Production data represent a source of information about the ongoing dynamic flow process in the reservoir. A proper analysis of these data might therefore give an indication of the relationships governing the fluid flow. Exploratory data analysis is one tool which can be used to extract the required information and to establish a statistical significance to the results. This paper presents a simple approach to examine the interwell communication and interference of a mature waterflood in order to identify and rank areas of potential improvement. Using this as a screening process we can identify areas with the best potential for further and more detailed studies. In addition some of the problems and limitations associated with analysis of production data are discussed.
Is there an easy way of evaluating waterflood performance based solely on surface measurements in the form of rates and pressures? Although production data have always been a key part of any waterflood evaluation, no good answer has been given to what we can expect to get from such an analysis.
Exploratory data analysis of production data is aimed at organizing and representing the data in an easy interpretive way. The analysis can take place at two levels, global and local. The global level will examine large scale trends (multiple wells), whereas the local analysis will focus on single well behavior and interwell relationships (well pairs). The use of production data as an analytical tool depends upon the quality and density of the data. More data of higher resolution and quality will yield significantly more information than irregular and poorly measured data. The quality of the analysis is therefore a direct function of the quality and the density of the measurements.
The analysis of production data poses a particular challenge since we have a 3-dimensional data set mixing both the spatial and the temporal (time) dimensions. Stationarity assumptions along the spatial plane do not translate into stationarity along the temporal axis. Special care is therefore necessary when dealing with both the spatial dimensions and the temporal dimension simultaneously. In addition, the surface measured production data constitute a 2-dimensional spatial plane, which represents the dynamic behavior of a 3-dimensional reservoir. Another challenge is the size of the data set which comprises the production data. An efficient handling and representation is therefore important to obtain a clear interpretation.
The primary objective of the methodology presented in this paper is to screen the field data to identify and rank areas of potential production improvement for a mature waterflood, in the form of changes to the individual wells, operational changes to the water injection allocation, infill drilling or pattern reorientation. This identification process depends upon the ability to extract information about the connectivity, sweep and interference among injection and production wells based on production data. Having identified potential areas of improvement, these areas can be further evaluated based on the geological knowledge and a more detailed reservoir study.
Microbial cultures can be used to reduce or eliminate scale formation in water flood injection systems. Several water floods have been treated for over a year with microbial culture products and the effects of the treatment on operating performance has been studied. The injection systems initially showed significant scale problems from calcium sulfate and calcium carbonate. Tracking water flood performance parameters such as injection pressures and volumes over the treatment period showed that, as compared with prior treatment historical data, a variety of positive results of treatment were evident. For example, the flood injection volumes were stabilized or increased, injection pressures were stabilized and oil and gas production was increased. Laboratory studies of injection water showed that filtration times could be significantly reduced by microbial treatment, indicating that near well-bore occlusion could be reduced by such treatments. The microbial treatments were considered more effective than prior alternative treatment technologies. This new technology offers a cost-effective superior alternative to conventional technologies.
Secondary recovery operations utilizing water flooding represent the majority of oil fields in the United States. The practice is particularly common in mature producing regions such as the Permian Basin. The handling of the large volumes of water used in such projects leads to specific operational problems related to water chemistry. Demand for water leads to the use of various aqueous sources which may contain differing amounts of cations and anions. These incompatible waters lead to the formation of mineral salt deposits such as calcium carbonate, calcium sulfate, barium sulfate or iron sulfide. These deposits may form at any point in the injection system from surface facilities to downhole. Occlusion of water injection wells by either preformed or in situ formed scale is a principal cause of reduced water flood efficiency. Skin damage and blockage of the formation by scale leads to decreased injection volumes and increased injection pressures. The former instance may reduce injection volumes and the latter increases operating costs.
