Search Documents

By L. H. Wolfe, J. O. Thieme

The tensile and shear strengths of frozen soil and the compressive strengths of ice and frozen soil were measured These tests showed that the strength of ice and of frozen soil increased as the temperature was decreased. A method is presented for the measurement of thermal conductivity and specific heat of earth and ice. Using the parameters thermal conductivity, specific heat and density, the thermal diffusivity was calculated. Results of thermal property tests establish water as the major variable contributing to thermal property values. INTRODUCTION During the last few years, a considerable interest has developed in underground storage of cryogenic liquids such as liquid natural gas. Because of its economy and safety, underground storage is being used. Before underground caverns could be seriously considered, however, the strength and thermal characteristics of the soil had to be obtained. Methods are presented herein for measuring tensile, compressive and shear strengths, specific heats and thermal conductivities. The tests were performed on two soils at temperatures from ambient to -195C. Because it is difficult to find data in the literature that pertain to a particular soil, ice was used to evaluate some of the experimental procedures. Some of our ice data are compared to the published values. PHYSICAI, PROPERTIES SOIL SELECTION AND SAMPLE PREPARATIONS The physical and thermal tests were performed on recompacted samples of two typical soils that might be encountered in underground caverns: a gray, fire clay that had been pulverized and mixed with water into a pliable mud; and a brown, sandy silt that was dug from a dry river bank and subsequent1y mixed with tap water. The moisture content of the test specimens was 17 to 22 per cent (based on the wet weight of the soil). This amount of water nearly saturated the silt, but the clay was well below the saturation point. The soils are partially described by the sieve analyses (Fig. 1) which show the particle size distribution of the coarser than 44 micron portion of the soils. The curves also show that approximately 75 per cent of the clay and 20 per cent of the silt are finer than 44 microns. The particle size seems to affect some of the physical properties, and the finer than 44 micron portion is important in the thermal conductivity test. Physical specimens were cut with a bandsaw from frozen blocks of the soil. The specimens were then sanded and kept frozen until after they were tested. They were stored at the approximate test temperature for several hours. Then, to equilibrate the samples to the test temperature, they were stored at the temperature at least an hour before they were tested. TEST EQUIPMENT An Instron, which is a constant speed gear-driven instrument, applied force to the samples and drew stress-strain plots of the test. All of the samples were strained at 0.02 in./min. All of the physical tests were conducted in a temperature cabinet that controlled the temperature to 1C. TENSILE STRENGTH Briquette or "dog-bone" shaped tensile specimens were pulled with suitably shaped jaws. This shape of specimen, which is usually used for tensile tests of concrete, is satisfactory for testing frozen soils except that the samples tend to

Jan 1, 1965

By R. V. Smith, M. H. Cullender

Rigorous equations for calculating subsurface pres.tures in flowing and static gas wells, and pressures along horizontal pipelines are presented in this paper. These general equations, based on the mechanical encrgy balance, contain no assumptions regarding temperature and can be used with either straight-line or curved temperature gradients. The friction factors recommended in this report are based on an absolute roughness of 0.0006 in. Flow is always considered to he turbulent. Although these general flow equations are solved by numerical means, the methods are as convenient as many of those used today to calculate pressures in gas wells and pipelines. These numerical methods are illustrated for flowing and static columns of gas in wells and for flow in pipelines. The authors believe that the use of these methods will result in more accurately calculated pressures in gas wells and pipelines and that [he methods are more flexible than any in use at the present time. Also, the methods are easily adaptable to automatic computers. INTRODUCTION The problem of calculating subsurface pressures in gas wells and along flowing pipelines has been studied by many investigators with the results that two widely-used methods have been presented in technical literature. One approximates conditions in a gas well or pipeline by assuming that temperature and compressi- bility are constants tor the entire gas column. The other method assumes that temperature is constant at some average value but permits compressibility to vary with pressure at the constant temperature. While these approximations may be justified for relatively shallow and low-pressure wells, they become more and more unrealistic as the calculations are applied to the deeper high-pressure wells. In wells or pipelines where there is an appreciable change in temperature between the inlet and outlet ends of the flow string, it is equally as important to consider the change in compressibility with temperature as it is to consider the change with pressure. Consequently, the methods presented in this paper were developed to provide a workable procedure for calculating pressures in gas wells and pipelines which makes no assumptions regarding either temperature or compressibility. The friction factors recommended for use with the equations were the result of study of published literature and extensive experimental work with gas wells. The F,. values in Appendix IV which contain the friction factor are equivalent to an absolute roughness of 0.0006 in. as calculated by the Nikuradse equation. The absolute roughness of 0.0006 in. was previously published by Cullender and Binckley2 for gas wells; it is in agreement with 0.00065 in. of Smith, Dewees, and Williams1 for gas wells and 0.00065 in. of Baker and Swerdloff4 for pipelines; it is in reasonable agreement with 0.0007 in. reported by the Bureau of Mines and the American Gas Assn.5 for large transmission lines. However, any type of friction factors may be used in the equations by a change in the F, values in Appendix [V].

Jan 1, 1957

By J. W. Wolfe

The Benedict-Webb-Rmbin equation of state was used in digital computer programs to make rapid determinations of natural gas equilibrium phase compositions. Mixture components were the nine hydrocarbons, methane through heptane. Comparative tests were made with data from field separators, plant exchangers and literature sources. For most tests made, the B-W-R equations predrcted component K-vallles, mixture enthalpies and phase aensities within tile accuracy of available correlations for rune-component mixtures: significant deviations were noted when mixtures contained components heavier than heptune. INTRODUCTION Rapid, accurate methods for calculating the properties of hydrocarbon mixtures are needed for predicting the behavior of natural gas processes. For example, because of the common use of two-phase separation methods in gas proccssing, calculations must be made for the density and composition of equilibrium vapor and liquid phases. In addition, predictions must often be made for the temperature and pressure changes associated with the flow of single- and two-phase natural gas fluids through heat exchangers, chokes and turbines. The complex nature of non-ideal mixtures makes rigorous calculations for these natural gas properties difficult and time-consuming. Natural gas properties have conventionally been determined with generalized data correlations. These correlations, such as the K-values presented in the NGAA Equilibrium Ratio Data Book, are usually based on some average property of the mixture and have proven to give satisfactory estimates for the properties of most oil and gas systems. Digital computers, however, have not only increased the utility of these correlations, but also have made practical the use of equations of state for calculating the properties of natural gases. Potentially, if basic pure component data are available, all the desired thermodynamic properties of mixtures can be derived from a single equation of state describing the mixture. Currently. the best equation of state for describing the behavior of pure, light hydrocarbons over single-and two-phase regions, both below and above critical pressure, is that of Benedict. Webb and Rubin.' The equation contains eight constants which are determined by empirically fitting the equation to pure-component PVT data. Benedict and his co-workers presented the clevelopment of their equation as early as 1940.2 The B-W-R equation is a modification of the Beattie-Bridgeman equation of state and gives better PVT predictions than the Beattie equation at densities above the critical density. Benedict et al. made the equation applicable to mixtures by relating the mixture constants to composition and the constants for pure compounds. These authors, in a serics of articles in 1951,' showed the accuracy of the B-W-R equation when applied to light hydrocarbons and their simple mixtures, and demonstrated the equation's utility for making vapor-liquid equilibria calculations. Since its inception, the B-W-R equation has been the subject of many papers. Numerous additions and modifications have been proposed,'-',"'-':' the purpose of most being to improve the equation's ability to predict PVT properties of specific mixture systems. In 1959, Opfell, Pings and Sage summarized much of the work on the B-W-R equation in their book on equations of state." Since natural gases are primarily mixtures of light hydrocarbons, the B-W-R equation is a natural choice for an equation of state to compute both phase equilibria and thermodynamic properties of natural gases. Consequent-

