| 基于异常识别和关联分析的桥梁数据复合诊断 |
| |
| 引用本文： | 梁栋,张宇峰,袁慎芳等.基于异常识别和关联分析的桥梁数据复合诊断[J].振动．测试与诊断,2012,32(3):402-407. |
| |
| 作者姓名： | 梁栋 张宇峰 袁慎芳等 |
| |
| 作者单位： | 1. 南京航空航天大学智能材料与结构航空科技重点实验室,南京,210016
2. 长大桥梁健康检测与诊断技术交通行业重点实验室,南京,211112 |
| |
| 基金项目： | 国家自然科学基金资助项目，江苏省自然科学基金资助项目，江苏省普通高校研究生科研创新计划资助项目 |
| |
| 摘 要： | 提出了一种基于异常识别和多传感器关联分析的桥梁数据复合诊断方法。该方法利用超球面一类支持向量机对传感器数据特征进行异常识别,通过ε-支持向量回归机对多传感器测量值进行位置关联分析,从而判定该传感器数据异常是由外部荷载还是传感器自身故障引起。通过江阴大桥主梁加速度传感器的相关测量数据,表明了该方法的有效性。
|
| 关 键 词： | 桥梁结构健康监测 异常识别 超球面一类支持向量机 关联分析 ε-支持向量回归机 |
Bridge Data Compound Diagnosis Based on Novelty Recognition and Correlation Analysis |
| |
| Liang Dong,Zhang Yufeng,Yuan Shenfang,Wu Jian.Bridge Data Compound Diagnosis Based on Novelty Recognition and Correlation Analysis[J].Journal of Vibration,Measurement & Diagnosis,2012,32(3):402-407. |
| |
| Authors: | Liang Dong Zhang Yufeng Yuan Shenfang Wu Jian |
| |
| Affiliation: | 1(1.The Aeronautic Key Laboratory of Smart Material and Structure,Nanjing University ofAeronautics and Astronautics Nanjing,210016,China)(2.Key Laboratory of Long-Span Bridge Health Inspection & DiagnosisTechnology of Ministry of Transport Nanjing,211112,China) |
| |
| Abstract: | The bridge data compound diagnosis method based on novelty recognition and multi-sensor correlation analysis is presented.First of all,the abnormal feature of the original signal is recognized by hyperspherical one-class support vector machine.Then,the multi-sensor position correlation analysis is achieved with ε-support vector regression machine to determine that the data novelty is from the events or sensor itself.The acceleration measurement data in Jiangyin bridge show that the method is effective. |
| |
| Keywords: | bridge structural health monitoring novelty recognition hyperspherical one-class support vector machine correlation analysis ε-support vector regression machine |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《振动．测试与诊断》浏览原始摘要信息 |
|
点击此处可从《振动．测试与诊断》下载全文 |