Conventional treatments for injection system scale include acid and chemical treatments. Both technologies suffer from significant drawbacks. Acid treatments can promote corrosion, utilize a hazardous chemical and require significant downtime. Acid alone is not effective on calcium or barium sulfate scale. Chemical treatments are expensive and may not be cost-effective or successful. They may also not be successful in treating downhole permeability problems. An alternative technology is the use of microbial culture products to improve water flood operations.
Large-scale CO2 projects, like those implemented more than a decade ago in the Wasson Field, are not as attractive today because of large capital investments up-front and lower and uncertain oil prices. However, building upon our experience in large-scale floods, as well as recent improvements in the reservoir characterization and simulation technology, a grass-roots CO2 project can still be made attractive, in today's environment, by high-grading the reservoir resources, staging the development, and re-injecting the produced gas without facilitating a large gas processing plant. The paper contrasts the characteristics of the Bennett Ranch CO2 project, a CO2 flood designed today, with traits of CO2 floods of yesteryear. The paper also highlights the innovative efforts in reducing and delaying the capital expenditures for wells and facilities, resulting in a less than 1 MM$ capital investment prior to the CO2 injection. Due to small flood patterns, significant production responses have been observed after only 3 months of CO2 injection.
The Bennett Ranch Unit (BRU) is located in the northeastern portion of the Wasson Field which produces oil from the San Andres formation. (See Figure 1) The unit was formed in 1964 to install a waterflood. Historically, the unit production reached 12,100 bopd in 1974, but by 1994 production had steadily declined to 3000+ bopd. Currently, the unit-wide-average oil cut is around 6% but it varies areally from below 4% in the most mature southern area to above 10% in the north. Many wells in the southern area have been shut-in due to high water cut. In June 1994, cumulative primary and secondary recovery reached 104 mmbo. CO2 flooding is considered the only viable option to extend the field life.
San Andres Reservoir Description
The San Andres formation is a mid-Permian aged dolomite located at depths ranging from approximately 4800' to 5600'. Structurally, the San Andres at BRU forms a northeast plunging anticline (see Figure 2) with an average elevation approximately 350' below the highest stratigraphic equivalent interval of the Wasson field (which is located in the Denver Unit). As a result of this low structural elevation, there is no gas cap at BRU. Within the unit, the San Andres formation dips approximately 4.5 degrees on the eastern flank and 1.5 degrees on the western flank.
The San Andres reservoir at BRU can be stratigraphically divided into two major intervals: "First Porosity" and "Main Pay." The deeper Main Pay interval consists of dolomitized open marine packstones and grainstones and possesses better reservoir properties than the First Porosity does. The First Porosity interval was deposited in a more restricted marine intertidal environment; it has a finer crystalline matrix and is less continuous than the Main Pay. Supratidal mudstones and anhydrites that lie above the First Porosity interval form the seal for the hydrocarbon accumulation. Based on the character of gamma ray and sonic logs, we have subdivided the First Porosity interval into a single interval (F1) and the Main Pay into 6 zones (M1-M6) (see Figure 3).
The application of cyclic CO2, often referred to as the CO2 Huff-n-Puff (H-n-P) process, may find its niche in the maturing waterfloods of the Permian Basin. Coupling the CO2 H-n-P process to miscible flooding applications could provide the needed revenue to sufficiently mitigate near-term negative cash flow concerns in these capital intensive projects. Initial results from a field demonstration of the process are provided. The paper briefly reviews the reservoir characterization associated with the project and the compositional simulation of the CO2 H-n-P process. Simulation results suggest that reservoir characterization of flow units is not as critical for a CO2 H-n-P process as for a miscible flood. Entrapment of CO2 by gas hysteresis is considered the dominant recovery factor for a given volume of CO2. The repetitive application of the process was found to be unwarranted in a waterflooded environment. Future history matching of the performance will allow better forecasts and evaluation of the economic impact available with this underutilized process.