Jan 1, 1967

By David Cornell, D. L. Katz

Procedures for computing turbulent flow of gas in steady state near the well bore and a graphical method for predicting unsteady state laminar flow at distances from the well have been combined to compute pressure gradients in gas reservoirs. Methods are discussed for predicting single and multiple transients at constant flow rate in an infinite reservoir, predicting constant How rate. in a finite reservoir, reproducing and interpreting back pressure test data, and prediction of the behavior of a closed-in gas well. An example of the graphical method is given for a single transient and the results are compared to a published analytical solution. A typical calculation of the pressure gradient in a reservoir and the production of gas is made .starting with data from a hack pressure test. INTRODUCTION The calculation of pressure gradients throughouse a ga-reservoir at any point in its production history is a complex problem involving turbulent flow near the well bore and unsteady state flow of the gas from the reservoir. The problem is complicated further by the variety of boundary conditions that may be imposed upon the flow. Such boundary conditions include finite or infinite reservoirs. constant or variable rates of production, constant or varying bottom hole pressures, complex initial pressure distributions throughout the reservoir, and pressure maintenance through cycling. In addition to these problems there are the practical difficulties associated with natural gas reservoir analysis These might include: (1) variation. in the permeability and porosity throughout the formation. (2) variations in the thickness of the formation, (3) communication of the producing formation with other producing zones. (4) radial variations in permeability due to influx of drilling fluid. acidizing. or the buildup of liquid in the formation around the well, (5) partial penetration of the producing zone. (6) water flood, and many others. These problems will be considered eventually. In the mean- time, it is necessary to present methods of handling the case of radial flow through a homogeneous, regular, producing stratum to a single well uncomplicated by other factors. The idealized case of steady state, radial flow from a natural gas reservoir was studied by Elenbaas and Katz5 and by Mac-Koberts.' An analysis of the unsteady flow of gases through porous media has been made by Aronofsky and Jenkins' for the one dimensional laminar case with the boundary conditions of constant downstream pressure and uniform initial pressure. Solutions for the partial differential equation for unsteady state radial flow of liquids through porous media have been given by Van Everdingen and Hurst" for the constant bottom hole pressure and constant production rate cases for laminar flow in a reservoir initially at a constant pressure. The analogous heat transfer equation has been treated by Perry and Berggren9 for the case of quenching a cylindrical hole in an infinite medium to a constant temperature. The back pressure curve, which involves the variation of the bottom hole pressure as it is determined by the behavior of the reservoir as a whole. has been studied by Rawlins and Schellhardt,10 Binck-lev,' Baumel and Breitung,' and others. No complete. adequate treatment has been given for the unsteady state, radial flow of gases into a cylindrical hole from a porous medium with or without the presence of turbulent flow and for any set of boundary conditions. If one relies only on analytical solutions of the basic equations, a new solution must be obtained for each new boundary condition. Furthermore, the complexity of the mathematical expressions for boundary conditions other than the very simple cases limits the use of analytical solutions. Graphical methods exist, however. that are general in nature, accurate, and readily employed. The analysis of steady and unsteady, laminar and turbulent, radial flow for various boundary conditions with application to specific natural gas well problems by means of graphical procedures forms the content of this paper. STEADY STATE RADIAL FLOW EQUATION FOR LAMINAR AND TURBULENT FLOW A steady state radial flow equation for laminar and turbulent flow through unconsolidated sands has been given by Muskat as discussion of the work by Elenbaas and Katz.5 This equation is based on the properties of unconsolidated

Jan 1, 1953

By A. B. Cook, C. J. Walker, G. B. Spencer

Although water will displace oil from a petroleum reservoir to a greater extent than gas will, there are some reservoirs in which gas rather than water should be used for pressure maintenance. This is indicated by the high-percentage oil recovery in the case history reports on the Pickton field1 in East Texas and the Raleigh field2 in Mississippi. In gas cycling, total oil recovery includes vaporized products from the immobile oil in addition to oil produced by displacement. If a large part of the immobile oil is vaporized, total oil recovery may be higher than that obtainable from pressure maintenance by water injection. However, the amount of immobile oil vaporized may range from almost 0 to 100 percent, and no simple and reliable method has been presented in the literature for calculating oil vaporization. Development of such a method was the purpose of this study. Several papers concerning the calculation of oil vaporization have been published,3 6 and each is based on the concept dealing with K values — equilibrium constants for the various components in reservoir oils. (K for a component in a system of vapor and liquid in equilibrium is the mole fraction in the vapor phase divided by the mole fraction in the liquid phase.) This appears to be the most logical approach. Yet, the problem is too great for a perfect solution, even with modem laboratories to analyze oils and high-speed digital computers to perform calculations; the number of components in reservoir oils is too large. Because of their complexity no two reservoir oils are exactly alike. Also, K values change not only with variations in pressure and temperature but also with composition of the reservoir oil. In addition, during gas cycling, the lighter hydrocarbons tend to vaporize first. Thus, the reservoir oil becomes more dense and less volatile as gas cycling continues. Furthermore, the greatest amount of oil vaporization occurs near the injection well. Therefore, a simplified method is required for calculating oil vaporization because a rigorous method is not practical. The present methods of calculating vaporization generally require knowledge of the reservoir oil composition in which the mole fractions of the lighter components through hexanes are given; the remaining oil is described as C, + (heptanes and others of greater molecular weights). Assigning one K value for the C,+ system provides the simplest solution, but this incorrectly assumes that K for C, + does not change as gas cycling continues. This assumption can cause large errors because the K value may become less than a thousandth of the original after a large amount of oil vaporization. Another way to solve the problem is to determine changing K values for C, + according to how much of the C, + has been vaporized. To determine these values a sample of the reservoir oil is injected into a pressure-volume cell, and dry gas is batched in and out of the cell. Appropriate measurements and tests are made to determine K

