Multiscale investigation of fluid transport and enhanced recovery in shale
Abstract/Contents
- Abstract
- In 2019, the U.S. produced 75% of its natural gas from shales and 59% of its oil from tight oil resources. Multistage hydraulic fracturing along with horizontal pad drilling enabled operators to increase significantly production from these resources. Despite the vastness of shale resources, recovery factors are small typically, amounting to 5-10% for oil and ~25% for gas. In this work we examine various enhanced recovery techniques across multiple length scales to gain a better understanding of enhanced resource recovery mechanisms resulting from injection of gas, such as carbon dioxide (CO2). In doing so, we develop in-house shale characterization experimental methods to quantify fluid flow, storage, and recovery in the laboratory. An experimental workflow is presented for rock characterization (porosity, permeability, and adsorption) to quantify accurately gas storage and flow needed for enhanced gas recovery (EGR) experiments. Both pulse decay and Computed Tomography (CT) were used independently to establish consistency between results derived from each method. New image processing routines for CT data were developed that better match mass balance derived porosity and storativity results compared to conventional CT methods. Measured porosity values using helium (He) for each sample proved to be constant at various equilibrium pore pressures justifying its use as a reference gas for excess adsorption computations for other gases studied. Nitrogen (N2), methane (CH4), krypton (Kr), and CO2 apparent permeability and storativity at different pore pressures were determined. All adsorptive gases, except CO2, exhibited monolayer Langmuir adsorption behavior. CO2 uniquely showed multilayer behavior that was observed in two cores (Eagle Ford (EF1) and Wolfcamp (WC2)). The impact of adsorption on gas permeability was captured in our experiments showing a negative correlation between adsorption affinity and permeability. For instance, Kr and CO2 reduced the liquid-like permeability value determined using He by factors of 2 and 8, respectively, for sample EF1. Finally, a persistent five-fold reduction in permeability was observed in sample WC2 after CO2 exposure that is attributed to kerogen swelling or matrix softening. The degree of kerogen swelling is impacted by the affinity of the gas to adsorb and its ability to dissolve into kerogen. Matrix softening, on the other hand, enhances compaction of the pore space under constant effective stresses. Diverse diagnostics across multiple scales were used to examine the impact of CO2-water fluids on oil recovery and matrix flow on both core and micron scales. Enhanced oil recovery (EOR) was investigated on a Utica (W2-2) core that was artificially split and saturated with crude oil for 3 months. The core was cut to create a conductive pathway and to increase surface area to help oil saturate the sample. Core-scale examinations using pulse decay, injection experiments, and CT showed no material enhancement to matrix fluid flow or oil recovery using dry supercritical CO2, water-saturated CO2, or carbonated water. Approximately 87% of the in-situ oil was recovered using dry supercritical CO2 initially without any further recovery. CT visualizations showed that most of the oil resided in the main fracture with small amounts of oil residing in the matrix. Potential enhancement in core-scale matrix flow was investigated by conducting He pressure pulses before and after a carbonate-rich Eagle Ford (EF-1) sample was exposed to carbonated water for 6 months. Measured permeability values were identical before and after exposure to the acidic fluid. Micron-scale findings, on the other hand, using scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS) and micro-CT showed vugs and pits, from calcite dissolution, ranging from 1 micro-m to 10's micro-m in size in samples exposed to carbonated water. These samples were exposed to carbonated water in either a batch reactor setting or a core-scale carbonated water injection experiment. The wet supercritical CO2 phase did not induce any observable carbonate dissolution in the shale sample tested. Finally, it was determined that gold coating of the sample (a preparation step needed for SEM imaging) has no impact on fluid-rock interaction during our experiments. A novel experimental setup was designed for investigating EGR in shale cores. The detailed sample characterization conducted on both samples (EF1 and WC2) was used to assess initial rock storativity, adsorption, and permeability that are vital for proper experimental planning given the small pore volumes in shales. Experiments were run with Kr or CH4 as in-situ gases and CO2 or N2 as injection gases. Continuous Kr gas