论文部分内容阅读
监测信息的数据失真识别是桥梁结构长期健康监测系统的核心问题之一。针对桥梁结构长期健康监测系统监测信息的海量数据中由于仪器设备失效或环境干扰而导致的数据失真问题,结合数据变化率和聚类变化等数学统计方法,提出基于数据变化率的识别算法。该方法主要是对比数据之间的结构特性,首先建立相应结构参数的结构特性数据库,然后训练出该类型参数的安全系数以及变化率阈值,并通过数据库的不断更新和扩大来提高安全系数和阈值的精确性,最后通过验算实时采集数据的变化率是否超过对应的样本阈值来判别该数据是否失真。通过对佛山平胜大桥长期健康监系统采集的主塔偏位和环境温度数据进行失真识别验证,表明了该算法在实际工程中运行的有效性和可靠性。
Data distortion identification of monitoring information is one of the core problems of long-term health monitoring system of bridge structure. According to the data distortion caused by the equipment failure or environmental interference in the massive data of monitoring information of bridge structure long-term health monitoring system, a recognition algorithm based on data rate of change is proposed based on mathematical statistics such as data change rate and cluster change. The method mainly compares the structural characteristics of the data. First, the structure characteristic database of the corresponding structural parameters is established, and then the safety factor and the rate of change threshold of the type parameters are trained. The safety factor and the threshold are improved through the continuous updating and expansion of the database The accuracy of the final by checking whether the rate of change of real-time acquisition data exceeds the corresponding sample threshold to determine whether the distortion of the data. Through the distortion identification verification of the main tower deviation and the ambient temperature data collected by the long-term health monitoring system of Pingsheng Bridge in Foshan, the validity and reliability of the algorithm in practical engineering are demonstrated.