基于支持向量机的桥梁挠度传感器自适应修正方法研究  

Adaptive Correction Method of Bridge Deflection Sensor Based on Support Vector Machine

作　　者：胡顺仁[1,2] 熊文超[1] 包明[1] 

机构地区：[1]重庆理工大学电子信息与自动化学院,重庆市400054 [2]重庆大学光电技术及系统教育部重点实验室,重庆市400044

出　　处：《公路》2013年第8期1-6,共6页Highway

基　　金：国家科技部863项目;项目编号2006AA04Z433;国家科技部重大攻关资助项目;项目编号2002BA105C;重庆市科技攻关重大专项;项目编号7289;重庆市自然基金资助项目;项目编号2005BB6081

摘　　要：通过分析连通管式挠度测量原理,对挠度修正方法进行了改进。为避免现有方法需要对初始状态参数求取,引入了支持向量机模型以求得挠度传感器更换后的理论值,进而得到挠度偏差Δf。本方法不受温度及工况的限制,具有一定的自适应特性。多次实验结果显示现有方法和本方法的均方误差分别为0.785312和0.329756,说明本方法明显优于现有求均值的方法。An improved method for deflection correction is presented by anglicizing the principle of Connected Pipe Opto-electronic Deflection Measurement System. In order to avoid calculating the parame- ters of initial state while using the existing method, Support Vector Machine （SVM） is used to calculate the theoretic values of deflection after replaced, and then the deviation of deflection can also be calculated. This method has an adaptive feature and does not consider whether the temperature and working eondition is similar. The mean square errors of the existing method and the new method are 0. 785 312 and 0. 329 756 from numerous experiments, which show that the new method is obviously better than the existing method.

关 键 词：桥梁监测 自适应修正 支持向量机 挠度 研究 

分 类 号：U446.2[建筑科学—桥梁与隧道工程]