eprintid: 17282
rev_number: 2
eprint_status: archive
userid: 1
dir: disk0/00/01/72/82
datestamp: 2023-12-19 03:23:42
lastmod: 2023-12-19 03:23:42
status_changed: 2023-12-19 03:07:47
type: conference_item
metadata_visibility: show
creators_name: Hassan, A.M.
creators_name: Ayoub, M.A.
creators_name: Mohyadinn, M.E.
creators_name: Al-Shalabi, E.W.
creators_name: Alakbari, F.S.
title: A New Insight into Smart Water Assisted Foam SWAF Technology in Carbonate Rocks using Artificial Neural Networks ANNs
ispublished: pub
keywords: Carbonates; Carbonation; Enhanced recovery; Error analysis; Forecasting; Learning algorithms; Machine learning; Neural networks; Offshore oil well production, Artificial neural-network based modeling; Carbonate rock; Input and outputs; Input parameter; Output parameters; Overfitting; Relative errors; Statistical error analysis; Trend analysis; Water assisted, Contact angle
note: cited By 10; Conference of 2022 Offshore Technology Conference Asia, OTCA 2022 ; Conference Date: 22 March 2022 Through 25 March 2022; Conference Code:187172
abstract: The smart water-assisted foam (SWAF) technology is a novel enhanced oil recovery (EOR) technique, which combines the synergistic effect of both smart water and foam-flooding methods. The smart water enables multilevel improvements, namely, stabilization of foam-lamella and wettability alteration of the carbonate rock, which leads to desirable oil relative-permeability behavior. Contact angle tests are the common approach for measurement of the preferential affinity of reservoir rocks to fluids. However, the laboratory methods for contact angle measurement are costly and time-consuming. Therefore, in this study, we propose a new approach to predict contact angle based on a machine learning technique. A model based on artificial neural network (ANN) algorithm was developed using 1615 datasets acquired from diverse published resources. The developed ANN-based model to predict contact angle was further evaluated by applying the trend analysis approach, which verify the correct relationships between the inputs and output parameters. The collected datasets were trifurcated into training, validation, and testing segments, so that the over-fitting and under-fitting issues are evaded. Furthermore, some statistical error analyses, namely, the average absolute percentage relative error (AAPRE), and the correlation coefficient (R) were performed to present the robustness and accuracy of the proposed model. The findings from the trend analysis showed the sound relationships between the inputs and output parameters. The statistical error analyses proved that the developed ANN-based model does not have any under-fitting or overfitting anomalies, and correctly determines the contact angle with high accuracy, substantiated by the R values of 0.9988, 0.9985, 0.9967, and AAPRE values of 1.68, 1.62, 1.81, for training, validation, and testing datasets, respectively. The proposed ANN-based model for contact angle prediction has many advantages including speed, reliability, and ease of usage. This work highlights the potential of machine learning algorithms in oil and gas applications, particularly in contact angle prediction from SWAF technology. The findings from this study are expected to add valuable insights into identifying the optimal conditions (i.e., optimum smart water and surfactant aqueous solution) for the operation sequence of SWAF technology, leading to successful field applications. Copyright © 2022, Offshore Technology Conference.
date: 2022
publisher: Offshore Technology Conference
official_url: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85147155840&doi=10.4043%2f31663-MS&partnerID=40&md5=7ad3e5176da0685cdae5e89e7a6e258c
id_number: 10.4043/31663-MS
full_text_status: none
publication: Offshore Technology Conference Asia, OTCA 2022
refereed: TRUE
isbn: 9781613998359
citation:   Hassan, A.M. and Ayoub, M.A. and Mohyadinn, M.E. and Al-Shalabi, E.W. and Alakbari, F.S.  (2022) A New Insight into Smart Water Assisted Foam SWAF Technology in Carbonate Rocks using Artificial Neural Networks ANNs.  In: UNSPECIFIED.