Jan 1, 1970

By L. G. Jones

The objective of this paper is to present a better method for determining reservoir reserves of gas from extended drawdown test data. Means are proposed for conducting the test with or without the requirement of prior build-up to a static reservoir pressure. The test is based upon solutions to unsteady-state gas flow developed in an earlier paper and is believed to be superior to previous methods for determining reserves from flow-test data. The limitatiom of such tests are discussed, with emphasis on the subject of test duration. The test can supply useful information if the mechanics and limitations of it are understood. This type of information can be particularly valuable in deciding whether or not to drill offsets to producing wells. INTRODUCTION In the life of any gas or oil field, the engineer in charge must decide if certain wells need stimulation and whether or not to drill more wells in the field. Well production tests can aid in making these decisions. For instance, the value of stimulation can be estimated by determining formation permeability and the amount of well damage from drawdown or build-up tests. These tests are freely used in the oil industry. A less well-known type of analysis uses data from extended drawdown tests for determining reservoir reserves. Into this category fits the "reservoir limit test" of Jones,'-3 and the "reservoir reserve test" presented in this paper. The chief value of these tests is their ability to "see" reservoir boundaries and thus prevent costly dry holes. Methods for determining reserves of both gas and oil reservoirs were proposed in the work of Jones. His proposals appear sound for single-phase liquid flow. The generality of his work on gas flow is open to question, however, since the analysis is based upon solutions to liquid-flow equations. The test for gas reserves presented here is derived from unsteady-state gas-flow solutions presented in a previous paper.' In deriving this reservoir reserve test, it has been assumed that all of the reservoir energy is provided by expansion of the reservoir fluid and that the reservoir is radially symmetrical. No mention has been made of the effects of deviations from Darcy's law such as quadratic (turbulent) flow and well damage. Since the terms representing these effects are function of rate only, since the rate is constant and since the data are plotted in a differential form, it seems reasonable that these terms will not affect the test results in any way. No attempt has been made to estimate the effects of water drive, faults, or other departures from an idealized and isolated reservoir. For a qualitative discussion of some of these effects, the reader is referred to a paper by Jones3

By C. E. Turner, J. R. Elenbaas, R. D. Grimm, J. A. Vary, D. L. Katz

A technique is presented by which well samples and core plugs of dolomite formations are classified by microscopic examination into seven different porosity grades. Quantitative values of porosity and permeability are determined for each grade by a statistical correlation of the core plug test data with the porosity grading system. These quantitative values are applied directly to the grades exhibited in the well samples for the purpose of estimating the reservoir void space for wells that were not cored. The procedure is described for estimating the gas reserves per unit area lor the South Hugoton gas field. but a reserve estimate for the field is not given. INTRODUCTION The miscroscopic examination of well sample; and the graphic recording of their lithologic qualities and other distinguishing characteristics of various geologic formations drilled is both a science and an art of long standing and wide application. Usually the primary objective of a geologist who "sits on the well" and examines the samples are: to identify the formation being drilled, determine the total depth, casing point. and completion interval. In most cases the porosity is described. if done at all, in general terms. such as: trace, scattered, fine, poor, fair, medium. good, excellent, or in some other relative terms. In fields where various geologists have examined samples and recorded observations on many wells considerable variations in lithologic terms and porosity descriptions occur unless there is primary effort to establish uniformity of logging observations and standards of recording observable porosity. When an estimate of the pore volume of a reservoir is made a geologic concept of the processes that control the magnitudes of porosity and permeability is developed by microscopic examination of well samples. The characeristics and appearances are then mentally related to rather general quantitative units of porosity based on physical core data from the same reservoir or on such data or experience in other reservoirs that have similar qualities. The reliability of such estimates depends largely on the variations of the lithology of the formations, the geometric properties of its void system. the extent of comparisons of sample appearances with porosity data, as well as the uniform recording of all relevant characteristics. This statement is particularly significant for dolo-mitized limestone formations of substantial thicknesses and heterogeneity such as the Permian Dolomites of the Hugoton gas field. Jn this field, as well as in most of the Permian Dolomite fields, the producing formations are of relatively great thicknesses in which the porosity and permeability of the reservoir varies substantially in all directions, depending on the crystalline structure. degree and kind of impurities, kind of fossils anti cementation thereof, degree of dissolution. and fracturing. The variations of the lithologic texture of the dolomites and post deposition alterations have resulted in porosities and permeabilities of such magnitudes that only a part of the gross thickness can be counted as "pay." At the time of this study insufficient gas production had been experienced to apply the pressure decline production method in the South Hugoton Field and the electric logs are not definitive enough. The problem of estimating gas reserves in the south part of the Hugoton Field is primarily one of determining the pay thickness and porosity from well samples and core data. The area studied embraced all that part of the field lying south of an east-west line through Guvmon, Okla., anti containing approximately 1.000,000 acres. This paper describes a technique of correlation of physical core data with well samples so that quantitative values of pay thickness, porosity. Permeability, and connate water may be assigned to well samples that are representative of a given interval, and thereby permitting the estimation of gas reserves lor a given unit area. The procedure was developed by a uniform microscopic qualitative porosity grading of the dolo. mite core plugs.. and relating these grades to the respective physical core data on a statistical basis The well sample-were also graded in a Similar manner in order that the quantitative values established lor the core plugs could be applied to the well Sample for wells that were not cored. GRADING OF DOLOMITE A group of experienced geologists was given the assignment of examining the samples on all wells in South Hugoton in order that they could log their observations in a uniform and standardized manner and grade the observed porosity so that it could be related quantitatively to the core data. The group initiated the study 011 chips from cores which bad been tested for porosity and permeability. This study continued until all of the geologists developed a common knowledge of lithologic terms and of the characteristic appearances of the samples and their relations to measured porosity. The characteristic appearances of the dolomite samples under twelve-power magnification as related to their qualitative porosities afforded a classification of the dolomite into :even grades of porosity, ranging from dolomite of no-visible porosity under twelve-power magnification to dolomite of excellent porosity. The assigned grade for a specific 10-ft interval is a weighted average of all visible grades of porosity exhibited by the cuttings representing that interval. The porosity characteristics were recorded by a color graph adjacent to the lithology column in conjunction with a numerical system for further definition of relative porosity as shown in Fig. 1. The three vertical lines to the right of the lithology column each represent 33 1/3 per cent. which lines were used to record the percentage of the samples, for any particular interval. that showed porosity under the microscope. The colors were used to denote actual pore size. i.e., orange. blue and I-ed for pore diameter of one-fourth or less. one-fourth to one-Ilalf. and greater than one-half millimeter, respectively. The area colored 1)). one or more colors represents the percentages of the samples exhibiting pores of the respective size or sizes. The numerals from one to six inclusive shown on the log in

