Optimization of Gas Assisted Gravity Drainage (GAGD) Process in a Heterogeneous Sandstone Reservoir Field-Scale Study

Abstract Gas injection has been widely used for IOR/EOR processes in oil reservoirs. Unlike the conventional modes of continuous gas injection (CGI) and water –alternative gas (WAG), the Gas Assisted Gravity Drainage (GAGD) process takes advantage of the natural segregation of reservoir fluids to provide gravity stable oil displacement. It has been proved that GAGD Process results in better sweep efficiency and higher microscopic displacement to recover the bypassed oil from un-swept reservoir regions. Therefore, CO2 was considered in this research for immiscible injection in the main pay/upper sandstone formation in South Rumaila oil field located in Iraq through the GAGD process application. This field, with a 60-year production history, has 40 production wells and is surrounded by an infiniteacting edge water aquifer from the east and the west flanks. Since the east flank is much less effective than the west one, 20 injection wells have been drilled at the east flank over the last 35 years to maintain the reservoir pressure. The formation depth is 10350 ft. sub-sea with a maximum vertical oil column of 350 ft. The GAGD process was adopted here using compositional reservoir simulation and PVT modeling to increase oil recovery. The GAGD process consists of placing a horizontal producer near the bottom of the payzone and injecting gas through existing vertical wells that have been used in prior waterfloods. As the injected gas rises to the top to form a gas zone, oil and water drain down to the horizontal producer. The location of horizontal wells is slightly above the oil-water contact. In the reservoir modeling, different reservoir and fluid properties have been investigated their effect on the flow response to implement sensitivity analysis, history matching through Design of Experiments. The operational design parameters of production/injection wells were considered to determine the optimal future reservoir performance through the GAGD process. Among many prediction scenarios, the GAGD process led to significant recovery incremental, especially in early future production years in comparison with the base case of no injection and 10,000 barrels of water injection per well in the same injection wells.

Optimization of Gas Assisted Gravity Drainage (GAGD) Process in a

Heterogeneous Sandstone Reservoir: Field-Scale Study

Watheq J Al-Mudhafar, and Dandina N Rao, Louisiana State University

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Enhanced Oil Recovery Conference held in Kuala Lumpur, Malaysia, 11–13 August 2015.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s) Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s) The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited Permission to reproduce in print is restricted to an abstract of not more than

300 words; illustrations may not be copied The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract

Gas injection has been widely used for IOR/EOR processes in oil reservoirs Unlike the conventional modes of continuous gas injection (CGI) and water –alternative gas (WAG), the Gas Assisted Gravity Drainage (GAGD) process takes advantage of the natural segregation of reservoir fluids to provide gravity stable oil displacement It has been proved that GAGD Process results in better sweep efficiency and higher microscopic displacement to recover the bypassed oil from un-swept reservoir regions Therefore, CO2 was considered in this research for immiscible injection in the main pay/upper sandstone formation in South Rumaila oil field located in Iraq through the GAGD process application This field, with a 60-year production history, has 40 production wells and is surrounded by an infiniteacting edge water aquifer from the east and the west flanks Since the east flank is much less effective than the west one, 20 injection wells have been drilled at the east flank over the last 35 years to maintain the reservoir pressure The formation depth is 10350 ft sub-sea with a maximum vertical oil column of 350 ft The GAGD process was adopted here using compositional reservoir simulation and PVT modeling to increase oil recovery The GAGD process consists of placing a horizontal producer near the bottom of the payzone and injecting gas through existing vertical wells that have been used in prior waterfloods As the injected gas rises to the top to form a gas zone, oil and water drain down to the horizontal producer The location of horizontal wells is slightly above the oil-water contact

In the reservoir modeling, different reservoir and fluid properties have been investigated their effect on the flow response to implement sensitivity analysis, history matching through Design of Experiments The operational design parameters of production/injection wells were considered to determine the optimal future reservoir performance through the GAGD process Among many prediction scenarios, the GAGD process led to significant recovery incremental, especially in early future production years in comparison with the base case of no injection and 10,000 barrels of water injection per well in the same injection wells

Introduction

Natural drive mechanisms are solution gas drive, water drive, gas cap drive, and gravity segregation The ending of the primary recovery stage is done by reaching very low reservoir pressure or high gas oil ratio (GOR) The water flooding is the most usual secondary recovery mechanism After that, the oil left behind

is being produced through injection of external parameters to increase the oil recovery such as gas, thermal, and chemical injection These processes are called EOR

