Search Documents

By N. G. Johnson

ALL of us are only too familiar with the fact that first the defense program, and finally the war, required vastly increased production from existing sources, and the discovery and development of new sources of minerals and metals. For this vast program, the explosives industry was well prepared. I use the term "well prepared" to cover all the phases of explosives service, which include formulation, design, manufacture, distribution, and application.

Jan 1, 1944

E. D. Peters, Jr., Dorchester, Mass.: This paper of Mr. Keller's seems to me a step in a direction that has been very little exploited, and iff likely to lead to valuable practical re-

Jan 1, 1899

By E. Topanelian

In recent years considerable progress has been made in the development and application of mathematical techniques for the solution of certain problems involving economic "strategies". Such a problem might involve, for example, the scheduling of shipments of a commodity from a number of sources to a number of destinations. The object would be to schedule the shipments in a manner so as to satisfy the destination requirements and at the same time minimize the transportation costs. The solution to such a problem is not necessarily intuitively obvious. The "obvious" solution is frequently far from optimum. If the shipments are to be made from, for example, only two sources to four destinations, the optimum schedule is readily found. However, if shipments are to be made from, say, 10 sources to several hundred destinations, even a competent and experienced scheduler may spend considerable time in finding a reasonable answer. Even then he is not sure that he has the optimum solution. Furthermore, he has no way of knowing how far from optimum the answer is. Consequently, he does not know whether he should accept this solution or seek a better one. Prior to the advent of large high-speed digital computers, little more could be done with such problems because of their great size and multiplicity of possible solutions. A problem involving 20 sources and 50 destinations would require choosing, from a very large number of possible combinations, the optimum combination of 1,000 variables. The best one could do was to utilize intuition, extrapolation from past experience, and other non-exact approaches. With a high-speed computer, however, such problems can be solved providing a reasonable computational procedure (or algorithm) can be utilized. The purpose of this paper is to describe such a procedure (linear programming) and to apply this procedure to a production scheduling problem. THE RESERVOIR PROBLEM A simple reservoir model is used throughout this analysis to illustrate the use of linear programming. Even the simplest reservoir behavior problem is nonlinear in both space and time, but if the geometry is fixed for a particular study and the time variable is quantized, the resultant system may be described by linear constraints on the variable production rates. Crude oil is available from five separate sources and is delivered to a pipeline. Sources 1, 2, 3 and 4 are ideal reservoirs and therefore are subjected to reservoir flow restrictions. Source 5 is labeled "Outside Source" and crude oil from this source is assumed to be available in unlimited quantities without considering reservoir conditions. The general problem is to determine the schedule of crude oil production from five sources which, over an eight-year period and subject to certain restrictions, will result in maximum profit. "Profit" is defined to encompass all economic factors and expenses involved in producing and selling crude oil to a pipeline facility. Such economic factors are lumped to the extent that the term represents the dollars profit from one barrel of crude oil from any of the five sources entering the pipeline system. Table 1 presents the assumed profit per barrel for oil from any of the five sources over each of the four time periods. In this chart the eight years is broken into four equal intervals of two years. The selection of four time periods of two years each and the values representing the potential profit are completely arbitrary selections for the purpose of illustrating this idealized reservoir problem. In any practical application of this work a careful economic study would have to be made in order to estimate potential unit profits for time periods as far in the future as eight years. In this particular study it is assumed that all of the potential unit profits (C) are known and fixed before the production schedule is attempted. For each reservoir arbitrary values are assigned for

By A. N. Crownover, H. T. Kennedy, E. P. Miesch, C. H. Bowman

The paper presents correlations of (l) molar volume of gaseous hydrocarbon mixtures with pressure, temperature, composition and properties of the C,-plus fraction; (2) shrinkage of oils during flash and differential liberation of gas, including the calculation of formation volume factor under various conditions; and (3) bubble-point pressure with temperature, composition and characteristics of C, plus. The data on which the correlations are based comprise 1,615 measurements on 900 hydrocarbon systems, including numerous systems containing nitrogen, hydrogen sul-fide and carbon dioxide. In each correlation, the number of data points covered and the accuracy is substantially greater than in previously available work. Thus, the equation yielding molar volumes of gases has an average deviation of 2.04 per cent, applied to mixtures having temperatures up to 313F and pressures up to 9,800 psis, compared to 2.37 per cent for the Benedict-Webb-Rubin equation applied to the same data, and 4.53 per cent for the method based on the law of corresponding states. he equations resented are all explicit in the dependent variable, and require no iteration on the digital cow puter. INTRODUCTION The ease and accuracy of determining the composition of hydrocarbon mixtures, compared to the difficulty of measuring their properties under reservoir conditions, makes it desirable to utilize composition as the key to physical behavior to the greatest possible extent. As a result, there are available correlations between composition, or easily measured characteristics dependent on composition, and practically every important engineering property of reservoir fluids. The task confronting us is one of finding more exact relationships between important variables rather than extending correlations to new properties. This paper describes new correlations of molar volumes of gases, formation volume factors, and bubble-point pressures with composition, temperature and where possible, pressure. Each correlation is obtained by employing a sufficiently large amount of data so the calculated properties are probably as least as accurate as the measurement on which they are based. MOLAR VOLUME OF GASES Although many equations of state have been proposed for pure gases, only a few methods are applicable to hydrocarbons at conditions comparable to those in petroleum reservoirs. Still fewer arc useful in describing the behavior of mixtures, with which the petroleum engineer is largely concerned. The correlation presented here involves procedures similar to those of Alani and Kennedy The van der Waals equation is modified to make a and b functions of temperature instead of being constants for each material.This is a cubic equation, which may have either one or three real roots. The lowest root corresponds to liquid volume, while the highest applies to gas. When the procedure was applied to 703 pressure-temperature points of 164 gases of known composition and volume, the average deviation was 12.08 per cent, and the standard deviation 8.15. The above calculations were made using the constants derived for liquids by Alani and Kennedy and the mixture equation developed by them. A closer approximation to measured molar volumes is obtained by employing different sets of constants for different areas on the pressure-temperature chart, and by changing the relationship between a,,, and b,,, for mixtures and the a, and b, for individual hydrocarbons. The various sets are shown in Table 1, and the areas for which they are recommended are plotted in Figs. 1 through 7. In these figures, the area designated by zero is in the critical region for pure materials, and the values obtained for them may be unreliable. For normal mixtures, however, the area may be added to any of the bounding areas without substantial error. The new relationships employed are

