Search Documents

By Thomas S. West

A gravel pack completion is described which, in addition to sand screening, permits the exclusion of water and gas from the oil production. A gravel pack composed of alternate layers of different particle sizes is employed which has horizontal but not overall vertical permeability to drilling mud. A packer is run on the tubing and set in the screen member, around which the gravel layers are placed. The hydrostatic pressure of a mud column prevents the entry of fluid above (or below) the level of the packer. The producing level may be changed by raising or lowering the tubing. Vertical coning is controlled by replacing the mud column with oil and continually injecting oil into the sand section above (or below) the packer. Three well histories with this completion are described in detail. Statistics on the first 100 completions of this general type are given. INTRODUCTION Gravel pack type completions have heretofore been almost exclusively for the purpose of preventing sand production. This paper describes and presents well data on a gravel pack procedure which is primarily a selective completion method. Although complete sand screening is achieved, the exclusion of gas and water is the primary objective. The application of this procedure is therefore not limited to unconsolidated sands. Application to hard sand or limestone reservoirs is equally feasible. GRAVEL PACK PROCEDURE Figs. 1 and 2 illustrate the gravel pack procedure employed. The gravel is placed by means of the frangible disc type dump bailer shown by Fig. 1.' This bailer has its lower end closed by a plate glass or steel rimmed cement disc. This disc breaks when the bailer passes over the liner allowing the gravel to be deposited around the liner. The liner employed is usually 2 3/8 in. to 2 7/8 in. OD and 1.995 in. or 2.07 in. ID (ID of 2-in. tubing or 2-in. line pipe). Two tvDes,. of slotted sections have been employed. One type employs horizontal hacksaw cut or milled slots of about 1/32 in. width opposite the producing section. The other type employs the same type,. of horizontal slot. but slots are placed in groups of six at intervals of about 8 in. A cage composed of half round vertical bars, spaced apart, is then welded to the pipe over the slotted sections. This liner has the bottom end closed, usually by bull plugging. The upper end is closed, while placing the gravel, by a cap which is held in place by means of a shear pin. This cap has a mandrel attached to it which extends downward inside of the liner to within a few inches of its lower end. The top of the liner cap is formed to fit a releasing type wire line core barrel overshot. The liner is lowered to bottom on the wire line by means of this overshot and released. The length of the liner is usually suffi- cient to extend to about 20 ft above the highest level to which gravel is to be placed, the upper end of the liner always being in the casing. The dimensions of the liner and bailer are such with respect to the casing size that the bailer will always pass over the liner. After releasing the liner the bailer is run to bottom. Passage of the empty

Jan 1, 1951

By E. S. Mardock, Robert E. Bush

The object of this paper is to report recent developments in the quantitative interpretation of radioactivity logs. The use of reference lines is described in the application of the new zero radioactivity reference line for the determination of neutron derived porosities is discussed. Illustrated examples are shown of logs from the reef limestone of the North Snyder Field and from the Lansing-Kansas City limestone of central Kansas. A method of using an interrelation factor calculated from the neutron and gamma ray curves, called the neutron Productivity number, for the estimation of productivity from sandstone formations is discussed. When used in a relative manner the neutron productivity number has aided in Predicting whether a given zone will produce gas or liquid or whether it is too shaly to produce. Examples of logs from wells in the Freites District, Venezuela, and in East Texas are shown to illustrate the application of the neutron productivitv number. INTRODUCTION The petroleum industry was introduced to some of the preliminary results of investigations into the quantitative interpretation of radioactivity logs in October, 1949." It was stated at that time that the methods presented were entirely relative because they involved the use of reference points obtained from the radioactivity log, i.e., shale reference lines, dense zone reference lines, or a combination of both shale and dense zone reference lines. These relative methods offered many advantages to the industry in interpreting radioactivity logs in terms of porosity with particular application to limestone and dolomite reservoirs. However, these relative methods did not offer the wide range of applicability desired because the esta1)lishment of reliable reference points from shale in some areas were difficult. It was recognized that a more reliable mean; of establishing reference lines was vitally needed. Such a method is now available through the introduction of an instrumeutal zero device. It is the purpose of this paper to discuss the application of this new zero reference to the quantitative interpretation of radioactivity logs. Previously all applications of quantitative interpretations of radioactivity logs have been made in limestone reservoirs with little or no emphasis placed on sandstone reservoirs. A brief resumé of investigations into the quantitative interpretation of radioactivity logs in sandstone reservoirs is presented here. This discussion deals primarily with the interrelation of the gamma ray and neutron curves and the response of the latter to the density of the fluids filling the pore space in relatively homogeneous sandstones of uniform porosity. The authors believe that the methods presented herein are only applicable locally to specific fields and reservoirs, and that they probably do not offer a wide range of applicability. However, tile authors believe that the wider use of the new instrumental zero references in sand areas will reveal more information about the specific response of the gamma ray and neutron curves to changes in lithology and saturation. Two applications of quantitative interpretation using the new instrumental zero references are presented. Examples are given which show the comparison of logs and porosity in the limestones and dolomites of West Texas and Kansas. Logs of wells in 'Venezuela and East Texas are compared with the drill stem test data, production data, etc.. to illurtrate the interrelation of the gamma ray and neutron curve: and the rise of methods involving these interrelations to determine certain saturation variables. THE QUANTITATIVE APPLICATION OF RADIOACTIVITY LOGS TO LIMESTONE AND DOLOMITE RESERVOIRS Tile theoretical basis for the gamma ray and neutron curves has been discussed elsewhere.9 he method of recording the logs and apparatus involved has also been thoroughly covered and will not be discussed here.'."".'" It has been shown that the neutron curve responds exponentially to liquid filled porosity over the 1 1/2 to 35 per cent

Jan 1, 1951

By R. H. Olds, B. H. Sage

The transient flow of gases in long pipe lines is a problem of industrial interest. The present discussion deals with the application of the conservation of momentum and material to the transient flow of gas in conduits which are long with respect to their diameter. Solutions of the resulting equations include consideration of variations in friction and specific volume of the fluid as functions of time and position along the conduit. The changes in state along the pipe line are established as a function of time. The methods proposed may be extended to other situations. The flow of compressible fluids under steady conditions' has been considered in detail. Primary emphasis was placed on the development of simple means of taking into account the non-uniform flow' realized in long conduits under such conditions. Binder2 discussed the essential differences between the flow of compressible and uoncompressible fluids. Bonilla" proposed direct methods of solution for equations describing isothermal flow in long pipes. Similar considerations limited to steady flow have been discussed by Ruth.' Palsgrove 5 sugrested a graphical means of solving specific problems relating to the flow of gases in conduits of uniform section. Some of the basic concepts of the flow of a compressible viscous fluid in straight tubes have been described by Kuo.6 Problems of this type require knowledge of the equation of state of the fluid. In many cases it liar been convenient to assume that the fluid follows the relationship ascribed to a perfect gas. The treatment of unsteady flow of compressible fluids is less complete. This state of affairs result* from the fact that the differential equations relating pressure and flow rate with time and position for unsteady flow are nonlinear in character.' Muskat and Wyckoff presented an excellent treatment of the situation in laminar Bow where inertia forces were not of controlling importance. Somewhat earlier Hurst9 discussed the uncteady flow of fluids in petroleum reservoirs. Helhering-ton. MarRoberts. and Huntington'" established experimentally the pressure distribution and accumulative flow rate of natural gas through an unconsolidated sand. Recently Joffe11,12 conhitlcretl the specific problem of the flow of natural gas in relatively long.,. pipe lines for unsteady conditions with emphasis upon the economic feusibility of the storage of gas in such a reservoir. These discussions were based upon approximations of the partial differential equations applying to the condititons of flow. The results were descriptive of the physical situation and indirated the effertiveness of large pipe lines as cyclic storage facilities for natural gas. The advent of analog computers" has increased materially the practicality of solving nonlinear partial differential expressions. Several iterative numerical and graphical methods14,15,16 have been developed to solve nonlinear equations. However, it appears that the pair of simultaneous nonlinear partial differential equations involved in problems of this nature are not directly susceptible to solution by presently available analog computer or formal iterative numerical techniques. Nevertheless. graphical analyses afford a means of obtaining data of engineering value without special equipment. MATHEMATICAL CONSIDERATIONS Since the length of the conduit is large in comparison to the cross section in the cases to be considered, it is possible to treat the characteristics of the flow as independent of position in the cross section of the conduit. This possibility follows since the energy- connected with the velocity profile is small compared to that required to overcome the friction associated with its movement throughout the length of the line. Average values of the pressure. velocity. kinetic energy. momentum, and specific volume may he assigned to the fluid at any particular cross section. From a mathematical viewpoint this may be expressed in the following way for velocity: µ = µ(?X)..........(1) The uncertainties introduced by such assumptions usually are negligible when the ratio of length to diameter is greater than the order of 1.000. On this basis a volume element ma!. be considered to occupy the entire cross section of the conduit. Such an element is presented in Fig. 1.

