InterPore2022

Fitting correlation-based and neural-network-based relative permeability models to a large dataset of forced and spontaneous imbibition experiments

31 May 2022, 17:15

15m

Oral Presentation (MS15) Machine Learning and Big Data in Porous Media MS15

Speaker

Helton Magno Ciriaco (Technical University of Denmark)

Description

Estimating multiphase flow properties from Special Core scale Analysis (SCAL) has been extensively applied to obtain multiphase flow parameters representing the reservoir scale. Core flooding experimental data can also be used to investigate the mechanisms of more complex multiphase flow systems such as modified salinity water flooding, which has been shown to increase the oil recovery in chalk formations. However, the complexity of physicochemical phenomena cannot be captured by core flooding experiments; thus, the available models and tools which are based on input from core flooding experiments create uncertainties unless the models are tuned to many core flooding experiments, which is expensive and time-consuming. Moreover, the core flooding experiments only cover a narrow range of saturation changes. Spontaneous imbibition tests, which are usually cheaper, can serve as a complementary source of data for obtaining multiphase flow parameters; however, simulating a spontaneous imbibition test requires a model that links the diffusion of ions to the core and to/from the water film between the oil and formation water, how these ions modify the wettability of the system and alter the mobility of the phases. Furthermore, unlike the viscous-force-driven core flooding experiments that can be simulated with a 1D model, the SI test requires a 2D or 3D model which can be computationally expensive when used with an optimization algorithm for parameter estimation. We develop a numerical tool that is capable of estimating multiphase flow parameters for modified salinity water flooding using a combination of core flooding and spontaneous imbibition data. Our model is capable of capturing all the rock-fluid interactions and can be run in a reasonable computational time. First, we address the modelling of the reactive flow problem, focusing on the surface interactions between the brine and chalk formation, and the implications of the physicochemical phenomena that can alter the flow properties. We use surface complexation models to describe the ionic interactions at the oil/brine/chalk interfaces. We also use kinetic and equilibrium models to describe the dissolution of chalk. We use an empirical parameter linking the surface reactions to the relative permeability and capillary pressure model parameters. Then, we discuss a solution to the inverse problem of the model described, which can obtain parameters using a large database of in-house and literature experimental data. Our model can accurately and rapidly estimate the relative permeability and capillary pressure curves by fitting a reactive multiphase flow model to the measurements from core flooding and spontaneous imbibition experiments. Additionally, we replace the correlation-based relative permeability curves with a neural network (NN) in the multiphase reactive flow model to create a physics-informed neural network model. We fit the model to our large database of experimental data and compare the NN-based model with the correlation-based relative permeability curves. Our results indicate that a machine learning tool that combines the governing physics equations and a large set of experimental data can predict with accuracy the multiphase flow properties at a reduced running time that reasonably matches with the expected trend of relative permeability curves.

