Experimental Studies of Oil Flow through Sands Citation Scoville, Loren P. (1930) Experimental Studies of Oil Flow through Sands. Master's thesis, California Institute of Technology. doi:10.7907/byrq-2c70. https://resolver.caltech.edu/CaltechTHESIS:08262024-210019529 Abstract No abstract. Item Type: Thesis (Master's thesis) Subject Keywords: (Chemical Engineering) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Lacey, William Noble Thesis Committee: Unknown, Unknown Defense Date: 1 January 1930 Record Number: CaltechTHESIS:08262024-210019529 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:08262024-210019529 DOI: 10.7907/byrq-2c70 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 16668 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 27 Aug 2024 22:59 Last Modified: 27 Aug 2024 23:03 Thesis Files PDF - Final Version See Usage Policy. 12MB Repository Staff Only: item control page

EXl?ERIMEBTAL STUDIES OF OIL FLOW THROUGH SANDS Thesis by Loren P. Scoville In partial fulfill ment of the re Quirements for the deg ree of Master of Science. CALIFORNI A IMSTITUTE OF TECHN OLOGY PASADENA , CALIFORNIA -- 1 9 3 0 -- 1. EXPERDi E:NTAL STUDIES OF OIL FLOW THROUGH S.AtrDS The percentage recovery of petrolewn from underground formations is rather varied. Opinions advanced by various experts range from 10 to 90~. Melcher estimated that ap- proximately 25% is recovered in the Bradford field in Pennsylvania. The estimates for the Texas and Louisana fields run as high as 50%. It can be seen by the above percentages that the recovery of the petroleum present in the formation is by no means complete. A great deal of work has been done to devise methods of increasing the yield. The Petrolewn Experimental Station of the u. S. Bureau of Mines is working on this problem. In an article in the .American Institute of .Mining and Ivletallurgical Engineering of October, 1928, R. v. A. Mills, J. Chalmers, and J. s. Desmond gave a short review of the questions which this station is attempting to answer on the basis of their experimental work. They are as follows: (1) What are the causes of excessive amounts of thick oil where re pressuring with compressed air and how can this thick oil be predicted? (2) Why is the oxygen often converted to carbon dioxide in re:pressuring? By the oxidation of the oil? How can the formation of low B. T. u. casinghead gas loaded with nitrogen be prevented? (3) ¼hat percentage of the oil can be recovered by ordinary methods when the gas originally dissolved in the oil is the propulsive agent? 2. (4) What percentage can be recovered by repressuring? (5) Wha t are the rela tive propulsion efficiencies of compressed air and natural gas? ( 6) Why should the re p ressuring be done early in the production rather than later? The flow tubes used in these investigations varied from 3 inches in diameter and 3 feet long to 12 inches in diameter and 26 feet long. The pressures used were from one pound to 800 pounds per square inch. The oil used was Bartlesville crude of 330 A.P.I. and Saybolt viscosity of 60 seconds at . 70 F and 47 seconds at 100° F. No results were given in the article. F. G. Tickell in the American Institute of Mining and Metallurgical Engineering of October, 1928, says that resistance to flow of oil in sand is proportional to its viscosity. In oil the viscosity is inversely proportional to the pressure since dissolved gas reduces the viscosity. Therefore the flow may be taken as proportional to the square of the pressure. Gas bubbles ca use increased r esistance due to deforma tion of the bubbles as they pass through the openings in the sand. This resistance is called the Jamin effect. All influences that restrain movements of li quids have the same effect as increasing t h e viscosity. L. c. Uren, Professor of Petroleum Eng ineering at the University of California, in an article in the Na t"ional