李争　李伟

摘要：针对土木工程监测方案中存在的人工成本高、检测精度较低和无法实时监控等问题，利用分布式光纤布里渊监测技术设计一套在线监测系统，实现土木工程结构的实时监控。利用小波变换对传感器采集的布里渊频移数据进行降噪处理，实现建筑损伤的精确识别;为验证本设计方案的可靠性，比较不同波基的算法的降噪效果。采用波基为N=3的小波滤波最大误差为4.21%，波基为N=5的小波濾波最大误差为3.38%;研究采用波基为db5的滤波方案，将其与孤立森林算法进行比较，得出采用db5小波滤波方案处理数据相对误差降低了3.8%。将光纤布里渊频移监测方案与拉压力传感器监测方案的检测数据进行比较，得出采用光纤布里渊频移监测方案其平均误差降低了5%以上。

关键词：土木工程结构;在线检测;布里渊频移;小波滤噪;建筑损伤识别

中图分类号：FU3

文献标识码：A文章编号：1001-5922（2022）06-0138-05

Research on online safety assessment and monitoring technology of civil engineering structure

Li Zheng Li Wei

（1.Zaozhuang Housing Development Center， Zaozhuang 277800， Shandong China;

2. School of Civil Engineering， Shandong University， Jinan 250061， China

）

Abstract：Aiming at the problems of high labor cost， low detection accuracy and inability to monitor in real time in the civil engineering monitoring scheme， this study uses distributed optical fiber Brillouin monitoring technology to design an online monitoring system， which can realize real-time monitoring of civil engineering structures. Wavelet transform is used to denoise the Brillouin frequency shift data collected by the sensor to achieve accurate identification of building damage. In order to verify the reliability of this design scheme， the noise reduction effects of algorithms with different wave bases are compared. The maximum error of wavelet filtering with wave base N=3 is 4.21%， and the maximum error of wavelet filtering with wave base N=5 is 3.38%; therefore， the filtering scheme with wave base db5 is adopted in this study. Compared with the isolated forest algorithm， it is concluded that the relative error of the data processed by the db5 wavelet filtering scheme is reduced by 3.8%. Comparing the detection data of optical fiber Brillouin frequency shift monitoring scheme with that of tension pressure sensor monitoring scheme， it is concluded that the average error of optical fiber Brillouin frequency shift monitoring scheme is reduced by more than 5%.

Key words：civil engineering structure; online detection; Brillouin frequency shift; wavelet denoising; building damage identification

随着全国范围内基础设施建设的逐步开展，对于土木工程设施的维护称为重中之重，需要一套对土木工程进行监控的方案。由于土木工程长期存在，该监控方案需要能够长时间提供稳定的监控能力，用于传统人工利用简易设备进行监测，这需要消耗大量人工成本且效率低下[1-2]。

由于传统建筑监测方案的不足，有学者提出一种利用有限元分析实现对建筑损伤进行识别，该方案通过将传感器采集到的数据进行降噪处理，然后进行模态参数识别，完成对……