References

Al-Shalabi, E.W., Sepehrnoori, K., Pope, G. et al. [2015] Mechanistic modeling of oil recovery due to
low salinity water injection in oil reservoirs. In: SPE Middle East Oil & Gas Show and Conference.
Society of Petroleum Engineers, –.
Bonto, M., Eftekhari, A.A. and M. Nick, H. [2020] Wettability Indicator Parameter Based on the Ther-
modynamic Modeling of Chalk-Oil-Brine Systems. Energy & Fuels, 34(7), 8018–8036.
Bonto, M., Eftekhari, A.A. and Nick, H. [2019] A calibrated model for the carbonate-brine-crude oil
surface chemistry and its effect on the rock wettability, dissolution, and mechanical properties. In:
SPE Reservoir Simulation Conference. OnePetro, –.
Dang, C., Nghiem, L., Nguyen, N., Chen, Z. and Nguyen, Q. [2016] Mechanistic modeling of low
salinity water flooding. J. Pet. Sci. Eng., 146, 191–209.
Eftekhari, A.A. [2019] PhreeqcMatlab: a Matlab wrapper for the PhreeqcRM C interface.
Eftekhari, A.A., Schüller, K., Planella, F.B. and Werts, M. [2015] FVTool: a finite volume toolbox for
Matlab.
Eftekhari, A.A., Thomsen, K., Stenby, E.H. and Nick, H.M. [2017] Thermodynamic Analysis of Chalk-
Brine-Oil Interactions. Energy and Fuels, 31(11), 11773–11782.
Evje, S. and Hiorth, A. [2011] A model for interpretation of brine-dependent spontaneous imbibition
experiments. Adv. Water Resour., 34(12), 1627–1642.
Fathi, S.J., Austad, T. and Strand, S. [2010] “Smart water” as a wettability modifier in chalk: the effect
of salinity and ionic composition. Energy & fuels, 24(4), 2514–2519.
Goldberg, S. [2013] Surface Complexation Modeling. In: Reference Module in Earth Systems and
Environmental Sciences, Elsevier, 1–14.
Katende, A. and Sagala, F. [2019] A critical review of low salinity water flooding: mechanism, labora-
tory and field application. Journal of Molecular Liquids, 278, 627–649.
Korrani, A.K. and Jerauld, G.R. [2019] Modeling wettability change in sandstones and carbonates using
a surface-complexation-based method. J. Pet. Sci. Eng., 174(December 2018), 1093–1112.
Ligthelm, D.J., Gronsveld, J., Hofman, J., Brussee, N., Marcelis, F., van der Linde, H. et al. [2009]
Novel waterflooding strategy by manipulation of injection brine composition. In: EUROPEC/EAGE
conference and exhibition. Society of Petroleum Engineers, –.
Loeve, D., Wilschut, F., Hanea, R.H., Maas, J., Hooff, P.M.E., den Hoek, P., Douma, S.G. and Doren,
J.F.M. [2011] Simultaneous Determination of Relative Permeability and Capillary Pressure Curves by
Assisted History Matching Several SCAL Experiments. In: Soc. Core Anal. Conf. Pap. SCA2011-35.
1–12.
Nia Korrani, A.K., Sepehrnoori, K. and Delshad, M. [2013] A novel mechanistic approach for modeling
low salinity water injection. SPE Annu. Tech. Conf. Proc., 7(1999), 3206–3223.
Nick, H.M., Raoof, A., Centler, F., Thullner, M. and Regnier, P. [2013] Reactive dispersive contaminant
transport in coastal aquifers: Numerical simulation of a reactive Henry problem. J. Contam. Hydrol.,
145, 90–104.
Parkhurst, D.L. and Wissmeier, L. [2015] PhreeqcRM: A reaction module for transport simulators based
on the geochemical model PHREEQC. Advances in Water Resources, 83, 176–189.
Qiao, C., Johns, R., Li, L. and Xu, J. [2015] Modeling low salinity waterflooding in mineralogically
different carbonates. Proc. - SPE Annu. Tech. Conf. Exhib., 2015-January, 4168–4187.
RezaeiDoust, A., Puntervold, T., Strand, S. and Austad, T. [2009] Smart water as wettability modifier in
carbonate and sandstone: A discussion of similarities/differences in the chemical mechanisms. Energy
& fuels, 23(9), 4479–4485.
Saw, R.K. and Mandal, A. [2020] A mechanistic investigation of low salinity water flooding coupled