Jan 1, 1951

By Herman Dykstra, R. L. Parsons

A numerical method for solving partial differential equations in steady state fluid flow is described. This method, known as the "relaxation method," has two advantages over analytical methods: (1) practically any problem can be solved, and (2) a solution can be obtained quickly. A disadvautage is that the solution is not general. The method is applied to core analysis and relative permeability measurement to calculate constriction effects and to calculate the true pressure drop measured by a center tap in a Hassler type relative permeability apparatus. Further applications are suggested. INTRODUCTION Many problems in fluid flow cannot be solved analytically because of the nature of the boundary conditions. For many problems, however. an exact answer is not necessary because boundary conditions are not exactly defined or the parameters describing the porous medium are not accurately known. The relaxation method can be used to obtain an approximate answer easily and quickly for the flow of incompressible fluids in porous media. The method can also be used for other types of problems, such as determining the stress in a shaft under load. or the temperature distribution during steady state heat flow. In this discussion only calculations concerned with the flow of fluids in porous media will be considered. The method was introduced by R. V. Southwell in 1935.' THEORY The treatment given here follows that given by Enimons.2 Consider a porous medium to be replaced entirely by a net of tubes of equal length and uniform cross-sectional area as shown in part in Fig. 1. Assume that the net of tubes behaves exactly like the porous medium which it replaces; that is, the net can be made fine enough to reproduce exactly the porous medium. Assume also that Darcy's Law can be used to calculate the flow from one point to another point through these tubes. The flow from point 1 to point 0 is KA . ------ P-P) .......(11 where a is the distance between points: K is the "permeability" of a tube; A is the cross-sectional area of a tube; is the viscosity of the liquid in the porous medium; and (P1 — P0) is the pressure difference between point 1 and point 0. In like manner the flow can be calculated from points 2, 3, and 4 to point 0. The net flow into point 0 is Qo = KA/µa (P1 + P2 + P33 + P4-4P0) . . (2) MB For an incompressible fluid the net flow into point 0 will be zero or, Q. = 0. This says that at point 0 fluid is neither being accumulated nor depleted. 'Therefore. P1 + P2 + P3 + P4 - 4P0 = 0 .... (3) . If. now. with specified boundary conditions. the pressure i.; known at a finite number of points in a given region, as at the points shown in Fig. 1, Equation (3) will be satisfied at every point. If, on the other hand, the pressure is not known, the pressure can be guessed at these points. Then. unless the guess is perfect. Equation (3) will not be satisfied at all of the points. When Equatiol~ (3,) is not satisfietl. let d = P1 + I?, + P, + P, - If' .,....(4) where 6 is an apparent error and is called the residual at point 0. Equation (4) shows how much the pressure guess is in error at point 0 with respect to the surrounding points. A positive residual means that the pressure is too low, and a negative residual means that the pressure is too high. To bring the residual, 6. to zero in order to satisfy Equation (3). it is necessary to make changes in the pressure guesses. Equation (4) shows that a +1 change in Po will change the residual at point 0 by -4. A +1 change in the pressure at any of the four surrounding points will change the residual at point 0 by +l. Thus it can be seen that a change at any point will affect the residual at that point and the four surrounding points. By changing the pressure from point to point, all of the residuals can eventually be brought nearly to zero and the problem will be solved. This procedure is the essence of relaxation methods and is used to relax the residuals so that Equation (3) is satisfied at every point. The procedure can be most easily explained in detail by solving a simple problem. as Southwell says, "To explain every detail of a practical technique is to risk an appearance of complexity and difficulty which may repel the reader. A

Jan 1, 1951

By D. W. Akins

The case history of a combination gas and water flood instituted early in the life of a field is described. It was recognized from the beginning that recovery would be low, under normal production methods, from the several separate, thin beds of porous limestone which had such poor native permeability that the wells would not produce without acidizing. It is thought that the combination flood gives excellent recovery efficiencies for this type of reservoir. The chief function of the gas flood now is to maintain bottom hole pressure in that part of the reservoir which will not be affected by the water injection program until later. Present estimates are that ultimate recovery will be approximately double the recovery expected under the original solution gas drive mechanism. Wide (80-acre) spacing has enhanced the economics of the over-all field program by the elimination of more than $5,000,000 worth of unnecessary wells. INTRODUCTION The Haynesville Pettit Field was discovered in November, 1941, with the completion of T. L. James Co.'s Callie Akin No. 1, located in the W 1/2 of the NE 1/4 Section 27-23N-8W, Claiborne Parish, La., at an approximate depth of 5,300 ft. The center of the field is about two miles north and a mile west of the town of Haynesville, La. A total of 189 producing wells have been drilled on 14,500 acres. An additional nine dry holes were drilled. The Louisiana-Arkansas state line divides the field so that 34 wells are in Arkansas and 155 are in Louisiana. Governmental sections are divided equally to contain eight production units making 80-acre spacing where the sections are regular. Along the state line in Louisiana there is an irregular tier of sections, each of which contains approximately 456 acres. When divided equally into eight production units the spacing is 57 acres per well for these three and a fraction productive sections. All Haynesville Pettit wells are identified by their section number followed by their position in the section as determined by numbering the production units counterclockwise beginning in the upper right corner of the section. All of the producing wells in the field except two are unit-ized and operated by the Haynesville Operators Committee. These two wells are located in the extreme northern part of the field in Arkansas. Efforts to bring them into the Unit are being continued. The Pettit operators organized the Haynesville Operators Committee and the field became a unit operation May 1: 1944. GEOLOGY The Pettit is Lower Cretaceous in age and is defined as the uppermost porosity in the Sligo formation. The unitized area known as the Haynesville Pettit Field is composed of two separate zones of porosity approximately 40 ft apart and without connection between the two, except in wells. These zones are known as the Upper Pettit and the Lower Pettit. The Upper Pettit is separated into two sections throughout most of its area by a relatively thin but quite impervious shale streak. These are called the "A" and "B" sections of the Upper Pettit. All of these lenticular zones are composed of coquina limestone and a small percentage of normal oolites. The large