Jan 1, 1966

By N. Parlee, V. C. Kashyap

The solubilities of nitrogen in liquid iron and liquid Fe-Ni, Fe-Mo, Fe-V, and Fe-Mo-V alloys were measured by the Sieverts method. Measurements were made at 1600° to 1800°C on binary alloys up to 20 pet Ni, 10 pet V, 10 pet Mo; and on ternary alloys containing 0 to 3 pet V with 0 to 5 pet Mo. The results on liquid iron confirm that Sieverts' law is obeyed in the range 50 to 750 mm pressure. The measured solubilities and their temperature coefficients are in general agreement with other modern authors. The logarithms of the activity coefficients of nitrogen in the binaries are approximately linear functions of the pet of alloying element. A method suggested by Wagner, Chipman, and others for predicting gas solubilities in complex alloys from solubility data on binaries was applied to the Fe-Mo-V alloys. Reasonably good agreement was obtained between calculated and experimental solubilities. NITROGEN'S solubility in liquid iron has been studied by a number of investigators. It is reasonably well established that the solubility of nitrogen follows Sieverts' law at least up to 1 atm partial pressure and that the solubility increases gradually with temperature. The solubility of nitrogen in Fe-X binary liquid systems has been investigated in only a few cases. Only the systems Fe-Cr,3-5, 7 Fe-V,3 Fe-Mn,3, 5 Fe-Ni,3, 5 Fe-Si,B, H Fe-C, L, H and Fe-P4 appear to have received any study. The only work on liquid Fe-V was done by Brick and Creevy.3 This was limited to 13.5 pet and 26.5 pet V, but no exact temperatures were recorded. No work whatsoever appears to have been done on liquid Fe-Mo alloys and even work on Fe-Ni alloys is rather scant. The work on liquid ternary systems is still more limited, and the only publication available appears to be that of Wentrup and Reif,7 who studied the Fe-Cr-Mn and Fe-Cr-Ni alloys. Rassbach, Saunders, and Harbrecht17 have reported work on the solubility of nitrogen in stainless steel but this was not a systematic study of any particular system. Commercial alloys usually contain several alloying elements. Thus, it often becomes important to be able to predict the nitrogen solubility in systems containing two or more added elements without the necessity of making measurements. Wagner,14 Chipman," Darken and Gurry," and Morris and Buehl 14 have suggested means by which this may be accomplished, and Langenberg14 has recently applied one of these methods to nitrogen solubility in complex liquid alloys. This paper reports the results of nitrogen solubility determinations on Fe-Ni, Fe-Mo, Fe-V, and Fe-Mo-V alloys and compares calculated and experimental solubilities in the ternary alloy. Attempted measurements on Fe-Cr alloys were not successful by the experimental method employed. Experimental Method The general method of solubility measurement was that of Sieverts.' The apparatus and procedure were essentially the same as used by Liang, Bever, and Floe."' The equipment consisted of a small vacuum furnace heated by a 6 kw induction unit, gas measuring apparatus, and a gas purification train. The crucibles used were Norton Alundum with alumina sand around them for insulation. The crucibles were charged with 100 to 200 g of material, about 180 g or nearly filling the crucible being found most satisfactory. Hot volumes were determined with argon and ranged from 95 to 130 cu cm (stp). Temperature measurements were made with a Leeds and Northrup optical pyrometer which was calibrated periodically against the melting point of iron in the same apparatus."-'* Good quality nitrogen, hydrogen, and argon gases were further purified and dried in standard purification trains. The sources of the high purity iron and other alloying elements used were as follows:

Jan 1, 1959

By A. S. Odeh

ABSTRACT A simplified model was employed to develop mathematically equations that describe the unsteady-state behavior of naturally fractured reservoirs. The analysis resulted in an equation of flow of radial symmetry whose solution, for the infinite case, is identical in form and function to that describing the unsteady-state behavior of homogeneous reservoirs. Accepting the assumed model, for all practical purposes one cannot distinguish between fractured and homogeneous reservoirs fram pressure build-up and/or drawdown plots. INTRODUCTION The bulk of reservoir engineering research and techniques has been directed toward homogeneous reservoirs, whose physical characteristics, such as porosity and permeability, are considered, on the average, to be constant. However, many prolific reservoirs, especially in the Middle East, are naturally fractured. These reservoirs consist of two distinct elements, namely fractures and matrix, each of which contains its characteristic porosity and permeability. Because of this, the extension of conventional methods of reservoir engineering analysis to fractured reservoirs without mathematical justification could lead to results of uncertain value. The early reported work on artificially and naturally fractured reservoirs consists mainly of papers by Pollard,l Freeman and Natanson,2 and Samara.3 The most familiar method is that of Pollard. A more recent paper by Warren and Root showed how the Pollard method could lead to erroneous results,4 Warren and Root analyzed a plausible two-dimensional model of fractured reservoirs. They concluded that a Horner-type pressure build-up plot of a well producing from a fractured reservoir may be characterized by two parallel linear segments. These segments form the early and the late portions of the build-up plot and are connected by a transitional curve. In our analysis of pressure build-up and drawdown data obtained on several wells from various fractured reservoirs, two parallel straight lines were not observed. In fact, the build-up and drawdown plots were similar in shape to those obtained on homogeneous reservoirs. Fractured reservoirs, due to their complexity, could be represented by various mathematical models, none of which may be completely descriptive and satisfactory for all systems. This is so because the fractures and matrix blocks can be diverse in pattern, size, and geometry not only between one reservoir and another but also within a single reservoir. Therefore, one mathematical model may lead to a satisfactory solution in one case and fail in another. TO understand the behavior of the pressure buildup and drawdown data that were studied, and to explain the shape of the resulting plots, a fractured reservoir model was employed and analyzed mathematically. The model is based on the following assumptions: 1. The matrix blocks act like sources which feed the fractures with fluid; 2. The net fluid movement toward the wellbore obtains only in the fractures; and 3. The fractures' flow capacity and the degree of fracturing of the reservoir are uniform. By the degree of fracturing is meant the fractures' bulk volume per unit reservoir bulk volume. Assumption 3 does not stipulate that either the fractures or the matrix blocks should possess certain size, uniformity, geometric pattern, spacing, or direction. Moreover, this assumption of uniform flow capacity and degree of fracturing should be taken in the same general sense as one accepts uniform permeability and porosity assumptions in a homogeneous reservoir when deriving the unsteady -state fluid flow equation. Thus, the assumption may not be unreasonable, especially if one considers the evidence obtained from examining samples of fractured outcrops and reservoirs. Such samples show that the matrix usually consists of numerous blocks, all of which are small compared to the reservoir dimensions and well spacings. Therefore, the model could be described to represent a "homogeneously" fractured reservoir. la this paper, the fundamental equation of flow

