BOLT LOOSENING STATE MONITORING BASED ON INNER PRODUCT MATRIX AND CONVOLUTIONAL AUTOENCODER
-
摘要: 螺栓连接结构中的螺栓松动容易导致结构失效，如何对结构中的螺栓松动状态进行监测是当前研究的一个热点。该文利用环境激励下结构振动响应的相关性分析，结合深度学习技术，研究了一种联合使用内积矩阵(inner product matrix, IPM)和卷积自编码器(convolutional autoencoder, CAE)的神经网络模型，即基于内积矩阵及卷积自编码器(inner product matrix and convolutional autoencoder, IPM-CAE)的深度学习模型。通过对螺栓连接搭接板的螺栓松动状态监测的试验研究，验证了该方法的可行性及有效性，并与使用 IPM的卷积神经网络(convolutional neural network, CNN)、堆栈自动编码器(stack autoencoder, SAE)及胶囊网络(capsule network, CapsNet)相比，IPM-CAE方法具有较快的网络训练收敛速度和较高的识别精度。Abstract: Bolt loosening of bolted connection structures can easily lead to structural failure. How to monitor the loosening state of bolts of a structure is a hot spot of current research. This paper uses the correlation analysis of structural vibration responses under environmental excitation and the deep learning technology, and studies a neural network model that uses both inner product matrix (IPM) and deep convolutional autoencoder (CAE), named inner product matrix and convolutional autoencoder (IPM-CAE) based deep learning model. This paper verifies the feasibility and effectiveness of the method through an experimental study on the bolt loosening state monitoring of a bolt connected plate. Compared with the convolutional neural network (CNN), stack autoencoder (SAE) and capsule network (CapsNet) using IPM, the proposed IPM-CAE method shows better network training convergence speed and recognition accuracy.
-
图 5 1号螺栓健康状态和部分损伤状态响应图
Figure 5. Responses of No. 1 bolt health status and partial damage status
表 1 实验编号设置
Table 1. Experiment number setting
螺栓编号 状态编号 预紧扭矩/(N·m) 健康状态 1 1-5 5.0 健康 1-4.5 4.5 损伤 1-4 4.0 损伤 1-3.5 3.5 损伤 1-3 3.0 损伤 1-2.5 2.5 损伤 1-2 2.0 损伤 1-1.5 1.5 损伤 1-1 1.0 损伤 1-0.5 0.5 损伤 1-0 0.0 损伤 
下载: 导出CSV
表 2 CAE模型结构对监测结果的影响
Table 2. Effect of the structure of different CAE models on the detection results
名称 卷积层1 卷积层2 卷积层3 卷积核大小 n m 准确率 CAE_1 28 − − (3,3) 8192 1024 0.964 CAE_2 56 − − (3,3) 8192 1024 0.972 CAE_3 128 − − (3,3) 8192 1024 0.984 CAE_4 28 14 − (3,3) 8192 1024 0.974 CAE_5 56 28 − (3,3) 8192 1024 0.981 CAE_6 128 56 − (3,3) 8192 1024 0.990 CAE_7 28 14 8 (3,3) 8192 1024 0.981 CAE_8 56 28 8 (3,3) 8192 1024 0.988 CAE_9 128 56 8 (3,3) 8192 1024 0.987
下载: 导出CSV
表 3 最佳的CAE网络结构
Table 3. The best structure of CAE network
类型 输入 输出 深度 滤波器大小 步幅 激活函数 InputLayer (None, 8, 8, 1) (None, 8, 8, 1) None None None None Convolution (None, 8, 8, 1) (None, 8, 8, 128) 128 (3,3) 1 relu Convolution (None, 8, 8, 128) (None, 8, 8, 56) 56 (3,3) 1 relu MaxPooling2D (None, 8, 8, 56) (None, 4, 4, 56) None (2,2) 2 None Flatten (None, 4, 4, 56) (None, 896) None None None None Fully Connected (None, 896) (None, 32) None None None relu softmax (None, 32) (None, 11) None None None softmax
下载: 导出CSV
表 4 6个螺栓测试集准确率
Table 4. Accuracy of 6 bolts test sets
编号 测试集准确率 1 0.99 2 0.98 3 0.98 4 0.98 5 0.97 6 0.98
下载: 导出CSV
表 5 不同网络的对比结果
