基于SVM的含缺陷20钢弯管爆破压力预测

doi:10.16265/j.cnki.issn1003-3033.2023.02.1269

摘要/Abstract

摘要：

为快速、精确预测含局部减薄缺陷的弯管爆破压力,首先验证显式非线性有限元模型的模拟精确性,然后以168组不同缺陷尺寸下20钢弯管爆破压力的有限元模拟数据作为学习样本,建立含局部减薄缺陷20钢弯管爆破压力预测的支持向量机(SVM)模型;其次利用交叉验证(CV)、遗传算法(GA)、粒子群算法(PSO)分别优化SVM模型;最后分析对比用于预测弯管爆破压力的3种优化SVM模型与ASME B31G-2009、DNV RP-F101、SHELL 92等3种通用规范的计算误差。结果表明:CV-SVM、GA-SVM、PSO-SVM等3种模型的预测误差均小于3种规范的计算误差,其最大相对误差分别为-2.33%、-3.4%和1.94%;说明SVM模型用于预测弯管爆破压力时操作简单、计算时间短、预测精度高、工程实用性好。

关键词: 支持向量机(SVM), 局部减薄缺陷, 20钢弯管, 爆破压力, 交叉验证(CV), 遗传算法(GA), 粒子群算法(PSO)

Abstract:

In order to quickly and accurately predict the burst pressure of 20 steel elbow with local wall-thinning defects, the SVM model for predicting burst pressure was established. After verifying the simulation accuracy of the explicit nonlinear finite element method, the 168 sets of data of explicit finite element simulation for burst pressure of 20 steel elbows with different defect sizes were used as learning samples of SVM model. CV, GA and PSO were used to optimize the SVM model. The prediction errors were analyzed by comparing the burst pressure calculated by the three optimized SVM model and 3 types of common criterions (ASME B31G-2009, DNV RP-F101 and SHELL 92). The results show that the prediction errors of the three optimized SVM models are less than that of the current common criterions at home and abroad. When CV-SVM and GA-SVM and PSO-SVM models are used to predict the burst pressure of 20 steel elbows with local thinning defects, and the maximum errors of CV-SVM, GA-SVM and PSO-SVM are-2.33%,-3.4% and 1.94% respectively. SVM model is easy to use, has high prediction accuracy, good engineering practicability and short time consumption.

Key words: support vector machine (SVM), local wall-thinning defects, 20 steel elbow, burst pressures, cross validation (CV), genetic algorithm(GA), particle swarm optimization (PSO)

郄彦辉, 郭涛, 周凌志, 王昱. 基于SVM的含缺陷20钢弯管爆破压力预测[J]. 中国安全科学学报, 2023, 33(2): 89-95.

QIE Yanhui, GUO Tao, ZHOU Lingzhi, WANG Yu. Prediction of burst pressure of 20 steel elbow with defects based on SVM[J]. China Safety Science Journal, 2023, 33(2): 89-95.

图/表 10

表1

图1

表2

图2

表3

图3

表4

表5

3种模型的性能对比

评价指标	CV-SVM	GA-SVM	PSO-SVM
MSE	1.802 9	1.122 1	1.047 8
最大误差/%	4.588 67	4.242 588	3.677 58
平均误差/%	0.845	0.82	0.669
决定系数 $R 2$	0.995 06	0.996 55	0.997 03

