基于堆叠GRU神经网络的桥梁动应变预测

doi:10.19678/j.issn.1000-3428.0069879

计算机工程 ›› 2026, Vol. 52 ›› Issue (3): 441-450. doi: 10.19678/j.issn.1000-3428.0069879

基于堆叠GRU神经网络的桥梁动应变预测

刘啸宇¹^,²^,*(), 廖志芳², 谈遂¹^,³^,⁴, 余志武¹^,³^,⁴

1. 高速铁路建造技术国家工程研究中心, 湖南长沙 410075
2. 中南大学计算机学院, 湖南长沙 410075
3. 中国中铁股份有限公司, 北京 100039
4. 中南大学土木工程学院, 湖南长沙 410075

收稿日期:2024-05-21 修回日期:2024-09-01 出版日期:2026-03-15 发布日期:2026-03-10
通讯作者: 刘啸宇
作者简介:
刘啸宇(CCF学生会员), 男, 硕士, 主研方向为代码注释自动生成
廖志芳, 教授、博士
谈遂, 教授、博士
余志武, 教授、博士
基金资助:
中国中铁股份有限公司科技研究开发计划(2020-专项-02); 中国中铁股份有限公司科技研究开发计划(2021-专项-08); 中国国家铁路集团有限公司科技研究开发计划(L2023G007); 中国国家铁路集团有限公司科技研究开发计划(L2022G007)

Bridge Dynamic Strain Prediction Based on Stacked GRU Neural Network

LIU Xiaoyu¹^,²^,*(), LIAO Zhifang², TAN Sui¹^,³^,⁴, YU Zhiwu¹^,³^,⁴

1. National Engineering Research Center of High-speed Railway Construction Technology, Changsha 410075, Hunan, China
2. School of Computer Science and Engineering, Central South University, Changsha 410075, Hunan, China
3. China Railway Group Limited, Beijing 100039, China
4. School of Civil Engineering, Central South University, Changsha 410075, Hunan, China

Received:2024-05-21 Revised:2024-09-01 Online:2026-03-15 Published:2026-03-10
Contact: LIU Xiaoyu

摘要/Abstract

摘要：

桥梁作为重要的基础设施, 由于长期受到自然环境和日常负荷的影响, 桥梁结构可能面临严重的安全隐患, 因此对桥梁结构健康状态进行实时监测和预测变得尤为迫切。在现有研究中, 对于复杂的桥梁结构健康状态预测, 仍然存在预测值易错、稳定性差和实时性不足等问题。为了解决上述问题, 提出基于堆叠门控循环单元(GRU)编解码器的SGRUA(Stacked GRU with Attention and Auto-Cycle)模型, 通过更好地捕捉时间序列数据中的长期依赖关系和重要特征来提高预测的准确性和稳定性, 并利用较少的参数量来提高预测速度使得预测实时化。首先对真实桥梁监测数据进行缺失值填补、异常值检测与处理, 以保证数据满足时序预测的完整性和可用性要求; 随后利用SGRUA模型对桥梁动应变指数进行时序预测, 通过对比实验和消融实验证明了模型的有效性。实验结果表明, 与TSMixer时序预测模型相比, SGRUA模型在桥梁B数据集上的均方根误差(RMSE)、平均绝对误差(MAE)、平均绝对百分比误差(MAPE)和对称平均绝对百分比误差(SMAPE)指标下降了11.07%、11.02%、11.00%和10.96%, 为桥梁结构健康监测与预测领域提供了一种新的有效方法, 同时为其他类似结构的健康监测问题提供了有益的探索与启示。

关键词: 时序预测, 深度学习, 桥梁监测, 自注意力机制, 健康监测

Abstract:

As important infrastructures, bridges may face considerable safety hazards due to the long-term influence of the natural environment and daily loads. Therefore, the health status of bridge structures must be monitored and predicted in real time. In existing studies, issues such as easy errors in prediction errors, poor stability, and lack of real-time performance monitoring have been identified that inhibit the the prediction of health status of complex bridge structures. To resolve the aforementioned issues, this study proposes a Stacked Gated Recurrent Unit (GRU) with Attention and Auto-Cycle (SGRUA) model based on a stacked GRU encoder-decoder. It improves the accuracy and stability of prediction by better capturing long-term dependencies and important features in time-series data and uses a smaller number of parameters to increase prediction speed, making predictions in real time. First, missing values are filled, and outliers are detected and processed for actual bridge monitoring data to ensure that the data meet the integrity and availability requirements for time-series prediction. Subsequently, the SGRUA model is used to predict the bridge dynamic strain index in the time-series, and the effectiveness of the model is verified through comparative tests and ablation experiments. The experimental results show that, compared with the TSMixer time-series prediction model, the SGRUA reduces the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Symmetric Mean Absolute Percentage Error (SMAPE) indicators by 11.07%, 11.02%, 11.00%, and 10.96% on the Bridge B dataset. The SGRUA provides a new and effective method for bridge structure health monitoring and prediction. Additionally, it provides useful solutions for health monitoring problems of other similar structures.

Key words: time-series prediction, deep learning, bridge monitoring, self-attention mechanism, health monitoring

刘啸宇, 廖志芳, 谈遂, 余志武. 基于堆叠GRU神经网络的桥梁动应变预测[J]. 计算机工程, 2026, 52(3): 441-450.

LIU Xiaoyu, LIAO Zhifang, TAN Sui, YU Zhiwu. Bridge Dynamic Strain Prediction Based on Stacked GRU Neural Network[J]. Computer Engineering, 2026, 52(3): 441-450.