Jan 1, 1951

By E. S. Messer

A knowledge of the magnitude of the irreducible inter.;titial water in a porous medium is so important to petroleum engineering that its determination has become routine in core analyses. The method of determination, being a production problem, should encompass the basic requirements of simplicity in technique and calculations, with reproducible results obtainable in a short interval of time. The results of the evaluation tests outlined in this report indicate that the evaporation method for determining the irreducible water is a technique which meets the requirements. The procedure consists, as the name implies. of permitting the saturant in the pore spaces to evaporate until only an irreducible volume remains. The determination of this volume can be made either graphically or by a mathematical comparison of fluid flows; the time required for each determination being dependent on the fluid used. When fluids other than those having reservoir characteristics were used, a volume factor had to be calculated which was based on the relative volume of various liquids adsorbed on grain surfaces and retained in pores. This factor made possible the calculation of an irreducible water volume when more volatile fluids such as toluene and benzene were used as the saturants. Also presented is the theoretical discussion necessary for the calculation of the capillary pressure as determined from the evaporation curve. A comparison is made between the calculated values and those obtained by experimental means. INTRODUCTION In all geological formations there exists, in the pore spaces of the rock structure, water that is held in a state of equilibrium between capillary and hydrostatic forces. "Interstitial water" is the term given to this water and is defined as that water coexisting in the pore space with the oil prior to exploitation. The term ''connate water" has often been used synonymously with this term; however, this can be true only by a specific definition since, geologically, it means the water in place at the time the rock structure was formed. The quantity of the interstitial water is a variable factor in any formation, since it depends on the hydrostatic forces present in any multiple-phase system. These forces may become unbalanced by the introduction of an extraneous force such as the raising or lowering of the "water table" or the migration of oil into a water-filled formation. Any unbalanced force results in a change in the interstitial water. There exists, however, an irreducible interstitial water. for a particular sand, that is the fraction of the pore space occupied by water when the capillary pressure at the particular point in question is at an equilibrium with the hydrostatic head of the oil sand in the reservoir. For this discussion the term "irreducible water saturation" will be used in place of "irreducible interstitial water saturation" for the sake of brevity; however, they are understood to be identical. A great amount of work has been devoted to the theory and methods for studying the irreducible water saturation and its related capillary pressure. As a result of the publications of Leverett;' Hassler, Brunner and Deahl;2 Calhoun and Lewis;3 and others, the role of capillary pressure studies is being accepted by the industry as a tool for studying suhsurface phenomena. Many techniques have been developed and published for determining the capillary pressure and irreducible water. In general, these techniques may be grouped into three classifications. One of the first was the capillary pressure method described by Leverett1 and expanded by Bruce and Welge.4 The experimental results were compared with water saturation of cores obtained using oil-base mud. Thornton and Marshall compared the irreducible water saturation of core samples determined by the capillary pressure method and by salinity and reported good agreement between the two methods. The second classification for determining the irreducible water and capillary pressure may be referred to as the "centrifugal force method." The general technique is similar to the capillary pressure method except that the force driving the reservoir fluid from the sample is of a centrifugal nature. A complete description of this method was presented by J. J. . McCullough and F. W. Albaugh.6 A process, the reverse of the capillary pressure method, was presented by W. R. Purcell.7 Mercury under pressure is driven into the pores of the rock and the saturation of the core determined at each applied pressure. The resulting capillary pressure curve is used to evaluate the irreducible water saturation. The techniques mentioned are singular in their approach to the irreducible water saturation. In all cases. an external force was applied to the core. The forces employed in the evaporation method are the vapor pressure of the liquid causing evaporation, the kinetic diffusion forces. adsorptive forces and. to a lesser degree, the viscous forces resisting flow to the surface. The basic definition of irreducible water is that water held in a state of equilibrium between capillary and hydrostatic forces This water has been described by previous investigators as being held in the microcapillaries too small to support fluid flow. Actually, this fluid volume is made up of the water in the microcapillaries and as a film adhering to the surface of the crystals. All capillaries. therefore, possess some liquid as a film, the thickness of the film being dependent on the properties of the fluid and solid. A discussion of experiments with references pertaining to the measurement of this immobile layer next to the solid surface can be found in the text by J. J. Bikerman.8 Eversole and Lahr calculated the thickness of this layer to be in the order of 10 ' to 10' cm for aqueous solutions and glass. Between two quartz surfaces they found the thickness to be 2 x 10 cm. The work of Volkova, on the capillary movement of water and toluene in quartz grains, indicated the thickness of the Immobile layers to be near 10' cm. Since any measurement is an average value, it is easy to understand that an absolute value would depend on the roughness of the surfaces involved and the complexity of the system. A calculated effective pore radius of 2 x 10 cm is obtained at the, irreducible saturation of a porous media in a water-air system when a capillary pressure of 100 psi is applied. Since the separation of the sand grains is of the same approximate magnitude as the immobile layer.

Jan 1, 1951

By A. D. K. Laird, John A. Putnam

This paper describes the application of x-ray theory to design procedures in connection with fluid saturation determinations during fluid flow experiments with porous media. A reliable and rapid method for calibrating the x-ray apparatuy is described. Extension of the method to fluid saturation determinations in three-fluid systems is described. INTRODUCTION In rerearch on oil production problems a method is required which will give quickly the quantity of each component of a fluid flow system present at any cross-section of a porous medium. The sample of porous medium under investigation is usually referred to as a core. The ratio of the volume of one component to the total fluid volume is defined as the saturation of the porous medium by that component. This ratio is generally given as per cent saturation. Some means of measuring saturation which have received consideration include: electrical conductivity of the fluids;1,2 emissions from radioactive tracers dissolved in the fluids; the radioactivity of silver caused by reflection of neutrons from hydrogen atoms in the fluids;' the attenuation of a microwave beam. the diminution and phase shift of ultrasonic wave trains.4,5 and the reduction in intensity of x-ray beams in passing through the fluids. X-rays have already been used with some success. Since every material has a different power to absorb x-rays, the reduction in intensity of an x-ray beam as it passes through a core depends on the fluids present. The strength of the emergent beam can be found by converting its energy into a measurable form such as heat or ionic current. or by its effect on a photographic plate or fluorescent screen. The beam strengths could be interpreted as quantities of known fluids in the core if, previously, these beam strengths had been identified with a known combination of the same fluids. With some fluid cornbinations it might be desirable to dissolve powerful x-ray absorbing materials in one or more of the fluids, to increase the differences in the beam strengths for various fluid saturations. Boyer, Morgan and Muskat6 have described a method of measuring two component fluid saturation. One component was air or water; the other. minerat seal oil in which was dissolved 25 per cent by weight of iodobenzene to increase its absorbing power. The x-ray source was a tungsten target tube operated at 43 kv potential. The beam emerging from the core was measured as ionic current flowing across an air-filled ionization chamber by means of an amplifying circuit and galvanometer. Another portion of the beam from the x-ray tube was passed through a metal plate and measured in another ionization chamber. This portion, called the monitor beam, was used as an indication of the performance of the x-ray tube. The galvanometer readings were calibrated against air-oil core saturations, gravimetrically determined. The method was apparently established by experimental means. In the present investigation the available theory of x-radia-tion was surveyed with a view to extending the usefulness of the method and to developing design procedures for its application to measurement of fluid saturation in porous media. Application of the theory permits prediction of relative meter readings to be expected for any combination of porous matrix, various saturating fluids and auxiliary filtering media. It is thus possible to calibrate the equipment in terms of fluid saturation by an indirect but rapid technique. The results of calculations based on x-ray theory indicate. and results of the saturation calibration technique confirm. that a valid measurement of the saturation of the core can be made for any two components and in some cases for three components. THEORY The strength of an x-ray beam, after it has passed through a distance. 1, of matter of density, p, and mass absorption coefficient, µ at a given wavelength, A, may be expressed by the absorption formula I = I0 e ...........(1) where I, represents the intensity of the incident x-ray beam and I is the intensity of the emergent beam. The expression e is called the transmission factor of the material. The variation of I,, with wavelength depends upon the materials through which the x-ray beam has previously passed and upon the spectral distribution of energy at the source of the x-radiation. A group of curves. called spectra. which show the variation of intensity with wavelength and x-ray tube voltage are given in Fig. 1. These curves represent the general radiation from a tungsten target tube. When the tube voltage is greater than 69.3 kv, the characteristic radiation of the tungsten is emitted and is superposed on the general radiation. At a given voltage the minimum wavelength A,,,,, at which energy can be emitted by an x-ray tube is given by the formula 12,340 xml. = ——..........(2) volts where A,.,,.. is in Angstrom units. The wavelength at which the spectra have maximum intensity a1so decreases with increasing x-ray tube voltaue. The area under each curve represents to an arbitrarv scale the total energy emerging from the x-ray tube for that voltage. The variation of µ with wavelength has been determined for many substances and may be found in such references as those by Compton and Allison7 and by Hodgman.8 The phenomenon of absorption is composed chiefly of the capture of photons by the atoms of the absorbing material with associated displacement of electrons, and of the scattering, or the deflection, of the photons by the atoms. Curves of these mass absorption coefficients show jump discontinuities. or absorption edges. at wavelengths which are short enough for the photons,