Jan 1, 1966

By Stuart S. Forbes

Changes and additions made to the Canadian Copper Refiners Ltd. electrolytic refinery between 1949 and 1955 are reviewed. The effect of high current density on current efficiency and section work is discussed. OPERATING and physical changes which have been made to the electrolytic tank house of Canadian Copper Refiners Ltd. suggest a review of present conditions. Previous papers1 a have described the general arrangement. Since then, expansion, closer anode spacing, and higher current densities have increased annual cathode capacity from 112,000 to 182,000 tons. A recent extension to the tank house, begun in the spring of 1954 and completed in August 1955, was undertaken to handle the monthly production of about 3,000 tons of anodes from Gasp6 Copper Mines. The anticipation of these additional receipts and large stocks of unrefined copper caused by the increase in custom smelting at Noranda made it necessary to increase production without immediately available additional cell capacity. Accordingly, from December 1954 through March 1955, the tank house was operated successfully at a current density of 24.0 amp per sq ft. Again, as of September 1955, one of the two tank house circuits is being operated at this high density. Although other refineries may operate at current densities higher than 24 amp per sq ft, none is doing so with anodes which contain silver and gold content as high as those being refined at Canadian Copper Refiners Ltd. General Arrangement The original building consisted of two parallel bays, each 660 ft long and 60 ft wide, with a 24 ft pump bay extending along the west side of the building. A previous expansion, completed in 1939, was to the west of the original building and consisted of two 60 ft parallel bays 240 ft long. The length of these two bays was extended in 1954 to 540 ft, adding one stripper and 12 commercial sections to No. 2 electrical circuit. Space is still available for an additional six commercial sections and one stripper section. Construction on these seven sections started in September 1955, and will be completed early in 1956. The tank house capacity will then be 206,000 tons of cathodes per year. In the two tank houses there is a total of 900 cells distributed as follows—commercial, 468; stripper, 36; total, 504 in the No. 1 tank house (No. 1 circuit): commercial, 360; stripper, 36; total, 396 in the No. 2 tank house (No. 2 circuit). Section arrangement consists of nine cells to a tier and two tiers or 18 cells to a section. Anodes, weighing 620 lb each, are cast with a Baltimore groove and are refined by the multiple system. Solution enters each cell through a bottom inlet at one end and overflows at the top of the opposite end. Rate of solution flow through the cells is controlled by valves at each section to 4.5 gpm. Electrical power to the tank house is now provided by nine 675 kw motor generator sets. Each set consists of a 135-12 v 5000 amp dc generator, driven by a 980 hp 2300 v synchronous motor. When No. 1 circuit is operated at 22,000 amp, five sets, at 4400 amp each, are connected in parallel. As the maximum allowable current on No. 2 electrical circuit has recently been set at 18,000 amp, only four sets are connected in parallel on No. 2 circuit. A tenth set, now being installed, will be reserved as a spare, for use in either circuit as required. Recent Changes In 1949, the annual capacity of the tank house was increased from 112,000 to 132,000 tons by closing the

Jan 1, 1957

By L. L. Handy

The most common method of identifying hydrocarbon-bearing strata in a well that penetrates many different formations involves measurement and interpretation of the electrical properties of the formations as determined by electrical logs. Even though this method is used extensively, and even though in a great many instances it is capable of indicating presence of oil or gas, situations arise for which it is extremely difficult, if not impossible, to deduce the presence of hydrocarbons. These situations may involve the following. 1. A thin formation, bounded by highly resistive formations, in which it is impossible to obtain the actual resistivity of the uninvaded zone with existing logging devices. 2. A formation in which invasion has been so extensive that a value for the uninvaded zone resistivity cannot be obtained. 3. A very shaly formation in which the resistivity index, I, is lower than that usually associated with productive formations. 4. Laminated formations comprised of thin productive sands separated by thin shale streaks in which the individual sand and shale streaks are too thin to permit measurement of uninvaded-zone resistivity with existing logging devices. 5. Productive formations in which the water saturation is high. To extend the utility of electric log interpretation to identification of hydrocarbons in all types of formations, there is strong incentive to find a method not subject to these limitations. Some time ago, in connection with research on the wettability of reservoir rock, an investigation was conducted in which the resistivities of cores were measured shortly after they were removed from a core barrel, and again after they had been extracted and restored to their original oil and brine saturation.' The resistivities after extraction were generally lower. Other tests made on the cores indicated that they were more nearly water wet after they were extracted; thus, it was assumed that the observed changes in resistivities were due to a change in wettability of the cores. Other experiments'," have shown that resistivities of rock samples are sharply dependent on wettability. These experiments have shown that oil-wet samples are more resistive than water-wet samples. To obtain an understanding of how the wetting properties of the surfaces of core material affect electrical resistivity, a series of experiments was conducted. Two groups of core samples were prepared for testing. One group contained brine, but no residual oil. The other group was saturated with brine, flooded with oil to a low water saturation, then flooded with brine to a final residual oil saturation. Resistivity measurements were made on each group. Both groups were then flooded with the original brine to which a chemical had been added that renders sand and clay surfaces preferentially oil wet, a so-called reverse-wetting agent. very little change in resistivity was observed in cores containing only water. The group containing residual oil, however, showed resistivity increases of 100 to 200 per cent. These experiments showed that the resistivity of a core containing oil could be altered by changing wet-tability of the core. Moreover, the possibility was introduced that reverse-wetting agents might be employed as the basis for a logging method for identification of oil-bearing strata. Since behavior of a porous rock containing gas and water might be expected to be similar to that of a rock containing oil and water, such a method should also be applicable to identification of gas-bearing zones. In principle the wettability of the invaded zone could be reversed without altering conductivity of the interstitial water or the hydrocarbon saturation therein. Those strata showing significantly increased invaded-zone resistivities would, therefore, contain hydrocarbons; those with no significant change would be filled only with water. Addition of a reverse-wetting agent to a hydro carbon -bearing zone which is, by nature, already preferentially oil wet would not result in an enhancement of its resistivity. It is generally believed, however, that most hydrocarbon-bearing strata are preferentially water wet.

