小波包与遗传算法优化BP神经网络相结合的井架钢结构损伤识别

韩东颖; 田伟; 黄岩; 朱国庆

doi:10.13433/j.cnki.1003-8728.20220215

小波包与遗传算法优化BP神经网络相结合的井架钢结构损伤识别

doi: 10.13433/j.cnki.1003-8728.20220215

燕山大学车辆与能源学院, 河北秦皇岛 066004

基金项目:

国家自然科学基金项目 51875500

河北省人社厅留学人员科技活动项目 C20190516

详细信息

作者简介:
韩东颖, 教授, 博士生导师, 博士, dongying.han@163.com

中图分类号: TG156
计量
- 文章访问数: 54
- HTML全文浏览量: 22
- PDF下载量: 16
- 被引次数: 0
出版历程
- 收稿日期: 2021-12-10
- 刊出日期: 2024-01-25

Identifying Damage of Derrick Steel Structure Based on BP Neural Network Optimized with Wavelet Packet and Genetic Algorithm

College of Vehicles & Energy, Yanshan University, Qinhuangdao 066004, Hebei, China

摘要

摘要: 井架钢结构损伤影响其承载安全性，为快速、准确对损伤位置进行识别，提出小波包与遗传算法优化BP神经网络相结合的井架钢结构损伤识别方法。首先利用小波包处理非平稳振动信号的优良性能对原始振动信号进行特征提取，获得表征井架钢结构损伤的信息；再通过特征参数建立数据集训练并测试井架钢结构损伤识别模型，该模型结合遗传算法自身特点改善了传统BP神经网络的不足。本文识别方法不需要损伤前的数据特征进行对比，便可对损伤位置进行确定。经过对石油井架钢结构模型实验验证：该方法对井架钢结构损伤识别准确率超过90%，相对于BP网络识别准确率以及识别速度均有所提高。
- 井架钢结构 /
- 损伤 /
- 小波包 /
- 遗传算法 /
- 优化的BP神经网络
Abstract: The damage of a derrick steel structure affects its bearing safety. In order to identify the damage location quickly and accurately, a method for identifying the damage of a derrick steel structure based on the BP neural network optimized with wavelet packet and genetic algorithm is proposed. Firstly, the excellent performance of the wavelet packet that processes non-stationary vibration signals is used to extract the original vibration signal features, and the information that characterizes the damage is obtained. Then, the data set is established through characteristic parameters to train and test the derrick steel structural damage identification method. Combined with the characteristics of the genetic algorithm, the identification method reduces the shortcomings of the traditional BP neural network. The identification method can determine the damage location without comparing the data characteristics before damage. The experimental results show that the accuracy of the derrick steel structural damage identification method is more than 90%, being higher than that of the traditional BP neural network.
- derrick steel structure /
- damage /
- wavelet packet /
- genetic algorithm /
- BP neural network optimization

HTML全文

图 1 小波包分解原理图

Figure 1. Schematic diagram of wavelet packet decomposition

下载: 全尺寸图片幻灯片

图 2 GA-BP神经网络算法流程

Figure 2. GA-BP neural network algorithm flow

下载: 全尺寸图片幻灯片

图 3 传感器布置及损伤设定位置

Figure 3. Sensor layout and damage setting position

下载: 全尺寸图片幻灯片

图 4 位置①处销钉损伤3号传感器时域信号图

Figure 4. Time domain signal diagram of pin damage No. 3 sensor at position ①

下载: 全尺寸图片幻灯片

图 5 位置①处销钉损伤3号传感器频域信号图

Figure 5. Frequency domain signal diagram of sensor No. 3 at position ①

下载: 全尺寸图片幻灯片

图 6 小波包三层分解前3频段信号图

Figure 6. Signal diagram of the first 3 frequency bands before the three-layer decomposition of the wavelet packet