Jan 1, 1951

By Irwin Politziner F. M. Townsend, L. S. Reid

Recently published data indicate that the water vapor content of a gas, as determined by dew point measurement, is inaccurate when the gas has been dehydrated with diethylene glycol. Water vapor contents and dew point measurements of gases dehydrated with triethylene glycol have been obtained in this investigation in order to determine the magnitude of a similar error, if one exists. Experimental data show a very low concentration of triethylene glycol vapor in gases dehydrated at atmospheric ternperatures and pressures ranging from 500 to 2,500 psia and that the accuracy of dew point measurements is not impaired by the presence of triethylene glycol vapor. INTRODUCTION Nearly all natural gas transported by pipe line to northern and eastern markets must be dehydrated to a low residual moisture content in order to assure maximum transmission efficiency and continuous delivery under peak load conditions. Without dehydration, water vapor often condenses in pipe lines in quantities sufficient to restrict the flow3 and. under low atmospheric temperature conditions, gas hydrates may form and plug gathering. transmission and distribution facilities. Gas may be dehydrated to pipe line specifications by a number of methods. The most widely used methods are (1) adsorption of water vapor on a solid dessicant such as activated bauxite, alumina or silica gel,' (2) by absorption of water vapor by either diethylene or triethylene glycol-water solutions of high glycol concentration,11,13 and (3) by simultaneous expansion-refrigeration of very high pressure gas in which gas hydrates are purposely formed and then quickly decomposed.'"" Of these three fundamental methods, the adsorption and absorption processes provide the bulk of the dehydrated gas moving to market at the present time. The expansion-refrigeration process. a relatively new development, is gaining in favor because of its simplicity but is sharply limited in its application by available differentials between source and gathering system pressures. Where gas sale or purchase contracts contain a maximum water vapor content specification, the water vapor content is usually determined by measuring the dew point of the gas at system pressure. Knowing he dew point temperature and the system pressure, the water vapor content of the gas is determined from a graphical correlation of saturation, or dew point, temperature us the water vapor content of the gas which is customarily expressed as pounds of water vapor per million cubic feet of gas measured at 14.7 psia and 60°F. Dew points are usually measured with the U. S. Bureau of Mines Dew Point Tester. shown in Fig. 1, which is well suited for use at temperatures lower than 120°F and at pressurei; ranging from atmospheric to 3,000 psi. A number of correlations of dew point temperatures vs water vapor content are available,3,5,15 the most recent of which is that by McCarthy et al.4 Where natural gas is dehydrated by absorhing contact with concentrated diethylene glycol-water solutions. considerable difficulty is experienced in observing water dew points due to condensation of liquid hydrocarbon and glycol films on the mirrored surface of the dew point tester. Riesenfeld and Frazierl2 report that the dew point method for determining the water vapor content of a dietllylene glycol-treated natural gas is in error due to the glycol content of the condensate formed on the tester mirror, and that the water vapor content of the gas tested is actually lower than that indicated by the standard' dew point-water vapor content correlation. Extensive experience in testing gases dehydrated by concentrated triethylene glycol-water solutions lias not revesled difficulties of the nature cited above for diethylene glycol. Water dew points are clear and sharp and condensation of

Jan 1, 1951

By H. G. Doll

A new electrical logging method called Laterolog is described which provides for better recording of formation resistivity. In this method a current, preferably of constant intensity, is forced into the formation perpendicularly to the wall of the hole as a sheet of predetermined thickness by means of a special electrode arrangement and of an autonlatic control system. The essential advantages of the Laterolog over the regular resistivity logs are explained. With the Laterolog, the mud column has much less influence than with conventional methods. The Laterolog is, therefore. particularly useful in wells drilled with highly conductive mud. Moreover. when the electrode system is located opposite a bed of moderate or small thickness, the disturbing effect of the adjacent formations becomes practically eliminated, provided the bed is thicker than the sheet of current. which itself is usually a few inches to a few feet thick. The sequences of beds are. therefore, much more sharply differentiated, and, in many cases, it is possible to read directly from the logs values close to the true resistivities of the formations without further corrections. Field examples are shown to illustrate the results of the method in various types of formations. INTRODUCTION As soon as it was introduced into the petroleum industry, electrical logging roved to be a powerful instrument for the delineation of the strata traversed by bore holes and for correlation of such strata." Research and engineering efforts were mostly devoted at this early time to the design of the techniqurs which could provide the best data for these purpoes. Later, attention was directed toward the quantitative analy5i.s of electrical logs, in order to obtain information on. reservoir characteristics, and, in particular, on oil and water saturation.3,4 Efforts have been made, accordingly, with a view to improving the already existing methods, or to finding new methods which could satisfy these new requirements." Quantitative analysis implies a determination as accurate as possible of the true resistivity of the formations. In fact, however; the "apparent resistivity" recorded in front of a given bed by a conventional electrical logging method is frequently quite different from the true resistivity of this bed. This is due to the combined influence of the mud column, of the adjacent formations both above and below the bed, and of the invaded zone, in which the original fluid has been more or less replaced by mud filtrate. When the beds are thick enough, and, as is frequently the case in sand and shale formations, when their resistivity is not much greater than that of the mud, it is generally possible to obtain their true resistivity directly on the electric log, or at least to obtain an approximate value which can be corrected without undue difficulty by means of recently published charts known as "departure curves."" When the beds are thin, and particularly when, in addition. their resistivity is substantially greater than that of the mud column, the apparent resistivities recorded on conventional logs are seriously affected, and it is generally very difficult, if at all possible, to estimate their true resistivity with a good enough accuracy. In the most severe cases, such as are encountered in holes drilled through hard formations with muds of high salinity, the conventional logs are so distorted and rounded that they do not even show clearly the houndaries of the different beds. In order to overcome the effect of the mud column and of the adjacent formations, it is possible to make use of appro. priate systems whereby the measurements involve a comp3ratjvelv thin slice of space at the level of the measuring device. A mention is due, in this respect, to a measuring system. called the "guarded electrode;" invented by C. Schlumberger over 20 years ago. The system, schematically represented in

Jan 1, 1951

By M. Muskat

Calculations are reported on the differential sensitivity of the updip invasion of oil strata of varving permeability to the driving pressure differential. It is assumed that the water-oil interfaces advance with sharp fronts, or that the microscopic displacement efficiency is independent of rate and capillary pressure effects. It is found that, under such conditions, limitations of withdrawal rates will not greatly inhibit the "fingering" and bypassing tendency through high permeability strata unless the differential fluid head across the original oil column is of the order of or exceeds about 80 per cent of the driving pressure differential. In practice, such restrictions will be feasible only for oil reservoirs and aquifers of high permeability and dip, but may develop automatically under combined gas injection and water drive operations. INTRODUCTION A commonly expressed view regarding the intrusion of edge-waters into oil producing formations is that the geometry of such advance is rate sensitive. Specifically, it has been suggested that high rates of advance, stimulated by rapid withdrawals, will lead to "fingering" and an irregular type of invasion, whereas at low withdrawal rates the water encroachment will be more uniform. Two factors tend to support this view. One is the resultant effect of capillary pressure and gravity gradients in the displacement processes' involved in water intrusion, which are magnified at low rates of flow so as to improve the recovery efficiency. The other is the direct retarding effect of the rising water head as it advances upslope into the producing area. Capillary pressure forces give rise to a rate sensitivity of the oil recovery by water displacement in uniform strata as well as differential effects in non-uniform formations. These are basically independent of the effect of gravity and would also occur in horizontal beds or when the oil and water have the same density. Unfortunately, however, the fundamental equation describing the displacement process under the action of capillary pressures can be solved only by numerical or graphical procedures, and moreover the solutions are sensitive to the character of the capillary pressure and relative permeability curves.* On the other hand, the direct effect of gravity can be given a formal analytical solution if the simplifying assumption is introduced that the water advances with sharp fronts in the individual uniform strata. It depends on the absolute permeabilities of the water invaded zones and hence results in a rate sensitivity of the differential invasion in stratified media. Although a tieatment of this factor alone will not give a complete evaluation of the effect of rate on water intrusion in non-uniform pays, it should provide at least a semiquantitative measure of the sensitivity of "fingering" to withdrawal rate. Such a treatment will be presented in the following sections. GENERAL THEORY The basic element of the theory is the history of water advance updip in a uniform stratum (Fig. 1). In contrast to the assumption of constant flow rate, generally made in the analysis of displacement processes, we shall assume here that