Jan 1, 1952

By E. G. Woods, A. G. Comer

A mathematical model is presented and solved for determinution of the saturation distribution and pressure in a radial gas-storage reservoir. The model consists basically of two parts: (I) the growing gas-bubble core, and (2) the surrounding aquifer. Since the total pressure at the injection well is a function of the two-phase flow in the gas bubble and the unsteady single-phase flow in the aquifer, the resistance to flow in both zones was taken into consideration. The assumptions involved for both the radial equivalent of the Buckley-Leverett two-phase-flow equation and the injection-pressure equation are as follows: (I) the geometry is radial, (2) the gas bubble is free to expand or contract, (3) compression or expansion of the gas within the bubble may occur at the beginning of a time step, (4) the fluids are immiscible, (5) water is incompressible within the gas-storage region whereas it is compressible outside of this region, (6) a stable gas-water interface exists and (7) gas injection occurs at a constant rate or a series of constant rates. The mathematical model was solved numerically using an IBM 650 computer. A comparison is presented between the predicted results of the model, the results assuming steady-state flow and the actual initial injection-pressure history of an operating reservoir. Using the initial field pressure for a basis, the average deviation between the predicted pressure and the actual field pressure was less than 4.3 per cent. INTRODUCTION Underground storage of natural gas in abandoned oil fields, in abandoned coal mines, in caverns and in aquifers has had varying degrees of success. The lack of these first three facilities in the vicinity of most major gas marketing areas is leading more and more to storage in virgin aquifers. For the most part, this storage in aquifers has led to the study upon which this paper is based. Several excellent papers have been published discussing the problems involved in the underground storage of natural gas; however, to these writers' knowledge, none of them considered the two-phase flow of fluids in the gas bubble and the fact that the gas bubble will grow or shrink in size depending upon the injection and withdrawal history of the reservoir. Not only will both of these factors influence the required injection pressure, but also they will have a definite bearing upon the amount of water production occurring upon withdrawal from the storage area. To solve this problem on an IBM 650 computer, several simplifying assumptions were made. Only the drainage portion of the relative-permeability curve was considered, thereby neglecting any hysteresis effects introduced during an imbibition cycle. The quantity of gas withdrawn during the period of the field study was only a small percentage of the total gas in place; therefore, this assumption is justified. For the problem studied in this case, the combined effects of capillarity and gravity were assumed to be negligible. This assumption becomes less valid as the formation thickness or the average pore diameter increases markedly. The justification of these assumptions should be studied in each case. THEORY This mathematical model for the prediction of saturation distribution and pressure for a radial gas-storage aquifer is based upon the equations for radial two-phase fluid flow'.' and for radial, unsteady-state, single-phase fluid flow.' The two-phase flow is considered to take place in a "core" whose radius is equal to the maximum radius that the gas zone will attain, as indicated in Fig. 1. The injection-withdrawal history is approximated by a series of constant flow rates. Within the "core", gas is assumed to behave as a semi-compressible fluid; that is, the gas is assumed to have a constant density based on the average gas-zone pressure for the flow period. Between each constant-rate time increment, the gas density is allowed to change. This involves an iterative solution of the pressure equation. The liquid phase in this zone is considered to be incompressible, whereas it behaves as a compressible liquid outside of this region. The equations of incompressible flow in the "core" may be written for the gas phase

By A. F. Bertuzzi, M. R. Tek, F. H. Poettmann

A method is presented for predicting pressure drop for two-phase fluid flow in horizontal pipes. A set of 267 experimental measurements randomly sampled from approximately 1,000 measurements from various literature sources was used. Pressure gradients calculated by the procedure developed and compared with the experimental values showed a bias of +0.82 per cent and a standard deviation of 20.8 per cent. The advantatges of this method over other available methods of predicting two-phase flow pressure drop are (a) its comparative simplicity of application, (b) its relative independence of flow patterns, (c) its accuracy, which on the basis of a statistical evaluation predicts pressure drop closer than other available methods, and, (d) its ability to satisfactorily correlate laboratory data from various sources while the correlations from these sources do not appear to agree with one another. The method has been reduced to a simplified graphical procedure, suitable for field use. A two-phase f factor is defined and correlated with parameters involving the flowing gas-liquid mass ratio, a Reynolds number for the gas phase, and a Reynolds number for the liquid phase. The choice of parameters allows the correlation to reduce to the usual f factor plot for the limiting conditions of (111 gas or all liquid. INTRODUCTION The mechanics and characteristics of two-phase flow systems have been of interest throughout the industry for some time. In numerous engineering installations such as pipe lines, chemical reactors, and heat exchangers, two-phase flow conditions are of every day occurrence. In oil production operations it has been desirable, in some cases, to consider transporting gas and oil together in a common pipe from oil field to process plant. The trend toward centrally located stock tank batteries in oil fields has resulted in longer gathering pipelines in which more than one fluid phase is flowing. The increase in the producing capacity of oil wells due to new production tech- niques has created the need for review and re-design of many surface gathering lines for properly handling the increased production. Optimum pipe size for the situations described above has become an important factor. The problem which is of interest here involves the ability to predict the relationship between pressure drop, fluid properties, fluid rates, pipe diameter, and pipe length. Although the literature1,5,10,13,18,19 contains various articles dealing with specific cases of the problem, Lock-hart and Martinelli17 proposed the most general solution. Their method has been later modified by Baker:1 Alves2 demonstrated that different flow patterns are possible for a flow mechanism defined by Martinelli such as gas turbulent, liquid turbulent. Baker explained deviations experienced with Martinelli's correlations on the basis that different flow patterns could exist for the same flow mechanism. Using Martinelli's method of correlatioil, Baker correlated data for each flow pattern for the turbulent-turbulent flow mechanism. Bergelin and Gazley6,7 presented a correlation for predicting gas phase

Jan 1, 1957

By R. D. Carter

Nurnerical .solutions are presented to some problems of unsteady-state radial flow of gas in which Darcy's law holds. These solutions are intended to aid in explaining the observed behavior of gas wells during drawdown tests. The variation of viscosity and Z-factor with pressure is accounted for in the solutions. Results are presented for three types of problems: (I) unfractured wells producing at a constant rate, (2) fractured wells producing at a constant rate and (3) unfractured wells producing through a critical flow prover. In the critical-pow-prover solutions, a correction of the calculated well-bore pressure to account for non-Darcy flow is made. A good correlation is shown between the solutions to the problem of Type 1 and the corresponding diffusivity-equation solution. Using this correlation, the diffusivity solution may be used to obtain gas-well solutions. The solutions to the problems of Type 2 illustrate the difference in drawdown-curve behavior to be expected between fractured and unfractured gas wells. Methods for estimatting fracture radius and flow capacity are presented and verified from solutions of Type 2. The solutions to the problems of Type 3 illustrate the differences to be expected between critical-flow-prover drawdown tests and constant-rate tests. An equation is presented for correcting critical-flow-prover data to constant-rate conditions. A description of the methods used in obtaining the solulions appears in the Appendix. INTRODUCTION It is necessary for many engineering purposes, such as predicting future deliverability performance, to be able to relate bottom-hole flowing pressure, formation pressure and flow rate for conditions of stabilized flow in gas wells. This information is customarily given as a plot of on logarithmic co-ordinates. It is desirable to be able to predict stabilized performance curves for gas wells based on data obtained from drawdown tests of short duration. Tests of long duration may be wasteful of gas, and low-permeability wells do not stabilize in a short-duration test. Although it is now commonly recognized that non-Darcy flow plays a significant part in gas-well behavior, it is believed that an understanding of transient behavior of systems obeying Darcy's law is still important in the interpretation of short-term test data, and the prediction of stabilized performance curves therefrom. In this connection, an equation for correcting Darcy flow It seems likely that solutions of the type presented herein, combined with the non-Darcy flow correction of the type of Eq. 1 and the usual laminar-flow skin factor, could provide a model which would adequately explain transient gas-well behavior for times after the non-Darcy flow region (which should be confined to a small area surrounding the well) has been established. The solutions presented in this paper are for three types of problems: (1) unfractured wells producing at a constant rate, (2) fractured wells producing at a constant rate and (3) unfractured wells producing through a critical flow prover. In the problems of Types I and 2, no corrections for non-Darcy flow were made in the results presented, since it was reasoned that the constant-rate boundary condition would facilitate such a correction if it was desired later. For the problems of Type 3, however, a condition of varying flow rate required that a non-Darcy flow correction of the type of Eq. 1 be included. This was done. Based on the numerical results presented, a number of approximations and a correlation are presented which may be useful in analyzing gas-well test data. It is assumed that the system can be described as a finite cylindrical gas reservoir produced by a single cylindrical well located at the center of the reservoir. Flow is strictly radial.