下载: 全尺寸图片幻灯片

图 7 误差与遗传代数变化图

Figure 7. Error and genetic algebraic variation diagram

下载: 全尺寸图片幻灯片

图 8 GA-BP网络识别结果

Figure 8. GA-BP network identification results

下载: 全尺寸图片幻灯片

图 9 BP网络与GA-BP网络识别结果对比

Figure 9. Comparison of recognition results between BP network and GA-BP network

下载: 全尺寸图片幻灯片

表 1 待测单损伤分布位置

Table 1. Distribution location of single damage to be measured

损伤位置	销钉1	销钉2	销钉3	斜撑1	斜撑2
位置编号	①	②	③	④	⑤

下载: 导出CSV

表 2 网络与GA-BP网络的识别结果统计

Table 2. Statistics of recognition results of GA-BP network and GA-BP network

损伤类型	销钉1	销钉2	销钉3	斜撑1	斜撑2
损伤标签	1	2	3	4	5
测试样本数	12	13	13	12	10
BP正确识别	9	11	12	9	8
GA-BP正确识别	12	12	13	11	8

下载: 导出CSV

参考文献(20)

[1]	韩东颖, 时培明. 基于频率和当量损伤系数的井架钢结构损伤识别[J]. 工程力学, 2011, 28(9): 109-114. HAN D Y, SHI P M. Damage identification of derrick steel structures based on frequency and equivalent damage coefficient[J]. Engineering Mechanics, 2011, 28(9): 109-114. (in Chinese)
[2]	刘习军, 孙良, 张素侠, 等. 改进的小波包能量指标在结构损伤识别中的应用[J]. 机械科学与技术, 2016, 35(5): 657-661. doi: 10.13433/j.cnki.1003-8728.2016.0501 LIU X J, SUN L, ZHANG S X, et al. Application of improved wavelet packets index to structural damage detection[J]. Mechanical Science and Technology for Aerospace Engineering, 2016, 35(5): 657-661. (in Chinese) doi: 10.13433/j.cnki.1003-8728.2016.0501
[3]	韩东颖, 时培明. 基于频率和BP神经网络的井架钢结构损伤识别[J]. 中国安全科学学报, 2012, 22(8): 118-123. HAN D Y, SHI P M. Identification of derrick steel structures damage based on frequency and BP neural network[J]. China Safety Science Journal, 2012, 22(8): 118-123. (in Chinese)
[4]	韩东颖, 周国强, 李子丰. 基于应力和当量损伤系数的石油井架损伤识别[J]. 石油钻采工艺, 2008, 30(1): 29-32. HAN D Y, ZHOU G Q, LI Z F. Damage identification of oil derrick based on stress and equivalent damage coefficients[J]. Oil Drilling & Production Technology, 2008, 30(1): 29-32. (in Chinese)
[5]	邹龙庆, 谢春强. 基于正则化频率变化率与神经网络的石油井架结构损伤识别[J]. 中国安全科学学报, 2008, 18(10): 30-33. ZOU L Q, XIE C Q. Damage identification of oil derrick based on normalized change ratio of frequency and neural network[J]. China Safety Science Journal, 2008, 18(10): 30-33. (in Chinese)
[6]	HAN D Y, WEI S M, SHI P M, et al. Damage identifica- tion of a derrick steel structure based on the HHT marginal spectrum amplitude curvature difference[J]. Shock and Vibration, 2017, 2017(2): 1-9. DOI: 10.1155/2017/1062949.
[7]	罗丹, 任敏. 应用小波包神经网络的塔机起重臂损伤识别[J]. 机械设计与制造, 2020(4): 208-211. LUO D, REN M. Study on damage identification of tower crane boom by using wavelet packet and neural network[J]. Machinery Design & Manufacture, 2020(4): 208-211. (in Chinese)
[8]	邱伟, 师培峰, 吴瑞斌, 等. 小波包-神经网络在桥梁损伤定位检测中的应用[J]. 遥测遥控, 2018, 39(3): 64-70. QIU W, SHI P F, WU R B, et al. Application of wavelet packet-neural network in the detection of bridge damage location[J]. Journal of Telemetry, Tracking and Command, 2018, 39(3): 64-70. (in Chinese)
[9]	王名月, 缪炳荣, 李旭娟, 等. 基于LMD样本熵和RBF网络的结构损伤识别研究[J]. 机械强度, 2018, 40(3): 522-527. WANG M Y, MIAO B R, LI X J, et al. Structural damage identification based on LMD sample entropy and RBF network[J]. Journal of Mechanical Strength, 2018, 40(3): 522-527. (in Chinese)