The CO2 H-n-P process is a proven enhanced oil recovery technology in Louisiana-Texas Gulf-Coast sandstone reservoirs.1,2 Application seems to mostly confine itself to low-pressure sandstone reservoirs.3 The process has been shown to be moderately effective in conjunction with steam on heavy California crude oils.4,5 A review of earlier literature1,6,7 provides an excellent introduction to the theory, mechanics of the process, and several case histories. Although the technology is proven in light oil sandstones, it continues to be a very underutilized enhanced recovery option for carbonates. However, the theories associated with the CO2 H-n-P process are not lithology dependent.
A significant amount of oil reserves are located in carbonate reservoirs. Specifically, the carbonates deposited in shallow shelf (SSC) environments make up the largest percentage of known reservoirs within the Permian Basin of North America. Many of these known resources have been under waterflooding operations for decades and are at risk of abandonment if crude oil recoveries cannot be economically enhanced.8,9
Miscible CO2 flooding is the process of choice for enhancing recovery of light oils10 and already accounts for nearly 12% of the Permian Basin's daily production.11 There are significant probable reserves associated with future miscible CO2 projects. However, many are marginally economic at current market conditions due to large up-front capital commitments for a peak response which may be several years in the future. The resulting negative cash-flow is sometimes too much for an operator to absorb. The CO2 H-n-P process is being investigated as a near-term option to mitigate the negative cash-flow situation - allowing acceleration of inventoried miscible CO2 projects.
This paper addresses mathematical modeling of free-fall gravity drainage which is believed to occur in naturally fractured reservoirs after depletion of oil in the fractures or gas injection into the fractured system. Comparison of wetting phase recoveries calculated using existing mathematical models with experimental data indicates the inaccuracy of these models. The causes of error are identified to be the unrealistic assumptions made in formulation of the models. Based on Darcy's law and film flow theory, we have developed a new mathematical model to describe the free-fall gravity drainage process. A simple non-linear governing equation for phase demarcator in dimensionless form was formulated and solved numerically as a function of dimensionless time. Based on the dimensionless demarcator, fluid recovery during free-fall gravity drainage is calculated. Comparisons of wetting phase recoveries given by the new model with 20 sets of experimental data obtained under thermodynamic equilibrium conditions for a variety of fluids and cores show much better accuracy of the model over the existing models. Using the dimensionless time, tD = ke gt/ L, fluid recovery obtained from laboratory studies can be scaled to field applications for estimation of projected oil recoveries in oil fields. We have also applied the new model to simulation of free-fall gravity drainage under non-equilibrium conditions where molecular diffusion between phases is considered. Experimental oil recovery data obtained from CO2 injection into a Berea core and a reservoir sandstone core, which were saturated with separator oil, have been matched by the model using empirical correlations for fluid properties. The objective of this paper is to provide reservoir engineers with a useful tool for estimating the oil recovery from fractured reservoirs after gas injection.
Because fractures are highly conductive to gas and gas is the nonwetting phase in the rock matrix, gas injection into fractured reservoirs has been traditionally considered as an inefficient method for enhancing oil recovery from fractured reservoirs. However, the Midale Pilot indicated that the efficiency of CO2 injection into fractured reservoirs is not as low as expected. The only explanation is that when a non-equilibrium gas is injected into the fractured system at elevated pressure, compositional effects become active between the gas in the fractures and oil in the matrix. Due to multi-contact mechanism, light hydrocarbons in the oil can be extracted from the virgin oil bank forming a "gas" -rich light liquid phase and an oil-rich heavy liquid phase. This kind of phase split has been reported by several investigators including Lansangan and Smith. The interfacial tension (IFT) between phases is low compared to that between virgin oil phase and gas phase. Therefore, the capillary pressure threshold may be overcome by gravity resulting in gravity drainage of the light oil from the matrix blocks. In order to understand the mechanism of gravity drainage and predict the response of fractured reservoirs to gas injection, a mathematical model of the process is desirable.