By R. P. Alger, M. P. Tixier, D. R. Tanguy

In the combination induction-electrical log used at present in the field, the induction logging tool is appropriate for the investigation of moderately invaded formations. A new induction sonde with a radius and investigation about twice as large has been developed recently for the case of deep invasion. It has very nearly the same vertical resolution as the present sonde so that thin beds are defined as accurately as before. The characteristics of the new tool are described, the corresponding interpretation charts are given and field examples are discussed. The design of the sonic logging tool has been modified to improve the calibration and the reliability. The fact that porosity can be accurately recorded by means of the sonic log has prompted new interpretation procedures for saturation estimation, wherein the data concerning the various permeable beds in a given well are correlated. One approach consists of plotting transit time vs true resistivity, with an appropriate scale. With this approach, saturations can be estimated conveniently even in cases where formation water resistivity is not well known. In another approach, a comparison is made of the values of the formation waters computed from the re- sistivity and sonic logs. Using the concept of continuity, this procedure makes possible a quick determination of zones of saturation in shaly sands and/or in case of appreciable variations of formation salinities with depth. It has been found that the comparison of porosity from the sonic log with the apparent porosity computed from a short-investigation resistivity log may reveal, in many cases, the presence of residual oil and thus detect potentially productive formations; this procedure is valuable when the true formation resistivity and the resistivity of the formation water are in doubt. INTRODUCTION During the past year, the efficiency of log interpretation has been vastly improved. The improvements have largely resulted from the introduction of a deep-investigation induction device and from the application of new interpretation techniques that utilize sonic vs resistivity readings. Since the new interpretation techniques depend, in part, upon good values of true formation resistivity, the new induction log will be discussed under Part I. The sonic interpretation techniques will be studied under Part 11. Early in 1959, the 6FF40 induction equipment was introduced in the field. This device was designed for a better approach to true formation resistivities in deeply invaded zones. The greatly improved radial investigation of the 6FF40 equipment has been achieved without sacrificing vertical resolution. The first combination induction-electrical log, the 5FF40, was introduced as a standard tool in 1956 for the logging of wells drilled with fresh muds. The tool has received wide industry acceptance in the United States. 'The 5FF40 induction log has a radial investigation sufficient to overcome average depths of mud filtrate invasion. At 5d invasion, for example, the 5FF40 induction log will read about 1.4 R. in a water sand where R., = 10R,. At 10d invasion, such induction log would read 2.45 R. in the same water sand. In either case, the effects of invasion would not be sufficiently great to cause a water sand to be mistaken for a shale-free oil- or gas-producing zone. Some formations, however, invade deeply — in excess of 10d. Such water zones could be mistaken for oil-or gas-saturated sands unless the porosity balance' can clearly make the distinction. It is for these deeply invaded formations that the 6FF40 was developed. CHARACTERISTICS OF THE 6FF40 Radial Investigation Characteristics To describe the comparative responses of the 5FF40

By D. L. Albright, G. A. Colligan

An investigation has been conducted to determine the nature of the crystallographic substructure of nickel and a 1.0 wt pct Ag-Ni alloy which had been undercooled 105°C prior to solidification. A rotating back-reflection X-ray technique and modified Weissmann diffractometer were used to orient single-crystal sections from these poly-crystalline ingots and to determine the substruc-tural features of these single crystals, The sub-grain size in the pure nickel ingot is 0.5 mm diam while the silver-doped nickel subgrain size is only 0.1 mm diam. The misorientation between adjacent subgrains is approximately 1 deg of arc in both ingots, and the spread of orientation of sub-grains about the mean orientation of the crystal is 5 deg of arc in both ingots. The addition of silver results in a stable impurity substructure which is preserved during isothermal solidification and subsequent cooling to room temperature. In the absence of silver the initial impurity substructure is destroyed by dislocation interactions which produce a secondary substructure which grows to a size five times larger than the initial substructure. THE process of undercooling a metal entails cooling of the molten sample below the equilibrium freezing point without the occurrence of solidification, i.e., by exercising proper constraint over the experimental variables contributing to nucleation in the melt. The current work of walker1 and Colligan Metal reports undercooling large continuous samples of nickel by imposing appropriate restrictions upon the parameters of solidification. The silver-nickel equilibrium diagram3 reveals the presence of a monotectic reaction at 1435°C. At this temperature the maximum solid solubility of silver in nickel is approximately 3 wt pct, but this solubility decreases with decreasing temperature and the solute is rejected as a silver-rich liquid phase at temperatures below 1435°C. Thus, the 1.0 wt pct solute should be held in solution during undercooling and solidification and subsequently precipitate as a silver-rich liquid phase at grain and subgrain boundaries following solidification during cooling below 1435. This assumes no solute segregation sufficient to cause the monotectic reaction. One would expect this second phase to inhibit grain and subgrain growth by retarding dislocation motion after solidification is complete. The effectiveness of silver addition in growth retardation of high-angle grain boundaries has been demonstrated by Colligan and suprenantS4 The principal [l00] dendrite growth directions in a 2 wt pct Ag-Ni alloy undercooled 53 all radiated from the nucleation site. A similar pure nickel ingot undercooled 59°C exhibited a random orientation of [l00] dendrite directions with respect to the nucleation site. Since undoped nickel ingots do not retain any evidence of the casting or solidification texture and silver-doped ingots do retain this texture, it is reasonable to assume that considerable grain growth has taken place in the pure nickel. These particular experiments were performed in order to reveal the differences, if any, in the features of the crystallographic substructure of undoped undercooled nickel and silver-doped undercooled nickel. 1) EXPERIMENTAL PROCEDURE Data are reported on selected specimens from two massive (approximately 260 g) undercooled ingots of nickel. One ingot consisted of electrolytic nickel and the other of this same material with 1.0 wt pct addition of high-purity silver. Powder patterns were taken of both the as-received and as-solidified materials; in all cases the reflections recorded were attributable to the individual elements nickel and/or silver. Fig. 1 is a. schematic diagram of the apparatus employed for the undercooling experiments. The equipment consists in part of a fused silica crucible for containing the charge. A fused silica tube encloses the system, which is covered by a brass cap to permit partial sealing of the inert atmosphere. This cap contains a hole through which a Pt-Pt 13 pct Rh thermocouple is inserted for temperature measurement. The charge is inductively heated under a purified argon atmosphere.