Jan 1, 1951

By H. E. Stamm, H. H. Spain, R. C. Rumble

This paper presents a reservoir analyzer study of the performance of the Woodbine formation in the East Texas basin. The study was made possible by the compilation of available information on the configuration. thickness and the pressure drawdown of the formation. The investigation was made of the pressure variations in the basin incident to production from the several Woodbine reservoirs. In addition, the apparent compressibility of Woodbine water was evaluated so that the potential water yield of the formation could be determined. The distribution of permeability of the Woodbine sand and the interference between producing areas were also investigated. In conjunction with this basic study of the Woodbine sand, an acceptable match of the production-pressure relationship in the East Texas Field was established on the analyzer. INTRODUCTION When production is taken from a reservoir contiguous to an aquifer, the resultant pressure gradient causes a water influx into the reservoir. For some years it has been the practice to predict the performance of water drive reservoirs with mathematical equations that relate the water influx into reservoirs to their pressure behavior. A rigorous solution of these equations is quite complex: hence, the conventional method requires the simplifying assumptions that the formation have constant thickness, permeability and poiosity, and that it have a known and regular shape. It is necessary to use performance records of the reservoir in question to determine the equation constants that make the results of the mathematical calculations duplicate the reservoir history. The values of the constants determined in this manner sometimes differ so greatly from theoretical or measured values that the constants appear unreasonable. The equation constants comprise such factors as the thickness, porosity and permeability of the formation, and the viscosity and compressibility of water. Usually the thickness, porosity and viscosity are known within a reasonable degree of certainty. whereas the formation permeability and the water compressibility are known with considerably less certainty. Hence. permeability and compressibility values are arbitrarily selected that will make the mathematical equations duplicate the reservoir behavior. In many reservoir studies the value of the compressibility of water thus selected has exceeded the reported value by a factor of from severalfold to manyfold. Whether this excessive value results from the inadequacy of the mathematical approach to take into account the shape irregularities and the heterogeneities of the aquifer, or is due to some type of formation compaction and decrease of porosity has not been established. Because of the significance of the potential water yield of formations with reference not only to oil reservoirs but to water resources as well, it is very important that a representative value for the effective water compressibility be determined. One of the best known examples of a field for which it has been possible in the past to predict the future reservoir pressure is the East Texas Field. Until about 1943 it was possible by means of mathematical equations to duplicate past pressure behavior and to predict with a high degree of accuracy the reservoir pressure of the field. Since 1943, however, the calculated pressures have deviated progressively from the observed pressures to such a degree that in recent years, it has

Jan 1, 1951

By Walton J. Greer, System By John E. Owen

The guard electrode system measures the resistivity of formations by employing a thin disk of current which is caused to flow perpendicular to the bore hole. The control of this current disk is obtained in a brute force manner through the use of relatively long equipotential electrodes above and below the measuring electrode. The log obtained from the system is free of "lag," "plateau," "shadow," "reflection" and other distortions evident on conventional logs. The proportionate contributions of the hole size and invaded zone resistivity to the apparent resistivity reading are reduced when the mud and filtrate resistivities are lower. Thus, the guard system favors the use of conductive muds. The mathematical development of the theory is given in an appendix. Equations are derived to evaluate the effects of mud resistivity, hole size, invaded zone resistivity and depth, and the true formation resistivity. Sample logs from several provinces are reproduced and compared with conventional electric logs to illsutrate the much greater detailed lithology that can be presented by the guard electrode system. INTRODUCTION With increased ability of those in the petroleum industry to analyze formation characteristics from electrical resistivity data, there has become apparent the need for more precise measures of that parameter. The measurement of the resistivity of a homogeneous and isotropic material can be made simply when the body of the material is infinite in all dimensions and the measuring electrodes can be placed within the body. In such a case, the current leaving an electrode follows a spherical distribution and the mathematical expression for the potential level at any distance from that electrode can readily be derived.'.' If it is considered that the active electrodes are within a bore hole filled with mud of a resistivity different from that of the formation. the current does not flow in a spherical pattern near the current source or sink, and the calculations of potential levels become more difficult. The additional consideration of a section of formation invaded by mud filtrates, and thus having resistivity different from either that of the mud, the adjacent formations or of the undisturbed formation, makes the mathematical problem extremely complex. Should the vertical dimensions of the formations be considered as finite. the problem becomes so complex as to preclude reasonable solution. Yet, even these considerations have not described adequately the conditions encountered in a large proportion of the wells to be studied. In efforts to evaluate the effects of the various conditions. combinations of twn, three. and four electrode logging con- figurations of the conventional types have been used. The resulting logs have left much to be desired in some provinces. It has become apparent that the use of such configurations will not result in adequate measurements being made, particularly in cases where the formation is composed of thin sections. If all the current of a resistivity measuring circuit is made to flow through the formation opposite the exploring electrode, many of the shortcomings of the conventional measuring system are eliminated. In the guard electrode method of electrical well logging to be discussed in this paper, the current is so controlled by causing it to flow in a thin disk normal to the axis of the logging tool. EARLY HISTORY The guard electrode method of electrical well logging in its early years was a joint development by the Geophysical Research Corp. and the Amerada Petroleum Corp. The method was first experimentally investigated by means of model studies. The results of these studies indicated that current focusing systems properly applied to electrical well logging

Jan 1, 1951

By Walter Rose

It is the purpose of this note to call attention to the circumstance complicating the attainment of uniform gas-liquid distributions in multi-phase flow systems, and especially in those of the so-called Hassler type.' Although the complication arises directly as a consequence of gas compressibility and the dependence of fluid distributions on interfacial curvature phenomena, little reference to the problem has yet appeared in the literature.' In fact, it is only the paper of Geffen et al.3 which gives any experimental evidence that a problem exists. On the other hand, previous authorsl,4,5 have made the tacit, albeit fallaceous assumption that uniform gas-liquid distributions automatically are established in linear flow systems by the expedient of maintaining the pressure gradients equal in the flowing gas and liquid phases. Such a device, it can be shown, will succeed only if the immiscible fluids are both equally compressible or both essentially non-compressible. Another necessary condition required before uniform fluid-fluid distributions can be achieved is that the porous matrix is isotropic. Uniform fluid saturation distribution conditions obtain in isotropic media only when the curvature of each interface of contact between wetting and non-wetting fluids is everywhere the same throughout the interspaces. This is a necessary condition which follows from the consideration that variation in interfacial curvature gives rise to finite capillary pressure gradients and therefore to variation in the saturation distribution. That this is not a sufficient condition follows from the consideration that hysteric possibilities allow for different saturations even though the capillary pressure gradient is zero. In any event, it will- be recalled that it is the intent in the Hassler scheme of relative permeability determination' to obtain initially (by the capillary pressure drainage or imhibition process) uniform conditions of fluid distribution in the core sample, and then to maintain this uniformity during mixture flow so that the resultant fluid mobilities which are calculated will refer to steady-state transfer under fixed conditions of uniform saturation. If the flowing fluids are incompressible, it is recognized that employment of the same value of the pressure gradient in each fluid will result in maintenance of the initially obtained condition of zero capillary pressure gradient. On the other hand, the consequence of gas compressibility is that the gradient in the gas pressure varies from point to point in the flowing stream. and since the pressure gradient in an incompressible liquid is constant it is thus impossible to maintain zero capillary pressure gradient during the simultaneous flow of gas-liquid mixtures. Therefore, one would expect to observe a variation in the gas-liquid saturation in the directions of the capillary pressure gradients, depending in magnitude on the particular way saturation changed with capillary pressure. In order to formulate the manner gas compressibility effects prevent the attainment of uniform gas-liquid distributions during mixture flow, consider a linear flow system of unit length. As noted, it is necessary to assume an isotropic medium (viz. one where permeability is independent of position and direction). One then can arbitrarily set the pressure

