[1]余移山,陆继东,董美蓉,等.T91金属管道表面特性对LIBS测量的影响[J].应用科技,2020,47(4):82-87.[doi:10.11991/yykj.202001002]
 YU Yishan,LU Jidong,DONG Meirong,et al.Study on the influence of surface characteristics of T91 metal pipeline on laser-induced breakdown spectroscopy measurement[J].Applied science and technology,2020,47(4):82-87.[doi:10.11991/yykj.202001002]
点击复制

T91金属管道表面特性对LIBS测量的影响(/HTML)
分享到:

《应用科技》[ISSN:1009-671X/CN:23-1191/U]

卷:
第47卷
期数:
2020年4期
页码:
82-87
栏目:
动力与能源工程
出版日期:
2020-07-05

文章信息/Info

Title:
Study on the influence of surface characteristics of T91 metal pipeline on laser-induced breakdown spectroscopy measurement
作者:
余移山 陆继东 董美蓉 陆盛资 黄健伟 张勇升
华南理工大学 电力学院,广东 广州 510640
Author(s):
YU Yishan LU Jidong DONG Meirong LU Shengzi HUANG Jianwei ZHANG Yongsheng
School of Electric Power , South China University of Technology, Guangzhou 510640, China
关键词:
激光诱导击穿光谱表面特性火电厂金属管道失效检测光谱分析机器学习支持向量机
Keywords:
laser-induced breakdown spectroscopysurface characteristicheat-engine plantmetal pipelinefailure detectionspectral analysismachine learningsupport vector machines
分类号:
O657.38
DOI:
10.11991/yykj.202001002
文献标志码:
A
摘要:
为了将激光诱导击穿光谱(laser-induced breakdown spectroscopy,LIBS)技术应用到火电厂运行的高温承压管道的失效检测中,对来自火电厂的4个不同老化等级的T91钢管道样品表面采用打磨与未处理2种处理方法,对比光谱与其作为预测集在老化等级模型中的表现。分析数据表明:打磨与未处理样品的代表性元素的谱线强度变化在前100的激光脉冲段有较大的差异;而在100~200的激光脉冲段打磨与未打磨的光谱强度趋于接近。在支持向量机(support vector machines, SVM)老化等级模型中,选取未打磨样品的激光脉冲段100~200的光谱数据输入SVM老化等级模型进行预测,得到了与打磨样品同样良好的预测结果,所有预测样品的准确率均达到了0.91以上。结果表明:选取合适的激光脉冲段,可以避免实际金属管道的表面特性对LIBS测量的影响,得到良好的老化等级预测结果。
Abstract:
In order to apply the laser-induced breakdown spectroscopy(LIBS) technology to the failure detection of high temperature pressure pipelines in the operation of thermal power plants, the surface of four T91 steel tube samples with different aging grades from thermal power plant was treated by grinding and untreated methods. The analysis data showed that the spectral line strength of the representative elements of polished and untreated samples was significantly different in the first 100 laser pulse segments, while the spectral strength of polished and untreated elements tended to be close in the first 100 to 200 laser pulse segments. In the SVM aging rating model, the spectral data from 100 to 200 laser pulse segments of unpolished samples were selected and input into the SVM aging rating model for prediction. The prediction results were as good as those of the polished samples, and the accuracy of all the predicted samples reached above 0.91. The results show that choosing appropriate laser pulse segment can avoid the influence of the surface characteristics of actual metal pipelines on the LIBS measurement, and obtain a better prediction result of aging grade.

