FAST AND ACCURATE PHASE EQUILIBRIUM CALCULATIONS FOR CONDENSATE SHALE GAS RESERVOIRS
-
摘要:对页岩油气藏中复杂流体的相平衡计算需要建立考虑毛细作用效应的先进的数值模型, 并设计出快速可靠的算法以应对实际工况中储层流体包含多达数十种组分的复杂情况. 本文将基于适合页岩油气藏常见组分的真实流体状态方程, 即Peng−Robinson状态方程构建具有热力学一致性的VT型孔观相平衡计算体系. 通过引入描述毛细压力做功的数学模型实现对页岩流体热力学性质更准确的刻画. 结合扩散界面模型建立动力学演化格式, 采用成熟的凸分裂方法求解摩尔数和体积分数的演变, 从而描述相平衡的动态过程. 在此基础上, 本文开发了一套具有自适应性的深度学习算法, 设计了独特的双网络结构以实现对不同流体中不同组分的广泛适用性. 该神经网络的输入和输出参数均在热力学分析的基础上选取关键的热力学性质参数, 并进行了全面的超参调试以确定最合适的网络架构和最后形成的预测模型的基本结构, 且通过多种深度学习技术解决了过拟合问题, 在显著加速了传统的基于迭代方法的闪蒸计算的同时保证了相平衡状态预测的准确性, 得到了较好的预测效果. 相分离判定自动整合在预测结果中, 且从最终预测结果可以显著地捕捉到毛细作用的影响. 这一套快速、准确、可靠地基于深度学习算法的页岩油气孔观相平衡计算体系可以为后续的多相流动模拟提供具有物理意义的相分布初场, 确定系统内各个阶段的相数, 并可以作为构建具有物理守恒性的多相数值模型的热力学基础.Abstract:Phase equilibrium calculations of complex fluids in shale gas reservoirs require the establishment of advanced numerical models that consider capillary effects, and the design of fast and reliable algorithms to handle the various components in the reservoir fluids in practical working conditions. In this study, we develop a thermodynamically consistent VT-type pore-scale flash calculation scheme based on realistic equations of state suitable for oil/gas reservoirs, e.g. the Peng-Robinson equation of state. The effect of capillarity has been incorporated in the scheme for a more accurate description of the thermodynamic properties of shale gas, and the diffuse interface model is applied to establish a dynamic evolution scheme in the phase equilibrium process, and a convex splitting method is used to model the evolution of compositional moles and volume. In order to accelerate the iterative flash calculations for realistic reservoir fluids containing a large number of components, a self-adaptive deep learning algorithm is developed in this paper with a novel structure to achieve wider applicability to various components in different fluids. The input and output features of the neural network are selected as the key thermodynamic features on the basis of thermodynamic analysis, and the network hyper-parameters have been carefully tuned to achieve a better performance on both accuracy and efficiency. Advanced deep learning technics resolving overfitting problems have been applied in our algorithm. The trained model significantly accelerates the conventional flash calculation based on iterative methods, while a good prediction accuracy has been preserved. Phase stability test and phase splitting calculations are automatically incorporated in our prediction, and we can significantly capture the effect of capillarity on phase equilibrium behaviors. Such a fast, accurate and reliable shale gas phase equilibrium calculation scheme using deep learning algorithms can provide an initial phase distribution field with physical meanings for subsequent multiphase flow simulations, while the number of phases can be also determined. The thermodynamic information and analysis can also be used as a thermodynamic basis for a multiphase numerical model with built-in physical conservation.