Jan 1, 1951

By Kefton H. Teague

Attempts to locate domestic supplies of sillimanite have been unsuccessful until recently. This paper describes recent discoveries of sil-limonite-bearing schists in the Southeastern States, with emphasis on geology of the deposits, mineral composition, origin, and reserves. Statements concerning possible methods of mining, milling, and utilization of sillimanite are given. Results from refractory tests on a laboratory scale are summarized. SINCE 1917, when synthetic mullite was first used in spark plug porcelains, there has been a continued search for natural sillimanite. During the recent war this search was intensified, not only for sillimanite but also for any materials which could be used satisfactorily as a substitute for the imported kyanite from India. The Indian massive kya-nite does not decrepitate during calcination as does our domestic bladed kyanite, thus a coarse grog can be obtained from it. Some of the most promising domestic materials suited for high temperatures are limited in quantity. Massive topaz from Chesterfield County, S. C., showed promise of possessing desirable refractory properties; but the action of combined fluorine, liberated during calcination, upon equipment and surroundings, as well as a limited supply of ore and the presence of impurities discouraged its use. Residual lumps and boulders of massive kyanite, similar to the material imported from India, occur in Georgia. This material, like the massive topaz, is of limited quantity. Andalusite from California and dumortierite from Nevada are satisfactory high-temperature refractory materials. Their uses have been restricted to specialized products since the location and character of the deposits are such that large production is impractical. Sillimanite, like kyanite and andalusite, has the theoretical chemical composition of Al2SiO5. It has a hardness of 6 to 7; vitreous, silky to subadamantine luster; specific gravity of 3.2 to 3.3; and is gray to bluish gray in color. Under the binocular microscope, the crystals are transparent. The mineral may occur as dense, fibrous mats (fibrolite) composed of fibrous, sometimes radiating hair-like crystals in schist which are associated with igneous intrusions. Deposits of this type have widespread distribution, but none discovered to date appears to have commercial possibilities. Sillimanite also occurs as bundles of crystals disseminated in biotite schist, as noted in the Hart-Elbert-Madison County, Ga., area and in some of the South Carolina and North Carolina deposits. This type of material has been called prismatic sillimanite by the U. S. Bureau of Mines. Deposits of this variety appear to offer commercial possibilities. When heated above 1650°C, sillimanite expands about 6.5 pct and is converted into a mixture of mullite (3Al2O3 . SiO2) and vitreous silica. This is a stable mixture which has a specific gravity of about 3.15. General Geology: Sillimanite (fibrolite) is a mineral common to the pre-Cambrian metasedi-ments of the Southeast. It occurs throughout the crystalline schists which lie between the Fall Line on the southeast and the established Paleozoic rocks on the northwest (fig. 1). In addition to the meta-sediments, this area contains pre-Cambrian igneous

Jan 1, 1951

By C. E. Lesher

THIS paper reports a study of the reactivity of 950°F and 1650°F cokes as measured by relative rates of reduction of iron oxides at temperatures up to 2200°F. Previous work cited shows general acceptance of the theory that reduction by carbon is a gaseous reaction, and that kind and character of carbon as well as particle size have measurable effect on the velocity of reaction. As will be shown, the data obtained in this study confirm those conclusions. The work was not designed to examine iron oxide reduction equilibrium, but if reaction velocity be defined as the speed with which "a reaction tends to approach conditions of equilibrium," the data here presented may be considered as a study of reaction rates, and the relative degree of reduction to metallic iron as the measure of reactivity. Three standardized combinations of Lake Superior brown iron ore with carbon were tested by similar procedures. One combination was a mechanical mixture of carefully sized high-temperature coke (1650°F) with the ore. The second was a mechanical mixture of the ore with Disco* obtained by carbonizing the identical coal at 950 °F. The third was an agglomerate prepared by carbonizing the coal and ore at 950°F, premixed in proportions to give as nearly as possible the same relative amounts of carbon and ore as the mechanical mixtures. This agglomerate, obtained by heating the finely divided ore (through 30 mesh) with coking coal through the plastic temperature range so as to form solid aggregates, gives a product in which the oxide particles are impregnated with, and intimately bound together with low-temperature coke. The agglomerate-—ore-Disco—was most active in oxide reduction; the mechanical mixtures of Disco and ore next in order, with coke the least reactive. General Discussion: Carbon exists in many forms and it is well known that the form or nature of the carbon used in reduction of oxides is related to the critical temperature of reduction. Sugar carbon, charcoal, and lampblack are forms of carbon that will reduce oxides at lower temperatures than high-temperature coke, and coke will, in turn, give a lower critical reduction temperature than graphite. There have been many investigations of this characteristic of carbons. Johnson' reported a difference of 130°F (70°C) in the critical reduction temperature of zinc oxide as between charcoal 1891 °F (1033°C) and Acheson graphite turnings 2021°F (1105°C) with zinc oxide. Bodenstein2 using charcoal and coke, found a difference of 138°F (77°C) comparing an experimental figure of 2066°F (1130°C) for coke and 1928°F (1053°C) for charcoal, in the reduction of zinc oxide. He concluded that this is very marked and observed that the "type of carbon merely raises or lowers the temperature at which rapid reaction takes place." Comparing the effectiveness of types of carbon in reduction of zinc oxide, it was found that a "brown coal coke" gave 97 pct zinc elimination at 1832°F (1000°C), as compared with 48 pct with "hard coal coke."' A wide range of metallic oxides was studied by Tammann and Sworykin,4 who found that the temperature at which decomposition of oxides begins depends on the nature of the carbon used. Carbon in the form of graphite, lampblack, and sugar carbon was investigated. Sugar charcoal will reduce Fe2O3 to Fe3O4 at 842°F (450°C) as compared with 1112°F (600°C) for coke, according to Meyer." Direct reduction of iron oxides by charcoal begins at 1382°F (750°C), but "first becomes intense" at 1652°F (900°C), whereas with coke, direct reduction begins at 1742°F (950°C), and "first becomes appreciable" at 2012°F (1100°C).6 he total reduction of the sample under certain conditions when heated in a current of CO with charcoal was about 100 pct for limonite and about 77 pct for magnetite. Using coke under the same conditions, the respective percentages were 75 and 47. In a study of processes for sponge iron7 by the Bureau of Mines, the conclusion was reached that a low-temperature char from noncoking subbituminous coal is the most satisfactory solid reducing agent. In a critical study of zinc smelting from a theoretical viewpoint Maier8 concluded that the reduction is by CO, that the reaction between ZnO and CO is intrinsically more rapid than the subsequent reduction of CO2 by C, which is limited by diffusion rates, which in part effectively limits the smelting process. Maier said that the operation is improved with the activity of the reducing carbon. An active carbon, he said, is one maintaining a low CO, content in the retort. Reactivity of Carbon: One form of carbon is more potent in reducing oxides than another. A carbon that reacts faster than another at a given temperature is said to be more reactive. Reactivity is measured by several methods, using carbon dioxide, air, or steam as reactants.9 ebastian and Mayers" have developed a method for the determination of absolute reaction rates between coke and oxygen by a study of ignition points under certain conditions. These and other investigators have established the relative reactivity of types of carbon. Lignite, charcoal, bituminous coal, cokes in the ascending order