参考文献/References:

[1] 郑晓红. 锅炉“四管”的失效机理研究与寿命预测[D]. 杭州: 浙江大学, 2002.
[2] 崔强. 锅炉受热面四管泄漏的原因分析及预防措施[J]. 中国设备工程, 2018(17): 58–59
[3] 李军. 激光诱导击穿光谱应用于电站锅炉受热面组织结构和性能诊断的研究[D]. 广州: 华南理工大学, 2016.
[4] YAO Shunchun, DONG Meirong, LU Jidong. Correlation between grade of pearlite spheroidization and laser induced spectra[J]. Laser physics, 2013, 23(12): 125702.
[5] 董美蓉, 韦丽萍, 陆继东, 等. 基于K-CV参数优化支持向量机的LIBS燃煤热值定量分析[J]. 光谱学与光谱分析, 2019, 39(7): 2202–2209
[6] LU Shengzi, DONG Meirong, HUANG Jianwei, et al. Estimation of the aging grade of T91 steel by laser-induced breakdown spectroscopy coupled with support vector machines[J]. Spectrochimica acta part B: atomic spectroscopy, 2018, 140: 35–43.
[7] 董璇. TP347H受热面材料的激光光谱特性分析[D]. 广州: 华南理工大学, 2015.
[8] LI Jun, LU Jidong, DAI Yuan, et al. Correlation between aging grade of T91 steel and spectral characteristics of the laser-induced plasma[J]. Applied surface science, 2015, 346: 302–310.
[9] GB/T4338-2006 金属材料高温拉伸试验方法[S].
[10] DL/T884-2004 火电厂金相检验与评定技术导则[S].
[11] HUANG Jianwei, DONG Meirong, LU Shengzi, et al. Estimation of the mechanical properties of steel via LIBS combined with canonical correlation analysis (CCA) and support vector regression (SVR)[J]. Journal of analytical atomic spectrometry, 2018, 33(5): 720–729.
[12] 屈华阳, 胡净宇, 赵雷, 等. 激光诱导击穿光谱法分析表面氧化钢铁样品中的碳硅锰磷硫铬镍铜铝[J]. 冶金分析, 2012, 32(7): 1–6
[13] 尚文利, 李琳, 万明, 等. 基于优化单类支持向量机的工业控制系统入侵检测算法[J]. 信息与控制, 2015, 44(6): 678–684
[14] 周中寒, 田雪咏, 孙兰香, 等. Fiber-LIBS技术结合SVM鉴定铝合金牌号[J]. 激光与光电子学进展, 2018, 55(6): 424–430
[15] SHU Xin, LAI Darong, XU Huanliang, et al. Learning shared subspace for multi-label dimensionality reduction via dependence maximization[J]. Neurocomputing, 2015, 168: 356–364.
[16] LIN Yaojin, HU Qinghua, LIU Jinghua, et al. Multi-label feature selection based on neighborhood mutual information[J]. Applied soft computing, 2016, 38: 244–256.
[17] 程玉胜, 宋帆, 王一宾, 等. 基于专家特征的条件互信息多标记特征选择算法[J]. 计算机应用, 2020, 40(2): 503–509
[18] PEDREGOSA F, VAROQUAUX G, GRAMFORT A, et al. Scikit-learn: machine learning in python[J]. Journal of machine learning research, 2011, 12(10): 2825–2830.
[19] LU Shengzi, SHEN Shen, HUANG Jianwei, et al. Feature selection of laser-induced breakdown spectroscopy data for steel aging estimation[J]. Spectrochimica acta part B: atomic spectroscopy, 2018, 150: 49–58.
[20] KERBAA T H, MEZACHE A, OUDIRA H. Model selection of sea clutter using cross validation method[J]. Procedia computer science, 2019, 158: 394–400.
[21] 孙健, 李琪, 陈明强, 等. 基于机器学习的油气水层随钻识别模型优选[J]. 西安石油大学学报(自然科学版), 2019, 34(5): 79–85

备注/Memo

备注/Memo:
收稿日期:2020-01-03。
基金项目:广东省自然科学基金项目(2017B030311009)
作者简介:余移山,男,硕士研究生;陆继东,男,教授,博士生导师
通讯作者:陆继东,E-mail:jdlu@scut.edu.cn
更新日期/Last Update: 2020-11-27