|本期目录/Table of Contents|

[1]唐泽斯,郭进,王金贵,等.基于人工神经网络的气体泄爆最大超压预测研究[J].中国安全生产科学技术,2020,16(4):56-62.[doi:10.11731/j.issn.1673-193x.2020.04.009]
 TANG Zesi,GUO Jin,WANG Jingui,et al.Study on prediction of maximum overpressure in gas explosion venting based on artificial neural network[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2020,16(4):56-62.[doi:10.11731/j.issn.1673-193x.2020.04.009]
点击复制

基于人工神经网络的气体泄爆最大超压预测研究
分享到:

《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]

卷:
16
期数:
2020年4期
页码:
56-62
栏目:
职业安全卫生管理与技术
出版日期:
2020-04-30

文章信息/Info

Title:
Study on prediction of maximum overpressure in gas explosion venting based on artificial neural network
文章编号:
1673-193X(2020)-04-0056-07
作者:
唐泽斯郭进王金贵张苏段在鹏
(福州大学 环境与资源学院,福建 福州 350116)
Author(s):
TANG Zesi GUO Jin WANG Jingui ZHANG Su DUAN Zaipeng
(School of Environment and Resources,Fuzhou University,Fuzhou Fujian 350116,China)
关键词:
人工神经网络气体泄爆优化算法参数选择模拟预测最大超压
Keywords:
artificial neural network gas explosion venting optimization algorithm parameter selection simulation and prediction maximum overpressure
分类号:
X932
DOI:
10.11731/j.issn.1673-193x.2020.04.009
文献标志码:
A
摘要:
为解决传统经验公式在预测气体泄爆中最大超压出现时的较大偏差或过于保守的问题,提出使用人工神经网络预测气体泄爆最大超压。基于124组实验数据,采用BP与RBF神经网络,通过优化算法计算与迭代循环对泄爆样本中的影响因素进行降维与选择,并确定2类神经网络本身在学习与计算气体泄爆样本时的相关参数。结果表明:PCA(主成分分析法)在当前样本条件下的降维效果较差,而通过迭代对比确认气体泄爆样本中的5类特征全部保留时神经网络的训练模拟效果最好;通过对124组实验数据进行随机挑选训练集与测试集的训练模拟结果发现,神经网络对气体泄爆中最大超压的预测效果较好;通过对比Molkov提出的和经Fakandu等改进的NFPA 68经验公式以及2类神经网络的预测结果表明,神经网络相比于传统气体泄爆经验公式具有明显优势。
Abstract:
In order to solve the problem of large deviation or too conservative results when predicting the maximum overpressure in gas explosion venting by using the traditional empirical formulas,it was proposed to predict the maximum overpressure in gas explosion venting by using the artificial neural network.Based on 124 groups of experimental data,the BP and RBF neural network were adopted to carry out the dimensionality reduction and selection on the influence factors of explosion venting samples through the optimization algorithm calculation and iteration loop,then the relevant parameters of two kinds of neural networks when learning and calculating the gas explosion venting samples were determined.The results showed that the effect of dimensionality reduction by PCA under the current sample conditions was poor,while through the iterative comparison,it was determined that the training simulation effect of neural network was the best when all the five kinds of features in the gas explosion venting samples were preserved.According to the training simulation results of training sets and testing sets randomly selected from 124 groups of experimental data,it was found that the neural network had good prediction effect on the maximum overpressure in the gas explosion venting.After comparing the prediction results of two empirical formulas,one was proposed by Molkov,the other was proposed in NFPA 68 and improved by Fakandu et al.,as well as the prediction results of two kinds of neural network,it showed that the neural network had obvious advantage compared with the traditional empirical formulas of gas explosion venting.