Table 5. Comparison results of different networks
网络名称 准确率 损失 SAE 0.97 0.07 Capsnet 0.90 0.02 CNN 0.96 0.09 CAE 0.99 0.05
下载: 导出CSV
表 6 不同采样点数和测点数量下的准确率对比
Table 6. Comparison of accuracy rates under different sample points and measurement points
采样点数 6测点 4测点 2测点 8192 0.955 0.910 0.531 9216 0.989 0.928 0.535 10240 0.993 0.936 0.542 11264 0.995 0.959 0.549
下载: 导出CSV
-
[1] 朱梦杰, 任亮, 李宏男, 等. 基于iBeam3单元逆有限元法的冻土区管道变形研究[J]. 工程力学, 2022, 39: 1. doi: 10.6052/j.issn.1000-4750.2021.05.0334ZHU Mengjie, REN Liang, LI Hongnan, et al. Research on pipeline deformation in frozen soil area based on iBeam3 element inverse finite element method [J]. Engineering Mechanics, 2022, 39: 1. (in Chinese) doi: 10.6052/j.issn.1000-4750.2021.05.0334 [2] 何定桥, 杨军. 基于HHT的结构模态参数自动化识别方法和试验验证[J]. 工程力学, 2021, 39: 1. doi: 10.6052/j.issn.1000-4750.2021.06.0466HE Dingqiao, YANG Jun. Automatic identification method and experimental verification of structural modal parameters based on HHT [J]. Engineering Mechanics, 2021, 39: 1. (in Chinese) doi: 10.6052/j.issn.1000-4750.2021.06.0466 [3] 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. doi: 10.6052/j.issn.1000-4750.2014.08.ST11LI Hui, BAO Yuequan, LI Shunlong, et al. Structural Health Monitoring Data Science and Engineering [J]. Engineering Mechanics, 2015, 32(8): 1 − 7. (in Chinese) doi: 10.6052/j.issn.1000-4750.2014.08.ST11 [4] 郭健, 顾正维, 孙炳楠, 等. 基于小波分析的桥梁健康监测方法[J]. 工程力学, 2006, 23(12): 129 − 135. doi: 10.3969/j.issn.1000-4750.2006.12.023GUO Jian, GU Zhengwei, SUN Bingnan, et al. Bridge health monitoring method based on wavelet analysis [J]. Engineering Mechanics, 2006, 23(12): 129 − 135. (in Chinese) doi: 10.3969/j.issn.1000-4750.2006.12.023 [5] 朱宏平, 璟余, 张俊兵. 结构损伤动力检测与健康监测研究现状与展望[J]. 工程力学, 2011, 28(2): 1 − 11. doi: 10.6052/j.issn.1000-4750.2010.06.ST03ZHU Hongping, JING Yu, ZHANG Junbing. Research status a nd prospects of structural damage dynamic detection and health monitoring [J]. Engineering Mechanics, 2011, 28(2): 1 − 11. (in Chinese) doi: 10.6052/j.issn.1000-4750.2010.06.ST03 [6] 戴玉婷, 严慧, 王林鹏. 基于非线性气动力的失速颤振计算与试验研究[J]. 工程力学, 2020, 37(8): 230 − 236. doi: 10.6052/j.issn.1000-4750.2019.03.0141DAI Yuting, YAN Hui, WANG Linpeng. Calculation and experimental study of stall flutter based on nonlinear aerodynamic force [J]. Engineering Mechanics, 2020, 37(8): 230 − 236. (in Chinese) doi: 10.6052/j.issn.1000-4750.2019.03.0141 [7] 秦国华, 林锋, 叶海潮, 等. 基于残余应力释放的航空结构件加工变形模型与结构优化方法[J]. 工程力学, 2018, 35(9): 214 − 222, 231. doi: 10.6052/j.issn.1000-4750.2017.06.0433QIN Guohua, LIN Feng, YE Haichao, et al. -Machining deformation model and structure optimization method of aviation structural parts based on residual stress release [J]. Engineering Mechanics, 2018, 35(9): 214 − 222, 231. (in Chinese) doi: 10.6052/j.issn.1000-4750.2017.06.0433 [8] 詹阳磊, 肖黎, 屈文忠. 温变工况下的EMI螺栓松动损伤识别[J]. 