Jan 1, 1963

By Z. F. Oszter

This paper discusses the materials handling problems encountered in cement plants in the areas of raw and finished product grinding and cement storage. It illustrates by examples taken from actual installations, how optimum feed rate is maintained in grinding circuits with properly applied bin dischargers (vibrating hoppers, Mexican feed boxes and aerated bin bottoms) for materials with high moisture content and/or difficult handling characteristics. Initial operating problems with the equipment will be discussed and maintenance requirements will be indicated for comparison purposes. In a modern cement plant, properly functioning bins and feeders are as essential as process control, for no process control system, however sophisticated, can perform efficiently if raw materials do not flow into the plant processing facilities at the required time and rate. Most materials used in the manufacture of portland cement - such as silica sand, iron ore fines, crushed limestone, gypsum, clinker or cement itself-have difficult handling characteristics under specific conditions. High moisture content or even dampness due to condensation may create bin hang-ups, arching and ratholing, all of which result in improper functioning of the feeders or proportioning equipment. The occurrence of these problems in existing cement plants spurred a search for the corrective devices discussed in this article. The equipment investigated included bin dischargers in the form of vibrating hoppers for moist raw materials, Mexican feed boxes for clinker and aerated bin bottoms for cement. Operating experience with different types of bin dischargers showed that the actual performance of some did not come up to the expected level under prevailing operating conditions. The aim of this paper is to record the results of this experience and there- by aid designers and/or operators in selecting the most suitable equipment for their requirements. BIN DISCHARGERS (GENERAL) A discharger is a device that causes stored solids to flow out of a hopper without metering the rate of flow as a feeder does. The investigation discussed here encompasses the following types of bin dischargers used in a cement plant: (1) vibrating hoppers (horizontal motion), (2) whirlpool hoppers (horizontal and vertical motion), (3) aerated bin bottoms (air fluidization), and (4) Mexican feed boxes. Most of the bins used in the cement industry are of the plug flow type. Raw materials with difficult handling characteristics such as high moisture content may cause flow problems when fed through small openings (12 x 18 in.), even if two walls of the bin are vertical.' The most appropriate solution to these flow problems would be the installation of separate mass flow bins for the different materials such as silica sand, iron ore fines, crushed limestone and gypsum, with each bin correctly designed from test data obtained by measuring the flow properties of representative samples. However, limited storage capacity and plant space usually preclude this approach. Quantitative design of bins for flow based on measured flow properties of representative samples of the bulk solid to be handled is no longer an "engineering art," thanks to the basic research done on the flow of bulk solids in bins during the last decade.2, Where space limitations exist, however, or where improvements are being sought for equipment already installed, bin dischargers offer a satisfactory solution to flow problems. They can be applied effectively under flat-bottom bins or mounted to existing plug-flow bins with very little modification to these facilities. In both the foregoing cases, plug flow will be maintained, but the flow channel will be considerably increased so that it is larger than the discharger's bin-side diameter, and material flow can be sustained if the discharger is correctly selected. The main purpose of a bin discharger is to draw the bulk solids across the whole area of the bin outlet, to make the opening fully effective. As long as the bin-side opening of the discharger is large enough

Jan 1, 1969

By I. R. Dunlap, R. F. Burdyn, F. D. Patchen

The composition and properties of a new type of cementing mixture are discussed. In these compositions, API Class A or C cements are added to n stable water-in-oil emulsion instead of the crsual water. The cement disperses in the oil phase and remains out of contact with the water under normal conditions of mixing and storage. At higher temperatcrres, such as those encountered down-hole during placement, the emulsion breaks and cement hydration can then proceed. An extremely low filtration rate is obtained; the major part of the filtrate is oil. Thickening times of the mixture may be increased to meet the 18,000-ft casing cementing schedule by varying the amount of emzilsifier. As in conventional Portland cement mixtures, high-temperature strength retrogression can be overcome by addition of silica flour. Results of several field app1ication.r show that mixing presents no diffculty if emulsion blenders are used and the emulsion is prepared beforehand. INTRODUCTION Water-in-oil emulsions., in which oil is the continuous or external phase and in which water exists as discrete droplets, have for some time been used successfully in drilling muds to reduce filtration loss.1,2 Applied to cement slurries, the same results are obtained; i. e., the fluid-loss is greatly reduced. In addition, thickening times may be extended, lower densities and strengths obtained, and water-base borehole fluids more easily displaced without contamination of the cement. A low filtration rate is often a desirable property in oil well cements since loss of fluid increases placement difficulties and may also cause damage to water-sensitive producing formations.' In the casing cementing operation where the cement slurry is pumped into the narrow annulus between open hole and casing, rapid loss of fluid to a porous formation can cause the slurry to thicken and plug the annulus before the desired fill is obtained. Where a casing or a liner is cemented against a productive formation and the well completed through perforations, the fluid lost between the time the cement first contacts the formation and the time a set is obtained may invade the zone past the depth of the perforations. If the sand contains a material which will swell upon contact with cement filtrate, permanent damage may result. Previous methods for reducing the fluid-loss of cements have employed bentonite4 as an additive, usually with a lignosulfonate as a dispersant, or a derivative of some naturally occurring polymeric material such as cellulose ether." Oil emulsion cements have also been proposedv or fluid-loss reduction, but these have been of the oil-in-water type. Bentonite cements have been used most widely and have been successful in reducing incidence of premature setting of cement attributable to loss of water by filtration. It is the purpose of this paper to describe preparation and properties of a water-in-oil emulsion cement and to relate procedures followed and results obtained in several field applications. DESCRIPTION OF SYSTEM AND MECHANISMS A water-in-oil emulsion cement is prepared by adding an API Class A or C cement to a stable water-in-oil emulsion. Fig. 1 is a photomicrograph of this emulsion before addition of cement. Here the droplets are water and the surrounding medium is diesel oil. High emulsion stability is indicated by the small droplet size [1-5 microns], and by their uniformity.' To form a stable water-in-oil emulsion as illustrated, several conditions must be met. A predominantly hydrophobic emulsifier must be present in sufficient concentration to produce a strong interfacial film; initially, the phase volume ratio (the water-to-oil ratio) must be low;

