杂钢结构的损伤预警中。所提出频率预测模型的准确性和有效性采用潍坊市白浪河摩天轮钢结构实测数据进行验证。验证结果表明:与基本SVR模型、SVR-ARMA模型和PSO-SVR模型相比,所提模型具有更高的泛化能力和预测精度;在白浪河摩天轮钢结构的损伤预警中,基于粒子群优化的SVR-ARMA组合模型可检出由损伤造成模态频率轻微的异常变化,具有较强的损伤敏感性。研究成果可为环境激励下复杂钢结构的损伤预警提供参考。In order to realize damage warning of complex steel structures under environmental conditions,SVR-ARMA combinatorial model based on particle swarm optimization(PSO)is proposed.The damage warning method of complex steel structure is established by mean control chart and the proposed model.The accuracy of the prediction model is verified by the monitoring data of the White-wave River Ferris wheel inWeifang city.The results show that compared with the SVR model,SVR-ARMA model and PSO-SVR model,the PSO-SVR-ARMA model has higher generalization ability and prediction accuracy.In the structural damagewarning,the PSO-SVR-ARMA model can detect the slight abnormal changes of the modal frequency caused by the damage,which greatly improves the accuracy of the damage warning.The results can provide reference for damage warning of complex steel structures under environmental conditions.

关 键 词：粒子群优化 模态频率 支持向量回归-时间序列组合模型 结构损伤预警 

分 类 号：TU391[建筑科学—结构工程] TU12