基于子结构模型降阶和数据驱动的杆塔结构损伤识别方法

邓茂; 严波; 高英博; 杨寒旭; 吕中宾; 张博; 刘光辉

doi:10.21656/1000-0887.450052

基于子结构模型降阶和数据驱动的杆塔结构损伤识别方法

doi: 10.21656/1000-0887.450052

邓茂^1,,
严波^1, ,,
高英博¹,
杨寒旭¹,
吕中宾²,
张博²,
刘光辉²

1.
重庆大学航空航天学院，重庆 400044
2.
河南省电力公司电力科学研究院，郑州 450052

(我刊编委严波来稿)

基金项目:

国家电网公司总部科技项目 5500-202324543A-3-2-ZN

详细信息

作者简介:
邓茂(1999—)，男，硕士生(E-mail: 1621534559@qq.com)

通讯作者:
严波(1965—)，男，教授，博士，博士生导师(通讯作者. E-mail: boyan@cqu.edu.cn)

中图分类号: O34
计量
- 文章访问数: 161
- HTML全文浏览量: 61
- PDF下载量: 78
- 被引次数: 0
出版历程
- 收稿日期: 2024-02-29
- 修回日期: 2024-03-19
- 刊出日期: 2024-07-01

A Damage Identification Method for Transmission Towers Based on Substructure Model Reduction and Data Driving

1.
College of Aerospace Engineering, Chongqing University, Chongqing 400044, P.R.China
2.
Henan Electric Power Research Institute, Zhengzhou 450052, P.R.China

(Contributed by YAN Bo, M.AMM Editorial Board)

摘要

摘要: 针对受静载作用的输电杆塔大型复杂结构，提出了一种基于子结构模型降阶和数据驱动的损伤回归识别方法. 根据杆塔框架结构特征及其在自重和覆冰静载作用下的变形特征划分子结构，确定结构可能出现的损伤状态，定义损伤指标. 采用子结构模型降阶方法对含损伤结构的有限元模型进行降阶，形成降阶模型库. 进一步，根据杆塔受载特征确定标定载荷，根据变形及破坏模式设计应变传感器布置方案，采用有限元方法计算降阶模型库中所有模型在标定载荷作用下的变形，构建数据集. 以传感器测点的应变数据作为输入，损伤指标作为输出，利用BP神经网络算法建立损伤回归识别模型，实现杆塔损伤位置识别和损伤指标预测，为杆塔结构健康状态实时监测技术开发奠定了基础.
- 子结构模型降阶 /
- 输电杆塔 /
- 数据驱动 /
- 损伤识别 /
- 回归模型
Abstract: A damage regression identification method for large and complex transmission tower structures subjected to static loads was proposed based on the substructure model reduction and data-driven method. According to the structural features of the transmission tower and its deformation under self-weight and ice loading, the full finite element model for the tower was reduced by means of the sub-structure method, the possible damage modes were predicted and the damage indexes defined. The substructure modeling method was used to reduce the orders of the structure with different damage states, and the order reduction model library was established. The calibration load was determined based on the loading characteristics of the tower, and the strain sensor layout was designed according to the deformation and failure modes. The deformations of all the reduced-order models under calibration loads were numerically simulated with the finite element method, and a dataset was then created. With the data measured by the strain sensors as input and the damage indexes as output, a damage regression identification model was built by the BP neural network algorithm. With the identification model, the damage locations can be recognized and the damage indexes can be quantified. This work lays a foundation for real-time health monitoring of transmission tower structures.
- substructure model reduction /
- transmission tower /
- data-driven method /
- damage identification /
- regression model
edited-by

edited-by
注释:

1) (我刊编委严波来稿)

HTML全文

图 1 数据驱动结构损伤识别流程图

Figure 1. The flow chart of data-driven structural damage identification

下载: 全尺寸图片幻灯片

图 2 BP神经网络算法示意图

Figure 2. The BP neural network algorithm diagram

下载: 全尺寸图片幻灯片

图 3 典型特高压直流塔线体系有限元模型

Figure 3. The finite element model for the typical UHV DC tower-line system

下载: 全尺寸图片幻灯片

图 4 典型覆冰载荷下杆塔的Mises应力分布和典型实际杆塔倒塌事故

Figure 4. The Mises stress distribution of the tower under typical ice load and the collapse mode of a real tower

