湍流状态下化学品扩散溯源中不同目标函数的影响分析

doi:10.11949/0438-1157.20191550

摘要/Abstract

摘要：

化学气体泄漏会对生命财产安全及环境造成重大破坏，建立有效的监测装置及评估系统能够在发生泄漏事故时为应急救援提供辅助决策，有效降低泄漏带来的潜在风险。通常会通过在风险区域排布的气体浓度传感器检测气体浓度，根据气象条件及气体扩散前向分布模型来进行泄漏估计。由于泄漏场景数据不易获得，因此本文基于fires dynamics simulator(FDS)的大涡模拟方法模拟含湍流下的泄漏扩散场景获取数据，在此基础上研究了目标函数对于建模过程及溯源结果的影响。通过溯源结果发现，以浓度偏差平方和为目标函数的溯源结果从总体效果来看比其他目标函数更优，但其仍不能避免受湍流影响在某些场景的溯源指标远离平均值，而综合考虑不同目标的溯源结果在一定程度上能改善该情况。

关键词: 泄漏, 扩散, 模型, 湍流, 溯源, 目标函数

Abstract:

Hazardous gas release will threaten social security and cause damage to property and environment. Establish an effective emergency response system with monitoring devices can help reduce the potential risk and provide evaluation for release accident emergencies. The gas concentration is detected by the gas concentration sensor arranged in the risk area, and the leakage assessment is carried out according to the meteorological conditions and the forward diffusion model. Because the data of leakage scenarios is not easy to obtain, the large eddy simulation method based on Fires Dynamics Simulator (FDS) is used in this paper to generate date through simulating release scenarios with turbulent condition. On this basis, the impacts of the cost function to the modeling process and source term estimation are studied. The results of source term estimation shows the function of the square sum of concentration deviation is better than other functions. However, the function cannot avoid the condition that the estimation index of some scenes is far from the average value because of turbulent flow. In addition, a combination of the estimation results of different cost functions can improve the situation to a certain extent.

Key words: release, dispersion, model, turbulent condition, source term estimation, cost function

中图分类号:

N 945.12

董吉开, 杜文莉, 王冰, 许乔伊. 湍流状态下化学品扩散溯源中不同目标函数的影响分析[J]. 化工学报, 2020, 71(3): 1163-1173.

Jikai DONG, Wenli DU, Bing WANG, Qiaoyi XU. Investigating impacts of cost functions to atmospheric dispersion modeling and source term estimation in turbulent condition[J]. CIESC Journal, 2020, 71(3): 1163-1173.

图/表 12

图1 泄露场景空间示意图

Fig.1 Sketch of release scenario

图2 FDS仿真泄露场景示意图

Fig.2 FDS simulation of release scenario

表1 高斯扩散系数的参数

Table 1 Parameters of Gaussian dispersion coefficients

Model	a	b	c	d
Model 1	0.28065	0.72681	0.38029	0.59995
Model 2	0.30835	0.70756	0.34819	0.61652
Model 3	0.42271	0.65488	0.40053	0.56755

表2 测试场景下模型评价指标

Table 2 Model index of test scenarios

Model	Scenario	$R 2$	$F A C 2$	$F B$	$N M S E$
Model 1	1	0.8844	0.8956	0.1301	0.2607
	2	0.9441	0.9095	-0.1074	0.1069
	3	0.7638	0.8080	-0.1030	0.2810
Model 2	1	0.8914	0.8903	0.1180	0.2418
	2	0.9415	0.9111	-0.1200	0.1104
	3	0.7558	0.8042	-0.1157	0.2868
Model 3	1	0.8954	0.8389	0.0542	0.2185
	2	0.9277	0.9173	-0.1881	0.1274
	3	0.7645	0.8232	-0.1794	0.2594

表2 测试场景下模型评价指标

Table 2 Model index of test scenarios

Model	Scenario	$R 2$	$F A C 2$	$F B$	$N M S E$
Model 1	1	0.8844	0.8956	0.1301	0.2607
	2	0.9441	0.9095	-0.1074	0.1069
	3	0.7638	0.8080	-0.1030	0.2810
Model 2	1	0.8914	0.8903	0.1180	0.2418
	2	0.9415	0.9111	-0.1200	0.1104
	3	0.7558	0.8042	-0.1157	0.2868
Model 3	1	0.8954	0.8389	0.0542	0.2185
	2	0.9277	0.9173	-0.1881	0.1274
	3	0.7645	0.8232	-0.1794	0.2594

图3 所建模型在各场景下评估指标箱体图

Fig.3 Box-plot of modeling indexes of all scenarios with different model

图4 测试场景中预测浓度与观测浓度对比散点图

Fig.4 Scatter plots of predicted concentrations compared to observed ones of test scenario