By R. C. Craze

This paper discusses the characteristics of the velocity logs now available to the petroleum industry, and some of their advantages and limitations. The velocity log was designed as an aid to the geophysicist to determine velocity layering in the earth and to provide fast, accurate and economical vertical travel time data. After a number of logs had been made, studies were carried out on their use ia correlating strata in adjacent wells. It was found that adjacent strata usually show sufficient velocity contrast to make it possible to correlate these from well to well, even in areas where electric or radioactive logs fail.' In addition, velocity logs have been useful in the determination of porosity and thickness of reservoir rock.' These uses for velocity logs require an understanding of the characteristics of the log in response to the changes in velocity and in borehole conditions encountered normally in logging. GENERAL DISCUSSION OF INSTRUMENTATION The measurement made in velocity logging is more accurately described as a measurement of time rather than velocity. Ideally the log is obtained by measuring the length of time required for compressional sound waves to travel through a fixed distance in the formation. In practice, the time measurement may be made accurately but the distance through which the waves have traveled may vary slightly. Changes in borehole diameter are responsible largely for the difficulty in maintaining a fixed distance, especially if the diameter variations are so abrupt and frequent that the borehole instrument cannot follow contours of the hole. Several service companies have velocity logging equipment available or under development which have different transmitters, receivers, and measuring systems. While these differences are important, there are actually only two systems in use at the present time (Fig. I)."' The first system initiates a pulse at a trans-mitter and detects it at a single receiver. The transmitter and receiver are separated by a fixed distance, called the spacing. The time measured is that between the initiation of the pulse and the first arrival of acoustical energy at the receiver. The second system uses two receivers and a transmitter. The receivers are separated by a fixed distance and the time between arrivals of energy at the two receivers is measured. Here, the spacing is the distance between receivers. The two systems have much in common for it is possible to record both logs simultaneously. Usually the transmitter is an electro-mechanical transducer which produces discrete

By M. P. Tixier, K. S. Kunz

A method for the interpretation of temperature curves recorded in gas producing wells is described. One essen-tial feature of the method is a simple graphical construction which conveniently reflects the amount of gas flowing from a given producing formation. This method, applied in conjunction with the Induction Log and radioactivity logs, is valuable to determine the points of entry of gas and to estimate the over-all thickness of the producing zones and their respective contributions to the total production. The method in particular is helpful in evaluating the effect of fracturing operations. An outstanding result of the application of the method is to bring to the light the relative importance of sand-shale interfaces in the production of gas. The paper is illustrated with field examp1es. INTRODUCTION The electrical logging tools involving conventional systems are not applied in wells which do not contain water base mud (empty holes, or wells filled with oil base mud), because of the absence of electrical connection between the electrodes and the surrounding formations. It is, therefore, common practice in this case to use other methods which do not require the presence of a conductive fluid in the borehole: resistivity log with scratcher electrodes, Induction Log, radioactivity logs, temperature log. . . . Such conditions are encountered in some regions, as the San Juan Basin, N. Mex., where a technique of drilling gas wells, using gas as a drilling fluid, has been introduced in the last two years. By the middle of 1953, an appropriate method for the logging of these wells was devised on the basis of field tests and of theoretical studies. It is the purpose of the present paper to give a description of this method. In these wells the logging operations are performed during the process of gas production, so that the temperatures in the boreholes are strongly affected by the cooling due to the expansion of the gas. On the other hand, a special thermometer, with a high sensivity and a short time constant, has been recently made available, which has enabled much more detailed and accurate recordings of the borehole temperatures than with the conventional instruments. As a result of these two circumstances, the temperature log constitutes the essential element of the method, in making possible the detection of the gas producing zones, the definition of their boundaries, and the estimation of the respective contributions of each separate zone to the total production. The method also involves the radioactivity and Induction Logs, as auxiliary documents. These logs in the present instance are helpful in the determination of the reservoirs. However, the information they give in this respect is limited. This comes from the fact that, due to the absence of the MicroLog and/or MicroLaterolog and of the SP log, the estimation of formation factors and connate water salinities is somewhat conjectural. Furthermore, little help is obtained from the Neutron log for the evaluation of porosity and of formation factor because the variations of gas saturation affect the Neutron log in a way similar to the variations of porosity. It should be said also that even an exact knowledge of the gas saturation of the reservoir beds would not

Jan 1, 1956

By L. Cichowicz, K. K. Millheim

The constant-rate drawdown test performance for a low-permeability, verticany fractured gas well was investigated. A series of gar wells were tested by flowing each well at constant rate until the data could be analyzed using con-ventional radial flow theory. Each well was then shut in to build up. After a sufficient buildup was obtained, another flow Iest commenced but at a higher flow rate than Ihe first test. Again, the well was shut in when radial fiow was obtained. his procedure was repeated for three to four different flow rates. Two wells in the San Juan basin were tested using this procedure. Both wellere fractured after completion, cleaned up and then shut in until flow testing commenced. Test designs of both wells permitted investigation of the most realistic values of egective permeability, wellbore radius and turbulence factor. Also, being able to determine the eflective fracture flow area and vertical fraclure efficiency was inherent with this testing approach. It was observed that fractures in both wells influenced the Pressure behavior for approximately Is to 40 hours (depending on the flow rate) before radial flow was evident. After this time, drawdown data were analyzed using radial pow theory. When a low-permeability gas well has vertitally oriented, induced fractures, the early flow geometry ic. essentially linear. It will be shown how to determine when a flow test has been conducted long enough so lha' the most representative values of effective permeahiiity. wellbore radius and turbulence factw can be calculated. From the linear pressure data, valuable information about The fracture treatment, such as the effective flow area and vertical fracture efficiency, can be determined for vertically froctured wells. Introduction During tests on gas wells in the Sari Juan basin7 initial transient behavior lasted for many days because of the low permeability of some porous media. As a result, stabilized flow performance could not be obtained. If these wells received some type of stimulation treatment, early pressure behavior deviated from conventional theoretical radial Row. When conventional radial flow theory was used to analyze these low-permeability fractured gas wells, larger values of flow capacity and absolute open flow potentials (AOF) sometimes resulted. Wells were assigned open flow poten- tials that proved to be 3 to 10 times higher than the well would sustain over a longer period of production. In some cases where the wells had flowed for longer periods of time during a constant-rate drawdown test, it was noticed that the effective flow capacity appeared to be decreasing with time until a certain value was reached. The early nonradial pressure behavior can be explained If linear flow is assumed. Rusell and Truitt mathematic. ally investigated the vertically fractured well in a bounded area. They showed that early flow behavior was essentially linear and, for x,/x. approximately less than 0.10 radial flow was obtained after short periods of time. Then realistic values of effective permeability and skin could be determined. Scotta experimentaliy studied the vertically fractured well with a heat flow analog. He showed that earb flow was linear, Both studies indicate that, for small values of x,/x,, linear flow approaches radial flow if the well is tested long enough, To help prove this concept of early linear flow caused by induced vertical fractures, two low-permeability gas wells were tested. Both wells received large fracture treatments prior to testing. A vertical fracture was indicated from the analysis of fracture treatments, As anticipated, tests of both wells indicated early linear flow that was later followed by a period of radial flow, Data collected from each well were analyzed. From the well tests, plus other information on each well, the effective permeability, wellbore radius and turbulence factor were calculated. Effective fracture flow areas calculated from test analyses for each well proved to be approximately one-fourth the created area calculated from classic hydraulic fracturing theory: other fractured wells that were tested but not presented in this paper also indicated that the effective fracture flow area was one-fourth to one-third the created area predicted from hydraulic fracturing theory. The vertical fracturing efficiency was estimated from the calculated values of effective wellbore radius and fracture flow area. For the two wells tested, calculated fracture lengths x, were 112 and 105 ft, and the vertical fracturing efficiencies E, were 122 and 183 percent. Development of Flow Model Agnew' showed that most induced fractures below 1,500 ft are vertical. Anderson and Stah15 indicated that most of the fractures they studied were vertical. Thc model proposed for early flow in most vertically fractured gas wells is shown by Fig. 1. This model should approximate early flow behavior until radial flow is reached, at which time a radial model with an effective wellbore radius of 0.5 they will