下载: 全尺寸图片幻灯片

图 5 损伤位置确定及应变传感器布置方案

Figure 5. Damage locations and the layout of strain sensors

下载: 全尺寸图片幻灯片

图 6 杆塔子结构划分及外部结点选择

Figure 6. Substructure division of the tower and selection of external nodes

下载: 全尺寸图片幻灯片

图 7 自重作用下杆塔原型和降阶模型的变形和应力

Figure 7. Deformations and stresses of prototype and the reduced order model for the tower under self-weight

下载: 全尺寸图片幻灯片

图 8 BP神经网络训练过程

Figure 8. The BP neural network training process

下载: 全尺寸图片幻灯片

图 9 测试集样本中不同损伤位置的损伤指标预测结果

Figure 9. Prediction results of damage indexes at different damage locations of the test dataset

下载: 全尺寸图片幻灯片

图 10 附加模型库样本中不同损伤位置的损伤指标预测结果

Figure 10. Prediction results of damage indexes at different damage locations of additional reduced-oreder models

下载: 全尺寸图片幻灯片

表 1 导线和地线物理参数

Table 1. Physical parameters of conductors and ground wires

model	Young’s modulus E/MPa	cross-sectional area A/mm²	mass m/(kg·m^-1)	diameter d/mm
conductor(JL/G1A-400/35)	65 400	973	3.071 2	40.6
ground wire(JLB20A-240)	147 200	238.76	1.595 5	20.01

下载: 导出CSV

表 2 线路结构参数和覆冰厚度取值

Table 2. Structural parameters of the transmission line and the ice thickness

span L/m	elevation ΔH/m	ice thickness h/mm
300, 400, 500	0, 25, 50	0, 10, 20, 30, 40, 50, 60

下载: 导出CSV

表 3 原模型和降阶模型所有结点位移与单元应力相对误差

Table 3. Relative errors of node displacements and element stresses in the prototype and reduced-order models

error type	relative error δ/%
error type	X-shift	Y-shift	Z-shift	axial normal stress
average error	0.478	0.516	0.462	1.025
maximum error	0.643	0.692	0.591	2.649

下载: 导出CSV

表 4 作用于杆塔上导地线挂点的标定载荷取值范围

Table 4. Calibration loads on hanging points of conductors and ground wires with the tower

hanging point	calibration loads F/N
hanging point	X-direction	Y-direction	Z-direction
1，2	0	[-273 946, -107 164]	[-346 193, -134 598]
3，4	[67 299, 173 096]	[-273 946, -107 164]	[116 565, 299 812]
5，6	0	[-115 348, -58 162]	[-145 770, -73 051]
7，8	[36 526, 72 885]	[-115 348, -58 162]	[63 264, 126 240]

下载: 导出CSV

表 5 典型损伤状态下回归和分类识别方法对比

Table 5. Comparison of regression and classification recognition methods

damage combination	realistic damage index δ_r/%				regression prediction of damage indicator δ_rp/%				classification prediction of damage indicator δ_cp/%
damage combination	①	②	③	④	①	②	③	④	①	②	③	④
C1	15	0	0	0	15.3	0.17	0.06	0.09	15	0	0	0
C2	0	17	0	0	0.05	16.9	0.08	0.03	0	15	0	0
C3	10	0	20	0	10.6	0	19.7	0.21	10	0	20	0
C4	0	6	13	0	0	6.32	13.2	0.16	0	10	10	0
C5	0	18	0	23	0.18	18.8	0.05	22.6	0	15	0	25
C6	5	0	5	0	5.26	0.15	4.49	0.04	0	0	0	0
C7	6	21	0	0	6.24	21.2	0.12	0.08	10	20	0	0
C8	27	0	0	11	25.3	0.26	0	10.91	20	0	0	15

下载: 导出CSV

参考文献(27)