参考文献/References:

[1]MOLKOV V,BRAGIN M.Hydrogen-air deflagrations:vent sizing correlation for low-strength equipment and buildings[J].International Journal of Hydrogen Energy,2015,40 (2):1256-1266.
[2]MOLKOV V,DOBASHI R,SUZUKI M,et al.Modeling of vented hydrogen-air deflagrations and correlations for vent sizing[J].Journal of Loss Prevention in the Process Industries,1999,12(2):147-156.
[3]RAZUS D M,KRAUSE U.Comparison of empirical and semi-empirical calculation methods for venting of gas explosions[J].Fire Safety Journal 2001,36(1):1-23.
[4]National Fire Protection Association.Standard on explosion protection by deflagration venting:NFPA 68 [S].London:IHS Markit,2018.
[5]British Standards.Gas explosion venting protective systems:EN 14994 [S].London:the Standards Policy and Strategy Committee,2007.
[6]FAKANDU B M,ANDREWS G E,PHYLAKTOU H N.Vent burst pressure effects on vented gas explosion reduced pressure[J].Journal of Loss Prevention in the Process Industries,2015,36:429-438.
[7]MOLKOV V,DOBASHI R,SUZUKI M,et al.Venting of deflagrations hydrocarbon-air and hydrogen-air systems[J].Journal of Loss Prevention in the Process Industries,2000,13(3):397-407.
[8]师喜林,王志荣,蒋军成.球形容器内气体的泄爆过程[J].爆炸与冲击,2009,29 (4):390-394. SHI Xilin,WANG Zhirong,JIANG Juncheng.The process of gas venting in a spherical chamber[J].Explosion and Shock Waves,2009,29 (4):390-394.
[9]李岳,王淑兰,丁信伟,等.气体爆炸泄放计算方法比较与实验[J].石油化工设备,2003,32 (1):6-10. LI Yue,WANG Shulan,DING Xinwei,et al.Comparison of calculation methods and experiment for venting of gas explosion[J].Petro-Chemical Equipment,2003,32 (1):6-10.
[10]陈晔,刘暄亚.受限空间氢泄爆过程压力计算方法研究[J].消防科学与技术,2018,37 (2):160-163. CHEN Ye,LIU Xuanya.Study on caculation methods for vented explosion overpressures from hydrogen-air deflagrations in confiend space[J].Fire Science and Technology,2018,37 (2):160-163.
[11]师喜林,蒋军成.连通器内预混气体泄爆过程的数值模拟[J].中国安全生产科学技术,2007,3(6):24-26. SHI Xilin,JIANG Juncheng.Numerical simulation of the vented explosion process of premixed gas in the communicating vessel[J].Journal of Safety Science and Technology,2007,3(6):24-26.
[12]袁曾任.人工神经元网络及其应用[M].北京:清华大学出版社,1999.
[13]焦李成,杨淑媛,刘芳,等.神经网络七十年:回顾与展望[J].计算机学报,2016,39 (8):1697-1710. JIAO Licheng,YANG Shuyuan,LIU Fang,et al.Seventy years beyond neural networks:retrospect and prospect[J].Chinese Journal of Computers,2016,39 (8):1697-1710.
[14]刘永,张立毅.BP和RBF神经网络的实现及其性能比较[J].电子测量技术,2007,30 (4):77-80. LIU Yong,ZHANG Liyi.Implementation of BP and RBF neural network and their performance comparison[J].Electornic Measurement Technology,2007,30 (4):77-80.
[15]JACKSON I R H.An order of convergence for some radial basis functions[J].IMA Journal of Numerical Analysis,1989,9 (4):567-587.
[16]PARK J,SANDBERG I W.Universal approximation using radial-basis-function networks[J].Neural Computation,1991,3 (2):246-257.
[17]WANG J G,GOU J,YANG F Q,et al.Effects of hydrogen concentration on the vented deflagration of hydrogen-air mixtures in a 1 m3 vessel[J].International Journal of Hydrogen Energy,2018,43(45):21161-21168.
[18]RUI S C,GUO J,LI G,et al.The effect of vent burst pressure on a vented hydrogen-air deflagration in a 1 m3 vessel[J].International Journal of Hydrogen Energy,2018,43 (45):21169-21176.
[19]BAUWENS C R,CHAO J,DOROFEEV S B.Effect of hydrogen concentration on vented explosion overpressures from lean hydrogen-air deflagrations[J].International Journal of Hydrogen Energy,2012,37 (22):17599-17605.
[20]BAUWENS C R,CHAFFEE J,DOROFEEV S B.Vented explosion overpressures from combustion of hydrogen and hydrocarbon mixtures[J].International Journal of Hydrogen Energy,2011,36 (3):2329-2336.
[21]DAUBECH J,PROUST C,JAMOIS D,et al.Dynamics of vented hydrogen-air deflagrations[C]//In:Proceedings of the 4th International Conference on Hydrogen Safety.San Francisco:Paper ICHS,2011:12-14.
[22]KUMAR R K.Vented combustion of hydrogen-air mixtures in a large rectangular volume[C]// In 44th AIAA Aerospace Sciences Meeting and Exhibit.Reno,Nevada:Paper AIAA:2006:9-12.
[23]柳小桐.BP神经网络输入层数据归一化研究[J].机械工程与自动化,2010,160 (3):122-126. LIU Xiaotong.Study on data normalization in BP neural network[J].Mechanical Engineering & Automation,2010,160 (3):122-126.
[24]周忠科,王立杰.基于BP神经网络的煤矿安全预警评估机制研究[J].中国安全生产科学技术,2011,7 (4):134-138. ZHOU Zhongke,WANG Lijie.Study on safety early-warning assessment in coal mine based on BP neural networks[J].Journal of Safety Science and Technology,2011,7 (4):134-138.
[25]沈花玉,王兆霞,高成耀,等.BP神经网络隐含层单元数的确定[J].天津理工大学学报,2008,24 (5):13-15. SHEN Huayu,WANG Zhaoxia,GAO Chengyao,et al.Determining the number of BP neural network hidden layer units[J].Journal of Tianjin University of Technology,2008,24 (5):13-15.
[26]CYBENKO G.Approximations by superpositions of a sigmoidal function[J].Mathematics of Control Signals and Systems,1989,2 (4):303-314.
[27]LIPPMANN R P.An Introduction to computing with neural nets[J].IEEE ASSP Magazine,1987,4 (2):4-22.