EOR is using external materials such as water, steam, nitrogen, carbon dioxide, and polymer for injection into the oil reservoirs in order to extract certain volumes of remaining oil after the primary and secondary recovery stage by rock-fluid interactions through different mechanisms to increase the oil mobility towards the producers (Tarek Ahmed, 2006) The main influential roles of EOR methods is either reducing the oil mobility ratio or increase the capillary number (Farouq Ali and Thomas, 1996) The gas injection processes lead to reduced interfacial tension (IFT) and decrease the oil viscosity resulting in high oil mobility towards the wellbore areas and incremental recovery of about (5–15)% of OIP (Lake et al., 1992) The selection criterion of gas injection type depends on whether it is miscible or immiscible, depth, pressure, and temperature of formation, and oil compositions Also, the selection process considers the availability of specific gas for injection CO2 is more favorable for injection than other gases because of its lower minimum miscibility pressure and lower compression cost The CO2 and hydrocarbon gases have been used in approximately 90% of EOR projects Selection the CO2 according

to whether it is miscible or immiscible depends on depth of the formation, oil compositions, and rock wettability (Kuoand Elliot, 2001) Because of its higher viscosity, lower mobility ratio, closer density to the light oil in its miscible condition, CO2 has lower injection problems than other gas types In addition, CO2 has higher gravity segregation in the high water saturation zones of the reservoir and this leads to efficient oil displacement Carbon dioxide or any other gas that is being injected for EOR operations in three distinct modes: Continuous Gas Injection (CGI), Water Alternating Gas (WAG), and finally through the Gas Assisted Gravity Drainage (GAGD) The conventional continuous gas injection has been considered in the 1920’s for secondary and tertiary recovery modes to inject 100% carbon dioxide (no water) to lower the interfacial tension between the injected gas and the reservoir oil leading to increased displacement efficiency and achieve higher oil recovery (Kulkrani, 2003; Hinderaker et al., 1996) Over most of the previous CGI projects, the incremental recovery was around 5% of Initial Oil in Place (IOIP) (Kulkarni and Rao, 2004) The CGI is preferred in water-wet reservoirs within immiscible mode (Rogers and Grigg, 2000) The Water Alternating Gas injection (WAG) process was first proposed

in 1958 to improve sweep efficiency by injecting water slugs after CO2; however, the incremental oil recovery from 59 WAG projects was 5–10 % of the IOIP (Christensen et al., 1998) The Miscible WAG projects have led to an average incremental recovery of 9.7% of IOIP; while 6.4% incremental recovery was for immiscible WAG Projects (Caudle and Dyes, 1958; Christensen et al., 1998) Compared to CGI, WAG provides better mobility control and higher CO2 exploitation efficiency (Mahmoud and Rao, 2007) According to the WAG process, WAG does not consider the gravity-stable mode and the water saturation increases in the reservoir leading to diminished gas injectivity and increase the fluid competition towards the wellbore Consequently, and in order to overcome all other limitations for both WAG and CGI, a new process of Gas Assisted Gravity Drainage (GAGD) has been recently introduced

as a potential alternative to WAG in order to take advantage of natural segregation, due to the distinct fluid densities, of the injected gas from crude oil in the reservoir (Kulkarni and Rao, 2004; Rao et al.,

2006; Mahmoud and Rao, 2007)

Background

The Gas Assisted Gravity Drainage (GAGD) process has been suggested for improved oil recovery in secondary and tertiary modes for both immiscible and miscible processes The GAGD consists of placing

a horizontal producer at the bottom of the payzone above the oil-water contact and injecting the gas, either immiscible or miscible, in a gravity-stable mode through the vertical wells from the top of the formation (Rao et al., 2004) Due to the gravity segregation resulting from the different fluid densities at the reservoir conditions, the injected gas accumulates at the top of the reservoir providing gravity stable oil displace- ment that drains towards the horizontal producer at the bottom of the payzone (Mahmoud and Rao, 2007) The schematic Drawing of GAGD process has been shown in Figure 1

Figure 1—Schematic Drawing of GAGD process (Rao, 2012)

In the GAGD process, the fluids gravity segregation and the oil drainage towards the bottom of payzone lead to better sweep efficiency and higher oil recovery The CO2 gas is preferred for injection because it attains high volumetric sweep efficiency with high microscopic displacement efficiency especially in miscible injection mode Additionally, the high volumetric sweep efficiency assures delaying in CO2 breakthrough to the producer (Rao et al., 2006) Delaying or eliminating the gas breakthrough results in diminished concurrent gas-liquid flow further leading to increased gas injectivity

to maintain reservoir pressure Since the GAGD process includes vertical gas injection with vertical oil drain towards the horizontal producers, many factors might be affecting its performance, especially the geological structure in addition to the petrophysical properties and rock facies