Jan 1, 1969

By G. Thodos, R. B. Grieves

A method has been developed for the accurate calculation of the cricondentberm and cricondenbar temperatures of multicomponent hydrocarbon mixtures of known composition. The mixtures may contain any number of components including paraffins and isoparaffins, olefins, acetylenes, naphthenes and aromatics. This approach is based upon the mole fraction of the low-boiling component in the mixture and graphically presents the ratios of the cricondentherm and cricondenbar temperatures to the pseudocritical temperature as {unctions of a boiling-point parameter. The only requirements are the critical temperature, normal boiling point and the approximate vapor pressure behavior of each component. For mixtures of more than two constituents a stepwise calculation procedure is necessary where the pseudocritical temperature is based upon the critical temperature of the pure low-boiling component and upon the actual cricondentherm or cricondenbar temperature of the mixture of all the remaining higher-boiling components. From an analysis of 22 binary systems (123 compositionsl, the average deviation of calculated cricondentherm temperatures from reported values is 0.87 per cent, based on degrees Rankine; and for 10 multicomponent mixtures containing from three to six components, the average deviation is 0.98 per cent. From an analysis of 18 binary systems (104 compositions) the average deviation of calculated cricondenbar temperatures from reported values is 0. 79 per cent; and for 15 multicomponent mixtures, 1.47 per cent. Equations, derived from the graphical relationships, are presented which enable rapid calculations for both binary and multicomponent systems. INTRODUCTION With the increasing tendency towards the use of high pressure processes, the need for an accurate method of calculation of the properties of a multicomponent hydrocarbon mixture in the critical region is becoming more essential. Critical region phase relations are also significant in gas condensate reservoirs, for which a knowledge of the fluid phase boundaries and regions of retrograde condensation make it possible to evaluate optimum operating conditions. The possibility of liquefaction in the underground structure may be established, and information on the feasibility of the use of repressurizing may be determined. A typical phase diagram for a multicomponent hydrocarbon mixture is presented as Fig. 1. This figure also shows the regions of retrograde behavior encountered by following lines AB and AC. If accurate methods of calculating the temperatures and pressures at the critical, cricondentherm (maximum temperature) and cricondenbar (maximum pressure) points of multicomponent hydrocarbon mixtures of known composition were available, it would be possible to estimate closely the entire phase diagram of any mixture. Several methods have been presented in the literature for the estimation of the critical temperatures and pressures of multicomponent mixtures containing all types of hydrocarbons and including the fixed gases.4-8,21 Considering the cricondentherm and cricondenbar points, the work is more limited. Etter and Kay5 have developed rather complex equations for

By A. M. Rowe

A considerable quantity of experimental hydrocarbon K-factor data has been correlated as a function of component identity, temperature, pressure and convergence pressure. To utilize these correlations effectively, convergence pressure must be determined accurately, particularly for volatile mixtures near their critical states. This paper presents phase diagrams that illustrate physically the meaning of convergence pressure. A new method, referred to hereafter as the "critical composition method", will be outlined for calculating convergence pressure. An example calculation has been included to illustrate how to use this new technique. INTRODUCTION The principle of hydrocarbon phase composition calculations as applied to such diverse problems as optimizing separator performance or predicting fluid compositions at various stages of reservoir depletion is the same. Usually, temperature, pressure and the numbers of moles of the various components of the fluid contained in a given volume are known. Questions answered by the phase calculation are what fraction of the total mass of fluid exists in each of the equilibrium phases, and what are the mole fractions of the various components in the two phases? To answer these questions the Natural Gasoline Supplymen's Assoc. (NGSMA) has correlated a considerable quantity of K-factor data as a function of temperature, pressure, component identity and convergence pressure.l To use these correlations to obtain the best answers possible, one must be able to calculate the convergence pressure. This is particularly true for over-all fluid compositions in the neighborhood of the critical state. STATEMENT OF THEORY Use of convergence pressure as a correlating parameter is based on a postulate similar to the law of corresponding states used in correlating PVT data of hydrocarbon gases. This postulate, which proposes convergence pressure as a correlating parameter, has been stated as follows: "The equilibrium vaporization constant for one component in a complex system is the same as the equilibrium constant at the same temperature and pressure for the same number or kind of components, providing only that the convergence pressures of the two systems are exactly the same at the same temperature and that the components are of the same homologous series".2 This law, as with all laws of physics, cannot be proven theoretically. It can only be justified by experimental data supporting its premises. Arguments have been made that this law violates Gibb's phase rule. For example, consider a four-component system. By Gibb's phase rule, which is thermodynamically rigorous, f = c — P + 2 =4, for a four-component, two-phase system. Thus, four independent intensive variables must be specified to establish completely all the intensive variables of the equilibrium phases. On the other hand, according to convergence pressure theory, only three variables need be specified for any mixture containing four different components. These variables are temperature, pressure and convergence pressure. Thus, the two laws appear to be in conflict. However, the convergence pressure postulate is more restrictive than Gibb's phase rule. It applies only to mixtures of the same homologous series. Hence, these two concepts are not in disagreement. PHASE DIAGRAMS DESCRIBING CONVERGENCE PRESSURE This paper presents phase diagrams grading from the simple two-component system to the more complex four-component system to illustrate convergence pressure. A less abstract definition of convergence pressure will be stated that is