By S. W. Kennedy, M. Cohen, L. D. Calvert

A high-temperature X-ray diffraction method has been used to study the composition and the kinetics of formation of oxide scales at 800 °C on iron and pure iron-nickel alloys containing 25.6, 75, and 84 pct Ni. This technique has been combined with chemical analysis, and metallographic examination, and determination of weight-gain time curves. The 75 and 84 pct Ni alloys form a three-layer scale consisting of porous NiO next to the metal, a spinel layer, and a -Fe203 on the outside. The 25 pct Ni alloy produces a scale of spinel and Fe,O,, in which, however, traces of FeO and NiO are observed near to the metal. On all the nickel alloys the Fe20, is orieqted with the (107) plane parallel to the surfaces. The exposed surface of the Fe20, is in the form of needles whose axis is perpendicular to the specimen surface. Although the alloys show nonuniform oxidation rates as determined from weight-gain measurements, the X-ray data indicate that the total scale and individml scale layers grow uniformly and parabolically. Deep subscale formation is observed in the 25 pct Ni alloy only. This appeared to be associated with cracking of the scale and penetration of the oxygen along the grain boundaries. ALTHOUGH film formation has been investigated by high-temperature electron-diffraction techniques, for example by Gulbransenl and Hickman,' there have apparently been no investigations of scaling by methods permitting study of the scale during formation at the reaction temperature. Methods involving quenching of scales are based on the working assumption that no significant changes take place on cooling. In general, however, compounds may disproportionate, and the different thermal expansions of constituents may cause disruption of the scale on cooling. In particular, some scales cannot be quenched without spalling. These difficulties are avoided by X-ray diffraction examination at the temperature of formation. High-temperature diffraction procedures also provide a method of determining the kinetics of formation of each phase separately in a multicom-ponent scale. The scaling of binary iron-nickel alloys Containing up to 30 pct Ni at a temperature of 800°C has been studied by Benard and Moreau using cooled specimen~,~ whilst Foley, Druck, and Fryxel14 made an electron-diffraction study of cooled oxide films formed on an iron 42 pci Ni alloy. In the present work the scaling of three iron-nickel alloys containing 25.6, 75, and 84 pct by weight of nickel, at 800°C in oxygen at atmospheric pressure was investigated by X-ray diffraction techniques using a Norelco Geiger-counter diffrac-tometer. The oxidation of iron was investigated for comparison. In addition normal metallographic examinations and chemical and X-ray analyses were made, and weight-gain time curves obtained. EXPERIMENTAL Materials Used—The analyses of the alloys were as follows: Percent by Weight Nickel 25.6 pct 75.4 pct 83.2 pct Aluminum 0.004 None found 0.002 Cobalt — 0.05 0.05 Chromium — — — Silicon — 0.01 None found Carbon All less than 0.05 pct The chemical analysis of the iron was: Al— 0.002 i 0.001 pct, C—0.007 pct, Cr less than 0.001 pct, Cu—0.002 ± 0.001 pct, Mn—0.003 ± 0.001 pct, N—0.0002 pct, 0-0.013 pct, P—less than 0.0002 pct, Pb—not detected, S—0.0017 pct, Si—0.0095 pct, Sn—none detected. This iron was the same as that used to make the alloys and was obtained from the same source-Vacuum Metals Corp. One specimen of pure iron from The British Iron and Steel Research Association was used for preliminary investigation. Specimens were machined to shape from cast rod, polished (the final polish being with 600 Sic paper), degreased in trichlorethylene, washed, and etched. The etching and neutralizing procedure adopted by

Jan 1, 1960

By F. N. Rhines, A. S. Nemy

The mode of high-temperature tertiary creep of 523-0 aluminum alloy was found to be strongly stress dependent. The occurrence of necking and/or fissures during tertiary creep exhibited a sequence with varying stress, wherein fissures with no necking occur at a minimum stress and necking with no fissures occur at a maximum stress. At an intermediate stress, tertiary creep begins without necking but with fissures in a region just beneath the external surface, followed by necking and the development of fissures along the axis of the specimen. The results are interpreted in terms of grain boundary gliding and the hydrostatic stresses produced by restraints to grain boundary gliding. THERE appear to be two possible reasons for the existence of a period of accelerating elongation prior to rupture during high-temperature creep in metals; namely, 1) a change in the structure of the metal itself, leading to a change in its response to loading, and 2) a reduction in the cross section of the metal resulting in a higher unit loading. Howe,' by annealing and retesting broken low-temperature creep specimens, found that irreparable damage is done once a metal is subjected to creep. Andrade2 first pointed out the association between the reduction in cross section and the increase in creep rate. The present research was designed to investigate both of these factors by retesting specimens machined from material which had been subjected to high-temperature creep testing short of rupture. EXPERIMENTAL METHOD An aluminum alloy, 52s-0, was selected for these studies because it is constitutionally simple, is strong enough to be relatively insensitive to mechanical damage during machining,3 and undergoes accelerating creep in a convenient temperature range. This alloy is hardened by 2.6 wt pct Mg in solid solution and is not subject to overaging effects; it has a very small quantity of an insoluble impurity phase, Mg2Si, finely dispersed throughout its micro-structure. Creep specimens, similar in design to ASTM 0.252-in. tensile specimens, were machined from 3/4-in. round stock and the surfaces were electro-polished. A 24-hr anneal at 500°C was then applied to stabilize the grain size to an average diameier of 0.27 mm. Subsequent heating for as long as 700 hr at the intended test temperatures produced no change in the tensile properties, indicating a well-stabilized state of the material. Corresponding to each of the typical testing conditions listed in Table I, one creep specimen was tested to destruction and this test was employed as a guide in the selection of times for subsequent tests. Duplicate creep tests were then run to various points along the creep curve to produce specimens which were used to determine changes in mechanical properties and microstructures. The latter were examined at the midsections of the specimens. Other specimens were tested to various stages of tertiary creep, after which they were remachined to new cylindrical test specimens, electropolished and re-tested in creep. RESULTS The major findings of this investigation may be illustrated by a comparison of the results of three sets of creep tests. These were run at 400°C at three stress levels, namely 750, 1000, and 1500 psi. In presenting these results in Figs. 1, 5, and 8, the creep curves are given in pairs: That curve of each pair which is drawn through filled circles corresponds to the initial test, and the curve drawn through open circles (and designated by - R) represents the result of retesting the same specimen after machining to a new cylindrical surface. The load was readjusted to the original unit stress at the beginning of each retest.

