基于BP人工神经网络预测地热井中流体的结垢位置

doi:10.16085/j.issn.1000-6613.2022-0216

摘要/Abstract

摘要：

地热井筒中常存在因地热流体结垢而导致的生产能力下降甚至无法生产的问题，因此研究地热流体在井筒中的结垢位置等行为具有重要的应用价值。人工神经网络（ANNs）可用于开发预测地热井筒中结垢位置新模型。由于其没有机理建模的性质，故只可作为一种新的代理模型。本文以地热流体在井口和井底的温度、压力以及井深等参数作为输入变量，成功训练了三层ANNs结构，以小于10%的相对误差实现了ANNs代理模型的合适精度。对ANNs代理模型预测的结垢位置进行了分析，并与现场测量的井筒结垢位置进行了比较，分析了产生误差的原因。结果表明，新建的ANNs代理模型可作为一种实用工具，能够可靠地预测地热流体在井筒中的结垢位置。

关键词: 地热流体, 结垢, 神经网络, 预测, 井筒, 沉淀

Abstract:

There are often problems in geothermal wellbore that its production capacity of geothermal fluids decreases or even cannot be produced due to the scaling of geothermal fluids in it. Therefore, it is particularly important to study the scaling behavior, such as to predict the scaling location of geothermal fluid in the wellbore. Artificial neural networks (ANNs) can be used to develop a new model for predicting the scaling location in geothermal wellbore. Because it has no nature of mechanism modeling, it can only be used as a new proxy model. The three-layer ANNs structure was successfully trained using the temperature, pressure and well depth data of the geothermal fluids at the wellhead and bottomhole as input variables, and the appropriate accuracy of the ANNs proxy model was achieved with a relative error of less than 10%. The scaling locations predicted by the ANNs proxy model were analyzed and compared with the actual measured wellbore scaling locations, and the reasons for the errors were analyzed. The results showed that the new ANNs proxy model could serve as a practical tool to reliably predict scaling locations in geothermal wellbores.

Key words: geothermal fluids, scaling, neural networks, prediction, wellbore, precipitation

中图分类号:

TK521

李帅, 刘明言, 马永丽. 基于BP人工神经网络预测地热井中流体的结垢位置[J]. 化工进展, 2022, 41(11): 5761-5770.

LI Shuai, LIU Mingyan, MA Yongli. Prediction of scaling location of fluid in geothermal well based on BP artificial neural network[J]. Chemical Industry and Engineering Progress, 2022, 41(11): 5761-5770.