By R. E. Holmes, T. W. Leland

The solution to the unsteady-state mass- and heat-transfer equations describing the adsorption of a dilute component from a gas stream flowing through a packed bed is readily applicable to the design of hydrocarbon recovery units. The solution to these equations depends on the assumption of a linear equilibrium equation to relate concentrations on the surface to concentrations in the gas. This assumption is shown to be justifiable for components which are very dilute in the gas stream. This approach permits the prediction of the effects of bed shape and size, cycle time, and gas mass velocity on the recovery of each component. It also allows the time and gas flow rate needed for regeneration to be determined. The equations involve a constant for each component and another constant characterized by the type of adsorbent. These constants must be evaluated from operating data. Values of these constants determined for silica gel are presented. This paper points out the type of experitrrental data needed for a thorough design of hydrocarbon adsorption units. INTRODUCTION The growing importance of adsorption processes to recover hydrocarbons from natural-gas streams has created an urgent need for better methods to design and predict the performance of these processes. Since the original use of solid adsorbents in natural-gas processing was for dehydration, most of the current design methods for hydrocarbon recovery re extensions of those developed earlier for dehydration. These methods are based on an adsorptivity or bed loading factor defined as the amount of a particular component adsorbed per unit weight or per unit volume of the packed bed. The recovery of a given component is then expressed as an empirical function of its adsorptivity on a given adsorbent. In spite of the simplicity and initial utility of these methods, they are not completely satisfactory in that they do not predict the effect of the dimensions of the bed in relation to flow rate and cycle time, and do not predict the distribution of various components along the length of the bed. It is possible, however, to make a considerable improvement in the design of these processes by applying the basic equations of mass and heat transfer applicable to gases containing small concentrations of adsorbing components flowing through granular beds. These equations have been developed and solved by Anzelius,' Furnas,% nd Hougen and Marshall." where G = mass velocity of the inert constituents (assumed to be methane), lb/hr-sq ft, y = weight of an adsorbing component in the gas per unit weight of inert, lb of a given constituent per lb of methane, x = bed length, ft, p, = packed bed density, lb/cu ft, P, = gas density, Ib/cu ft, f = void fraction in the bed, t = time, hours, and w = weight of a given component adsorbed per unit weight of adsorbate free solid, Ib/lb pure adsorbent. The term on the left side of Eq. 1 represents the flow rate per unit bed volume of a component entering less the flow rate per unit bed volume of that component leaving the differential bed length. The terms on the right side represent, respectively, the accumulation rates per unit bed volume on the solid and in the gas phase within the differential bed section. The last term is negligible for this application since pgfv is quite small in comparison with p, and G. The partial differential is determined by the mass-transfer rate between gas and solid in the differential section.

By R. Al-Hussainy, P. B. Crawford, H. J. Ramey

The effect of variations of pressure-dependent viscosity and gas law deviation factor on the flow of real gases through porous media has been considered. A rigorous gas flow equation was developed which is a second order, non-linear partial differential equation with variable coeficients. This equation was reduced by a change of variable to a form similar to the diffusivity equation, but with potential-dependent diffusivity. The change of variable can be used as a new pseudo-pressure for gas flow which replaces pressure or pressure-squared as currently applied to gas flow. Substitution of the real gas pseudo-pressure has a nrtmber of important consequences. First, second degree pressure gradient terms which have commonly been neglected under the assumption that the pressure gradient is small everywhere in the flow system, are rigorously handled. Omission of second degree terms leads to verious errors in estimated pressure distributions for tight formations. Second, flow equations in terms of the real gas pseudo-pressure do not contain viscosity or gas law deviation factors, and thus avoid the need for selection of an average pressure to evaluate physical properties. Third, the real gas pseudo-pressure can be determined numerically in term of pseudo-reduced pressures and temperatures from existing physical property correlations to provide generally useful information. The real gas pseudo-pressure was determined by numerical integration and is presented in both tabular and graphical form in this paper. Finally, production of real gas can be correlated in terms of the real gas pseudo-pressure and shown to be similar to liquid flow as described by diffusivity equation solutions. Applications of the real gas pseudo-pressnre to radial flow systems under transient, steady-state or approximate pseudo-steady-state injection or production have been considered. Superposition of the linearized real gas flow solutions to generate variable rate performance was investigated and found satisfactory. This provides justification for pressure build-up testing. It is believed that the concept of the real gas pseudo-pressure will lead to improved interpretation of results of current gas well testing procedures, both steady and unsteady-state in nature, and improved forecasting of gas production. INTRODUCTION In recent years a considerable effort has been directed to the theory of isothermal flow of gases through porous media. The present state of knowledge is far from being fully developed. The difficulty lies in the non-linearity of partial differential equations which describe both real and ideal gas flow. Solutions which are available are approximate analytical solutions, graphical solutions, analogue solutions and numerical solutions. The earliest attempt to solve this problem involved the method of successions of steady states proposed by Muskat.' Approximate analytical solutions' were obtained by linearizing the flow equation for ideal gas to yield a diffusivity-type equation. Such solutions, though widely used and easy to apply to engineering problems, are of limited value bemuse of idealized assumptions and restrictions imposed upon the flow equation. The validity of linearized equations and the conditions under which their solutions apply have not been fully discussed in the literature. Approximate solutions are those of Heatherington et al.. MacRobertsl and Janicek and Katz.' A graphical solution of the linearized equation was given by Cornell and Katz. Also, by using the mean value of the time derivative in the flow equation, Rowan and Clegg' gave several simple approximate solutions. All the solutions were obtained assuming small pressure gradients and constant gas properties. Variation of gas properties with pressure has been neglected because of analytic difficulties. even in approximate analytic solutions. Green and Wilts8 used an electrical network for simulating one-dimensional flow of an ideal gas. Numerical methods using finite difference equations and digital computing techniques have been used extensively for solving both ideal and real gas equations. Aronofsky and Jenkins"I " and Bruce et al.11 gave numerical solutions for linear and radial gas flow. Douglas et al." gave a solution for a square drainage area. Aronofsky 13 included the effect of slippage on ideal gas flow. The most important contribution to the theory of flow of ideal gases through porous media was the conclusion reached by Aronofsky and Jenkins" that solutions for the liquid flow case'" could be used to generate approximate solutions for constant rate production of ideal gases. An equation describing the flow of real gases has been solved for special cases by a number of investigators using numerical methods. Aronofsky and Ferris 10 onsidered linear flow, while Aronofsky and Porter 17 considered radial gas flow. Gas properties were permitted to vary as linear functions of pressure. Recently, CarteP 18 proposed an empirical correlation by which gas well behavior can be estimated from solutions of the diffusivity equation using instantaneour values of pressure-dependent gas

Jan 1, 1967

By R. F. Hinds

A pressurizing system was designed and built to apply a radial pressure of 5.000 psi to rock samples. Samples of the Bradford, Weir and Kirkwood sandstones were subjected to radial pressures parallel and perpendicular to the bedding in the system and the changes in conductivity, porosity and permeability were deter-mined. Asymptotic decreases in conductivity, porosity and Klinkenberg permeability were observed over the 0- to 3,500-psi range. From these data, the increase of formation factor over the same range was calculated. The cementation exponent expressed in the Archie equation was found to increase with pressu,re. The rate of change varied from formation to formation. Only small differences were observed for samples taken parallel to the bedding as comnared to those taken perpendicular to it except for the Bradford sand. Above 3,500 psi the data are not good enough to tell if the properties continue to fall-off asymptotically or in a different manner. A possible explanation for the probable litnit of the asymptotic rate around 3,500 to 4,500 psi is that this represents the maximum pressure which the formations have unrdergone during their geological life as a result of burial. INTRODUCTION For many years measurements have been made in the laboratory of the properties of oil-bearing sandstones with little consideration being given to whether these properties had been changed when the samples were moved from their original position in the ground to the laboratory. Relatively little has been reported in the literature on the effect of pressure on the properties of sandstones. The changes in permeability for eight typical consolidated oil-bearing sandstone samples with overburden pressure (0- to 15,000-psi range) was presented by Fatt and Davis' in graphical form. In 1953 Fatt' also reported the relationship between porosity and overburden pressure for four typical consolidated oil-bearing sandstone samples (0- to 5,000-psi range). Fatt later described the changes of formation factor and porosity of 20 typical oil-bearing sandstone samples with overburden pressure changes (0- to 5,000-psi range) giving the values of the coefficients and exponents in the Archie' equation, and the more general Winsauer equation', where C, k and m are empirical constants, F is the formation factor, and + is the porosity. During the experiments of this investigation the core was contained in a neoprene rubber sleeve as shown in Fig. 1. It was held between two steel end members which were mounted in a steel cylinder. The unassembled pressure cylinder with its component parts is shown in Fig. 2. The core was insulated from the two steel end members by a Teflon sleeve with a small hole in the

