Modeling viscosity of CO2 at high temperature and pressure conditions

Menad Nait Amar,Mohammed Abdelfetah Ghriga,Hocine Ouaer,Mohamed El Amine Ben Seghier,Binh Thai Pham,Pål Østebø Andersen

doi:10.1016/j.jngse.2020.103271

Abstract

The present work aims at applying Machine Learning approaches to predict CO2 viscosity at different thermodynamical conditions. Various data-driven techniques including multilayer perceptron (MLP), gene expression programming (GEP) and group method of data handling (GMDH) were implemented using 1124 experimental points covering temperature from 220 to 673 K and pressure from 0.1 to 7960 MPa. Viscosity was modelled as function of temperature and density measured at the stated conditions. Four backpropagation-based techniques were considered in the MLP training phase; Levenberg-Marquardt (LM), bayesian regularization (BR), scaled conjugate gradient (SCG) and resilient backpropagation (RB). MLP-LM was the most fit of the proposed models with an overall root mean square error (RMSE) of 0.0012 mPa s and coefficient of determination (R2) of 0.9999. A comparison showed that our MLP-LM model outperformed the best preexisting Machine Learning CO2 viscosity models, and that our GEP correlation was superior to preexisting explicit correlations.

Highlights

Carbone dioxide (CO2) is the main contributor to greenhouse gas (GHG) emissions with up to 72% of the total GHG emissions recorded in 2010
The main purpose of this study is to advance the research on development of high exactness and simple-to-use machine learning ap proaches that can predict the viscosity of CO2
Several machine learning techniques were applied to establish robust and simple-to-use models to accurately predict the viscosity of CO2 under wide ranges of pressure and temperature condi tions, using density and temperature as input parameters

Summary

Introduction

Carbone dioxide (CO2) is the main contributor to greenhouse gas (GHG) emissions with up to 72% of the total GHG emissions recorded in 2010. As per CCS, the CO2 is stored in under ground geological formations including saline aquifers, depleted oil/gas reservoirs, and other geological options (Aminu et al, 2017). Developments of the last decade have indicated that CO2 injection can be a viable option for enhanced oil and gas recovery in tight shale reservoirs (Yu et al, 2014; Sheng, 2015; Eshkalak et al, 2014; Hoffman and others, 2012; Jin et al, 2017). Alternative industrial applications include the use of CO2 as a refrigerant in heating and refrigerating processes (Sawalha et al, 2017; Li et al, 2016), as feedstock in the production of chemicals (Ampelli et al, 2015; Chen et al, 2016) and carbon source for micro algae to produce biofuels (Taher et al, 2015; Aslam et al, 2018)

Objectives

Methods

Results

Conclusion