Jan 1, 1951

By Fred C. Bond, Jen-Tung Wang

Comminution energy is principally energy of deformation before breakage, which appears as heat An empirical equation is presented which covers the entire comminution range. The new strain-energy theory considers comminution from the known principles of mechanics and the reduction ratio. Energy requirements according to the different theories the romnnred. THE present status of the theory of comminution is extremely unsatisfactory. The amount of various ores and rock products crushed and ground annually approaches the staggering sum of perhaps one half billion tons. Yet the basic concepts underlying an operation of such magnitude are unknown, and actual knowledge of what takes place in comminution is almost entirely empirical. The responsibility for this situation cannot be charged to apathy or lack of appreciation of the importance of the problem. Many attempts have been made to evolve a workable hypothesis, but none has been completely successful. General Considerations In order to break rock, it must be subjected to a stress which strains the rock beyond its critical breaking point. The stress imparts energy to the rock, most of which is released in the form of heat when the stress and resulting strain are removed. If the strain induced exceeds the critical strain, the energy is released by breaking; if the induced strain does not reach the critical point, the energy is released with removal of the stress as energy of resilience. Rock is a brittle material and it is assumed that it breaks approximately at its yield point, so that no permanent deformation results from strains below the critical. When rock is broken the total energy input is accounted for by the heat liberated, and by the surface energy of the new surface produced. A small and probably negligible amount is released as noise. Rock is commonly broken under compression and the applied stress is ordinarily compressive. However, for breakage to occur, it is only necessary that the induced strain exceed the critical value, and this strain may represent the resultant of compressive, shearing, and tensile forces. The stress-strain diagram of rocks under compression, and the modulus of elasticity as determined therefrom, should be fundamental considerations in the development of any theories of rock breakage. Crushing in jaw or gyratory crushers results primarily from a squeezing action, and reduction in a hammer mill is the result of impact. The action in rod and ball mills is a combination of impacting, squeezing, and wearing away by attrition, or rubbing. In ordinary crushing and grinding, the forces are applied at protruding points and are not distributed evenly throughout the rock. This is one of the features of comminution which theoretical considerations have not covered adequately. Another feature is the effect of impact velocities, which may greatly influence the total energy required for breakage. Impact velocities may vary from 1 or 2 fps in crushers to 10 or 20 fps for ball mills, and as high as 100 or more for hammer mills. However, according to the findings of geophysicists, the velocity of sound and of compression waves in stone is many times greater; the primary longitudinal compression waves travel at perhaps 15,000 fps, and the secondary transverse waves, which may reach the surface and cause cracks to form thereon, travel at perhaps 5000 fps. When a crack tip forms, the total surrounding stresses are concentrated in this tip, which rapidly extends throughout the rock particle.1 It would seem that the energy required to deform the rock beyond its critical strain, resulting in the formation of a crack tip, represents practically all of the energy required. Most of the primary crack tips presumably form on the surface. A simple method of designating feed and product size is necessary for the evaluation of crushing and grinding results. Taggart's suggestion2 that the 80 pet passing size be used has been found very practical. In this paper, the square opening screen size which 80 pct of the feed passes is designated as the feed size F, and the size which 80 pct of the product passes, or the product size, is designated as P. The reduction ratio at the 80 pet passing size, or F/P is designated as n. The feed and prod-

Jan 1, 1951

By B. H. Caudle, R. L. Slobod, E. R. Brownscombe

An earlier publication has discussed three methods lor obtaining relative permeability data on small core sample-: and the apparatus and technique for the capillary pressure displacement method. This paper describes the apparatus and technique for the solution gas and the dynamic displacement methods and presents a routine procedure for obtaining oil-gas and water-oil relative permeability data. Theoretical and experimental considerations are presented to show that the end effect commonly associated with the dynamic flow mechanism is extremely small where constant rates are employed in the flowing phase. An effect of flow rate on relative permeabilities obtained by the dynamic system is found only when gas is one of the flowing phases and this effect is ascribed to a form of channeling in the capillary system. The apparatus and procedure* used to obtain relative permeability data with three phases flowing are described and some preliminary results of the use of this method are shown. INTRODUCTION In a previous publication from this laboratory there appeared a brief discussion of the concepts behind three basic methods for obtaining relative permeability data.' These three methods were called the capillary pressure displacement method. the solution gas displacement method, and the dy. namic displacement method — the names being suggestive of the type of process used for obtaining the desired saturation prior to making the permeability measurements. In the same publication, the apparatus and technique for the capillary pressure displacement method were described and some typical results obtained by this method were presented. This paper will present: 1. The routine procedure used to obtain complete relative permeabilitv data (In small core samples with two phasec flowing; 2. The experimental techniques for the solution gas and dynamic displacement methods for obtaining relative permeability data; 3. The results of some studies on the mechanism of fluid flow through consolidated porous media; and, 4. The preliminary results on the determination of relative permeability for a system in which three phases are flowing. TECHNIQUE AND APPARATUS FOR THE DETERMINATION OF RELATIVE PERMEABILITY WITH TWO PHASES FLOWING In order to obtain relative permeability data on a small core sample which will describe the various phases of reservoir depletion. it is necessary that the methods used duplicate, as closely as possible, the various mechanisms involved in rescbrvoir depletion. It is important that. in any type of relative permedbility determination. the core be originally at virgin reservoir conditions with respect to the fluids within the pore *paces. (for example. oil used to displace water from core until connate water saturation estahlished). and that the subsequent saturation changes in the core be analagous to tile reservoir depletion mechanism for which the data are to he used. PROCEDURE OF DETERMINING RELATIVE PERMEABILITY DATA WITH TWO PHASES FLOWING A standardized procedure has been set up for obtaining relative permeability data and other information on small core samples. This procedure essentially consists of 11 steps as shown below: 1. Determination of tile capillary pressure curve.'! This information is used to help select samples for relative permeability determinations and to charactrrize. in general. the pore size distribution of the sample. 2. Determination of the specific permeabilities to air, oil. and water. These values are those on which the relative pern~eahility calculations are based. 3. Reduction of the core to connate water by an oil drive using the high speed centrifuge. 1. Determination of the critical gas saturation value and the associated relative permeability to oil by the solution gas drive method. 5. Determination of the relative permeabilities to oil and gab in the region where both phases are flowing by the capillary presqure method. 6. Reduction of the sample to residual oil in the centrifuge by an air drive. and the determination of the relative permeability to gas at this point.

Jan 1, 1951

By P. G. Nahin, A. Grenall, R. S. Crog, W. C. Merrill

A progress report of an experimental investigation into the role of clay in reservoir performance is presented. The Paper gives some of the reasons for considering clay as a significant component and outlines the objectives of a broad field of stud) which it is intended to pursue. Descriptions of the analytical methods used are given; these include X-ray diffraction. elec tron miscroscopy, thin section petrography, infrared spec-troscopy, and cation exchange analysis. A suite of the more important clay minerals has been assembled and characterized l~y these methods for use as standards in core analysis. From the data obtained it appears that although no one method of analysis is diagnostic for all of the clay minerals the infrared technique shows considerable promise in this direction. For the present, one or more supplementary methods should be used to confirm the clay mineral identifications. The methods of analysis are applied to field cores taken from repesentative and widely differing strata especially as regards their susceptibility to damage by fresh water. well.; completed in the stevens and Gatchell zones in San Joaquin valley are I,articularly clear-cut examples of this behavior with stevens zone wells being more adversely affected by fresh water. cores from these zones have been studied and are discussed. It appears that differences in this behavior can be ascribed to differences in the nature of the contained clays. The value of the infrarecl spectra of the clay fractions in establishing the identity of the predominant clay minerals is given particular emphasis. INTRODUCTION It is a challenge to the technical resources of the petroleum industry that when the economic limit of production is reached, from 40 to 70 per cent of the oil in California reservoirs remains unproduced even by use of the best presently known methods of recovery. The magnitude of this abandoned volume of oil can be appreciated when it is considered that to 1950 in excess of 8 billion bbll has been produced from California reservoirs with estimated economically recoverable reserves in known fields and pools totaling nearly 4 billion bbl.24 If for every barrel of oil produced there is at least another barrel still in place, it is evident that the revenue obtained from the recovery of only a .few per cent of this volume would repay the cost of the required research manyfold. From well completion experience. production behavior, and a growing body of laboratory data it now appears certain that the mineral composition of a producing stratum has an important bearing on the productivity and ultimate yield. In addition to the organic component and water, the cores con,ist of gravel, sand. silt, and clay" in diverse variety of (a, composition and (b) texture. It is the composite effect of these two factors which is probably responsible in large measure for the way in which the oil flows to the well. The role of the clay and fine-size accessory minerals is not clear but there is a growing opinion, based on their physical and chemical properties, that it is a significant one. of particular importance are the prime facts: 1. The silt and clay fractions of the reservoir matrix possess the highest surface area per gram, and 2. The silt and especially the clay fractions are the most chemically reactive of the inorganic constituents present. Only within the last few years has the knowledge of clay mineralogy and the techniques of identifying the clay minerals reached such a stage as to enable reliable inquiry into the composition of argillaceous sediments.2,8,10,11,12,16,26 It is the purpox of this and succeeding papers to add to the fund of information on the role which these materials play in the production of petroleum from California formations by correlating their presence and associated properties with observed reservoir behavior. In the present paper attention is directed to their possible influence on damage by fresh water. OBJECTIVES The attack on this problem divides naturally into two broad phases: 1. Determination of the nature of the clays and their relationships to the other mineral components, and 2. Determination of the physico-chemical relationships between the clays and the interstitial fluids. In the work described in this paper the emphasis has been on phase 1, which stems logically from the necessity of identifying and understanding the materials to be dealt with in Phase 2. Based on the authors' present opinion that not all of the minerals which occur in oil-bearing formation are of equal importance in their effects on the flow and recovery of oil, it was decided to focus attention first upon the clay minerals content and then. later perhaps. work into the field of the normally larger size non-clay minerals and fractions. The