https://www.ecice06.com/CN/Y2026/V52/I3/441

图/表 10

图1 GRU内部结构

Fig.1 GRU internal structure

图2 缺失值填补前DCAM02传感器部分数据

Fig.2 Partial data of DCAM02 sensor before filling the missing values

图3 缺失值填补后DCAM02传感器部分数据

Fig.3 Partial data of DCAM02 sensor after filling the missing values

图4 异常值处理前后DCAM02传感器数据箱线图

Fig.4 DCAM02 sensor data box diagram before and after processing outliers

图5 特征热力图

Fig.5 Characteristic heat map

图6 SGRUA模型架构

Fig.6 Architecture of the SGRUA model

参考文献 31

1	KO J M , NI Y Q . Technology developments in structural health monitoring of large-scale bridges. Engineering Structures, 2005, 27 (12): 1715- 1725. doi: 10.1016/j.engstruct.2005.02.021
2	《中国公路学报》编辑部. 中国桥梁工程学术研究综述·2024. 中国公路学报, 2024, 37 (12): 1- 160. doi: 10.19721/j.cnki.1001-7372.2024.12.001
	Editorial Department of China Journal of Highway and Transport . Review on China's bridge engineering research: 2024. China Journal of Highway and Transport, 2024, 37 (12): 1- 160. doi: 10.19721/j.cnki.1001-7372.2024.12.001
3	宋航, 周凤, 熊伟. 基于自相关-变分对抗学习的物理系统异常检测. 计算机工程, 2024, 50 (12): 358- 366. doi: 10.19678/j.issn.1000-3428.0068358
	SONG H , ZHOU F , XIONG W . Anomaly detection of physical systems based on autocorrelation-variance adversarial learning. Computer Engineering, 2024, 50 (12): 358- 366. doi: 10.19678/j.issn.1000-3428.0068358
4	梁宏涛, 刘硕, 杜军威, 等. 深度学习应用于时序预测研究综述. 计算机科学与探索, 2023, 17 (6): 1285- 1300. doi: 10.3778/j.issn.1673-9418.2211108
	LIANG H T , LIU S , DU J W , et al. Review of deep learning applied to time series prediction. Journal of Frontiers of Computer Science and Technology, 2023, 17 (6): 1285- 1300. doi: 10.3778/j.issn.1673-9418.2211108
5	马朝忠, 朱建青, 韩松辉. 基于ARIMA模型的卫星钟差异常值探测的模型选择方法. 武汉大学学报(信息科学版), 2020, 45 (2): 167- 172. doi: 10.13203/j.whugis20180230
	MA C Z , ZHU J Q , HAN S H . Model selection method based on ARIMA model in outliers detection of satellite clock offset. Geomatics and Information Science of Wuhan University, 2020, 45 (2): 167- 172. doi: 10.13203/j.whugis20180230
6	张越, 臧海祥, 程礼临, 等. 基于自适应时序表征和多级注意力的超短期风电功率预测. 电力自动化设备, 2024, 44 (2): 117- 125. doi: 10.16081/j.epae.202306018
	ZHANG Y , ZANG H X , CHENG L L , et al. Ultra-short-term wind power forecasting based on adaptive time series representation and multi-level attention. Electric Power Automation Equipment, 2024, 44 (2): 117- 125. doi: 10.16081/j.epae.202306018
7	DHIMAN H S , DEB D , GUERRERO J M . Hybrid machine intelligent SVR variants for wind forecasting and ramp events. Renewable and Sustainable Energy Reviews, 2019, 108, 369- 379. doi: 10.1016/j.rser.2019.04.002
8	YE L , LI Y L , PEI M , et al. A novel integrated method for short-term wind power forecasting based on fluctuation clustering and history matching. Applied Energy, 2022, 327, 120131. doi: 10.1016/j.apenergy.2022.120131
9	杨京渝, 罗隆福, 阳同光, 等. 基于气象特征挖掘和改进深度学习模型的风电功率短期预测. 电力自动化设备, 2023, 43 (3): 110- 116. doi: 10.16081/j.epae.202208019
	YANG J Y , LUO L F , YANG T G , et al. Wind power short-term forecasting based on meteorological feature exploring and improved deep learning model. Electric Power Automation Equipment, 2023, 43 (3): 110- 116. doi: 10.16081/j.epae.202208019
10	林勋. 时间序列分析在建筑物变形监测中的应用[D]. 长春: 吉林大学, 2005.
	LIN X. Application of time series analysis in building deformation monitoring[D]. Changchun: Jilin University, 2005. (in Chinese)
11	邓永春, 徐跃, 徐丹丹, 等. 基于支持向量机的GNSS时间序列预测. 全球定位系统, 2019, 44 (2): 70- 75. doi: 10.13442/j.gnss.1008-9268.2019.02.010
	DENG Y C , XU Y , XU D D , et al. GNSS time series prediction based on support vector machine. GNSS World of China, 2019, 44 (2): 70- 75. doi: 10.13442/j.gnss.1008-9268.2019.02.010
12	冯建强, 宋昆仑. 基于CEEMDAN-LSTM的桥梁变形时间序列预测研究. 地理空间信息, 2023, 21 (7): 40- 43. doi: 10.3969/j.issn.1672-4623.2023.07.010
	FENG J Q , SONG K L . Research on time series prediction of bridge deformation based on CEEMDAN-LSTM. Geospatial Information, 2023, 21 (7): 40- 43. doi: 10.3969/j.issn.1672-4623.2023.07.010
13	刘曦, 乔升访, 陈航, 等. 基于LSTM神经网络的桥梁监测数据预测方法. 广州建筑, 2022, 50 (4): 33- 38.
	LIU X , QIAO S F , CHEN H , et al. Prediction method of bridge monitoring data based on LSTM neural network. Guangzhou Architecture, 2022, 50 (4): 33- 38.
14	李世佳, 侯丽娟, 汤斌, 等. 基于ICEEMDAN-BiLSTM-ARIMA组合模型的桥梁健康监测数据预测模型. 计算机与现代化, 2023 (7): 36- 42.
	LI S J , HOU L J , TANG B , et al. Bridge health monitoring data prediction model based on ICEEMDAN-BiLSTM-ARIMA combined model. Computer and Modernization, 2023 (7): 36- 42.