机械科学与技术, 2019, 38(6): 821 − 827.ZHAN Yanglei, XIAO Li, QU Wenzhong. EMI Bolt Loose Damage Identification under Temperature Variation Conditions [J]. Mechanical Science and Technology, 2019, 38(6): 821 − 827. (in Chinese) [9] 杜飞, 张子涵, 徐超. 法兰螺栓松动的超声导波监测方法[J]. 压电与声光, 2019, 41(5): 679 − 684.DU Fei, ZHANG Zihan, XU Chao. Ultrasonic Guided Wave Monitoring Method for Flange Bolt Looseness [J]. Piezoelectric and Acousto-Optic, 2019, 41(5): 679 − 684. (in Chinese) [10] NA W S. Bolt loosening detection using impedance based non-destructive method and probabilistic neural network technique with minimal training data [J]. Engineering Structures, 2021, 226: 111228. doi: 10.1016/j.engstruct.2020.111228 [11] 王刚, 肖黎, 屈文忠. Lamb波高斯混合模型螺栓松动损伤检测[J]. 机械科学与技术, 2020, 39(4): 493 − 500.WANG Gang, XIAO Li, QU Wenzhong. Lamb wave Gaussian mixture model bolt loosening damage detection [J]. Mechanical Science and Technology, 2020, 39(4): 493 − 500. (in Chinese) [12] TU Y, HUANG Y, TU S. Real-time monitoring of bolt clamping force at high temperatures using metal-packaged regenerated fiber Bragg grating sensors [J]. International Journal of Pressure Vessels and Piping, 2019, 172: 119 − 126. doi: 10.1016/j.ijpvp.2019.03.006 [13] 白鲁帅, 李钢, 靳永强, 等. 一种隔离损伤的桁架结构性态识别方法[J]. 工程力学, 2019, 36(1): 53 − 60. doi: 10.6052/j.issn.1000-4750.2017.09.0691BAI Lushuai, LI Gang, JIN Yongqiang, et al. A damage-isolated truss structure behavior identification method [J]. Engineering Mechanics, 2019, 36(1): 53 − 60. (in Chinese) doi: 10.6052/j.issn.1000-4750.2017.09.0691 [14] 赵林鑫, 江守燕, 杜成斌. 基于SBFEM和机器学习的薄板结构缺陷反演[J]. 工程力学, 2021, 38(6): 36 − 46. doi: 10.6052/j.issn.1000-4750.2020.06.0416ZHAO Linxin, JIANG Shouyan, DU Chengbin. Defect invers ion of thin plate structure based on SBFEM and machine learning [J]. Engineering Mechanics, 2021, 38(6): 36 − 46. (in Chinese) doi: 10.6052/j.issn.1000-4750.2020.06.0416 [15] 徐浩, 沙刚刚, 李腾腾, 等. 基于二维连续小波变换与数据融合技术的逆有限元法-伪激励法结构损伤识别方法[J]. 复合材料学报, 2021, 38(10): 3564 − 3572. doi: 10.13801/j.cnki.fhclxb.20210115.001XU Hao, SHA Ganggang, LI Tengteng, et al. Inverse finite element method-pseudo-excitation method based on two-dimensional continuous wavelet transform and data fusion technology for structural damage identification [J]. Journal of Composite Materials, 2021, 38(10): 3564 − 3572. (in Chinese) doi: 10.13801/j.cnki.fhclxb.20210115.001 [16] LU Y, TANG J. On time-frequency domain feature extraction of wave signals for structural health monitoring [J]. Measurement:journal of the International Measurement Confederation, 2018, 114: 51 − 59. doi: 10.1016/j.measurement.2017.09.016 [17] 周广东, 操声浪, 刘定坤. 基于自适应动态惩罚遗传算法的桥梁监测无线测点优化研究[J]. 