[1]	李惠, 鲍跃全, 李顺龙, 等. 结构健康监测数据科学与工程[J]. 工程力学, 2015, 32 (8): 1-7. https://www.cnki.com.cn/Article/CJFDTOTAL-GCLX201508002.htm LI Hui, BAO Yuequan, LI Shunlong, et al. Data science and engineering for structural health monitoring[J]. Engineering Mechanics, 2015, 32 (8): 1-7. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-GCLX201508002.htm
[2]	BAO Y, CHEN Z, WEI S, et al. The state of the art of data science and engineering in structural health monitoring[J]. Engineering, 2019, 5 (2): 234-242. doi: 10.1016/j.eng.2018.11.027
[3]	郑栋梁, 李中付, 华宏星. 结构早期损伤识别技术的现状和发展趋势[J]. 振动与冲击, 2002, 21 (2): 1-6. doi: 10.3969/j.issn.1000-3835.2002.02.001 ZHENG Dongliang, LI Zhongfu, HUA Hongxing. A summary review of structural initial damage identification methods[J]. Journal of Vibration and Shock, 2002, 21 (2): 1-6. (in Chinese) doi: 10.3969/j.issn.1000-3835.2002.02.001
[4]	LI H, SPENCER B F, BAO Y. Machine learning paradigm for structural health monitoring[J]. Structural Health Monitoring, 2021, 20 (4): 1353-1372. doi: 10.1177/1475921720972416
[5]	FARRAR C R, WORDEN K. Structural Health Monitoring: a Machine Learning Perspective[M]. John Wiley Sons, 2012.
[6]	YING Y, GARRETT J H, OPPENHEIM I J, et al. Toward data-driven structural health monitoring: application of machine learning and signal processing to damage detection[J]. Journal of Computing in Civil Engineering, 2013, 27 (6): 667-680. doi: 10.1061/(ASCE)CP.1943-5487.0000258
[7]	MANSON G, WORDEN K, ALLMAN D. Experimental validation of a structural health monitoring methodology, part Ⅰ: damage location on an aircraft wing[J]. Journal of Sound and Vibration, 2002, 259 (2): 323-343.
[8]	MANSON G, PAPATHEOU E, WORDEN K. Genetic optimization of a neural network damage diagnostic[J]. Aeronautical Journal, 2008, 112 (1131): 267-274. doi: 10.1017/S0001924000002219
[9]	HAKIM S J S, RAZAK H A, RAVANFAR S A. Fault diagnosis on beam-like structures from modal parameters using artificial neural networks[J]. Measurement, 2015, 76 : 45-61. doi: 10.1016/j.measurement.2015.08.021
[10]	SONY S, DUNPHY K, SADHU A, et al. A systematic review of convolutional neural network-based structural condition assessment techniques[J]. Engineering Structures, 2021, 226 : 111347. doi: 10.1016/j.engstruct.2020.111347
[11]	KHODABANDEHLOU H, PEKCAN G, FADALI S M. Vibration-based structural condition assessment using convolution neural networks[J]. Structural Control and Health Monitoring, 2019, 26 (2): e2308.
[12]	黄斌, 纵瑞芳, 杨涛. 基于静力测量数据的随机梁式结构损伤识别[J]. 计算力学学报, 2013, 30 (2): 180-186. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJG201302003.htm HUANG Bin, ZONG Ruifang, YANG Tao. Damage identification of random beam structures based on static measurement data[J]. Chinese Journal of Computational Mechanics, 2013, 30 (2): 180-186. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSJG201302003.htm
[13]	LAM H, YANG J. Bayesian structural damage detection of steel towers using measured modal parameters[J]. Earthquakes and Structures, 2015, 8 (4): 935-956. doi: 10.12989/eas.2015.8.4.935
[14]	QU W, SONG W, XIA Y, et al. Two-step method for instability damage detection in tower body of transmission structures[J]. Advances in Structural Engineering, 2013, 16 (1): 219-232. doi: 10.1260/1369-4332.16.1.219