图5 测试场景FDS仿真与各模型预测浓度在传感器平面浓度分布

Fig.5 Plane concentrations of FDS simulated compared to model calculated

图6 所有场景溯源平均距离偏差及源强相对偏差

Fig.6 Mean TD and ARDQ of all scenarios with different models and cost functions

表3 总60个场景溯源偏差结果的平均值及标准差

Table 3 Mean and standard deviation of result of source term estimation for all scenarios

模型	优化目标	x轴偏差/m		y轴偏差/m		z轴偏差/m		总距离差/m		源强相对偏差/%
模型	优化目标	平均值	标准差	平均值	标准差	平均值	标准差	平均值	标准差	平均值	标准差
模型1	1	8.35	6.36	0.96	0.66	0.53	0.61	8.61	6.17	9.22	6.73
	2	5.11	4.12	0.81	0.49	0.42	0.57	5.32	4.02	8.83	6.92
	3	4.98	3.86	0.75	0.46	0.32	0.49	5.15	3.79	10.68	8.47
	4	5.45	4.65	0.81	0.49	0.35	0.54	5.63	4.58	9.61	7.37
	5	5.45	4.65	0.81	0.49	0.35	0.54	5.63	4.58	10.46	8.38
	6	5.45	4.65	0.81	0.49	0.35	0.54	5.63	4.58	10.75	8.34
	7	13.28	7.82	1.25	0.99	1.45	1.12	13.55	7.73	10.42	8.75
模型2	1	8.13	6.38	0.96	0.66	0.56	0.64	8.42	6.16	8.99	6.22
	2	5.14	4.11	0.81	0.49	0.50	0.61	5.36	4.02	9.46	7.35
	3	5.20	3.88	0.75	0.46	0.39	0.55	5.37	3.81	11.62	8.71
	4	5.54	4.57	0.81	0.49	0.40	0.58	5.74	4.48	10.06	7.74
	5	5.54	4.57	0.81	0.49	0.40	0.57	5.73	4.48	11.47	8.76
	6	5.54	4.57	0.81	0.49	0.40	0.58	5.73	4.48	11.48	8.73
	7	13.21	7.95	1.26	0.99	1.65	1.12	13.53	7.82	11.28	9.54
模型3	1	6.66	5.34	0.96	0.66	0.63	0.70	6.95	5.19	9.61	7.52
	2	7.47	4.32	0.80	0.49	0.63	0.63	7.64	4.23	14.51	8.88
	3	8.19	4.05	0.75	0.46	0.55	0.63	8.29	4.02	15.68	10.55
	4	8.39	4.29	0.81	0.48	0.55	0.62	8.51	4.23	14.05	9.13
	5	8.38	4.29	0.81	0.49	0.54	0.62	8.50	4.23	16.49	10.20
	6	8.39	4.29	0.81	0.48	0.55	0.62	8.52	4.23	15.87	10.44
	7	10.08	7.67	1.27	0.97	1.69	0.98	10.58	7.41	9.26	8.15