This study examines the application of the GAGD process on the main pay of the upper sandstone member of South Rumaila oil field, located in Iraq in order to improve oil recovery and decrease the water cut levels which have reached 90% in many wells (Al-Mudhafer, 2010) CO2 can be obtained from natural resources such as from the associated gas production from oil fields Also, it can be captured from the large stations such as the refineries and thermal power plants As it is known, large quantities of natural gas are being flared from South Rumaila oil field currently However, re-injecting the produced flue gas for EOR will also lead to reduce CO2 emissions in the atmosphere

Field Study Description

Basrah Petroleum Company has discovered the giant South Rumaila oil field in October 1953 The Rumaila oil field is located in the south of Iraq about 50 km west of Basrah city and about 30 km to the west of the Zubair field (Al-Ansari, 1993) Rumaila field is about 100 km long, ranges between 12 to 14

km in width and its depth extends 3 km below sea level Dip angles on the flanks do not exceed 3, whereas in the crest’s parts they are about 1 The initial oil in place in Rumaila field/Main Pay is 19.5 Billion STB (Al-Mudhafer, 2010)

Geological Description

The South Rumaila oil field is composed of many oil-producing reservoirs Zubair is one of the oil reservoirs that are represented by Late Berriasian-Albian cycle and its sediments, which belong to Lower Cretaceous age, rich in organic deposition and accumulation of sedimentary matter (Al-Obaidi, 2010) The Zubair formation thickness ranges between 280 – 400 m it’s thickness increases towards the north-east and decreases towards the south-west (Al-Obaidi, 2009) Zubair formation encompasses five members based on sand to shale ratio and these have been named from top to bottom: Upper shale member, Upper sandstone member (main pay) where the GAGD process will be applied, Middle shale member, Lower sand member, and Lower shale member (Al-Ansari, 1993)

The upper sandstone member of the Zubair formation is the main pay zone of South Rumaila Oil Field (Mohammed et al., 2010) The main pay comprises five dominated sandstone units, separated by two shale units The shale units act as good barriers impeding vertical migration of the reservoir fluids except

in certain areas where they disappear The five unit zones have been denoted from top to bottom as AB, DJ1, DJ2, LN1, and LN2 and there are two shale layers C & K between AB and DJ1 & DJ2 and LN1 (Mohammed et al., 2010) More details are shown in Figure 2

Figure 2—Geological Column of the Main Pay/ South Rumaila Oil Field (Mohammed et al., 2010)

Sector Study Description

The South Rumaila Field is divided into four production sectors From the north to the south, the sectors are Qurainat, Shamiya, Rumaila, and Janubia The choice of this sector was made mainly because it has the least lack of data and it is the largest sector in which the production and injection operations are carried out The sector area is more than a third of the entire field The reservoir has five layers with about 80 m total thickness

Production and Injection Schedules

The primary production started in this field in 1954 and the water injection was started in 1980s During this period, 40 production wells were opened to flow in the simulated domain There is an infinite acting edge-aquifer located at the boundary of the reservoir For more than two decades, the natural depletion and water drive have been the only production mechanisms (Al-Mudhafer, 2013) Twenty injection wells have been drilled only at the east flank in order to maintain the huge aquifer support from the west flank, which

reaches to 20 times the influx from the east flank (Kabir, 2004) The production of some layers was ceased because of the higher water cuts of above 85% (Al-Mudhafer et al, 2010)

The 20 injection wells are primarily arranged in two parallel downdip rows The inner row of injectors

is completed only in the LN interval, while the most downdip row of injectors is completed only in the AB/DJ interval The cumulative water injection through 2004 was approximately 1.1 billion barrels Injection rates have varied widely with a maximum of nearly 426,000 BPD for two months in 1988 Artificial lift has been installed in the main pay wells recently, which has been suggested to handle the wells being incapable of flowing to surface after water cuts reach approximately 80% The estimated original oil in place (OOIP) for the main pay is 19.5 billion barrels The approximate current recovery factor from primary deletion and waterflood (AL-Mudhafer, 2013) is 50% The peak oil production rate

of 600,000 BPD occurred in May 1979 for the sector under study However, the peak oil production for the entire field was approximately 1.25 MMBPD during the same period Figure 3shows the production history for the field sector

Figure 3—Field Production History

Boundary conditions treatment

There are two types of boundary conditions that are encountered in the reservoir under study, a no-flow boundary and an aquifer The northern and southern boundaries are assumed to be of no-flow This assumption may be considered realistic since balanced production and injection rates are adopted for the reservoir and the isobaric lines crossing the northern streamlines are perpendicular to these boundaries Thus, the direction of flow will be parallel to northern and southern boundaries The flow boundaries at the east and west are natural water drive boundaries (Al-Mudhafer et al., 2010)