with ion tuning for enhanced oil recovery. RSC Advances, 10(69), 42570–42583.
Seccombe, J., Lager, A., Webb, K., Jerauld, G. and Fueg, E. [2008] Improving waterflood recovery:
LoSal EOR field evaluation, SPE. In: DOE Symposium on Improved Oil Recovery (Tulsa, Oklahoma,
USA), SPE, 113480. –.
Song, J., Zeng, Y., Wang, L., Duan, X., Puerto, M., Chapman, W.G., Biswal, S.L. and Hirasaki, G.J.
[2017] Surface complexation modeling of calcite zeta potential measurements in brines with mixed
potential determining ions (Ca2+, CO2−3 , Mg2+, SO2−4 ) for characterizing carbonate wettability. J.
Colloid Interface Sci., 506, 169–179.
Sun, Y., Petersen, J.N. and Clement, T.P. [1999] Analytical solutions for multiple species reactive trans-
port in multiple dimensions. Journal of Contaminant Hydrology, 35(4), 429–440.
Taheri, M., Bonto, M., Eftekhari, A.A., M Nick, H. et al. [2019] Towards Identifying the Mechanisms
of the Modified-Salinity Waterflooding by a Novel Combination of Core flooding and Mathematical
Modeling. In: SPE Middle East Oil and Gas Show and Conference. Society of Petroleum Engineers,
1–30.
Taheriotaghsara, M., Bonto, M., Eftekhari, A.A. and Nick, H.M. [2020] Prediction of oil breakthrough
time in modified salinity water flooding in carbonate cores. Fuel, 274(January), 117806.
Webb, K., Black, C., and Al-Ajeel, H. [2004] Low salinity oil recovery-log-inject-log. In: SPE/DOE
Symposium on Improved Oil Recovery. OnePetro, –.
Yu, L., Evje, S., Kleppe, H., Kårstad, T., Fjelde, I. and Skjaeveland, S.M. [2009] Spontaneous imbibition
of seawater into preferentially oil-wet chalk cores - Experiments and simulations. J. Pet. Sci. Eng.,
66(3-4), 171–179.
u, L., Kleppe, H., Kaarstad, T., Skjaeveland, S.M., Evje, S. and Fjelde, I. [2008] Modelling of wetta-
bility alteration processes in carbonate oil reservoirs. Networks Heterog. Media, 3(1), 149–183.
Zhang, P. and Austad, T. [2006] Wettability and oil recovery from carbonates: Effects of temperature
and potential determining ions. Colloids Surfaces A Physicochem. Eng. Asp., 279(1-3), 179–187.
Zhang, P., Tweheyo, M.T. and Austad, T. [2007] Wettability alteration and improved oil recovery by
spontaneous imbibition of seawater into chalk: Impact of the potential determining ions Ca2+, Mg2+,
and SO42-. Colloids Surfaces A Physicochem. Eng. Asp., 301(1-3), 199–208.
Gasmi, Cedric Fraces, and Hamdi Tchelepi. 2021. “Physics Informed Deep Learning for Flow and Transport in Porous Media,” April. http://arxiv.org/abs/2104.02629.
Fuks, Olga, and Hamdi A Tchelepi. 2020. “LIMITATIONS OF PHYSICS INFORMED MACHINE LEARNING FOR NONLINEAR TWO-PHASE TRANSPORT IN POROUS MEDIA.” Journal of Machine Learning for Modeling and Computing. Vol. 1. www.begellhouse.com.
Kadeethum, Teeratorn, Thomas M. Jørgensen, and Hamidreza M. Nick. 2020. “Physics-Informed Neural Networks for Solving Nonlinear Diffusivity and Biot’s Equations.” PLoS ONE 15 (5). https://doi.org/10.1371/journal.pone.0232683.
Kadeethum, Teeratorn, Thomas M. Jørgensen, and Hamidreza M. Nick. 2020. “Physics-Informed Neural Networks for Solving Inverse Problems of Nonlinear Biot’s Equations: Batch Training.” ArXiv. http://onepetro.org/ARMAUSRMS/proceedings-pdf/ARMA20/All-ARMA20/ARMA-2020-1134/2255578/arma-2020-1134.pdf/1.
Almajid, Muhammad Majid, and Moataz Omar Abu-Alsaud. 2020. “Prediction of Fluid Flow in Porous Media Using Physics Informed Neural Networks.” In Society of Petroleum Engineers - Abu Dhabi International Petroleum Exhibition and Conference 2020, ADIP 2020. Society of Petroleum Engineers. https://doi.org/10.2118/203033-ms.