15	王成菀. 基于时序数据分析的桥梁结构健康状态识别[D]. 重庆: 重庆大学, 2018.
	WANG C W. Identification of health status of bridge structure based on time series data analysis[D]. Chongqing: Chongqing University, 2018. (in Chinese)
16	罗宇, 罗林艳, 范嘉智, 等. 基于深度GRU神经网络的逐小时气温预报模型. 湖北农业科学, 2021, 60 (6): 119-122, 126. doi: 10.14088/j.cnki.issn0439-8114.2021.06.025
	LUO Y , LUO L Y , FAN J Z , et al. An hourly prediction model of air temperature based on deep GRU neural network. Hubei Agricultural Sciences, 2021, 60 (6): 119-122, 126. doi: 10.14088/j.cnki.issn0439-8114.2021.06.025
17	郝晴, 迟涛, 于政军, 等. 基于门控循环单元神经网络的浆液pH的预测建模. 科学技术与工程, 2023, 23 (18): 7824- 7830. doi: 10.3969/j.issn.1671-1815.2023.18.027
	HAO Q , CHI T , YU Z J , et al. Predictive modeling of slurry pH based on gated recirculating unit neural network. Science Technology and Engineering, 2023, 23 (18): 7824- 7830. doi: 10.3969/j.issn.1671-1815.2023.18.027
18	陈伟, 陈克松, 纪青春, 等. 基于1D-CNN+GRU的光伏阵列故障诊断方法研究. 自动化仪表, 2022, 43 (6): 13- 17. doi: 10.16086/j.cnki.issn1000-0380.2021110057
	CHEN W , CHEN K S , JI Q C , et al. Research on fault diagnosis method of photovoltaic array based on 1D-CNN+GRU. Process Automation Instrumentation, 2022, 43 (6): 13- 17. doi: 10.16086/j.cnki.issn1000-0380.2021110057
19	程艳, 孙欢, 陈豪迈, 等. 融合卷积神经网络与双向GRU的文本情感分析胶囊模型. 中文信息学报, 2021, 35 (5): 118- 129. doi: 10.3969/j.issn.1003-0077.2021.05.016
	CHENG Y , SUN H , CHEN H M , et al. Text sentiment analysis capsule model combining convolutional neural network and bidirectional GRU. Journal of Chinese Information Processing, 2021, 35 (5): 118- 129. doi: 10.3969/j.issn.1003-0077.2021.05.016
20	张艳君, 丁德平, 沈平, 等. 基于GRU神经网络的桥梁变形预测方法研究. 建筑结构, 2022, 52 (S2): 1938- 1943.
	ZHANG Y J , DING D P , SHEN P , et al. Research on bridge deformation prediction method based on GRU neural network. Building Structure, 2022, 52 (S2): 1938- 1943.
21	刘欢, 李富年, 颜永逸, 等. DBSCAN和GRU算法在桥梁监测系统的研究. 现代电子技术, 2022, 45 (20): 114- 118. doi: 10.16652/j.issn.1004-373x.2022.20.023
	LIU H , LI F N , YAN Y Y , et al. Research on DBSCAN and GRU algorithm in bridge monitoring system. Modern Electronics Technique, 2022, 45 (20): 114- 118. doi: 10.16652/j.issn.1004-373x.2022.20.023
22	SQUARTINI S, HUSSAIN A, PIAZZA F. Preprocessing based solution for the vanishing gradient problem in recurrent neural networks[C]//Proceedings of the 2003 International Symposium on Circuits and Systems. Washington D.C., USA: IEEE Press, 2003: 1-8.
23	GERS F A , SCHMIDHUBER J , CUMMINS F . Learning to forget: continual prediction with LSTM. Neural Computation, 2000, 12 (10): 2451- 2471. doi: 10.1162/089976600300015015
24	LIU J L , GONG X Q . Attention mechanism enhanced LSTM with residual architecture and its application for protein-protein interaction residue pairs prediction. BMC Bioinformatics, 2019, 20 (1): 609. doi: 10.1186/s12859-019-3199-1
25	CHO K, MERRIENBOER B V, GÜLÇEHRE Ç, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]//Proceedings of Conference on Empirical Methods in Natural Language Processing. Washington D.C., USA: IEEE Press, 2003: 1-10.
26	WEN Z P , ZHOU R L , SU H Z . MR and stacked GRUs neural network combined model and its application for deformation prediction of concrete dam. Expert Systems with Applications, 2022, 201, 117272. doi: 10.1016/j.eswa.2022.117272
27	ZHOU R J , QIAN H , ZHANG J L , et al. Self-attention mechanism enhanced user interests modeling for personalized recommendation services in cyber-physical-social systems. IEEE Transactions on Network Science and Engineering, 2022, 9 (6): 3835- 3846. doi: 10.1109/TNSE.2021.3093782
28	LU M R , XU X R . TRNN: an efficient time-series recurrent neural network for stock price prediction. Information Sciences, 2024, 657, 119951. doi: 10.1016/j.ins.2023.119951
29	ZHOU H Y, ZHANG S H, PENG J Q, et al. Informer: beyond efficient Transformer for long sequence time-series forecasting[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto, USA: AAAI Press, 2021: 11106-11115.
30	YU C Q, WANG F, SHAO Z Z, et al. DSformer: a double sampling Transformer for multivariate time series long-term prediction[C]//Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. New York, USA: ACM Press, 2023: 3062-3072.
31	EKAMBARAM V, JATI A, NGUYEN N, et al. TSMixer: lightweight MLP-mixer model for multivariate time series forecasting[C]//Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2023: 459-469.

