基于DBSCAN-LERP-LSTM的桥梁静力水准垂直位移监测异常值检测与分析

doi:10.3969/j.issn.1674-0696.2025.08.04

重庆交通大学学报（自然科学版） ›› 2025, Vol. 44 ›› Issue (8): 25-32.DOI: 10.3969/j.issn.1674-0696.2025.08.04

• 桥梁与隧道工程 • 上一篇

基于DBSCAN-LERP-LSTM的桥梁静力水准垂直位移监测异常值检测与分析

潘国兵1,虞洪兵1,宿林2,张顺涛3,吴畏1

(1. 重庆交通大学智慧城市学院，重庆 400074；2. 四川省铁路建设有限公司，四川成都 500643； 3. 四川九一五工程勘察设计有限公司，四川眉山 620020)

收稿日期:2024-08-28 修回日期:2024-11-12 发布日期:2025-09-05
作者简介:潘国兵（1976—），男，四川眉山人，教授，博士，主要从事智能测绘方面的研究。E-mail：50030350@qq.com 通信作者：虞洪兵（1998—），男，安徽安庆人，硕士，主要从事自动化监测方面的研究。E-mail：1092831200@qq.com
基金资助:
国家自然科学基金项目（42074004）

Detection and Analysis of Anomalous Values of Static Level Instrument Vertical Displacement Monitoring of Bridges Based on DBSCAN-LERP-LSTM

PAN Guobing1，YU Hongbing1，SU Lin2，ZHANG Shuntao3，WU Wei1

(1. Smart Cities Institute，Chongqing Jiaotong University，Chongqing 400074， China； 2. Sichuan Railway Construction Co., Ltd.，Chengdu 500643， Sichuan， China； 3. Sichuan Province Bureau of Geology and Mineral Resources,Geological Team 915，Meishan 620020，Sichuan， China)

Received:2024-08-28 Revised:2024-11-12 Published:2025-09-05

摘要/Abstract

摘要： 针对桥梁沉降数据受环境变化和传感器故障影响而产生噪声的问题，提出了一种基于DBSCAN-LERP-LSTM的分析方法，以提高数据可靠性和分析准确性。以某高速公路斜拉桥2021—2023年的静力水准仪监测数据为例，先用DBSCAN算法，邻域半径ε为40，领域内最少点数M为20，剔除9.8%异常值并线性插值填补缺失值，再通过时间序列分解发现2022年底沉降值约-0.4 mm，最后构建LSTM模型并用PSO、SSA、ACO的3种方法优化参数。结果表明：PSO-LSTM模型最优，均方根误差（RMSE）为0.419，平均绝对误差（MAE）为0.337，平均绝对百分比误差（MAPE）为0.142%，为静力水准仪监测系统提供了有效的数据处理流程，对桥梁长期安全运营意义重大。

关键词: 桥梁工程；桥梁健康监测；DBSCAN模型；LSTM模型；参数优化

Abstract: To address the issue of noise of bridge settlement data caused by environmental changes and sensor failures, a DBSCAN-LERP-LSTM-based analysis method was proposed to enhance data reliability and analysis accuracy. Taking the 2021—2023 static level instrument monitoring data of a cable-stayed bridge on a highway as an example, the DBSCAN algorithm (ε=40, M=20) was firstly used to remove 9.8% of the outliers and fill in the missing values through linear interpolation. Then, it was found that the settlement value reached approximately -0.4 mm at the end of 2022 by time series decomposition. Finally, an LSTM model was constructed, and the parameters were optimized by three methods such as PSO, SSA and ACO. The results show that the PSO-LSTM model performs the best, with an RMSE of 0.419, an MAE of 0.337, and an MAPE of only 0.142%, which provides an effective data processing procedure for static level instrument monitoring systems and is of great significance for the long-term safe operation of bridges.

Key words: bridge engineering; bridge health monitoring; DBSCAN model; LSTM model; parameter optimization

中图分类号:

U442.3

潘国兵1,虞洪兵1,宿林2,张顺涛3,吴畏1. 基于DBSCAN-LERP-LSTM的桥梁静力水准垂直位移监测异常值检测与分析[J]. 重庆交通大学学报（自然科学版）, 2025, 44(8): 25-32.

