|Verification Of Capillary Pressure Functions And Relative Permeability Equations For Modeling Gas Production From Gas Hydrates
||Last Reviewed 6/3/2016
The goal of this project is to verify and validate the capillary pressure functions and relative permeability equations that are frequently used in hydrate numerical simulators. In order to achieve this goal, numerical simulation using a network model will be used to suggest fitting parameters, modify existing equations or, if necessary, develop new equations for better simulation results. Experimental measurements of water retention curves in Tetrahydrofuran (THF) hydrate-bearing sediments will validate the numerical results.
Arizona State University, Tempe, AZ 85281
Numerical simulation is used to estimate and predict long-term behavior of hydrate-bearing sediments during gas production [Kurihara et al., 2008; Moridis et al., 2009; Moridis et al., 2005; Moridis and Regan, 2007a; Moridis and Regan, 2007b; Anderson et al., 2011, Myshakin et al., 2011; Myshakin et al., 2012]. Numerical simulators for gas hydrate are very complicated programs that include many equations and parameters, and two of the most important are the capillary pressure function and relative permeability equation. Permeability is the most important characteristic for predicting the gas production rate during gas hydrate development [Johnson et al., 2011; Minagawa et al., 2004; Minagawa et al., 2007; Kleinberg et al., 2003]. Permeability governs the production rate of water as well; therefore, enhancing the ability of hydrate simulators to predict gas and water production rates is predicated on determining the proper parameters for a capillary pressure function and generating a relative permeability equation.
Capillary pressure functions and relative permeability equations originate from unsaturated soil mechanics [Corey 1954; Brooks and Corey, 1964; Stone, 1970; van Genuchten, 1980]. These equations require empirical parameters, and several studies have been conducted to experimentally determine these parameters in the laboratory [Wösten et al., 1999].
However, in all experiments performed in those conventional studies, water and gas were injected from one boundary of the specimen to the other (a completely different gas generation mechanism from that observed during hydrate dissociation). When gas hydrate dissociates, gas nucleates from several pores inside sediments. In other words, gas is generated from within sediments instead of being pushed into the sediments from without. This different gas generation mechanism may result in completely different gas permeabilities during gas invasion and nucleation.
A laboratory experiment to obtain fitting parameters for capillary pressure functions and relative permeability equations is very complex, as it is difficult to control hydrate saturation and measure gas and water permeability at different saturations under high-pressure conditions [Kneafsey et al. 2011]. Conducting experiments under high-pressure conditions necessitates large-scale experimentation in a large, high-pressure chamber to produce more reliable data for gas flow.
An alternative method of estimating fitting parameters for capillary pressure functions and relative permeability equations during hydrate dissociation is history matching to in situ tests. A few short-term field-scale gas hydrate production tests were performed to evaluate depressurization and thermal stimulation methods at Mallik [Kurihara et al., 2005; Kurihara et al., 2008; Dallimore and Collett, 2005; Dallimore et al., 2008; Yamamoto and Dallimore, 2008]. Short-term field tests conducted in permafrost hydrate-bearing sediments such as Mallik [Hancock et al., 2005] and Mt. Elbert [Anderson et al., 2011] provided valuable information needed to derive parameters for relative permeability and characteristic curve (capillary pressure function) [Myshakin et al., 2011]. Because each hydrate reservoir has unique properties that affect gas production [Myshakin et al., 2012], it is not economical to conduct in situ testing at every hydrate-bearing reservoir to determine the fitting parameters. However, the parameters for relative permeability embedded in several numerical simulators could be verified to determine whether they correctly represent hydrate dissociation conditions.
This project will include a pore-network model simulation to predict the parameters for capillary pressure functions and relative permeability equations appropriate for simulating hydrate dissociation. The results of this research will support the collaborative efforts [e.g., Wilder et al., 2008; Anderson et al., 2011] to compare several existing numerical simulators.
The tools and values for numerical simulators produced through this research will help to determine bottomhole pressure, predict more accurate production rates of methane and water, and facilitate the selection of hydrate reservoirs for economic development. In addition, parameters obtained by numerical simulation could reduce the cost to perform in situ testing to calibrate numerical simulators.
Accomplishments (most recent listed first)
Hydrate pore habit is observed in micromodel experiment. Water retention curves for THF hydrate-bearing sediment have been measured for several hydrate saturation conditions. Fitting parameter m values for water retention curves have been suggested. A research paper including this study has been recently accepted in Geophysical Research Letters.
Fitting parameters for several initial hydrate saturation cases and for several sizes of patch formation cases are suggested based on the results of pore-network model simulation and a manuscript is under review for journal publication (Geochemistry, Geophysics, Geosystems).
Micro CT images were taken of natural sandy sediment recovered from Mallik site in collaboration with the DOE/NETL research team and used to extract a pore-network model. The effects of hydrate saturation and morphology were explored using the extract pore-network model. Preliminary results were presented at the American Geophysical Union meeting in San Francisco in 2014.
The project team completed an algorithm for calculating gas expansion and relative permeability during depressurization. The team investigated the effect of pore-network size on permeability data as well as hydrate habit (distributed versus patchy formation) and hydrate saturations on the shape of relative permeability curves. Fitting parameters for the Stone-type relative permeability equation were suggested for different hydrate saturations. Study results were published in Geochemistry, Geophysics, Geosystems.
The grain size distribution and effective stress of hydrate-bearing sediments were compiled from literature and used as input parameters for discrete element model (DEM) simulations to generate a three-dimensional particle packing similar to in situ hydrate-bearing sediments.
The project team numerically extracted an image of pore space from the three-dimensional particle packing generated by DEM simulation. Then, by using the maximal ball algorithm developed by Al-Kharusi and Blunt (2007) and Dong and Blunt (2009), they generated a pore-network model based on the image of the pore space. The generated pore-network models consist of several pores linked by connecting tubes.
Current Status (June 2016)
The project team is investigating the effect of pore size distribution and gas viscosity change on water retention curves and relative permeabilities. Pore-network model simulation results are compared with experimental results.
Project Start: October 1, 2012
Project End: July 31, 2016
Project Cost Information:
DOE Contribution: $241,735.86
Performer Contribution: $61,321.14
NETL – John Terneus (John.Terneus@netl.doe.gov or 304-285-4254)
Arizona State University – Jaewon Jang (firstname.lastname@example.org or 480-727-4309)
Quarterly Research Performance Progress Report [PDF-430KB] October - December, 2014
Quarterly Research Performance Progress Report [PDF-1.51MB] July - September, 2014
Quarterly Research Performance Progress Report [PDF-1.34MB] April - June, 2014
Quarterly Research Performance Progress Report [PDF-606KB] January - March, 2014
Quarterly Research Performance Progress Report [PDF-11.0MB] October - December, 2013
Quarterly Research Performance Progress Report [PDF-6.81MB] July - September, 2013
Quarterly Research Performance Progress Report [PDF-2.66MB] April - June, 2013
Quarterly Research Performance Progress Report [PDF-1.89MB] January - March, 2013
Quarterly Research Performance Progress Report [PDF-1.50MB] - October - December, 2012