Jan 1, 1951

By W. B. Lilienthal, J. C. Melrose

The application of Bingham's law to the behavior of drilling fluids in a rotational viscometer permits the expression of viscometric data in terms of plastic viscosity and yield value, the flow properties of a plastic fluid. A commercially available rotational viscometer is described, and when modified to a multispeed type viscometer, is shown to provide a simple and convenient instrument for the measurement of these properties both in the laboratory and in the field. The data obtained are shown to be useful in defining and understanding mud control problems relating to chemical treatment and to the hydro-dynamic behavior of muds. INTRODUCTION The highly complex drilling fluids which are required for deep drilling often give rise to new and unusual mud control problems. Rapid and economic solutions to these problems may require, on the one hand, better understanding of the changes which contaminants and chemical treating agents produce in the colloidal and inert solids of the mud, or, on the other hand, closer control of the hydrodynamic behavior of the mud. The latter objective obviously can be achieved only if a correct rheological analysis of the flow behavior of drilling muds is available and if this is accompanied by the appropriate rheological measurements. The purpose of this paper is to describe such measurements in the field, and to show how the resulting data can be of value in solving difficult mud control problems. It is now generally recognized that Bingham's law of plastic flow can be utilized in describing the hydrodynamic behavior of drilling fluids in the non-turbulent flow range. Beck, Nuss, and Dunn' have recently applied this law to the flow of mud in small pipes, and Rogers2 has reviewed the rather extensive literature on this subject. So far, however, the use of Bingham's law has been restricted to the analysis of mud flow in pipes or capillary tubes, and it has not been directly applied to the flow in rotational viscometers. In the work to be reprted, the Reiner-Riwlin3 equation for the flow of a plastic fluid in a rotational viscometer has been utilized to permit the expression of multispeed viscometric data in terms of plastic viscosity and yield value. the two absolute flow properties of a plastic fluid. With regard to the application of these measurements, the calculation of the relationship between pumping rate and pressure drop, both in the drill pipe and annular space, has long been a subject of interest. Beck, Nuss, and Dunn,' following Caldwell and Babbitt: base their calculations for non-turbulent flow on Buckingham's integration of Bingham's law for pipe flow and measurements of the plastic viscosity (rigidity in their terminology) and yield value. In the case of turbulent flow, Fanning's equation is employed, and the pressure drop is relatively insensitive to the flow properties of the mud. Since flow in the drill pipe is likely to be turbulent at usual circulation rates, the plastic flow properties will chiefly influence the pressure drop in the annular space. As pointed out by Beck,' the control of this component of the total pressure drop may be of special importance where lost circulation problems are encountered. Other hydrodynamic problems to which it should be possible to apply measurements of the plastic flow properties include predictions of the velocity distribution in non-turbulent flow and the critical velocity for transition to turbulence. Plastic viscosity and yield value. as abmlute flow propertie.;, will reflect the colloidal or surface-active behavior of the solids present in drilling fluids. Measurements of these properties should therefore find application in developing a better understanding of such behavior and in characterizing the type and condition of these solids. Garrison and ten Brink have utilized multispeed viscometric data in this manner. although their measurements were not expressed in terms of the absolute flow properties. In connection with the application of these measurements, it should be recognized that the presently used one-point viscosity measurements are relative in nature. The API Stormer 600-rpm measurement, for example. is a function of both plastic viscosity and yield value, as well as mud weight, and will often be misleading when its application to mud control problems is attempted. NOMENCLATURE, UNITS, AND DEFINITIONS In Fig. 1 an idealized plot is given of the flow variables involved in any viscometric measurement. It is seen that the flow behavior of plastic fluids is characterized by two constants — plastic viscosity, µp, and yield value, F. Other workers hate used the term rigidity for plastic viscosity or the term mobility for its reciprocal. The term plastic viscosity, however, emphasizes the close relation this property bears to the viscosity of a true fluid and is expressed in the familiar viscosity units of centipoises. The yield value is expressed in lbs/100 sq ft, the units adopted for gel strength measurements with the APT shearometer. Definitions of these properties based on rheological or macrc)scopic flow considerations follow from Fig. 1. The plastic viscosity of a substance obeying Bingham's equation is defined

Jan 1, 1951

By B. E. Morgan, C. K. Dumbauld

The need for a low-strength cementing composition for use in well cementing is reviewed and results are presented of laboratory and experimental field tests of a modified cement having a controlled ultimate tensile strength of approximately 200 psi. The modified cementing composition may he prepared from either high early strength or normal portland cements by the addition of bentonite clay and a suitable agent for dispersing and controlling the set of the slurry. Substitution of the modified cement for conventional slow-set cements may give better completion results in many wells because the modified cementing composition has lower set strength, lower slurry density, and greater slurry stability than conventional cement slurrieh. The lower ultimate strength allows greater penetration with less shattering of the set cement when perforating casing and cement. The lower elurry density allows the placement of longer columns of cement slurry, and the greater slurry stability reduces the possibility of having an uncemented section caused by the settling of cement particles before the cement set.. INTRODUCTION High strength lias always been one of the accepted criteria of a good cement. During the early use of portland cement in well cementing. emphasis was placed upon securing cements with higher strengths. In 1931, Barkis' reported that, "Normal oil well cements have been improved to develop greater strengths and uniformity of product, which has aided in producing successful jobs in cementing the deeper strings." As long as most wells were completed by the open-hole method. the use of cements having high strengths seemed desirable, and there was no objection raised against high-strength cementing compositions. For a number of years. how-ever, the industry has been completing a large numher of wells by setting and cementing casing through productive horizons and then obtaining production by gun-perforating the section of casing opposite the desired interval. Although this method has been generally successful, difficulties have been experienced in some cases in completing or recompleting wells because of apparent lack of adequate penetration by the bullets through the casing and surrounding sheath of cement and into producing formation. In addition to the penetration trouble: there have been indications that fracturing and shattering of the set cement by perforating might he a contributing factor in causing the failure of some jobs to exclude water or gas from oil producing zones. The possibility that cements having high set strengths were contributing to til difficully ill obtaining satihfartory perforating results has re. ceived attention during recent years. Gun perforating tests conducted in 1944 showed that the depth of bullet penetration into set cement varied with the hardness of the cement, the greater the strength of the cement the less being the penetration. In 1946, Farris2 pointed out that high strength cements were not needed in well cementing, and in 1947, data published by Oliphant and Farris3 showed that set cement was perforated without shattering at approximately 150 psi tensile strength. whereas at approximately 300 psi tensile strength severe cracking and shattering occurred. Oliphant and Farris suggested that wells be perforated at the proper time interval after placement of the cement so that the set cement would have the desired strength. Several difficulties may he encountered in trying to perforate a cement job at the correct time to catch the tensile strength near 150 psi. The rate of strength development of different cements varies considerably. This fact is illustrated by results of tensile strength measurements presented in Fig. 1. These data show that at 175°F the tensile strengths of three conventional slow-set cements varied from 75 to 235 psi at the end of 12 hours. After 24 hours, the tensile strengths varied from 200 to 455 psi. The rate of strength development is affected. also. by the temperature of the forniation, and this adds to the uncertainty of perforating at the rorrect time, since accurate well cementing temperatures may not he known in many cases. Furthermore, in the recomple-tion of wells it is sometimes necessary to perforate cement which has set for a long period and has developed maximum or final strength. In view of the apparent need for a cementing composition having a controlled ultimate strength, an investigation was

Jan 1, 1951