PAN Guobing1，YU Hongbing1，SU Lin2，ZHANG Shuntao3，WU Wei1. Detection and Analysis of Anomalous Values of Static Level Instrument Vertical Displacement Monitoring of Bridges Based on DBSCAN-LERP-LSTM[J]. Journal of Chongqing Jiaotong University(Natural Science), 2025, 44(8): 25-32.

参考文献

［1］李静佳, 陶雄, 李东平. 桥梁健康监测与评估中的静力水准技术研究［J］. 测绘与空间地理信息, 2021, 44(4): 195-197.
LI Jingjia, TAO Xiong, LI Dongping. Research on static leveling technology in bridge health monitoring and evaluation［J］. Geomatics & Spatial Information Technology, 2021, 44(4): 195-197.
［2］贾军, 张涛. 高精度静力水准测量系统在桥梁变形监测中的应用［J］. 公路工程, 2018, 43(2): 172-176.
JIA Jun, ZHANG Tao. Application of high accuracy static level automatic system in bridge deformation monitoring［J］. Highway Engineering, 2018, 43(2): 172-176.
［3］钟国强,柳尚,徐润,等.考虑异常监测数据影响的桥梁拉索振动频率识别方法研究［J］.中南大学学报(自然科学版),2023,54(12):4870-4881.
ZHONG Guoqiang, LIU Shang, XU Run, et al. Research on bridge cable vibration frequency identification method considering the influence of abnormal monitoring data ［J］. Journal of Central South University (Science and Technology), 2023, 54(12): 4870-4881.
［4］ DESJARDINS S, LAU D. Advances in intelligent long-term vibration-based structural health-monitoring systems for bridges［J］. Advances in Structural Engineering, 2022, 25(7): 1413-1430.
［5］ ENTEZAMI A, SARMADI H, BEHKAMAL B, et al. On continuous health monitoring of bridges under serious environmental variability by an innovative multi-task unsupervised learning method［J］. Structure and Infrastructure Engineering, 2024, 20(12): 1975-1993.
［6］ CIVERA M, SIBILLE L, ZANOTTI FRAGONARA L, et al. A DBSCAN-based automated operational modal analysis algorithm for bridge monitoring［J］. Measurement, 2023, 208: 112451.
［7］李元梦, 李登华, 丁勇. 基于DBSCAN的大坝安全监测异常数据检测算法［J］. 水电能源科学, 2024, 42(1): 149-152.
LI Yuanmeng, LI Denghua, DING Yong. Abnormal data detection algorithm for dam safety monitoring based on DBSCAN［J］. Water Resources and Power, 2024, 42(1): 149-152.
［8］戴领, 李少林, 刘光彪, 等. 基于LV-DBSCAN算法的大坝安全监测数据异常检测［J］. 人民长江, 2024, 55(1): 236-241.
DAI Ling, LI Shaolin, LIU Guangbiao, et al. Detection of abnormal values in dam safety monitoring data based on LV-DBSCAN algorithm［J］. Yangtze River, 2024, 55(1): 236-241.
［9］汤凯.我国临空经济区对地区经济增长的影响——基于离散型空间溢出效应［J］.中国流通经济,2020,34(8):81-90.
TANG Kai. The impact of Chinas airport economic zones on regional economic growth-Based on discrete spatial spillover effects ［J］. Chinas Circulation Economy, 2020, 34(8): 81-90.
［10］唐娟, 秦放鸣, 唐莎. 中国经济高质量发展水平测度与差异分析［J］. 统计与决策, 2020, 36(15): 5-8.
TANG Juan, QIN Fangming, TANG Sha. Measurement and difference analysis of China’s high-quality economic development level［J］. Statistics & Decision, 2020, 36(15): 5-8.
［11］王晓玲,王成,王佳俊,等.大坝渗压混合预测的STL分解-集成学习模型［J］.水力发电学报,2024,43(9):106-123.
WANG Xiaoling, WANG Cheng, WANG Jiajun, et al. STL decomposition-ensemble learning model for hybrid prediction of dam seepage pressure ［J］. Journal of Hydroelectric Power, 2024, 43(9):106-123.
［12］ DOROUDI R, LAVASSANI S H H, SHAHROUZI M. Optimal tuning of three deep learning methods with signal processing and anomaly detection for multi-class damage detection of a large-scale bridge［J］. Structural Health Monitoring, 2024, 23(5): 3227-3252.
［13］唐菲菲,唐天俊,朱洪洲,等.结合注意力机制和Bi-LSTM的降雨型滑坡位移预测［J］.测绘通报,2022 (9):74-79.
TANG Feifei, TANG Tianjun, ZHU Hongzhou, et al. Rainfall-type landslide displacement prediction combined with attention mechanism and Bi-LSTM ［J］. Surveying and Mapping Bulletin, 2022 (9): 74-79.
［14］许江波,侯鑫敏,吴雄,等.基于MIC-XGBoost-LSTM模型的边坡位移预测研究［J］.中国公路学报,2024,37(10):38-48.
XU Jiangbo, HOU Xinmin, WU Xiong, et al. Research on slope displacement prediction based on MIC-XGBoost-LSTM model ［J］. China Journal of Highway and Transport, 2024, 37(10): 38-48.
［15］莫林海.基于静力水准和GNSS的多传感器数据融合地基沉降监测技术［D］.西安：长安大学,2021.
MO Linhai. Research on Foundation Settlement Monitoring Technology Based on Multi-sensor Data Fusion of Static Leveling and GNSS ［D］. Xian: Changan University, 2021.