Petroleum News of February 9th, 1927, discourses at length concerning resistance to flow of oil through reservoir s ands. 3. ~re even offers equations by which to calculate the amount of flow. He gives a formula worked out by ~'l Difrawf from the results of King and Slichter who worked on the flow of underground water. q = (0.383 x 10- 0 ) x p x d2 x s U X 1 where x o3.3 q is discharge in gallons. p is effective pressure in feet of water. d is effective diameter of sand grains in mm. s is cross section of sand colwnn. u is absolute viscosity of oil in poises. l is length of sand colwnn in feet. 0 is porosity of sand in percentage. Total flow must be found by integration. R. s. McIntyre derived the following equation by using the above, stating, however, that it is not complete. He gave no indication of what additional factors or terms are needed. p = C ,c q x V X LoglO L • • • • • • • t x d2 .>< 0 3.3 where R p is pressure loss in feet of water. q is q_uantity of oil in barrels per day. V is absolute viscosity in poises. t is thickness of sand in feet. d is effective diameter of sand in mm. e is percentage porosity. L is radius of drainage in feet. R is radius of well through productive sand in feet. 4o e is a constant. Uren then sums up that the flow varies (a) Directly as pressure. (b) Directly as the thickness of productive sand. (c) Directly as the sq_uare of the measured diameter of the sand grains. (d) Directly as the 3.3 power of the percentage porosity. (e) Inversely as the viscosity. (f) Inversely as the log of the ratio of radius of drainage to the ii'adius of the well. These eg_ua.tions, however, do not take into account the Jamin effect. They are worked out on the flow of water. With a change in viscosity u, they should work for oil which is not saturated with gas. As far as the writer knows there is nothing in the literature concerning definite data on flow measurements of oil and gas through sands. The ~uestions which such measurements should answer are the following: 1. Does the decrease in viscosity due to dissolving gas in oil incr~ase the flow or does the Jamin effect which enters when gas bubbles are present more than counter balance this decrease? 2. What would be the effect of back pressure on 3. Does the gas which comes out of the oil when the the flow? :pressure is reduced in a sand formation rise and flow in separate channels from the oil or continue mixed with the oil? 4. How does the pressure gradient vary with the distance from the well? 5. What effect does the porosity of the sand have upon resistance due to the Jamin effect? In order to be able to get data with which to obtain some light on these q_uestions a flow tube was constructed with glass sides so that the :passage of the gas and oil through the sand could be observed. Means of reading pressures at various points were also incorporated. Description of Apparatus. The complete set-up of apparatus exclusive of the carbon dioxide cylinder is shown in the photograph. It consists of three main :parts: the mixing and. storage chamber (I) , the flow tube (II), and the sepa rating and weighing chamber (III). These w1i ts were built by Kobe Inc., manufacturers of slotted casings and other oil well supplies. The mixing and storage chamber is 7 inches inside diameter, 14 inches inside height, and has a capacity of about 3 gallons. It is built of one half inch stock and has a flanged. cover to facilitate cleaning. Dis the oil inlet. A is the gas inlet which is divided so that the gas may be let in over the oil through B or bubbled. through the oil th rough C. The exit to the flow tube is through E. The flow tube is constructed of steel with glass sides for observation purposes. The glass is very slightly under one inch in thicJ{ne ss and is sup:ported by beams every 2-1/ 4 inches. In tests, it was found to withstand a p ressure of 250 pounds per s~uare