Equilibrium Gravity Drainage. Studies on gravity drainage were conducted a century ago when King investigated the principles and conditions of aquifer motion. Investigations of gravity drainage of oil in oil reservoirs were initiated in early 40's of this century. Leverett and Katz presented data and discussed the theory relating capillary and gravitational forces acting on liquids contained in a sand body. Stahl et al. conducted experiments to investigate behavior of free-fall gravity drainage of water and oil in an unconsolidated sand. Elkins et al. presented a simplified theory of regional drainage of oil from upstructure location to downstructure location due to gravity assuming zero capillary pressure gradient. Cardwell and Parsons presented a governing equation for the free-fall gravity drainage process. They could not solve the equation because of its non-linearity. P. 23
Two fields, Ozona (Canyon) and Sawyer (Canyon), with more than 4,000 total wells, contain the majority of wells in the Canyon Sand trend of Crockett, Edwards, Schleicher and Sutton Counties in Southwest Texas. Average ultimate recovery for the fields decreased from 1,100 MMCF/well for wells drilled in 1970 to 400 MMCF/well for wells drilled in 1985, then remarkably reversed the trend. Recently drilled wells (since 1991) are expected to recover an average 625 MMCF. Initial monthly production decreased from 15,000 mcf/month to 5,000 mcf/month in 1985 and then increased to the present 12,000 mcf/month. Average first year wellhead pressure continued to decrease throughout the 25 years studied to a present 1300 psig in Ozona and 700 psig in Sawyer. The data suggest that drainage interference began to be a factor in about 1985 in the more densely developed areas. Although the average well density was about 200 acres per well in 1985, the true density in areas of active infill drilling probably approached 120 acres per well. Present average density is 143 acres per well in Ozona Field and 131 acres per well in Sawyer Field, but many leases are developed on density of 80 acres or less.
During the 1970's, the Canyon Sands of Crockett, Edwards, Schliecher and Sutton Counties, Texas became a significant source of new gas production (Figure 1). More than 4,000 productive wells have been completed in the Sawyer and Ozona Fields. Drilling activity since 1970 has varied, peaking in the mid to late 1970's and falling to a low in 1986 as gas demand from these fields fell.
As development of the trend proceeded and fields merged, the Texas Railroad Commission ordered a number of the adjacent fields reclassified into either Ozona (Canyon Sand) Field or Sawyer (Canyon) Field. The two fields now contain more than 65% of the wells in the entire 6,000 well Canyon Sand gas trend.
Sawyer (Canyon) Field was discovered in 1967 and produces gas from Pennsylvanian age deltaic or turbidite lenticular sands at about 5,200 to 6,500 feet. The field covers portions of Edwards, Schleicher and Sutton Counties, and contains some 2,200 active and abandoned wells. Ozona (Canyon Sand) Field is located almost entirely in Crockett County, west of the Sawyer Field, and trending deeper into the Val Verde Basin at about 6,500 to 7,500 ft. It was discovered in 1962, but, like Sawyer, was slow in development until the rise of air drilling in the early 1970's. Figures 2 and 3 show the development history of each field group.
Both fields produce dry gas with small amounts of water from tight sands of 8-10% porosity and less than .1 md permeability1. There is core evidence that 1-2 inch streaks of significantly greater permeability exist within the larger sand lenses and these may enable the sands to produce in economic quantities.
A study of estimated ultimate recovery, initial production rates, wellhead pressures, and well density was completed for each field. The study encompassed all 3,200 wells drilled since 1970 in the two Fields. The Ozona group of 1,270 wells included minor numbers in the adjacent fields of Ozona, NE (Canyon 7520), Ozona, North (Canyon), Ozona NW (Canyon) and Ozona, SW (Canyon, Lower). The Sawyer Field study of 1,950 wells included only those classified in Sawyer (Canyon) Field.
An oxygen activation log was modified for use in obtaining injection profile surveys in four short radius horizontal injection wells. The tool was run inside a workstring that was temporarily placed in the lateral. Injection was established down the well annulus and the tool measured flow rates at points in the annulus between the workstring and open hole lateral. From this data, a distribution of injection along the lateral could be determined. Three different stimulation techniques were used on the horizontal wells and the profile results provided valuable insight into the effectiveness of these techniques.
