Physics-Aware Deep Learning on Multiphase Flow Problems

A physical description and the operating characteristics for a multiphase flow test facility are given. The facility is designed for wet-gas conditions where the gas-void-fraction (GVF) is typically greater than 0.95. However under many conditions, the liquid flowrate can be increased which results in a lower GVF. Lean natural gas, whose typical energy content is less than 1100 BTU/ft3, is used as the flowing gaseous media. The flowing liquid can range from a pure hydrocarbon liquid (such as decane) to a mixture of water and hydrocarbon liquids (condensate). Several investigations into the performance of various single-phase flowmeters and gas-liquid separators have been conducted for wet-gas flowing conditions. Present work includes the modification of the test facility to study hydrate formation and methods that can be employed to inhibit the hydrate formation. Visual images obtained with a high-pressure viewing section will be presented which show the different flow patterns that can exist within pipes that are contain multiphase fluids.

Download Full-text

Impact of solutal Marangoni convection on oil recovery during chemical flooding

Petroleum Science ◽

10.1007/s12182-020-00451-z ◽

2020 ◽

Vol 17 (5) ◽

pp. 1298-1317

Author(s):

Sepideh Palizdan ◽

Jassem Abbasi ◽

Masoud Riazi ◽

Mohammad Reza Malayeri

Keyword(s):

Multiphase Flow ◽

Oil Recovery ◽

Marangoni Convection ◽

Alkaline Solutions ◽

Dimensionless Number ◽

Chemical Flooding ◽

Spontaneous Movement ◽

Static Tests ◽

Flow Problems ◽

Marangoni Effects

Abstract In this study, the impacts of solutal Marangoni phenomenon on multiphase flow in static and micromodel geometries have experimentally been studied and the interactions between oil droplet and two different alkaline solutions (i.e. MgSO4 and Na2CO3) were investigated. The static tests revealed that the Marangoni convection exists in the presence of the alkaline and oil which should carefully be considered in porous media. In the micromodel experiments, observations showed that in the MgSO4 flooding, the fluids stayed almost stationary, while in the Na2CO3 flooding, a spontaneous movement was detected. The changes in the distribution of fluids showed that the circular movement of fluids due to the Marangoni effects can be effective in draining of the unswept regions. The dimensional analysis for possible mechanisms showed that the viscous, gravity and diffusion forces were negligible and the other mechanisms such as capillary and Marangoni effects should be considered in the investigated experiments. The value of the new defined Marangoni/capillary dimensionless number for the Na2CO3 solution was orders of magnitude larger than the MgSO4 flooding scenario which explains the differences between the two cases and also between different micromodel regions. In conclusion, the Marangoni convection is activated by creating an ultra-low IFT condition in multiphase flow problems that can be profoundly effective in increasing the phase mixing and microscopic efficiency.

Download Full-text

Prediction of Discretization of GMsFEM Using Deep Learning

Mathematics ◽

10.3390/math7050412 ◽

2019 ◽

Vol 7 (5) ◽

pp. 412 ◽

Cited By ~ 3

Author(s):

Min Wang ◽

Siu Wun Cheung ◽

Eric T. Chung ◽

Yalchin Efendiev ◽

Wing Tat Leung ◽

...

Keyword(s):

Neural Networks ◽

Porous Media ◽

Deep Learning ◽

Multiscale Model ◽

Porous Media Flow ◽

Scale Parameters ◽

Flow Problems ◽

Local Problems ◽

Nonlinear Relation ◽

Model Reduction Technique

In this paper, we propose a deep-learning-based approach to a class of multiscale problems. The generalized multiscale finite element method (GMsFEM) has been proven successful as a model reduction technique of flow problems in heterogeneous and high-contrast porous media. The key ingredients of GMsFEM include mutlsicale basis functions and coarse-scale parameters, which are obtained from solving local problems in each coarse neighborhood. Given a fixed medium, these quantities are precomputed by solving local problems in an offline stage, and result in a reduced-order model. However, these quantities have to be re-computed in case of varying media (various permeability fields). The objective of our work is to use deep learning techniques to mimic the nonlinear relation between the permeability field and the GMsFEM discretizations, and use neural networks to perform fast computation of GMsFEM ingredients repeatedly for a class of media. We provide numerical experiments to investigate the predictive power of neural networks and the usefulness of the resultant multiscale model in solving channelized porous media flow problems.