图7 所有场景下溯源结果指标的箱型图

Fig.7 Box-plot of results of source term estimation of all scenarios

图8 不同场景下溯源位置与真实位置分布图

Fig.8 Distribution of real source location and estimated ones in three scenarios

图9 综合预测位置距离偏差指标

Fig.9 Aggregative indicator of TD

参考文献 36

1	Bhattacharjee Y J. Combating terrorism: new efforts to detect explosives require advances on many fronts[J]. Science, 2008, 320(5882): 1416-1417.
2	Efthimiou G C, Andronopoulos S, Tavares R, et al. CFD-Rans prediction of the dispersion of a hazardous airborne material released during a real accident in an industrial environment[J]. Journal of Loss Prevention in the Process Industries, 2017, 46: 23-36.
3	Kumar P, Feiz A A, Ngae P, et al. CFD simulation of short-range plume dispersion from a point release in an urban like environment[J]. Atmospheric Environment, 2015, 122: 645-656.
4	Lin Y, Chen G, Chen T, et al. The influence of advertisement boards, street and source layouts on CO dispersion and building intake fraction in three-dimensional urban-like models[J]. Building and Environment, 2019, 150: 297-321.
5	Dharmavaram S, Hanna S, Hansen O R. Consequence analysis—using a CFD model for industrial sites[J]. Process Safety Progress, 2005, 24(4): 316-272.
6	Mouilleau Y, Champassith A. CFD simulations of atmospheric gas dispersion using the fire dynamics simulator (Fds)[J]. Journal of Loss Prevention in the Process Industries, 2009, 22(3): 316-323.
7	Kumar P, Singh S K, Ngae P, et al. Assessment of a CFD model for short-range plume dispersion: applications to the fusion field trial 2007 (FFT-07) diffusion experiment[J]. Atmospheric Research, 2017, 197: 84-93.
8	Onodera N, Aoki T, Shimokawabe T, et al. Large-scale les wind simulation using lattice Boltzmann method for a 10 km× 10 km area in metropolitan Tokyo[J]. TSUBAME e-Science Journal Global Scientific Information and Computing Center, 2013, 9: 2-8.
9	Hanna S R, Hansen O R, Ichard M, et al. CFD model simulation of dispersion from chlorine railcar releases in industrial and urban areas[J]. Atmospheric Environment, 2009, 43(2): 262-270.
10	Long K J, Zajaczkowski F J, Haupt S E, et al. Modeling a hypothetical chlorine release on a college campus[J]. Journal of Computers, 2009, 4(9): 881-890.
11	Baalisampang T, Abbassi R, Garaniya V, et al. Fire impact assessment in FLNG processing facilities using computational fluid dynamics (CFD)[J]. Fire Safety Journal, 2017, 92: 42-52.
12	Hutchinson M, Oh H, Chen W H J. A review of source term estimation methods for atmospheric dispersion events using static or mobile sensors[J]. Information Fusion, 2017, 36: 130-148.
13	陈增强. 危险化学品泄漏源的定位研究[D]. 北京: 北京化工大学, 2013.
	Chen Z Q. Research on source identification approaches for hazardous chemical releases[D]. Beijing: Beijing University of Chemical Technology, 2013.
14	Lauret P, Heymes F, Aprin L, et al. Atmospheric dispersion modeling using artificial neural network based cellular automata[J]. Environmental Modelling & Software, 2016, 85: 56-69.
15	Wang B, Qian F. Three dimensional gas dispersion modeling using cellular automata and artificial neural network in urban environment[J]. Process Safety and Environmental Protection, 2018, 120: 286-301.
16	Zhou Z, Chen J, Jin G. Prediction of lagrangian dispersion of fluid particles in isotropic turbulent flows using large-eddy simulation method[J]. Acta Mechanica, 2017, 228(9): 3203-3222.
17	Trini C S, Armand P, Tinarelli G, et al. Validation of a lagrangian particle dispersion model with wind tunnel and field experiments in urban environment[J]. Atmospheric Environment, 2018, 193: 273-289.
18	Holmes N S, Morawska L J. A review of dispersion modelling and its application to the dispersion of particles: an overview of different dispersion models available[J]. Atmospheric Environment, 2006, 40(30): 5902-5928.
19	Bosanquet C, Pearson J L. The spread of smoke and gases from chimneys[J]. Transactions of the Faraday Society, 1936, 32: 1249-1263.
20	Green A, Singhal R, Venkateswar R. Analytic extensions of the gaussian plume model[J]. Journal of the Air Pollution Control Association, 1980, 30(7): 773-776.
21	Weil J, Brower R P. An updated Gaussian plume model for tall stacks[J]. Journal of the Air Pollution Control Association, 1984, 34(8): 818-827.
22	Horst T W. A surface depletion model for deposition from a Gaussian Plume[J]. Atmospheric Environment, 1977, 11(1): 41-46.
23	de Visscher A. Air Dispersion Modeling: Foundations and Applications[M]. John Wiley & Sons, 2013: 141-200.
24	Li H, Zhang J, Yi J. Computational source term estimation of the Gaussian puff dispersion[J]. Soft Computing, 2018, 23(1): 59-75.
25	Qiu S, Chen B, Wang R, et al. Atmospheric dispersion prediction and source estimation of hazardous gas using artificial neural network, particle swarm optimization and expectation maximization[J]. Atmospheric Environment, 2018, 178: 158-163.
26	Cui J, Lang J, Chen T, et al. Investigating the impacts of atmospheric diffusion conditions on source parameter identification based on an optimized inverse modelling method[J]. Atmospheric Environment, 2019, 205: 19-29.
27	Efthimiou G C, Kovalets I V, Venetsanos A, et al. An optimized inverse modelling method for determining the location and strength of a point source releasing airborne material in urban environment[J]. Atmospheric Environment, 2017, 170: 118-129.
28	Wang Y, Huang H, Huang L, et al. Source term estimation of hazardous material releases using hybrid genetic algorithm with composite cost functions[J]. Engineering Applications of Artificial Intelligence, 2018, 75: 102-113.
29	Ma D, Tan W, Zhang Z, et al. Parameter identification for continuous point emission source based on tikhonov regularization method coupled with particle swarm optimization algorithm[J]. J. Hazard. Mater., 2017, 325: 239-250.
30	Hutchinson M, Oh H, Chen W H. Entrotaxis as a strategy for autonomous search and source reconstruction in turbulent conditions[J]. Information Fusion, 2018, 42: 179-189.
31	Ma D, Tan W, Wang Q, et al. Location of contaminant emission source in atmosphere based on optimal correlated matching of concentration distribution[J]. Process Safety and Environmental Protection, 2018, 117: 498-510.
32	Cervone G, Franzese P. Monte Carlo source detection of atmospheric emissions and error functions analysis[J]. Computers & Geosciences, 2010, 36(7): 902-909.
33	Hansen N. The CMA Evolution Strategy: A Tutorial[M]. New York: Springer, 2005.
34	Auger A, Hansen N. Tutorial CMA-Es: evolution strategies and covariance matrix adaptation[C]// Conference Companion on Genetic & Evolutionary Computation. New York: Association for Computing Machinery, 2012.
35	Rood A S. Performance evaluation of aermod, calpuff, and legacy air dispersion models using the winter validation tracer study dataset[J]. Atmospheric Environment, 2014, 89: 707-720.
36	Tartakovsky D, Stern E, Broday D M. Comparison of dry deposition estimates of aermod and calpuff from area sources in flat terrain[J]. Atmospheric Environment, 2016, 142: 430-432.