The productive reservoir in the Slaughter field is the San Andres, a dolomite formation at a depth of 5000 feet. The field is located 100 miles north of Midland, Texas. Solution gas drive served as the primary recovery mechanism for the initially undersaturated crude. The reservoir has been under waterflood since the 1960's and other operators began CO2 flooding in the 1980's. Texaco initiated a miscible CO2 flood in 1994.
Horizontal wells can improve waterflood and CO2 flood sweep efficiency.1,2 In order to realize this improvement, it is imperative that injection be evenly distributed along the entire lateral. Texaco has now drilled seven horizontal injectors on the Sundown Slaughter Unit, an 8684-acre property in the Slaughter field. These wells have lateral sections ranging from 589 to 864 feet. They were drilled at a slight angle and designed to penetrate all layers in the pay.
This paper discusses part of the evaluation process for these wells. The lack of available equipment and techniques designed to work in a short radius, 4 1/2 inch open hole completion created a need for a new method to determine the injection profile. Figure 1 shows the location of the four wells where this new method has been applied.
The Slaughter field is located on the Northwest Shelf of the Permian Basin. Cyclic deposition resulted in a stratified reservoir with significant variation in porosity and permeability within the pay section. Core analysis shows that porosity ranges from two to 20 percent. Permeability varies from nearly zero to 50 millidarcies. A dense layer composed of mostly anhydrite with interspersed dolomite layer covers the reservoir and forms an excellent seal. The lower limit of the resesrvoir is defined by the top of a transition zone. Rock properties in the transition zone are similar to those in the pay zone.
Due to the variation in reservoir properties between and within zones in the pay section, and the existence of layers within the pay that may act as flow barriers, it was important to verify that all zones are being flooded. Also, we desire to confirm that no thief zones existed in the curved section of the well above the pay zone.
Conformance control is any action taken to improve the injection or production profile of a well. It encompasses procedures that enhance recovery efficiency, improve wellbore/casing integrity, and satisfy environmental regulations. Unwanted fluid production in oil- and gas-producing wells is a limiting factor that controls the productive life of a well. The cost of produced water disposal in an environmentally non-threatening fashion may be a major concern for many producers. In addition, control of excess water and gas production improves profitability by allowing additional oil to be produced.
Application of water control technology assists to minimize water production and maintain the oil flow rate of a well. Reservoir engineering and well testing play essential roles in characterizing and detecting the problems associated with producing formations and wells.
This paper starts with problem identification in both producing and injection wells. It then covers the behavior and control of several types of reservoirs with different drive mechanisms. Applications of reservoir simulation and well testing in the control of water and gas coning, determination of the amount and the best appropriate treatment application, effect of formation layering, control of relative permeability to water, application of permeability blocking agents, water channeling, secondary recovery, viscous fingering, and polymer flooding are detailed in this paper. In additional several techniques to evaluate the overall success of a conformance control procedure are presented.
The purpose of this paper is to identify reservoirs associated with excessive gas or water production and to recommend the proper remediation. Conformance control begins with identifying the source of problem. This requires a thorough investigation of all aspects of well and reservoir parameters. The following is a list of some of the parameters required for identification of the reservoir performance problems.
Reservoir permeability and porosity
Relative permeability to oil, water, and gas
Connate water and irreducible oil and gas saturations
Net formation height
Oil, gas, and water production rates versus time
Original water-oil contact
Length of time before water production began
Location of all perforations
Fraction of the productive interval completed
Identification of fluid entry locations and the type of fluid entering wellbore (production logging)
Location and continuity of any shale layers
Identification of cement to casing bonding (bond log)
A new, alternative method, referred to as "direct synthesis", is proposedfor interpretation of pressure tests in naturally fractured reservoirs.