Download Full-text

Intelligent Production Monitoring with Continuous Deep Learning Models

10.2118/206525-ms ◽

2021 ◽

Author(s):

Anton Gryzlov ◽

Sergey Safonov ◽

Muhammad Arsalan

Keyword(s):

Machine Learning ◽

Deep Learning ◽

Multiphase Flow ◽

Pressure Sensors ◽

Production Optimization ◽

Learning Models ◽

Learning Methods ◽

Machine Learning Methods ◽

Production Monitoring ◽

Oil Gas

Abstract Monitoring of production rates is essential for reservoir management, history matching, and production optimization. Traditionally, such information is provided by multiphase flow meters or test separators. The growth of the availability of data, combined with the rapid development of computational resources, enabled the inception of digital techniques, which estimate oil, gas, and water rates indirectly. This paper discusses the application of continuous deep learning models, capable of reproducing multiphase flow dynamics for production monitoring purposes. This technique combines time evolution properties of a dynamical system and the ability of neural networks to quantitively describe poorly understood multiphase phenomena and can be considered as a hybrid solution between data-driven and mechanistic approaches. The continuous latent ordinary differential equation (Latent ODE) approach is compared to other known machine learning methods, such as linear regression, ensemble-based model, and recurrent neural network. In this work, the application of Latent ordinary differential equations for the problem of multiphase flow rate estimation is introduced. The considered example refers to a scenario, where the topside oil, gas, and water flow rates are estimated using the data from several downhole pressure sensors. The predictive capabilities of different types of machine learning and deep learning instruments are explored using simulated production data from a multiphase flow simulator. The results demonstrate the satisfactory performance of the continuous deep learning models in comparison to other machine learning methods in terms of accuracy, where the normalized root mean squared error (RMSE) and mean absolute error (MAE) of prediction below 5% were achieved. While LODE demonstrates the significant time required to train the model, it outperforms other methods for irregularly sampled time-series, which makes it especially attractive to forecast values of multiphase rates.

Download Full-text

Simulating Multiphase Flow in Reservoirs with Generative Deep Learning

10.2118/206126-ms ◽

2021 ◽

Author(s):

Abdullah A. Alakeely ◽

Roland N. Horne

Keyword(s):

Deep Learning ◽

Multiphase Flow ◽

Flow Behavior ◽

Training Data ◽

Data Set ◽

Operational Conditions ◽

Heterogeneous Reservoirs ◽

Reservoir Flow ◽

Latent Space ◽

Reservoir Simulator

Abstract This study investigated the ability to produce accurate multiphase flow profiles simulating the response of producing reservoirs, using Generative Deep Learning (GDL) methods. Historical production data from numerical simulators were used to train a GDL model that was then used to predict the output of new wells in unseen locations. This work describes a procedure in which data analysis techniques are used to gain insight into reservoir flow behavior at a field level based on existing historical data. The procedure includes clustering, dimensionality reduction, correlation, in addition to novel interpretation methodologies that synthesize the results from reservoir simulation output, characterizing flow conditions. The insight was then used to build and train a GDL algorithm that reproduces the multiphase reservoir behavior for unseen operational conditions with high accuracy. The trained algorithm can be used to further generate new predictions of the reservoir response under operational conditions for which we do not have previous examples in the training data set. We found that the GDL algorithm can be used as a robust multiphase flow simulator. In addition, we showed that the physics of flow can be captured and manipulated in the GDL latent space after training to reproduce different physical effects that did not exist in the original training data set. Applying the methodology to the problem of determining multiphase production rate from new producing wells in undrilled locations showed positive results. The methodology was tested successfully in predicting multiphase production under different scenarios including multiwell channelized and heterogeneous reservoirs. Comparison with other shallow supervised algorithms demonstrated improvements realized by the proposed methodology, compared to existing methods. The study developed a novel methodology to interpret both data and GDL algorithms, geared towards improving reservoir management. The method was able to predict the performance of new wells in previously undrilled locations without using a reservoir simulator.

Download Full-text