[10]	XIONG S C, ZHOU H D, HE S, et al. A novel end-to-end fault diagnosis approach for rolling bearings by integrating wavelet packet transform into convolutional neural network structures[J]. Sensors, 2020, 20(17): 4965.
[11]	李雪松, 马宏伟, 林逸洲. 基于卷积神经网络的结构损伤识别[J]. 振动与冲击, 2019, 38(1): 159-167. LI X S, MA H W, LIN Y Z. Structural damage identification based on convolution neural network[J]. Journal of Vibration and Shock, 2019, 38(1): 159-167. (in Chinese)
[12]	CHIRKOV D, FILATOVA A, POLOKHIN S. Multi- objective shape optimization of Francis runner using metamodel assisted genetic algorithm[J]. IOP Conference Series: Earth and Environmental Science, 2021, 774: 012109.
[13]	陈果. 基于遗传算法的支持向量机分类器模型参数优化[J]. 机械科学与技术, 2007, 26(3): 347-350. CHEN G. Optimizing the parameters of support vector machine's classifier model based on genetic algorithm[J]. Mechanical Science and Technology for Aerospace Engineering, 2007, 26(3): 347-350. (in Chinese)
[14]	GUAN X Y, XIE S J, CHEN G, et al. Modal parameter identification by adaptive parameter domain with multiple genetic algorithms[J]. Journal of Mechanical Science and Technology, 2020, 34(12): 4965-4980.
[15]	张泽宇, 惠记庄, 石泽. 小波包最优基分解树的降噪滤波方法研究[J]. 机械科学与技术, 2020, 39(1): 28-34. doi: 10.13433/j.cnki.1003-8728.20190239 ZHANG Z Y, HUI J Z, SHI Z. Research on denoising and filtering method based on wavelet packet optimal base decomposition tree[J]. Mechanical Science and Technology for Aerospace Engineering, 2020, 39(1): 28-34. (in Chinese) doi: 10.13433/j.cnki.1003-8728.20190239
[16]	胡蓉, 谭宏斌, 冯志宇, 等. GA-BP网络与D-S证据相结合的多传感器信息融合与传信[J]. 重庆邮电大学学报(自然科学版), 2011, 23(2): 220-223. HU R, TAN H B, FENG Z Y, et al. Multi-sensor information fusion based on integration of GA-BP network and D-S evidence theory[J]. Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition), 2011, 23(2): 220-223. (in Chinese)
[17]	郭秀才, 刘冰冰, 王力立. 基于小波包和CS-BP神经网络的矿用电力电缆故障诊断[J]. 计算机应用与软件, 2021, 38(9): 105-110. GUO X C, LIU B B, WANG L L. Fault diagnosis of mining power cable based on wavelet packet and CS-BP neural network[J]. Computer Applications and Software, 2021, 38(9): 105-110. (in Chinese)
[18]	吕明珠, 苏晓明, 陈长征, 等. 小波包能量熵与EMD结合分析法在风机滚动轴承故障诊断中的应用[J]. 机械与电子, 2018, 36(6): 8-12. LYU M Z, SU X M, CHEN C Z, et al. Application of wavelet packet energy entropy and EMD conjoint analysis in fault diagnosis of wind turbine bearing[J]. Machinery & Electronics, 2018, 36(6): 8-12. (in Chinese)
[19]	高玉明, 张仁津. 基于遗传算法和BP神经网络的房价预测分析[J]. 计算机工程, 2014, 40(4): 187-191. GAO Y M, ZHANG R J. Analysis of house price prediction based on genetic algorithm and BP neural network[J]. Computer Engineering, 2014, 40(4): 187-191. (in Chinese)
[20]	苏崇宇, 汪毓铎. 基于改进的自适应遗传算法优化BP神经网络[J]. 工业控制计算机, 2019, 32(1): 67-69. SU C Y, WANG Y D. BP neural network optimized by improved adaptive genetic algorithm computer engineering and applications[J]. Industrial Control Computer, 2019, 32(1): 67-69. (in Chinese)