图/表 12

参考文献 38

1	《中国地热能发展报告(2018)》白皮书发布[J]. 地质装备, 2019, 20(2): 3-6.
	China geothermal energy development report (2018) [J]. Equipment for Geotechnical Engineering, 2019, 20(2): 3-6.
2	蔡义汉. 地热直接利用[M]. 天津: 天津大学出版社, 2004.
	CAI Yihan. Geothermal direct-use[M]. Tianjin: Tianjin University Press, 2004.
3	ODDO J E, TOMSON M B. Why scale forms in the oil field and methods to predict it[J]. SPE Production & Facilities, 1994, 9(1): 47-54.
4	BOCH Ronny, LEIS Albrecht, HASLINGER Edith, et al. Scale-fragment formation impairing geothermal energy production: interacting H₂S corrosion and CaCO₃ crystal growth[J]. Geothermal Energy, 2017, 5(1): 1-19.
5	刘明言. 地热流体的腐蚀与结垢控制现状[J]. 新能源进展, 2015, 3(1): 38-46.
	LIU Mingyan. A review on controls of corrosion and scaling in geothermal fluids[J]. Advances in New and Renewable Energy, 2015, 3(1): 38-46.
6	刘明言, 朱家玲. 地热能利用中的防腐防垢研究进展[J]. 化工进展, 2011, 30(5): 1120-1123.
	LIU Mingyan, ZHU Jialing. Progress of corrosion and fouling prevention in utilization of geothermal energy[J]. Chemical Industry and Engineering Progress, 2011, 30(5): 1120-1123.
7	TOPCU Gokhan, KOÇ Gonca A, BABA Alper, et al. The injection of CO₂ to hypersaline geothermal brine: a case study for Tuzla region[J]. Geothermics, 2019, 80: 86-91.
8	LEDÉSERT Béatrice A, HÉBERT Ronan L, MOUCHOT Justine, et al. Scaling in a geothermal heat exchanger at soultz-sous-forêts (upper Rhine graben, France): a XRD and SEM-EDS characterization of sulfide precipitates[J]. Geosciences, 2021, 11(7): 271.
9	李义曼, 庞忠和. 地热系统碳酸钙垢形成原因及定量化评价[J]. 新能源进展, 2018, 6(4): 274-281.
	LI Yiman, PANG Zhonghe. Carbonate calcium scale formation and quantitative assessment in geothermal system[J]. Advances in New and Renewable Energy, 2018, 6(4): 274-281.
10	蔡正敏, 李刚, 李源, 等. 肯尼亚地热电站结垢问题的日常维护[J]. 科技视界, 2018(25): 41-43.
	CAI Zhengmin, LI Gang, LI Yuan, et al. Maintenance of geothermal power plant in Kenya[J]. Science & Technology Vision, 2018(25): 41-43.
11	SATMAN Abdurrahman, UGUR Zuleyha, ONUR Mustafa. The effect of calcite deposition on geothermal well inflow performance[J]. Geothermics, 1999, 28(3): 425-444.
12	Gabriella STÁHL, György PÁTZAY, László WEISER, et al. Study of calcite scaling and corrosion processes in geothermal systems[J]. Geothermics, 2000, 29(1): 105-119.
13	SPINTHAKI Argyro, MATHEIS Juergen, HATER Wolfgang, et al. Antiscalant-driven inhibition and stabilization of “magnesium silicate” under geothermal stresses: the role of magnesium-phosphonate coordination chemistry[J]. Energy & Fuels, 2018, 32(11): 11749-11760.
14	Stefán ARNÓRSSON, SIGURDSSON Sven, Hördur SVAVARSSON. The chemistry of geothermal waters in Iceland. Ⅰ‍. Calculation of aqueous speciation from 0° to 370℃[J]. Geochimica et Cosmochimica Acta, 1982, 46(9): 1513-1532.
15	Gültekin TARCAN, Tuğbanur ÖZEN, Ünsal GEMICI, et al. Geochemical assessment of mineral scaling in Kzldere geothermal field, Turkey[J]. Environmental Earth Sciences, 2016, 75(19): 1317-1335.
16	REED M H, SPYCHER N F, PALANDRI J. Users guide for CHIM-XPT: a program for computing reaction processes in aqueous-mineral-gas systems and MINTAB guide[M]. version 2.43. Eugene, Oregon: University of Oregon, 2012.
17	Stefán ARNÓRSSON. Deposition of calcium carbonate minerals from geothermal waters—Theoretical considerations[J]. Geothermics, 1989, 18(1/2): 33-39.
18	O'SULLIVAN Michael J, PRUESS Karsten, LIPPMANN Marcelo J. State of the art of geothermal reservoir simulation[J]. Geothermics, 2001, 30(4): 395-429.
19	BÄCHLER D, KOHL T. Coupled thermal-hydraulic-chemical modelling of enhanced geothermal systems[J]. Geophysical Journal International, 2005, 161(2): 533-548.
20	RYLEY D J. The mass discharge of a geofluid from a geothermal reservoir—Well system with flashing flow in the bore[J]. Geothermics, 1980, 9(3/4): 221-235.
21	CHADHA P K, MALIN M R, PALACIO-PEREZ A. Modelling of two-phase flow inside geothermal wells[J]. Applied Mathematical Modelling, 1993, 17(5): 236-245.
22	BARELLI A, CORSI R, Del PIZZO G, et al. A two-phase flow model for geothermal wells in the presence of non-condensable gas[J]. Geothermics, 1982, 11(3): 175-191.
23	BANKOFF S G. A variable density single-fluid model for two-phase flow with particular reference to steam-water flow[J]. Journal of Heat Transfer, 1960, 82(4): 265-272.
24	PÁTZAY G, STÁHL G, KÁRMÁN F H, et al. Modeling of scale formation and corrosion from geothermal water[J]. Electrochimica Acta, 1998, 43(1/2): 137-147.
25	AKIN Taylan, Aygün GÜNEY, KARGI Hulusi. Modeling of calcite scaling and estimation of gas breakout depth in a geothermal well by using PHREEQC[C]//Fortieth Workshop on Geothermal Reservoir Engineering. Stanford University, Stanford, California, 2015:1-8.
26	HAIZLIP Jill, Aygün GÜNEY, HAKLIDIR Fusun Servin Tut, et al. The impact of high noncondensible gas concentrations on well performance Kizildere geothermal reservoir, TURKEY[C]// Thirty-Seventh Workshop on Geothermal Reservoir Engineering. Stanford University, Stanford, California 2013:1-6.
27	BJORNSSON G. A multi-feedzone geothermal wellbore simulator[R]. Office of Scientific and Technical Information (OSTI), 1987.
28	GUNN Calum, FREESTON Derek. An integrated steady-state wellbore simulation and analysis package[C]//Proceedings of the 13th New Zealand Geothermal Workshop. Auckland, NZ: New Zealand Geothermal Workshop,1991: 161-166.
29	GARG Sabodh K, PRITCHETT John W, ALEXANDER James H. A new liquid hold-up correlation for geothermal wells[J]. Geothermics, 2004, 33(6): 795-817.
30	王龙洋, 蒙西, 乔俊飞. 基于改进集合经验模态分解和深度信念网络的出水总磷预测[J]. 化工学报, 2021(5): 2745-2753.
	WANG Longyang, MENG Xi, QIAO Junfei. Prediction of effluent total phosphorus based on modified ensemble empirical mode decomposition and deep belief network[J]. CIESC Journal, 2021(5): 2745-2753.
31	ZHANG Guoqiang, EDDY Patuwo B, HU Michael Y. Forecasting with artificial neural networks[J]. International Journal of Forecasting, 1998, 14(1): 35-62.
32	AYDIN Hakki, AKIN Serhat, SENTURK Erdinc. A proxy model for determining reservoir pressure and temperature for geothermal wells[J]. Geothermics, 2020, 88: 101916.
33	A Álvarez del CASTILLO, SANTOYO E, GARCÍA-VALLADARES O. Α new void fraction correlation inferred from artificial neural networks for modeling two-phase flows in geothermal wells[J]. Computers & Geosciences, 2012, 41: 25-39.
34	赵黎丽. 两种人工智能方法应用于地热热泵系统辨识[J]. 系统仿真学报, 2004, 16(7): 1376-1379.
	ZHAO Lili. Two artificial intelligent methods applied in the identification of geothermal heat pump system[J]. Acta Simulata Systematica Sinica, 2004, 16(7): 1376-1379.
35	梁海军, 郭啸峰, 高涛, 等. 河北博野某地热井结垢位置预测及影响因素分析[J]. 石油钻探技术, 2020, 48(5): 105-110.
	LIANG Haijun, GUO Xiaofeng, GAO Tao, et al. Scaling spot prediction and analysis of influencing factors for a geothermal well in Boye County, Hebei Province[J]. Petroleum Drilling Techniques, 2020, 48(5): 105-110.
36	BASSAM A, ORTEGA-TOLEDO D, HERNANDEZ J A, et al. Artificial neural network for the evaluation of CO₂ corrosion in a pipeline steel[J]. Journal of Solid State Electrochemistry, 2009, 13(5): 773-780.
37	张冰, 唐和礼, 黄冬梅, 等. 人工神经网络及智能算法在膜污染研究中的应用[J]. 膜科学与技术, 2021, 41(4): 160-169.
	ZHANG Bing, TANG Heli, HUANG Dongmei, et al. Applications of artificial neural networks and intelligent algorithms in the research of membrane fouling: a critical review[J]. Membrane Science and Technology, 2021, 41(4): 160-169.
38	张驰, 郭媛, 黎明. 人工神经网络模型发展及应用综述[J]. 计算机工程与应用, 2021, 57(11): 57-69.
	ZHANG Chi, GUO Yuan, LI Ming. Review of development and application of artificial neural network models[J]. Computer Engineering and Applications, 2021, 57(11): 57-69.