建筑科学与工程学报, 2018, 35(5): 86 − 92. doi: 10.3969/j.issn.1673-2049.2018.05.012ZHOU Guangdong, CAO Shenglang, LIU Dingkun. Research on wireless measurement point optimization of bridge monitoring based on adaptive dynamic penalty genetic algorithm [J]. Journal of Building Science and Engineering, 2018, 35(5): 86 − 92. (in Chinese) doi: 10.3969/j.issn.1673-2049.2018.05.012 [18] KANARACHOS S, CHRISTOPOULOS S G, CHRONEOS A, et al. Detecting anomalies in time series data via a deep learning algorithm combining wavelets, neural networks and Hilbert transform [J]. Expert Systems with Applications, 2017, 85: 292 − 304. doi: 10.1016/j.eswa.2017.04.028 [19] CHANG C, LIN T, CHANG C. Applications of neural network models for structural health monitoring based on derived modal properties [J]. Measurement, 2018, 129: 457 − 470. doi: 10.1016/j.measurement.2018.07.051 [20] CHEN L, XU G, ZHANG S, et al. Health indicator construction of machinery based on end-to-end trainable convolution recurrent neural networks [J]. Journal of Manufacturing Systems, 2020, 54: 1 − 11. doi: 10.1016/j.jmsy.2019.11.008 [21] DENG W, MOU Y, KASHIWA T, et al. Vision based pixel-level bridge structural damage detection using a link ASPP network [J]. Automation in Construction, 2020, 110: 102973. doi: 10.1016/j.autcon.2019.102973 [22] 王慧, 郭晨林, 王乐, 等. 基于内积矩阵及深度学习的结构健康监测研究[J]. 工程力学, 2021, 39(2): 14 − 22, 75. doi: 10.6052/j.issn.1000-4750.2020.12.0935WANG Hui, GUO Chenlin, WANG Le, et al. Structural health monitoring based on inner product matrix and deep learning [J]. Engineering Mechanics, 2021, 39(2): 14 − 22, 75. (in Chinese) doi: 10.6052/j.issn.1000-4750.2020.12.0935 [23] CHENG T, LI N, CONSELICE C J, et al. Identifying strong lenses with unsupervised machine learning using convolutional autoencoder [J]. Monthly Notices of the Royal Astronomical Society, 2020, 494: 3750 − 3765. doi: 10.1093/mnras/staa1015 [24] ADEM K, KILIÇARSLAN S, CÖMERT O. Classification and diagnosis of cervical cancer with stacked autoencoder and softmax classification [J]. Expert Systems with Applications, 2019, 115: 557 − 564. doi: 10.1016/j.eswa.2018.08.050 [25] 卓德兵, 曹晖. 基于小波时频图与轻量级卷积神经网络的螺栓连接损伤识别[J]. 工程力学, 2021, 38(9): 228 − 238. doi: 10.6052/j.issn.1000-4750.2021.02.0116ZHUO Debing, CAO Hui. Bolt connection damage identification based on wavelet time-frequency diagram and light weight convolutional neural network [J]. Engineering Mechanics, 2021, 38(9): 228 − 238. (in Chinese) doi: 10.6052/j.issn.1000-4750.2021.02.0116 [26] ZENG S, GOU J, DENG L. An antinoise sparse representation method for robust face recognition via joint l1 and l2 regularization [J]. Expert Systems with Applications, 2017, 82: 1 − 9. doi: 10.1016/j.eswa.2017.04.001 [27] JAIS I K M, ISMAIL A R, NISA S Q. Adam Optimization Algorithm for Wide and Deep Neural Network [J]. Knowl. Eng. Data Sci., 2019, 2: 41 − 46. doi: 10.17977/um018v2i12019p41-46 [28] 王乐. 结构边界条件参数识别及损伤检测方法研究[D]. 西安: 西北工业大学, 2010.WANG Le. Research on the method of structural boundary condition parameter identification and damage detection [D]. Xi'an: Northwestern Polytechnical University, 2010. (in Chinese) -
点击查看大图
图(8) / 表(6)
计量
- 文章访问数: 239
- HTML全文浏览量: 120
- PDF下载量: 64
- 被引次数: 0
