珠江黄埔大桥模态频率连续监测中的温度影响Ⅱ:温度影响及建模分析  

Temperature Influence in Modal Frequency Continuous Monitoring of Huangpu Suspension Bridge Ⅱ on the Zhujiang River: Temperature Influence and Model Analysis

作　　者：王立新[1,2,3] 朱嘉健 姜慧[1,2,3] 赵贤任[1,2,3] 李雪艳[4] 

机构地区：[1]广东省地震局,广州510070 [2]中国地震局地震监测与减灾技术重点实验室,广州510070 [3]广东省地震预警与重大工程安全诊断重点实验室,广州510070 [4]暨南大学力学与土木工程系,广州510632

出　　处：《震灾防御技术》2016年第2期251-260,共10页Technology for Earthquake Disaster Prevention

基　　金：地震科技星火计划(XH16031);广东省重大科技专项(2012A080102008);国家自然科学青年基金(51208230);广东省地震预警与重大工程安全诊断重点实验室建设项目(2011A060901006)

摘　　要：本文在《珠江黄埔大桥模态频率连续监测中的温度影响Ⅰ:频率识别》的基础上,利用2013年10月至11月的识别频率和温度数据,针对环境温度对珠江黄埔大桥频率监测的影响问题进行研究。首先通过时程曲线对比以及线性拟合,对温度与识别频率进行相关性分析;然后分别采用多项式回归和支持向量机回归方法(SVR)进行建模分析,并利用拟合模型对测试样本进行频率预测;最后通过残差分析,评价了两种非线性模型的拟合效果。研究结果表明,大桥模态频率与环境温度之间存在明显的非线性的负相关关系,两种非线性模型均能很好地对频率与温度的关系进行模拟,得到了较好的预测结果。Based on ＂Temperature Influence in Modal Frequency Continuous Monitoring of Huangpu Suspension Bridge I on the Zhujiang river： Frequency Identification＂, we applied the monitoring data from October to November in 2013 to study the relationship between identified frequencies and temperature of Huangpu bridge on the Zhujiang river in this paper. The correlation analysis was conducted firstly. Then two nonlinear models, i.e. the polynomial and Support Vector Machine （SVM） models, were set up to simulate the relation between the identified frequencies and temperature. And the prediction qualities of these two models were estimated by the residual analysis. The results show that there is a clear negative nonlinear correlation between the identified frequencies and temperature. Both the polynomial and SVM models can predict frequency variations caused by temperature of Huangpu bridge on the Zhujiang river quite well.

关 键 词：珠江黄埔大桥 频率变化 温度影响 支持向量机方法 

分 类 号：U446[建筑科学—桥梁与隧道工程]