The use of chromatographic retention volume data to obtain vapor-liquid equilibrium coefficients has been extended to a ternary system with all components present at finite concentrations in the liquid phase and essentially a light binary gas comprising the vapor phase. The data are obtained by observing the retention time of the solute of interest when that solute is eluted by a flowing gas stream through a column packed with an inert porous firebrick on which is impregnated a fixed, relatively nonvolatile liquid. The retention times are used to compute retention volumes which may be related t o the K values of the solute.A previous mathematical description of the elution process is modified to include the case of an N component elution gas, all constituents of which are soluble in the fixed liquid phase. A general solution is derived for the multicomponent elution gas from which K values for each component may be calculated i f retention volume data are taken for each, provided molecules of the solute samples used are distinguishable from others of the same species present in the eloution gas. For the case of indistinguishable sample molecules, the rate theory development of Stalkup and Deans is verified for a binary elution gas.Sufficient retention data were taken to completely define the concentration dependence of the K value for propane in the system methane-propane-n-decane at -20", 0", 40°, and 70°F. from 20 to 1,OOO Ib/sq.in.abs. Six methane-propane binary mixtures were used as elution gases, consisting of 2.08, 4.31, 6.90, 9.44, 13.09, 16.27 mole % propane. Retention data were also taken for methane under the same conditions. Distinguishability of sample molecules was achieved by using radioactive solute samples tagged with carbon-14. Propane retention data were also obtained for the system methane-propane-n-heptane at In this paper the application of gas-liquid partition chromatography ( GLPC) to the study of vapor-liquid equilibria is generalized to include multicomponent systems. Establishing the validity of such application greatly increases the spectrum of systems subject to study by GLPC. The rapidity and simplicity of the elution process compared with conventional methods for studying phase behavior make such an increase in applicability desirable.In the GLPC process a vapor-liquid system is established by passing a gas through a column packed with an inert, porous solid material on the surface and in the pores of which an immobile liquid phase is fixed. A small perturbation with the solute of interest is eluted through the column by the carrier gas. The time required for the perturbation to travel the length of the column is determined by and related to the manner in which the solute distributes itself between the two phases. GLPC was first introduced by Martin and James (21 ) .It was quickly recognized by Martin (20) technique represented potentially the easiest method of measuring the physical properties of solutes in solvent systems. His suggestion was verified by Porter, Deal, and Stross ...

Vapor Press., Torr. 1.50 xlO"6 5.16 xlO"6 2.32x10"* 6.37x10"* 1.90 xlO"3 5.46 xlO"3 2.91 xlO"2 5.56 x 10"2 0.108

The results of the experimental solubility data are tabulated in Table II. In Figure 3 the methane K-values are presented. Figure 4 shows the extrapolation of the solubility data to the vapor pressure of re-decane. A temperature crossplot of the K-values for methane obtained from the solubility data in this work along with those obtained by Sage and Lacey (6) at higher temperatures is presented in Figure 5. The agreement of results from the experiments outlined herein with those of more complex static equilibrium studies is quite favorable. ADVANTAGE OF THE SOLUBILITY TECHNIQUEThe validity of the volumetric, static experiments described herein for determining the solubility of a highly volatile material in another which is quite nonvolatile is established by the consistency of the results with those of other investigators. The technique has the distinct advantage of furnishing vapor-liquid equilibria information from P-V-T measurements, with no need for sampling or analysis of either phase when the assumption that the vapor contains essentially none of the nonvolatile components is valid.

Field tests have shown that compass orientation and the length of hydraulic fractures can be determined by pulse testing in different directions from wells before and after fracturing. The method determines orientation by sampling a large portion of the reservoir, is applicable to cased holes, and provides an estimate of fracture length. Introduction It is often desirable to know the flow patterns created after a number of wells in a reservoir have been fractured. This requires a knowledge of the compass orientations, lengths, and conductivities of hydraulic fractures. Standard pressure interference testing has been used to determine the orientation of natural fractures and inflatable impression packers and television cameras have been used to locate hydraulic fractures in open-hole completions. None of these methods, however, determines both the compass orientation and the length of a hydraulic fracture, and only one method samples a large portion of the reservoir. This paper describes a way to use pulse testing to determine both the compass orientation and the length of such fractures. Results from two field tests are presented. presented. Theory Johnson et al. give a complete description of the procedure and technology of pulse testing. The method procedure and technology of pulse testing. The method involves changing the rate of flow at one well and measuring the pressure response at one or more offset wells. The response, or pressure pulse, is characterized by two parameters, the time lag and the pulse amplitude. These and other pulse-testing parameters are illustrated in Fig. 1. Pulse amplitude depends on flow rate, pulse interval, between-pulse interval, and, to some extent, on reservoir properties, transmissibility, and storage. Time lag properties, transmissibility, and storage. Time lag primarily depends on transmissibility and storage. It has primarily depends on transmissibility and storage. It has been shown that the presence of a high-transmissibility zone or of a zone of very low transmissibility (a barrier) can be detected by pulse testing. A fracture, of course, creams a zone of high transmissibility with an insignificant change in storage as compared with that of the unfractured matrix. Fig. 2 shows the relationship between transmissibility (for a fixed value of storage) and both time lag and response amplitude. Clearly, a change in time lag is sensitive to a change in transmissibility. Pulse amplitude, on the other hand, varies directly with Pulse amplitude, on the other hand, varies directly with changes in transmissibility over part of the range; but for greater than 3 x 10 md-ft/cp, the amplitude responds weakly to increases in transmissibility. For these reasons, changes in time lag should be most effective for determining the direction and length of a hydraulically induced fracture. To evaluate the feasibility of this procedure, a model of a reservoir (see Appendix) was constructed with one pulsing well in the center and several responding wells pulsing well in the center and several responding wells around the pulser. A pulse test was then simulated by producing and shutting in the pulsing well intermittently producing and shutting in the pulsing well intermittently for equal increments of time. The pressure response was computed for all responders and was then analyzed using the tangent method to determine the time lags. A hydraulic fracture was then simulated by narrow blocks with high but finite permeability extending an assumed length from the well. The same pulsing sequence was repeated and the results were analyzed for time lags after fracturing. The ratio of the time lag before fracturing to the time lag after fracturing was plotted as a function of direction (angle) around the pulser. JPT P. 1433

One of the more important problems to be solved in designing a miscible flood is related to the size of the solvent bank used. Size of the bank may be critical to economic success. Too large a bank loses money; too small a bank may deteriorate and fail to maintain the miscibility needed for high recovery.An important factor in deterioration of a small bank is permeability channeling. In a highly stratified reservoir, solvent speeds ahead in the more permeable zones and mixes laterally with fluids bypassed in adjacent, low-permeability strata. Numerical solutions have been obtained .for the differential equations that describe the movement of a slug through a two-layer system in which mixing occurs both in the direction of flow and transversely. The solvent slug is assumed to have the same density and viscosity as the resident fluid and the pushing fluid. These solutions have been verified by comparing them with similar concentration profiles obtained in the laboratory in a 36-ft stratified model packed with glass beads.The theoretical study revealed that when the dominant mechanism causing a bank to fail is lateral mixing the bank size needed for a given recovery may increase with length rather than decreasing as the square root of reservoir length, as suggested by one-dimensional mixing theory. From a comprehensive examination of the variables, a generalized correlation is developed that relates strata thicknesses, bank size, fluid velocity, mixing coefficients, system length and simple solventresident fluid phase behavior to the area miscibly swept.