[15]	KARAMI-MOHAMMADI R, MIRTAHERI M, SALKHORDEH M, et al. Vibration anatomy and damage detection in power transmission towers with limited sensors[J]. Sensors, 2020, 20 (6): 1731. doi: 10.3390/s20061731
[16]	滕辉, 康帅, 尹俊红, 等. 基于不同机器学习模型的框架结构损伤识别对比研究[J]. 四川建筑科学研究, 2023, 49 (3): 44-53. https://www.cnki.com.cn/Article/CJFDTOTAL-ACZJ202303005.htm TENG Hui, KANG Shuai, YIN Junhong, et al. Comparative study of frame structure damage identification based on different machine learning models[J]. Sichuan Building Science, 2023, 49 (3): 44-53. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ACZJ202303005.htm
[17]	KOUCHAKI M, SALKHORDEH M, MASHAYEKHI M, et al. Damage detection in power transmission towers using machine learning algorithms[J]. Structures, 2023, 56 : 104980. doi: 10.1016/j.istruc.2023.104980
[18]	魏佳恒, 郭惠勇. 基于贝叶斯优化BiLSTM模型的输电塔损伤识别[J]. 振动与冲击, 2023, 42 (1): 238-248. https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ202301028.htm WEI Jiaheng, GUO Huiyong. Damage identification of transmission tower based on BO-BiLSTM model[J]. Journal of Vibration and Shock, 2023, 42 (1): 238-248. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ202301028.htm
[19]	QUARTERONI A, MANZONI A, NEGRI F. Reduced Basis Methods for Partial Differential Equations: an Introduction[M]. Springer, 2015.
[20]	VALLAGHÉ S, HUYNH P, KNEZEVIC D J, et al. Component-based reduced basis for parametrized symmetric eigenproblems[J]. Advanced Modeling and Simulation in Engineering Sciences, 2015, 2 (1): 1-30. doi: 10.1186/s40323-014-0017-1
[21]	张珺, 李立州, 原梅妮. 径向基函数参数化翼型的气动力降阶模型优化[J]. 应用数学和力学, 2019, 40 (3): 250-258. doi: 10.21656/1000-0887.390187 ZHANG Jun, LI Lizhou, YUAN Meini. Optimization of RBF parameterized airfoils with the aerodynamic ROM[J]. Applied Mathematics and Mechanics, 2019, 40 (3): 250-258. (in Chinese) doi: 10.21656/1000-0887.390187
[22]	赖学方, 王晓龙, 聂玉峰. 基于Mori-Zwanzig格式和偏最小二乘的非线性模型降阶[J]. 应用数学和力学, 2021, 42 (6): 551-561. doi: 10.21656/1000-0887.410230 LAI Xuefang, WANG Xiaolong, NIE Yufeng. Nonlinear model reduction based on the Mori-Zwanzig scheme and partial least squares[J]. Applied Mathematics and Mechanics, 2021, 42 (6): 551-561. (in Chinese) doi: 10.21656/1000-0887.410230
[23]	罗振东, 张博. Sobolev方程基于POD的降阶外推差分算法[J]. 应用数学和力学, 2016, 37 (1): 107-116. doi: 10.3879/j.issn.1000-0887.2016.01.009 LUO Zhendong, ZHAO Bo. A reduced-order extrapolation finite difference algorithm based on the POD method for Sobolev equations[J]. Applied Mathematics and Mechanics, 2016, 37 (1): 107-116. (in Chinese) doi: 10.3879/j.issn.1000-0887.2016.01.009
[24]	GUYAN R J. Reduction of stiffness and mass matrices[J]. AIAA Journal, 1965, 3 (2): 380. doi: 10.2514/3.2874
[25]	BEATTIE C, GUGERCIN S. Model reduction by rational interpolation[M]//Model Reduction and Approximation: Theory and Algorithms. Philadelphia: Society for Industrial and Applied Mathematics, 2017.
[26]	HURTY W. Dynamic analysis of structural systems using component modes[J]. AIAA Journal, 1965, 3 (4): 678-685. doi: 10.2514/3.2947
[27]	刘营, 李鸿光, 李韵, 等. 基于子结构的参数化模型降阶方法[J]. 振动与冲击, 2020, 39 (16): 148-154. https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ202016021.htm LIU Ying, LI Hongguang, LI Yun, et al. A component-based parametric model order reduction method[J]. Journal of Vibration and Shock, 2020, 39 (16): 148-154. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDCJ202016021.htm