By R. G. Agarwal, Ramey Jr. H. J., Al-Hussainy R.

Although it has long been realized that gas recovery from a water-drive gas reservoir may be poor because of high residual saturations under water drive, it appears that only limited infomlation on the subject has been available until recently. This study was performed to show the qiiantitative potential importance of water influx. Results indicate that gas recovery may be very low in some cases: perhaps as low as 45 per cent of the initial gas in place. Gas recovery under water drive appear to depend in an important was on: (I) the prodirction rate and manner of production; (2) the residual gas saturation; (3) aquifer propertie.); and (4) the volumetric displacement effciency of water invading the gas reservoir. The manner of estimating water-drive gas reservoir recovery can vary considerably. Examples are: the steady-state tnethorl. the Hurst modified steady-state method, and various unsteady-state methods such ac. those of van Ever-dingen-Hurst, Hurst, and Carter-Tracy. The Carter-Tracy water influx expression was used in this study. In certain cases, it appears that gas recovery can be increased significantly by controlling the production rate and manner of production. For this reason, the potential importance of water influx in particular gas reservoirs should he investigated early to permit adequtrtr planning lo optirtize the pay reserves. INTRODUCTION In recent years, the economic importance of natural gas production has become increasingly apparent. This has been evidenced by more intensive exploration efforts aimed at gas production, and exploitation of both deep, as well as low-permeability gas reservoirs. Technical developments such as deep-penetration fracturing have made development of such formations economically feasible. Unfortunately, water influx has forced abandonment of a number of gas reservoirs at extraordinarily high pressures. Although reservoir engineering methods for estimating water influx have long been available, it appears that application of these methods to the water-drive gas reservoir has been sporadic.'a Available methods for estimating water influx which can be applied to the water-drive gas reservoir problem include the steady-state method,1 the Hurst modified steady-state method and various unsteady-state methods such as those of van Everdingen-Hurst. Hurst, and Carter-Tracy. Interesting applications of these solu- tions to gas reservoir and the aquifer gas-storage problems have appeared recently.3,12,14 The experimental study of residual gas saturations under water drive by Geffen et al. in 1952 indicated that residual gas saturations could be extremely high. A value of 35 per cent of pore volume is often used in field practice when specific information is not available. The study of Geffen et al. showed that residual gas saturation might be much higher in some cases. Naar and Henderson concluded that the residual non-wetting phase saturation under imbibition should be about half of the initial non-wetting phase saturation. The Naar and Henderson result that residual gas saturation under water influx should be about half the original gas saturation is recommended as an estimate if laboratory measurements are not available. Thus, it is clear that a considerable portion of the initial gas in place might be trapped in a water-drive gas reservoir as residual gas at high pressure. A full water-drive would result in loss of residual gas trapped at initial reser.voir pressure. Consideration of transient aquifer behavior leads to the conclusion that high-rate production of water-drive gas reservoirs could result in improved gas recovery by reduction of the abandonment pressure. However, there appears to be little quantitative information on this possibility. One of the few advantages of water-drive gas production appears to be improved deliverability through water-drive support of the reservoir pressure. There may also be an advantage in higher condensate recovery caused by pressure maintenance for gas-condensate water-drive reservoirs. In view of the preceding, this study was made to assess the potential importance of water-drive in gas reservoir engineering. The Carter-Tracy approximate water-influx expression was used because this equation offers some advantages in hand-calculation which do not appear to have been generally recognized.' However, calculations were performed in the main with a high-speed digital computer to permit evaluation of the effect of water-drive under a large variety of conditions. CALCULATION METHOD Water-drive gas reservoir performance can be estimated in a manner completely analogous to oil reservoir calculations: a materials balance is written for the reservoir, and a water influx equation is written for the aquifer. Siniltaneous solution provides the cun~ulative water influx and the reservoir pressure. When reservoir performance data (gas produced and reservoir pressures) are available, it is usually possible to match performance data to determine the initial gas in place and the water influx parameters —

Jan 1, 1966

By M. H. Cullender

The performance characteristics 01 gas wells producing from formations which fail to stabilize within a relatively short period of time are obscured by the interrelation of the coefficient (C) and the slope "n" of the conventional back-pressure curve q = C(P2 f- p2s)n An empirical method is presented whereby the characteristic or true slope (11) of the back-pressure curve may be determined for a particular gas well. For those gas wells which do not stabilize within a relatively short period of time, the coefficient (C) is considered as a variable with respect to time and as a constant only with respect to a specific time. Thus, the back-presrure performance of such a well is presented as a series of parallel curve5, each curve representing the performance of the well at the end of a given time interval. Through use of the method presented, a simple presture gradient is developed and maintained within the drainage area around a producing gas well during the test period in order that the variation of the coefficient (C) with respect to time does not obscure the true value of the lope. INTRODUCTION The back-pressure method of testing gas wells as set forth by Schellhardt and Rawlins1 is dependent upon the requirement that a series of flow rates and corresponding pressure data be obtained under stabilized conditions. These data arc then plotted on logarithmic coordinates of flow rate (Q) vs difference in squared pressures (P2f — P2s) in order to determine the constants (C) and (n) for Equation 1. With respect to this method of presenting the performance characteristics of a gas well, it should be pointed out that although Equation 1 is an empirical relationship, the form of the equation was justified to a considerable extent by an independent development by Muskat and Botset.2 This development was theoretical in nature; however, certain assumptions with respect to the type of flow existing (viscous vs non-viscous) were madc which may or may not be correct. As the use of the method presented by Schellhardt and Rawlins spread through the industry, it became evident that the method of testing was applicable for those wells which approached stabilized prbducing conditions within a relatively short period of time. Performance characteristics could not be determined by this method, however, for wells which approached stabilized protlucing conditions slowly over a considerable period of time. This characteristic of "slow stabilization" has generally been associated with gas wells producing frorn reservoirs of low permeability. As the demand for natural gas continued to grow through the years, a great number of wells were completed in reservoirs of low permeability which exhibited "slow stabiliz:~tion" characteristics. Consequently, the problem of determining the true performancc characteristics of these wells became more apparent. It is with this problem in mind that the isochronal performance method of determining the flow characteristics of gas wells is presented.

Jan 1, 1956