Jan 1, 1960

By E. W. Dewing

The nature of solid solutions of sodium in non-paphitic carbon at temperatures near 1000°C has been investigated by an electrolytic technique. The activity coefficient is found to vary strongly with the heat-treatment temperature of the carbon, and an analysis of the energy terms involved suggests that this is due to variations in the Fermi level. The diffusion coefficient for sodium is in the range implying that the sodium atoms (or ions) are not bound to specific sites on the carbon, in agreement with lack of evidence for specific compounds in the activity -composition curves. In the light of These results phenomena occurring in aluminum reduction cell linings are dis-cussed. It is concluded that sodium migrates by diffusion through the carbon lallice and not the pores, and that A14C9 is formed within the lining hy The excess of NaF invariably found in used linings is formed by this reaction and by air oxidation: There has recently been much interest in the swelling of carbon which occurs when it is made cathodic in an aluminum electrolysis cell, and there is general agreement that the phenomenon is due to the penetration of sodium atoms, derived from the reaction on the carbon lattice. (NaF and AlF3 are constituents of the fused-salt electrolyte.) The low-temperature lamellar compounds of potassium with graphite are well-known, and one compound between sodium and graphite has been reported.4 There is, however, virtually no information on the basic chemistry of the interaction of sodium with nongraphitic carbon at high temperatures, although Dell3 has made some fundamental studies of rates of penetration. At the time the present work was carried out these were not available. Rapoport and samoilenkol had measured rates of physical expansion in a dilatometer, using the fractional increase in length of a carbon specimen after electrolysis in a cryolite bath for 2 hr as an index of susceptibility to cathodic swelling. Work here with a similar apparatus showed that expansion (of a 1.5-in.-diam cylindrical specimen) ceased in less than 20 hr, and that the final expansion was much more reproducible than the value at some intermediate time. This suggested that an equilibrium was being set up, and that a thermodynamic approach was required. The results obtained are not as accurate as had been hoped, but they are the only ones of their type available and have enabled useful conclusions to be drawn. The experimental work was divided into two parts: an investigation of the activity-concentration diagram for sodium in carbon, and a rough determination of the diffusion rate. For the former, most of the experiments were made with petroleum coke even though this is not a form of carbon used in cell linings. It has the advantages, however, that it is a homogeneous material, that it is low in ash, and that its structure is reasonably well understood. Anthracites and coal cokes gave results of the same type as obtained with petroleum coke, but the reprodu-cibility from specimen to specimen was extremely poor. EXPERIMENTAL 1) The Activity-Composition Diagram for Sodium in Carbon. The petroleum coke was ground, pressed at 40,000 psi into plates 19 by 9 by 2 mm without added binder, and heat-treated. The sodium activities were established by an electrolytic technique. The carbon was made anodic in a cryolite bath with an aluminum cathode, and a controlled potential, E, applied so that the sodium activity at the anode was given by ln Na = 1n asa - EF/RT [21 where as, is the sodium activity generated by Reaction [I]. The relative activity is sufficient for many purposes and is simply given by The value of Na not at present known, but when the thermodynamics of the Na-A1 system have been worked out it can be derived from the equilibrium

Jan 1, 1963

By H. H. Kellogg, T. R. Ingraham

Three anhydrous zinc sulfates have been identified. They are: ZnSO,(a), stable below 1007°K; ZnS04(/3), stable above 1007OK; and ZnO.ZZnSO,. The decomposition pressure of each sulfate has been measured and the following relations calculated from the results: The measured data have been combined with the known themzodynamic properties of ZnS, ZnO, Zn, SO,, and SO, to yield univariant and bivariant phase diagrams for the system Zn-S-0. The application of these diagrams to problems of roasting zinc concentrates is discussed. PRESENT knowledge of the equilibrium chemistry of the system Zn-S:O, so essential to scientific control of zinc-blende roasting, is deficient in several respects. A basic sulfate of zinc has been reported by many investigators, but its composition is in doubt and its thermodynamic properties have not been measured. Our scanty knowledge of the properties of the normal sulfate at roasting temperatures is based on old and inexact studies. A crystal transformation in the normal sulfate, involving a sizable heat effect, has been noted by several investigators but has never been quantified. The studies reported herein were designed to fill these gaps in knowledge and to make possible a reasonably complete thermodynamic description of the system Zn-S-0 at roasting temperatures. The results make possible a far more exact understanding of zinc roasting chemistry than was hitherto possible. EXPERIMENTAL The principal experimental technique employed was the measurement of the total gas pressure developed from decomposition of either ZnSO, or ZnO2ZnS0, in a closed system. The apparatus employed was essentially the same as that described by Warner and lngrahaml and Flengas and tngraham.' The unusual feature is a flexible Pyrex bellows used to separate the reaction gas mixture (S03, SOz, and 02) from the mercury in the manometer and thus prevent corrosion of the mercury by SO3. The bellows operates without evidence of hysteresis and, barring accidents, may be used for several months. The response of the bellows varies from manometer to manometer, but is generally of the order of 0.6-cm displacement for a l-cm change in pressure. The response for all manometers has been linear over the range 0 to 1 atm. To prevent condensation and polymerization of sulfur trioxide, the bellows-manometer was maintained at 100.0" * 0.5" by a circulating bath of silicone oil. The connecting tubing between the manometer and the furnace was maintained at 100.0" * 0.2"* also, with a stirred air bath. The open end of the manometer was connected with a chamber which was maintained at a known pressure. The pressure could be varied so that readings of the height of the mercury column, measured with a cathetometer, could be made with both rising and falling mercury column. By using the average of the readings, any error due to sticking of the mercury column was minimized. The sample was placed in a 2-in.-long platinum boat formed into the shape of a half-cylinder. The sample was prepared as a fine powder and pressed against the inner wall of the platinum boat. A silica thermocouple well, covered with a platinum sleeve, was placed lengthwise along the center of the boat. The temperature of the sample was read on a calibrated Pt-10 pct Rh theirnocouple. The temperature profile of the furnace was constant within ± 0.2 C over the length of furnace occupied by the sample. The horizontal furnace used to contain the sample and the quartz reaction tube was electrically heated, with the control thermocouple placed 1/4 in. from the windings. A Brown Pyro-O-Vane temperature controller with a 3-sec response time was used to maintain the furnace temperature. Before beginning a run, the apparatus, including the bellows manometer, was thoroughly washed in sib with acid and distilled water to remove any