-
图 6Bakken储层流体在60 mol/m3摩尔浓度时平衡态相数随温度的改变
Figure 6.Number of phases existing at equilibrium for Bakken reservoir fluids under constant overall mole concentration as 60 mol/m3
图 7Bakken储层流体在60 mol/m3摩尔浓度时甲烷组分在气相的摩尔分数在平衡态随温度的改变
Figure 7.Mole fraction of C1 component in the vapor phase at equilibrium for Bakken reservoir fluids changing with temperature and under constant overall mole concentration as 60 mol/m3
图 814组分Eagle Ford储层流体在343 mol/m3摩尔浓度时甲烷和庚烷组分在气相的摩尔分数在平衡态随温度的改变
Figure 8.Mole fraction of C1, C7components in the vapor phase at equilibrium for 14-component Eagle Ford reservoir fluids changing with temperature and under constant overall mole concentration 343 mol/m3
表 1Bakken储藏流体性质数据
Table 1.Fluid properties of Bakken reservoir
Component ${ {\textit{z}} }_{,i}$ ${ {T} }_{{\rm{c}},i}$/K ${ {P} }_{{\rm{c}},i}$/MPa $ {\omega }_{i} $ C1 0.250 6 190.606 4.600 0.008 C2~ C4 0.220 0 363.30 4.310 0.143 C5~ C7 0.200 0 511.56 3.421 0.247 C8~ C9 0.130 0 579.34 3.132 0.286 C10+ 0.199 4 788.74 2.187 0.687 
下载:
导出CSV
表 2深度学习算法的表现
Table 2.Performance of deep learning algorithm
Fluid tflash/s tdl/s $ \mathrm{\varepsilon } $ with capillarity 2214 7.8 0.086 without capillarity 2015 7.5 0.091 下载:
导出CSV
-
[1] 刘建亮, 杨涵, 王小彩. 天然气在英国能源转型中的作用及启示. 国际石油经济, 2021, 29(4): 74-82 (Liu Jianliang, Yang Han, Wang Xiaocai. Effect of natural gas in the transformation of UK energy and the inspiration.International Petroleum Economics, 2021, 29(4): 74-82 (in Chinese)doi:10.3969/j.issn.1004-7298.2021.04.010 [2] 赵文智, 贾爱林, 位云生等. 中国页岩油气勘探开发进展及发展展望. 中国石油勘探, 2020, 25(1): 31 (Zhao Wenzhi, Jia Ailin, Wei Yunsheng, et al. Progress in shale gas exploration in China and prospects for future development.China Petroleum Exploration, 2020, 25(1): 31 (in Chinese)doi:10.3969/j.issn.1672-7703.2020.01.004 [3] 何骁, 李武广, 党录瑞等. 深层页岩油气开发关键技术难点与攻关方向. 天然气工业, 2021, 41(1): 118-124 (He Xiao, Li Wuguang, Dang Lurui, et al. Key technological challenges and research directions of deep shale gas development.Natural Gas Industry, 2021, 41(1): 118-124 (in Chinese) [4] Feng Q, Xu S, Xing X, et al. Advances and challenges in shale oil development: A critical review.Advances in Geo-Energy Research, 2020, 4(4): 406-418doi:10.46690/ager.2020.04.06 [5] Eren T, Polat C. Natural gas underground storage and oil recovery with horizontal wells.Journal of Petroleum Science and Engineering, 2020, 187: 106753doi:10.1016/j.petrol.2019.106753 [6] Zhao H, Jing H, Fang Z, et al. Flash calculation using successive substitution accelerated by the general dominant eigenvalue method in reduced-variable space: comparison and new insights.SPE Journal, 2020, 25(6): 3332-3348doi:10.2118/202472-PA [7] Li Y, Zhang T, Sun S, et al. Accelerating flash calculation through deep learning methods.Journal of Computational Physics, 2019, 394: 153-165doi:10.1016/j.jcp.2019.05.028 [8] Zhang T, Li Y, Sun S, et al. Accelerating flash calculations in unconventional reservoirs considering capillary pressure using an optimized deep learning algorithm.Journal of Petroleum Science and Engineering, 2020, 195: 107886doi:10.1016/j.petrol.2020.107886 [9] de Azevedo Medeiros F, Stenby EH, Yan W. State function‐based flash specifications for open systems in the absence or presence of chemical reactions.AIChE Journal, 2021, 67(1): e17050 [10] Li Y, Qiao Z, Sun S, et al. Thermodynamic modeling of CO2solubility in saline water using NVT flash with the cubic-Plus-association equation of state.Fluid Phase Equilibria, 2020, 520: 112657doi:10.1016/j.fluid.2020.112657 [11] Zhang T, Li Y, Sun S. Phase equilibrium calculations in shale gas reservoirs.Capillarity, 2019, 2(1): 8-16doi:10.26804/capi.2019.01.02 [12] Rauter MT, Galteland O, Erdős M, et al. Two-phase equilibrium conditions in nanopores.Nanomaterials, 2020, 10(4): 608doi:10.3390/nano10040608 [13] Xiong W, Zhao YL, Qin JH, et al. Phase equilibrium modeling for confined fluids in nanopores using an association equation of state.The