[1]	房世龙，陈红，施小飞. 桥墩清水局部冲刷减速不冲防护技术试验研究[J]. 重庆交通大学学报（自然科学版）, 2016, 35(3): 71-77.
[2]	任辉，周亦唐，李睿. 双向流固耦合的圆柱式桥墩绕流的三维数值模拟[J]. 重庆交通大学学报（自然科学版）, 2013, 32(5): 926-929.
[3]	彭李立, 邹毅松, 车竞. 对最小二乘拟合法计算桥涵设计流量的改进[J]. 重庆交通大学学报（自然科学版）, 2008, 27(2): 185-187,212.
[4]	郝艳广，肖盛燮，周菊芳，范海军. 基于理想流的群围堰绕流解析理论[J]. 重庆交通大学学报（自然科学版）, 2007, 26(3): 133-135.
[5]	曹芳，肖盛燮. 洪水主流变异对桥梁结构的破坏分析及其防患[J]. 重庆交通大学学报（自然科学版）, 2006, 25(6): 27-30.
[6]	刘保军，陈明栋，杨忠超，杨胜发. 异形桥墩冲刷试验研究[J]. 重庆交通大学学报（自然科学版）, 2006, 25(6): 31-34.
[7]	胡江,董玉文,杨胜发. 新疆宽浅变迁河流调治构造物护脚抗冲试验研究[J]. 重庆交通大学学报（自然科学版）, 2004, 23(3): 118-122.
[8]	田伟平，李惠萍，武炎. 沿河公路的丁坝群冲刷防护试验研究[J]. 重庆交通大学学报（自然科学版）, 2003, 22(3): 111-115.
[9]	田伟平，李惠萍，高冬光. 弯道环流与沿河路基冲刷试验研究[J]. 重庆交通大学学报（自然科学版）, 2002, 21(3): 94-97.
[10]	杨斌，杨胜发. 重庆地维长江大桥对长江行洪能力影响研究[J]. 重庆交通大学学报（自然科学版）, 2002, 21(3): 98-102.
[11]	周玉利, 王亚玲, 罗娜. 桥长调查和桥长公式[J]. 重庆交通大学学报（自然科学版）, 1999, 18(3): 30-34.

基于DBSCAN-LERP-LSTM的桥梁静力水准垂直位移监测异常值检测与分析

Detection and Analysis of Anomalous Values of Static Level Instrument Vertical Displacement Monitoring of Bridges Based on DBSCAN-LERP-LSTM

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 11

编辑推荐

Metrics