井的开度	井深/m	套管直径/mm	壁厚/mm
一开	550.00	406.0	9.65
二开	1950.00	273.0	8.94
三开	3275.00	177.8	9.19
四开	3860.00	152.0	—

井的开度	井深/m	套管直径/mm	壁厚/mm
一开	550.00	406.0	9.65
二开	1950.00	273.0	8.94
三开	3275.00	177.8	9.19
四开	3860.00	152.0	—

算法	函数
梯度下降法	traingd
有动量的梯度下降法	traingdm
自适应lr梯度下降法	traingda
自适应lr动量梯度下降法	traingdx
弹性梯度下降法	trainrp
Fletcher-Reeves共轭梯度法	traincgf
Ploak-Ribiere共轭梯度法	traincgp
Powell-Beale共轭梯度法	traincgb
量化共轭梯度法	trainscg
拟牛顿算法	trainbfg
一步正割算法	trainoss
Levenberg-Marquardt	trainlm

算法	函数
梯度下降法	traingd
有动量的梯度下降法	traingdm
自适应lr梯度下降法	traingda
自适应lr动量梯度下降法	traingdx
弹性梯度下降法	trainrp
Fletcher-Reeves共轭梯度法	traincgf
Ploak-Ribiere共轭梯度法	traincgp
Powell-Beale共轭梯度法	traincgb
量化共轭梯度法	trainscg
拟牛顿算法	trainbfg
一步正割算法	trainoss
Levenberg-Marquardt	trainlm

序号	预测结果/m	实际结果/m	绝对误差	相对误差/%
1	1.580	1.500	0.080	5.06
2	33.069	30.000	3.069	9.28