injection experiments showed consistent results between mass balance and CT-derived results establishing reliability in our CT depictions. CO2 gas injection had a better initial displacement efficiency compared to N2 when displacing in-situ Kr. Homogeneous sample WC2 required approximately four times fewer pore volume injections to produce the entire original gas in place compared to sample EF1 that had two CT-visible conductive pathways or microcracks. Finally, core-scale findings reveal that continuous gas injection is more effective than huff-n-puff for enhancing gas recovery on a pore volume injected basis. Core-scale simulations using CMG GEM were created to mimic and validate lab pulse decay and EGR experiments. Porosity, permeability, and adsorption values were validated for various pressure pulses across both cores (EF1 and WC2) using all the gases investigated (He, N2, Kr, CH4, and CO2). Coal bed methane modeling in CMG GEM was utilized for matching highly adsorptive gases (Kr and CO2) due to a delayed downstream response given the experimentally determined porosity, permeability, and adsorption values. Another critical parameter, diffusion characteristic time (t*), was identified using this model during the history matching process that quantifies a mass transfer resistance to fracture flow due to fracture-matrix gas exchange. Although our experiments were not designed to measure directly t*, various pressure pulses for CO2 and Kr required a diffusion time of 1.44-1.92 hrs (0.06-0.08 days) to match our pressure pulses using coal bed methane modeling. A continuous gas injection experiment was simulated in CMG GEM that matched the experimental pressure history, recovery results, and CT visualizations for sample EF1. Sensitivity studies on diffusion time revealed its strong influence on recovery in low permeability areas that are predominant during late production. A huff-n-puff experiment was simulated given the same model parameters as the history matched continuous injection experiment. Huff-n-puff had a poorer recovery curve compared to continuous injection due to gas entrapment away from the microcracks with each cycle. Finally, core-scale simulations show that long diffusion times are favorable for huff-n-puff but disadvantageous for continuous injection emphasizing the importance of sample characterization, including transport properties, before evaluating the different EGR techniques. Learnings from core-scale experiments and simulations were translated to assess EGR applicability at field scale. Multiple reservoir uncertainties (porosity, stimulated permeability, diffusion time) and operational decisions (e.g. injection and soak times) were explored to understand their influence on CH4 recovery and CO2 storage for continuous injection and huff-n-puff. A simplified CMG GEM field model was created that utilized 1300 m horizontal wells that have 13 fracture stages with 4 clusters per stage. Field continuous injection scenarios yielded a loss in cumulative CH4 production compared to cases with primary production only over a 20 year period. Injection started after 10 years of primary production; however, the economic benefits from CO2 storage outweighed CH4 losses in cases with short diffusion times (< 1 day). A conservative and simple cost structure was used with a gas price of $1.7/MCF and a CO2 tax credit of $35/tonne. Huff-n-puff scenarios conducted over a single horizontal well exhibited minimal CO2 storage but significant enhancement in CH4 production over a 5 year period. The performance of huff-n-puff EGR was most sensitive to the permeability of the stimulated reservoir volume, injection time and soak time. This highlights the impact of the original hydraulic fracturing job and operational decisions on the success of huff-n-puff EGR field operations.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2020; ©2020 |
| Publication date | 2020; 2020 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Elkady, Youssef Magdy Abdou Mohamed |
|---|---|
| Degree supervisor | Kovscek, Anthony R. (Anthony Robert) |
| Thesis advisor | Kovscek, Anthony R. (Anthony Robert) |
| Thesis advisor | Benson, Sally |
| Thesis advisor | Horne, Roland N |
| Degree committee member | Benson, Sally |
| Degree committee member | Horne, Roland N |
| Associated with | Stanford University, Department of Energy Resources Engineering |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Youssef Elkady. |
|---|---|
| Note | Submitted to the Department of Energy Resources Engineering. |
| Thesis | Thesis Ph.D. Stanford University 2020. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2020 by Youssef Magdy Abdou Mohamed Elkady
- License
- This work is licensed under a Creative Commons Attribution 3.0 Unported license (CC BY).
Also listed in
Loading usage metrics...