Direct synthesis utilizes the characteristic intersection points and slopesof various straight lines from a log-log plot of pressure and pressurederivative data. Values of these points are linked directly to the exact,analytical solutions to obtain reservoir or well parameters. The directsynthesis method offers the following advantages: (1) accurate results fromusing the analytical equations to calculate reservoir parameters, (2)independent verification is frequently possible from a third unique point, and(3) useful information is obtained when not all flow regimes are observed, as adirect result of the additional characteristic values developed by themethod.
Application of this technique is presented for single-well pressure tests inan infinite-acting naturally fractured reservoir with pseudosteady stateinterporosity flow, including the effects of wellbore storage and skin. Newanalytical and empirical expressions were developed as a result of this work.These expressions are an integral part of the technique, providing the desiredaccuracy and versatility. Several field examples are given to clarify thetechnique and also illustrate the accuracy of the method. When possible, acomparative analysis with other methods is included.
The basic equations for flow in naturally fractured reservoirs of dualporosity were originally formulated by Barenblatt, et al. Using continuummechanics, the rock and flow parameters of the two media, fractures and matrix,are defined at each mathematical point. The transfer of fluid between the twomedia is maintained in a source function, where the flow is assumed to bepseudosteady state in the matrix system. Warren and Root used thisapproach to develop a comprehensive and applicable solution to pressuredrawdown or buildup tests in a dual porosity, naturally fractured reservoir.From their work several flow regimes could be identified from semilog analysis.In chronological order there exist an early time, straight line representingfracture depletion only, a transition period when the matrix contribution toflow is dominant, and a late time, straight line which corresponds to the timewhen the entire reservoir produces as an equivalent homogenous reservoir.This late time, semilog straight line is parallel to the first straightline.
The Petroleum Industry is experiencing an increased awareness and understanding of sucker rod compression. This awareness has evolved during the past twelve years, from initial recognition of rod buckling, to a current desire to quantify (measure) the amount of compression required to initiate rod buckling.
Recent attention has focused on true or effective compressive loads in sucker rod strings. Measurement of these loads is being documented (1) and data collection is ongoing with improved technology.
This paper will provide a more accurate understanding of the amount of compression required to buckle sucker rods and sinkerbars of various diameters. This will be accomplished by presenting the following;
Predictive compressive loads that buckle various diameters and lengths of sucker rods and sinkerbars utilizing Euler loads.
Measured compressive loads that are required to buckle various diameter sucker rods and sinkerbars.
A comparison of predictive compressive loads to measured compressive loads.
Knowledge regarding the amount of rod string compression required to buckle various diameter sucker rods will provide the industry better rod string design guidelines.
Use of these guidelines will help identify dangerous compressive rod string loads which initiate rod-tubing contact, provide associated wear, and result in rod and/or tubing failure.
Euler Load (Critical Load, Pcr)
The famous mathematician Leonhard Euler (1707-1783), although hindered by loss of sight in one of his eyes in 1735 and the other in 1766, was the first to investigate buckling of slender columns. His investigations led to the determination of the critical load (Pcr) required to buckle an ideal, elastic column.
An ideal column is assumed to be perfectly straight and compressed by a centrally applied load (a load acting through the centroid of the cross section). An elastic column is assumed to reach stresses below the proportional limit before the critical load is reached.
The phenomenon of axial compression can be described by one of three (3) forms of equilibrium: stable, neutral or unstable buckling.
If the axial load (P) is less than the critical load (Pcr), the column remains straight and undergoes only axial compression. This straight form of equilibrium is called stable, which means that, if a lateral force is applied and a small deflection is produced, the deflection will disappear, and the column will return to its straight form when the lateral force is removed.
As the axial load (P) is gradually increased, a condition of neutral equilibrium is reached as the axial load (P) equals the critical load (Pcr). Under these conditions a small lateral force will produce a deflection which does not disappear when the lateral force is removed.