表5

表6

图4

参考文献 21

[1]	李艳, 赵均海, 曹雪叶, 等. 内压作用下弯管爆破压力的三剪统一解[J]. 应用力学学报, 2015, 32(4): 530-536,700.
	LI Yan, ZHAO Junhai, CAO Xueye, et al. Triple-shear unified solution for burst pressure of elbows under internal pressure[J]. Chinese Journal of Applied Mechanics, 2015, 32(4): 530-536,700.
[2]	陈钢, 张传勇, 刘应华. 内压和面内弯矩作用下含局部减薄弯头塑性极限载荷的有限元分析[J]. 工程力学, 2005, 22(2): 43-49.
	CHEN Gang, ZHANG Chuanyong, LIU Yinghua. Finite element analysis of plastic limit loads of locally thinned elbows under internal pressure and in-plane bending moment[J]. Engineering Mechanics, 2005, 22(02): 43-49.
[3]	李建, 周昌玉, 薛吉林. 含内局部减薄缺陷高温弯管蠕变极限载荷及其安全评定[J]. 机械强度, 2015, 37(2): 348-354.
	LI Jian, ZHOU Changyu, XUE Jilin. Creep limit load and safety assessment of pipe bends with inner local wall thinning defect at high temperature[J]. Journal of Mechanical Strength, 2015, 37(2):348-354.
[4]	LEE G H, SEO J K, PAIK J K. Condition assessment of damaged elbow in subsea pipelines[J]. Ships and Offshore Structures, 2017, 12(1): 135-151. doi: 10.1080/17445302.2015.1116245
[5]	张新生, 张玥. 基于Lasso-PSO-BP神经网络的腐蚀管道失效压力的预测[J]. 材料保护, 2020, 53(04):46-52.
	ZHANG Xinsheng, ZHANG Yue. Prediction of failure pressure of corroded pipeline based on Lasso-PSO-BP neural network[J]. Materials Protection, 2020, 53(04): 46-52.
[6]	KIM J W, YOON M S, PARK C Y. The effect of load-controlled bending load on the failure pressure of wall-thinned pipe elbows[J]. Nuclear Engineering and Design, 2013, 265(12): 174-183. doi: 10.1016/j.nucengdes.2013.07.027
[7]	LEE G H, POURARIA H, SEO J K, et al. Burst strength behaviour of an aging subsea gas pipeline elbow in different external and internal corrosion-damaged positions[J]. International Journal of Naval Architecture and Ocean Engineering, 2015, 7(3):435-451. doi: 10.1515/ijnaoe-2015-0031
[8]	王佳音, 帅健, 刘道乾, 等. 内压作用下含体积型缺陷弯管极限载荷研究[J]. 中国安全生产科学技术, 2019, 15(9):158-163.
	WANG Jiayin, SHUAI Jian, LIU Daoqian, et al. Research on ultimate load of elbow containing volumetric defects under effect of internal pressure[J]. Journal of Safety Science and Technology, 2019, 15(9): 158-163.
[9]	WANG Qingguo, ZHOU Wenxing. Burst pressure models for thin-walled pipe elbows[J]. International Journal of Mechanical Sciences, 2019, 159(5): 20-29. doi: 10.1016/j.ijmecsci.2019.05.027
[10]	WANG Qingguo, ZHOU Wenxing. A new burst pressure model for thin-walled pipe elbows containing metal-loss corrosion defects[J]. Engineering Structures, 2019, 200: DOI:10.1016/j.engstruct.2019.109720. doi: 10.1016/j.engstruct.2019.109720
[11]	贾思奇, 郄彦辉, 李煜彤, 等. 基于遗传-神经网络算法的含均匀腐蚀缺陷油气管线爆破压力预测研究[J]. 中国安全生产科学技术, 2020, 16(12): 105-110.
	JIA Siqi, QIE Yanhui, LI Yutong, et al. Research on burst pressure prediction of oil and gas pipelines with uniform corrosion defects based on GA-BPNNs algorithm[J]. Journal of Safety Science and Technology, 2020, 16(12): 105-110.
[12]	毕傲睿, 骆正山, 宋莹莹, 等. 内腐蚀海底管道剩余强度的FOA-GRNN模型[J]. 中国安全科学学报, 2020, 30(6): 78-83. doi: 10.16265/j.cnki.issn1003-3033.2020.06.012
	BI Aorui, LUO Zhengshan, SONG Yingying, et al. Residual strength analysis of internally corroded submarine pipeline based on FOA-GRNN model[J]. China Safety Science Journal, 2020, 30(6): 78-83. doi: 10.16265/j.cnki.issn1003-3033.2020.06.012
[13]	青树勇, 代卿, 郭鸿雁. 基于BP神经网络预测含腐蚀缺陷管道的极限压力[J]. 腐蚀与防护, 2018, 39(8): 634-637.
	QING Shuyong, DAI Qing, GUO Hongyan. Prediction of ultimate pressure of pipeline with corrosion defects based on BP neural network[J]. Corrosion and Protection, 2018, 39(8): 634-637.
[14]	靳文博, 肖荣鸽, 田震, 等. 基于广义回归神经网络的海底腐蚀管道极限承载力预测[J]. 热加工工艺, 2020, 49(8): 58-61.
	JIN Wenbo, XIAO Rongge, TIAN Zhen, et al. Prediction of ultimate bearing capacity of submarine corrosion pipeline based on generalized regression neural network[J]. Hot Working Technology, 2020, 49 (8): 58-61.
[15]	骆正山, 田珮琦. RS-PSO-ELM下腐蚀管道失效压力预测[J]. 中国安全科学学报, 2021, 31(3): 28-34. doi: 10.16265/j.cnki.issn1003-3033.2021.03.004
	LUO Zhengshan, TIAN Peiqi. Prediction of failure pressure of corrosion pipelines based on RS-PSO-ELM[J]. China Safety Science Journal, 2021, 31(3):28-34. doi: 10.16265/j.cnki.issn1003-3033.2021.03.004
[16]	马钢. 基于相关性分析和SVM算法的单缺陷管道剩余强度预测[D]. 西安: 西安石油大学, 2020.
	MA Gang. Residual strength prediction of single defect pipeline based on correlation analysis and SVM algorithm[D]. Xi'an: Xi'an Shiyou University, 2020.
[17]	VAPNIK V N. The nature of statistical learning theory[M]. NewYork: Springer-Verlag, 1995: 52-123.
[18]	骆正山, 王浩, 毕傲睿. 基于RS-SVM 的城市埋地燃气管道外腐蚀情况评价[J]. 中国安全科学学报, 2017, 27(6):109-114. doi: 10.16265/j.cnki.issn1003-3033.2017.06.019
	LUO Zhengshan, WANG Hao, BI Aorui. Evaluation of external corrosion of urban buried gas pipeline based on RS-SVM[J]. China Safety Science Journal, 2017, 27(6):109-114. doi: 10.16265/j.cnki.issn1003-3033.2017.06.019
[19]	周凌志. 基于有限元模拟的含局部减薄缺陷弯头的爆破压力研究[D]. 天津: 河北工业大学, 2021.
	ZHOU Lingzhi. Study on burst pressure of elbow with local thinning defects based on finite element simulation[D]. Tianjin: Hebei University of Technology, 2021.
[20]	李晓帅. 压力容器与压力管道元件的爆破压力研究[D]. 天津: 河北工业大学, 2017.
	LI Xiaoshuai. Study of bursting pressure of pressure vessel and pressure pipeline components[D]. Tianjin: Hebei University of Technology, 2017.
[21]	周启超, 刘剑, 刘丽, 等. 基于SVM的通风系统故障诊断惩罚系数与核函数系数优化研究[J]. 中国安全生产科学技术, 2019, 15(4): 45-51.
	ZHOU Qichao, LIU Jian, LIU Li, et al. Research on fault fiagnosis penalty coefficient and kernel function coefficient optimization of ventilation system based on SVM[J]. Journal of Safety Science and Technology, 2019, 15(4): 45-51.