Journal of Supercritical Fluids, 2021, 169: 105118doi:10.1016/j.supflu.2020.105118 [14] Abutaqiya MIL, Sisco CJ, Khemka Y, et al. Accurate Modeling of Asphaltene Onset Pressure in Crude Oils Under Gas Injection Using Peng–Robinson Equation of State.Energy&Fuels, 2020, 34(4): 4055-4070 [15] Epelle EI, Bennett J, Abbas H, et al. Correlation of binary interaction coefficients for hydrate inhibition using the Soave-Redlich-Kwong Equation of State and the Huron-Vidal mixing rule.Journal of Natural Gas Science and Engineering, 2020, 77: 103259doi:10.1016/j.jngse.2020.103259 [16] Whitson CH, Brulé MR. Phase behavior. Monograph series, Society of Petroleum Engineers, Richardson, 20, 2000. [17] Michelsen ML. Simplified flash calculations for cubic equations of state.Industrial&Engineering Chemistry Process Design and Development, 1986, 25(1): 184-188 [18] Jensen BH, Fredenslund A. A simplified flash procedure for multicomponent mixtures containing hydrocarbons and one non-hydrocarbon using two-parameter cubic equations of state.Industrial&Engineering Chemistry Research, 1987, 26(10): 2129-2134 [19] Hendriks EM. Reduction theorem for phase equilibrium problems.Industrial&Engineering Chemistry Research, 1988, 27(9): 1728-1732 [20] Hendriks EM, van Bergen ARD. Application of a reduction method to phase equilibria calculations.Fluid Phase Equilibria, 1992, 74: 17-34doi:10.1016/0378-3812(92)85050-I [21] Firoozabadi A, Pan H. Fast and robust algorithm for compositional modeling: Part i-stability analysis testing.SPE Annual Technical Conference and Exhibition. OnePetro, 2000. [22] Li Y, Johns RT. Rapid flash calculations for compositional simulation.SPE Reservoir Evaluation&Engineering, 2006, 9(5): 521-529 [23] Nichita DV, Petitfrere M. Phase stability analysis using a reduction method.Fluid Phase Equilibria, 2013, 358: 27-39doi:10.1016/j.fluid.2013.08.006 [24] Gaganis V. Rapid phase stability calculations in fluid flow simulation using simple discriminating functions.Computers&Chemical Engineering, 2018, 108: 112-127 [25] Petitfrere M, Nichita DV. A comparison of conventional and reduction approaches for phase equilibrium calculations.Fluid Phase Equilibria, 2015, 386: 30-46doi:10.1016/j.fluid.2014.11.017 [26] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks.Advances in Neural Information Processing Systems, 2012, 25: 1097-1105 [27] Ren S, He K, Girshick R, et al. Faster r-cnn: Towards real-time object detection with region proposal networks.Advances in Neural Information Processing Systems, 2015, 28: 91-99 [28] Jung YH, Hong SK, Wang HS, et al. Flexible piezoelectric acoustic sensors and machine learning for speech processing.Advanced Materials, 2020, 32(35): 1904020doi:10.1002/adma.201904020 [29] Font JM, Mahlmann T. Dota 2 bot competition.IEEE Transactions on Games, 2018, 11(3): 285-289 [30] Vasilyeva M, Tyrylgin A. Machine learning for accelerating effective property prediction for poroelasticity problem in stochastic media.ArXiv PreprintArXiv: 1810.01586, 2018. [31] Dang C, Nghiem L, Fedutenko E, et al. AI based mechanistic modeling and probabilistic forecasting of hybrid low salinity chemical flooding.Fuel, 2020, 261: 116445doi:10.1016/j.fuel.2019.116445 [32] Li Y, Zhang T, Sun S. Acceleration of the NVT flash calculation for multicomponent mixtures using deep neural network models.Industrial&Engineering Chemistry Research, 2019, 58(27): 12312-12322 [33] Zhang T, Li Y, Li Y, et al. A self-adaptive deep learning algorithm for accelerating multi-component flash calculation.Computer Methods in Applied Mechanics and Engineering, 2020, 369: 113207doi:10.1016/j.cma.2020.113207 [34] Shen Y, Ge H, Zhang X, et al. Impact of fracturing liquid absorption on the production and water-block unlocking for shale gas reservoir.Advances in Geo-Energy Research, 2018, 2(2): 163-172doi:10.26804/ager.2018.02.05 -
下载:
点击查看大图
图(8)/
表(2)
计量
- 文章访问数:733
- HTML全文浏览量:232
- PDF下载量:127
- 被引次数:0