[1]	曹继卫, 罗飞, 丁炜超. BS-YOLO: 基于BSAM注意力机制和SCConv的小目标检测算法[J]. 计算机工程, 2026, 52(3): 119-127.
[2]	张永宏, 孙书林, 龚蒙, 王俊飞, 马光义. 基于多尺度运动记忆模型的遥感云图预测方法[J]. 计算机工程, 2026, 52(3): 128-140.
[3]	苏建华, 池云仙, 许云峰, 高凯. 基于注意力模态融合的多模态意图识别[J]. 计算机工程, 2026, 52(3): 234-242.
[4]	张志, 尹昱凯, 孙奕灵, 孟雯锦, 彭畅. 基于多模态特征融合的Android恶意软件检测模型研究[J]. 计算机工程, 2026, 52(3): 243-254.
[5]	秦颖鑫, 张可佳, 潘海为, 巨亚昊. 计算机视觉对抗攻击研究综述[J]. 计算机工程, 2026, 52(2): 46-68.
[6]	赵旭东, 吴洪越, 孟柯, 许小龙, 窦万春. 服务推荐方法的研究进展与展望(特邀)[J]. 计算机工程, 2026, 52(1): 61-75.
[7]	陈亮, 赵英, 史晟辉, 尹琳. 基于超图神经网络的链路预测方法[J]. 计算机工程, 2026, 52(1): 136-143.
[8]	黄金贵, 刘朋, 唐文胜. MMD-YOLOv7:黑暗条件下车辆检测方法[J]. 计算机工程, 2025, 51(9): 340-349.
[9]	周晨阳, 刘雪宇, 梁少华, 吴永飞. 基于Swin Transformer的肾动脉血管检测分割与定量分析[J]. 计算机工程, 2025, 51(9): 252-267.
[10]	翟志鹏, 曹阳, 沈琴琴, 施佺. 基于多时空图融合与动态注意力的交通流预测[J]. 计算机工程, 2025, 51(9): 139-148.
[11]	徐瀅, 傅紫薇, 张伟, 陈云芳. 基于抽象语法树嵌入的智能合约漏洞检测技术[J]. 计算机工程, 2025, 51(9): 149-157.
[12]	武东辉, 王金凤, 仇森, 刘国志. 基于EWBiLSTM-ATT的数据手套手语识别[J]. 计算机工程, 2025, 51(8): 107-119.
[13]	林帆, 李建华. 基于多阶门控聚合网络的光学化学结构识别[J]. 计算机工程, 2025, 51(8): 364-372.
[14]	郝宏达, 罗健旭. 基于多尺度区域特征融合的多器官语义分割模型[J]. 计算机工程, 2025, 51(8): 270-280.
[15]	武东辉, 王金凤, 仇森, 刘国志. 基于EWBiLSTM-ATT的数据手套手语识别[J]. 计算机工程, 2025, 51(8): 107-119.

选择文件类型/文献管理软件名称

选择包含的内容

基于堆叠GRU神经网络的桥梁动应变预测

Bridge Dynamic Strain Prediction Based on Stacked GRU Neural Network

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 31

相关文章 15

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于堆叠GRU神经网络的桥梁动应变预测

Bridge Dynamic Strain Prediction Based on Stacked GRU Neural Network

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 31

相关文章 15

编辑推荐

Metrics

本文评价