相似文献/References:

[1]袁朋伟,宋守信,董晓庆.基于灰色神经网络优化组合模型的火灾预测研究[J].中国安全生产科学技术,2014,10(3):119.[doi:10.11731/j.issn.1673-193x.2014.03.020]
 YUAN Peng wei,SONG Shou xin,DONG Xiao qing.Study on fire accident prediction based on optimized grey neural network combination model[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2014,10(4):119.[doi:10.11731/j.issn.1673-193x.2014.03.020]
[2]陈锡进,潘勇.应用电性拓扑状态指数预测烷烃自燃点*[J].中国安全生产科学技术,2009,5(6):16.
 CHEN Xi jin,PAN Yong.Prediction of autoignition temperatures of alkanes by electrotopological state indices[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2009,5(4):16.
[3]孙勇,钟卓玲,李璐.基于班组安全氛围的建筑工人不安全行为预测[J].中国安全生产科学技术,2019,15(8):156.[doi:10.11731/j.issn.1673-193x.2019.08.025]
 SUN Yong,ZHONG Zhuoling,LI Lu.Prediction of construction workers’ unsafe behavior based on team safety climate[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2019,15(4):156.[doi:10.11731/j.issn.1673-193x.2019.08.025]

备注/Memo

备注/Memo:
收稿日期: 2019-11-15
* 基金项目: 国家社会科学基金项目(17CGL049);国家自然科学基金项目(51604083,51704079)
作者简介: 唐泽斯,硕士研究生,主要研究方向为气体泄爆与神经网络等。
通信作者: 郭进,博士,教授,主要研究方向为爆炸安全与防护。
更新日期/Last Update: 2020-05-11