Compositional Reservoir Modelling

An EOS-compositional reservoir model was developed to evaluate the reservoir through GAGD process implementation and predict its future performance The constructed grid in the current study includes the

reservoir and parts of aquifer along the eastern and western flanks The grid dimensions for this study are

3300 grids: 30 grids in I-direction, 22 grids in J-direction, and 5 grids in K-direction as illustrated in

Figure

4.The dimensions of grid are constant: 500 m × 433 m The gridding system includes the reservoir and the infinite-acting aquifer (edge-water drive)

Figure 4 —Well Locations and Grid System of Reservoir Model

For field-scale reservoir modeling, there are some factors affecting the grid size selection such as Cost and time available to conduct the study and the processing speed of the CPU Since the field has 60 years

of production history with 40 producers and 20 injectors, it requires much time for running the model The time step sizes normally used are in the range of 1– 6 months; therefore, 1-month time step has been considered because of the monthly production and injection rates A preliminary evaluation of GAGD process in the South Rumaila oil field was conducted by running the compositional reservoir model for

12 years of future production after installing a series of 10 horizontal producers at the crest of the reservoir The same future prediction period was considered for the base case of no-injection and 10,000 barrels water per day per each of the existing vertical injection wells (Al-Mudhafar and Al-Khazraji, 2014) GAGD process has led to significant incremental oil recovery (1.4 Billions STB and 1.1 Billions STB, respectively in comparison to primary production and water injection cases as shown in Figure 5

Figure 5—Comparison of GAGD Process with primary depletion & Water Injection Scenarios

Design of Experiments

The Design of Experiments (DoE) is a statistical tool that is being used for the purpose of identifying the most sensitive factors that affect the response through sensitivity analysis procedure Furthermore, DoE helps to obtain the most likely scenario that achieves optimal response to a given recovery process Meanwhile, the DoE approach has the possibility to evaluate the interaction terms between the selected factors to find out their extent of influence on the process performance (Lazic, 2006)

To create a population of observations, either binary or decimal sampling techniques should be considered Usually, full factorial design requires many reservoir simulation runs for identifying the most significant modeling parameters impacting the response especially when there are more than five factors to test (Box, et al., 2005) Furthermore, it is hard to adopt the conventional design of experiments approaches to handle more than three levels for each factor as it leads to very large number of experiments (Montgomery, 1997; Montgomery, et al., 2003) Consequently and in order to capture many levels of variation for each factors, Latin Hypercube Sampling approach (LHS) has been suggested to provide more uniform and low discrepancy space-filling design Latin hypercube sampling attempts to keep the sampling data distribution uniform over the space with less random distribution especially in high dimension numerical simulation as shown in Figure 6 LHS is a modern sampling technique that can handle any number of parameters with mixing levels (Kalla and White, 2007) LHS was considered in this study for GAGD Sensitivity Analysis, History Matching, and Recovery Optimization

Figure 6 —Latin Hypercube Sampling of the Reservoir Simulation Runs

Sensitivity Analysis

The prior step for conducting history matching and recovery optimization is Sensitivity Analysis (SA) The main purpose of SA is to identify the most influential factors affecting the flow response in order to reduce the running time by reducing the total of number of computer experiments required to attain history matching The response that was considered for Sensitivity Analysis, History Matching, and Recovery Optimization is the cumulative oil production Meanwhile, 13 geological uncertain factors that were tested for SA: permeability and porosity of the five reservoir layers, vertical to horizontal permeability ratio (Kv/Kh), the rock compressibility, and the aquifer size Moreover, the controllable operational factors regarding CO2 injection and oil production constraints were also considered All these factors were combined through multilevel selection of Latin Hypercube Sampling The cumulative oil production was obtained from the compositional reservoir model given all the generated simulation runs

of LHS as depicted in Figure 7 Figure 8 represents the basic diagnostic plots of the linear model fitting

of the generated runs and their oil production outcomes Normal q-q, and residual analysis have indicated good model fitting to be considered as an accurate modeling to represent the process validation and precisely conclude the most influential factors affecting the GAGD process After building and validating the statistical linear model of the these factors given the obtained data, the SA results showed that porosity and horizontal permeability in addition to the operational design factors are the most influential factors affecting the flow response through the GAGD process application as illustrated in Figure 9 It is not surprised that Kv/Kh ratio did not come out to be important because the reservoir model was not incorporated with the lithology modelling, especially the reservoir has discontinuous shale distribution between the first and second & third and fourth layers (AL-Ansari, 1993) Figure 10shows the effect of Kv/Kh on the field cumulative oil production and field oil production rate for the entire prediction period The figure supports the outcome of limited effect of Kv/Kh ratio on the field flow response

Figure 7—Sensitivity Analysis of the Reservoir Flow Response

Figure 8 —Basic Diagnostic Plots of the Reduced Linear Model