Jan 1, 1963

By W. F. Domis, F. von Gemmingen, F. Garofalo

The effect of rain size on the creep behavior of an austenitic iron-base alloy has been studied at 1300° F under conditions of constant stress. The average grain diameter varied between 9 and 190 p (ASTM 10.7-2). The subgrain size and apparent activation energy for creep were not found to depend in any systematic manner on gain size. Grain hozmdary serrations were observed in all specimens tested. The variation of the secondary-creep rate, 2,, with the pain diameter, 1, is given by: es = K(213m + 1 )/l.]; 1, is the grain diameter zohere is reaches a minimum. Results in the literature shoujthat 1,increases as the temperature increases. At low creep temperatures is is therefore proportional to 12, at intermediate temperatures <, exhibits a minimum, and at high creep temperatures is is proportional to 1/1. The dependence ofis on the stress, u, is given by: t, = A" (sinh acrln, a and n show little variation with 1. The dependence of A" on 1 is given by: A" = KA[(21k + 13)/1]. For 1 » >>lmn = 4, and at constant values 01- isat lozu stress levels, the steady-state stress is given by the Hall-Petch relation. The variation of c, with grain size is believed to he associated particularly with grain boundary generation of dislocations which becomes more important with increasing temperatures. SEEMINGLY inconsistent experimental findings have led to a certain amount of confusion concerning the effect of grain size on creep behavior. The confusion is attributed to a lack of test results obtained over sufficiently wide ranges in grain size, stress, and temperature. Experimental results presented in this paper and results available in the literature1-5 show a consistent behavior and lead to a unified concept for the dependence of secondary-or steady-state creep rate on grain size. In accordance with this concept, which is substantiated experimentally, secondary-creep rate should increase with increasing grain diameter at low creep temperatures, leading to greater creep strengths for fine-grained materials. At intermediate temperatures the secondary-creep rate should decrease to a minimum and then increase as the grain size is increased. At high creep temperatures the secondary-creep rate should decrease with increasing grain size, leading to greater creep strengths for coarse-grained materials. Whether a temperature is high or low for creep in a metal or alloy depends primarily on its melting temperature. MATERIAL AND TEST PROCEDURES The material tested was an austenitic iron-base alloy of the following composition (wt pct): C, 0.005; N, 0.017; Mn, 1.26; Ni, 14.21; and Cr, 17.18. Two 15 -lb ingots were forged and hot-rolled to 5/8-in.-diam bars which were then cold-rolled and cold-swaged to a diameter of 0.49 in. These were sectioned into 3-in.-long blanks and heat-treated at various temperatures to obtain different grain sizes. The procedures employed and the results obtained are summarized in Table I. The secondary treatment at 1400° F followed the primary treatment without intermediate cooling to room temperature. This final treatment was employed to minimize various effects arising from heating to different temperatures during the primary treatment, particularly differences in intragranular dislocation structure, segregation of interstitial atoms at grain boundaries, and density of grain boundary ledges. The intragranular dislocation structure was examined by electron transmission metallography after each of the treatments described in Table I. Isolated dislocations and a few tangles were observed but no evidence of subgrains was found after any of the treatments. No grain-size effect was evident on the density and distribution of dislocations. Stabilization of grain boundary structure is particularly important because, as will be discussed later, grain boundaries influence high-temperature creep behavior. Magnetic-permeability measurements after each treatment and

Jan 1, 1964

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,&apos; (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.&apos;"" 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&apos; 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 L. S. Reid, Irwin Politziner F. M. Townsend

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,&apos; (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.&apos;"" 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&apos; 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 C. Wagner

When an alloy solidifies and the composition of the solid differs from the composition of the liquid, atoms of the alloying elements rejected at the solid-liquid interface have to diffuse toward the bulk liquid. Diffusion may be supported by convection. Theoretical calculations have been made for a liquid without convection, for a liquid involving natural convection, and for solidification of a liquid alloy at the surface of a rotating disk as an example of forced convection. WHEN an alloy solidifies, the equilibrium composition of the originating solid phase differs in general from the composition of the liquid phase. Thus, segregation may take place, i.e., the composition of the solid phase formed at different times differs from the bulk composition. The degree of segregation depends on several factors such as the partition ratio of the alloying elements between the liquid and the solid phase, the rate of solidification, and convection. To clarify, a theoretical analysis has been worked out. The following presuppositions are made: 1—Calculations are confined to the solidification of binary alloys in which the concentration of component A (solvent) is much greater than that of component B (solute). 2—The equilibrium concentration of the alloying element B is assumed to be lower in the solid phase than in the coexisting liquid phase. 3—At the solid-liquid interface, partition equilibrium is assumed. 4—The dependence of the partition ratio of the solute between the liquid and the solid phase on concentration and temperature is disregarded since only small concentrations of the solute are considered. 5—The concentration of the alloying element B in the solid phase is supposed to be lower than the saturation concentration for the formation of a second solid phase. 6—diffusion in the solid phase is disregarded, since diffusion rates in the solid state are comparatively low. When the partition ratio of the solute between the liquid and the solid phase is greater than unity, a part of the solute is rejected at the solid-liquid inter- face and has to diffuse toward the bulk liquid. Thus there is necessarily a concentration difference in the liquid between the solid-liquid interface and the bulk liquid as has been pointed out by Hayes and Chipman,&apos; Rutter and Chalmers,&apos; Tiller, Jackson, Rutter, and Chalmers, and others (see Fig. 1). In view of the concentration gradient in the liquid phase, the ratio of the concentration of the solute in the solid phase to that in the bulk liquid differs from the equilibrium ratio, although at the solid-liquid interface, partition equilibrium is assumed. Under certain limiting conditions, the solid may even have virtually the composition of the bulk liquid as is shown below. General Equations Let u be the linear crystallization velocity in cm per sec, D the diffusion coefficient of component B in the liquid phase in cm2 sec-&apos;, c&apos;,, and c", the concentrations of component B in the liquid and the solid phase, respectively, in mol per cm", and s the distance from the solid-liquid interface toward the bulk liquid at a given time t. The partition ratio is assumed to be equal to a partition constant K. Thus. The difference in the specific volume between the liquid and the solid phase may be disregarded. In view of the discontinuity of the concentration of the solute at the interface according to Eq. 1, u(c&apos;, — c",), , , = tic&apos;, (S = 0)(1 — K) gram-atoms B per unit area per unit time are rejected at the interface and have to diffuse toward the bulk liquid. From Fick&apos;s first law it follows that uc&apos;,,(s = 0) (1 ~K) == - D(dc&apos;,,/?>s). _-„. [2] The concentration profile in the liquid near the interface depends on the interplay of diffusion and convection as is discussed in the following sections. A schematic concentration profile is shown in Fig. 1. The "effective thickness of the diffusion boundary

Jan 1, 1955