[1]	宋嘉豪, 王文. 斯特林发动机与高温热管耦合运行特性研究[J]. 化工学报, 2023, 74(S1): 287-294.
[2]	连梦雅, 谈莹莹, 王林, 陈枫, 曹艺飞. 地下水预热新风一体化热泵空调系统制热性能研究[J]. 化工学报, 2023, 74(S1): 311-319.
[3]	金伟其, 吴月荣, 王霞, 李力, 裘溯, 袁盼, 王铭赫. 化工园区工业气体泄漏气云红外成像检测技术与国产化装备进展[J]. 化工学报, 2023, 74(S1): 32-44.
[4]	金正浩, 封立杰, 李舒宏. 氨水溶液交叉型再吸收式热泵的能量及分析[J]. 化工学报, 2023, 74(S1): 53-63.
[5]	李科, 文键, 忻碧平. 耦合蒸气冷却屏的真空多层绝热结构对液氢储罐自增压过程的影响机制研究[J]. 化工学报, 2023, 74(9): 3786-3796.
[6]	王浩, 王振雷. 基于自适应谱方法的裂解炉烧焦模型化简策略[J]. 化工学报, 2023, 74(9): 3855-3864.
[7]	于旭东, 李琪, 陈念粗, 杜理, 任思颖, 曾英. 三元体系KCl + CaCl₂ + H₂O 298.2、323.2及348.2 K相平衡研究及计算[J]. 化工学报, 2023, 74(8): 3256-3265.
[8]	诸程瑛, 王振雷. 基于改进深度强化学习的乙烯裂解炉操作优化[J]. 化工学报, 2023, 74(8): 3429-3437.
[9]	闫琳琦, 王振雷. 基于STA-BiLSTM-LightGBM组合模型的多步预测软测量建模[J]. 化工学报, 2023, 74(8): 3407-3418.
[10]	尹刚, 李伊惠, 何飞, 曹文琦, 王民, 颜非亚, 向禹, 卢剑, 罗斌, 卢润廷. 基于KPCA和SVM的铝电解槽漏槽事故预警方法[J]. 化工学报, 2023, 74(8): 3419-3428.
[11]	李锦潼, 邱顺, 孙文寿. 煤浆法烟气脱硫中草酸和紫外线强化煤砷浸出过程[J]. 化工学报, 2023, 74(8): 3522-3532.
[12]	岳林静, 廖艺涵, 薛源, 李雪洁, 李玉星, 刘翠伟. 凹坑缺陷对厚孔板喉部空化流动特性影响研究[J]. 化工学报, 2023, 74(8): 3292-3308.
[13]	郭雨莹, 敬加强, 黄婉妮, 张平, 孙杰, 朱宇, 冯君炫, 陆洪江. 稠油管道水润滑减阻及压降预测模型修正[J]. 化工学报, 2023, 74(7): 2898-2907.
[14]	刘春雨, 周桓宇, 马跃, 岳长涛. CaO调质含油污泥干燥特性及数学模型[J]. 化工学报, 2023, 74(7): 3018-3027.
[15]	李艳辉, 丁邵明, 白周央, 张一楠, 于智红, 邢利梅, 高鹏飞, 王永贞. 非常规服役超临界锅炉的微纳尺度腐蚀动力学模型建立及应用[J]. 化工学报, 2023, 74(6): 2436-2446.