编号	外径/ mm	壁厚/ mm	角度/ (°)	失效位置	模拟/ MPa	试验/ MPa	误差/ %
1	219	8	90	内拱	34.0	34.4	1.16
2	325	10	90	内拱	29.8	29.7	0.34
3	219	10	90	内拱	33.8	34.1	0.68

样本编号	减薄长度/ (°)	减薄宽度/ (°)	减薄深度/ mm	爆破压力/MPa
样本编号	减薄长度/ (°)	减薄宽度/ (°)	减薄深度/ mm	外拱内部	中性线内部	内拱内部
1-3	5	5	1.8	140.3	139.2	131.7
4-6	5	20	2.7	137.8	135.2	130.4
7-9	5	20	4.5	118.4	116.3	115.1
︙	︙	︙	︙	︙	︙	︙
79-81	10	10	3.6	113.2	111.2	109.1
82-84	10	15	1.8	137.2	133.4	129.1
85-87	10	35	5.4	82.2	80.7	80.3
︙	︙	︙	︙	︙	︙	︙
160-162	20	20	5.4	74.7	74.2	73.3
163-165	20	35	1.8	136.2	131.3	127.7
166-168	30	30	1.8	135.3	129.8	125.7

优化方法	归一化范围	迭代次数	C	g
CV	[-1,1]	—	1 024	0.015 6
GA	[0,1]	100	49.79	0.314 7
PSO	[0,1]	150	100	0.357 3

数据编号	减薄长度/ (°)	减薄宽度/ (°)	减薄深度/ mm	缺陷位置	ASME B31G- 2009/MPa	DNV RPF101/ MPa	SHELL 92/ MPa	CV-SVM/ MPa	GA-SVM/ MPa	PSO-SVM/ MPa	显式非线性有限元/ MPa
1	5	20	3.6	内拱	81.95	109.80	88.11	127.34	129.44	128.53	129.6
2	5	20	3.6	中性线	81.91	109.76	88.02	125.01	127.60	125.05	126.2
3	5	20	3.6	外拱	81.87	109.70	87.92	122.29	124.30	122.55	124.2
4	10	15	5.4	内拱	80.50	107.53	83.74	90.20	90.74	90.67	89.8
5	10	15	5.4	中性线	80.26	107.15	83.05	89.69	89.75	89.74	90.7
6	10	15	5.4	外拱	79.99	106.73	82.33	88.70	88.77	89.86	89.2
7	20	35	4.5	内拱	77.89	103.92	78.16	87.94	87.47	90.34	89.3
8	20	35	4.5	中性线	77.36	103.09	77.01	86.145	85.20	88.54	88.2
9	20	35	4.5	外拱	76.83	102.23	75.89	84.19	83.73	87.26	85.6
10	30	20	1.8	内拱	79.65	106.88	83.59	133.17	133.10	132.89	132.7
11	30	20	1.8	中性线	79.39	106.47	83.10	129.12	129.90	129.60	129.2
12	30	20	1.8	外拱	79.13	106.07	82.64	124.09	124.48	122.81	123.2