inch without breaking . chamber is 27 inches. The length of the sand Its height is 4 inches and thickness l/2 inch plus the thickness of gaskets which would average about l/16 inch, ma.king the total distance between glass plates about 9/16 inch. There is a chamber 3-1/2 inch in length at each end of the sand chamber. placeable. The slots are 1~- Eight pressure tubes are inserted at P1 , P2 , P3 , etc. to obtain differential pressure readings. The oil enters at G from the mixing chamber and leaves at H going from there to the separating and weighing chamber. A gauge is put on P1 while P2 , P3 , etc. are connected by copper tubing and valves to either gauge N, or a manometer O, reading up to 20 pounds per square inch. The separating chamber is made of 1/4 inch stock and has a 7 inch inside diameter and is 19 inches high. In the upper half are three screens of 8 mesh to break down the foam. gas-oil mixture enters at J. The The oil may be removed at K while the gas goes through L to a trap and then to an American Meter Co. dry gas meter reading to .001 cubic feet. The sepa- rating and weighing chamber is on a 50 pound :platform scale so that weights may be taken initially and finally. Mis a water manometer to read the pressure within the chamber. It was found that the vaporization of the carbon dioxide upon expansion caused a cooling of the oil in the mixing chamber so the lead, Q, from the cylinder was sent through a water bath R before going to the mixing chamber. The water was A. B. in Ga 1.in • entrance to top of J. Inlet to separating ohatn L. Gas exit :fro c. Gas entrance to botto of !{ . m1Xi.ng ch robe r. o. D. Oil inlet. E. Oil outlet from ixing "·R. tube. s. ber to fl.o ~eparating ohamb r to trap and mixing ah mb.er. G. Inlet to f'lo ~ tu &. n. Ou.tl t from flo tub. T. r. eter. G u.ge . nometer. Gas line from C 2 e~lind r. ter bath. eetr1o heater. Gas et r. FLO TUBE. G_. Inl.et to fl tub • F1 and F • Sore n • 2 Pl , P2 , P3 , etc. Outl ts tor p readi H. Outlet from f'lo s. tube. - ---- - - ---------------------- .ICL 0 D. Oil inlet. N. Gauge. H. Outlet from now tube. o. Meroury manometer. J. Inlet to separating Pl, P2, P3, etc. ohamber. L. M. Outlets for pressure readings. Gas exit from separating R. Water bath. chamber to flow tube. s. u. Eleotric heater. Water manometer. Gas trap. heated by the electric heater s. The oil used was Union Crystal Oil of a specific gravity of 0.850 or 35° A.P.I. This oil was used because it has approximately the same general properties as a crude and has the advantage of being clear and thus facilitating observation through the glass. The gas used was co 2 and was selected because of its large solubility in oil thus emphasizing the effects of the gas bubbles and also because of its safety. White washed Monterey Beach sand was used. was from 14-20 mesh. Sand No. l Sand No. 2 was from 150-200 mesh. Run No. l - Sand No. 1 - No Gas. p Gal/Min 4 a 10 9.5 s.o 5.75 4.5 2.75 1.6 .0437 b 15 14.1 11.6 9.25 6.7 4,3 2.5 .0620 C 20 18.7 14.4 11.7 8.7 5.6 3.3 1.75 .0638 d 25 22.5 17.7 14.3 10.8 1.2 4.2 2.25 .1005 e 30 27.5 23.0 17.9 13.4 8.7 5.3 .1195 f 35 31.0 27.5 22.0 16.0 10.4 5.7 .1410 g 40 37.0 30.5 24.0 18.0 12.0 6.5 .1762 h 45 41.0 34.0 27.5 20.0 13.5 7.7 4.2 .1837 i 50 45.0 37.0 30.5 22.5 14.75 8.6 4.7 .2113 s. RWl No. 2 - Sand No. l - Saturated at P1 with co 2 • RWl P 1 p2 p3 p4 P5 p6 P7 PS *a 10 9.0 7.4 5.75 4.3 2.7 1.6 .'75 *b 10 *o 20 17.7 14.9 12.1 *d 30 27.5 22.5 18.0 l4o5 *e 40 37.5 32.0 25.0 18.2 *f 40 *g 40 38.0 33.0 27.0 19.2 12.7 'l. 2 *h 40 39.0 35.0 28.0 23.0 18.0 *i 20 18.0 15.2 13.0 10.3 j 40 37.5 32.0 27.0 *k 50 46.0 39.0 *l 30 27.5 m 20 *n Gal/Min Gas/Oil -----.0352 -----.0378 .0705 .02405 3.0 .1410 .1280 4.0 .2113 .0358 .1878 .1432 4.7 .1793 .1297 12.25 7.25 .1832 .5110 7.3 4.7 2.2 .0563 .1500 22.0 16.3 11.0 '7. 3 .1538 .2670 33.0 27.0 18.0 .2290 .2592 23.0 18.2 14.5 10.75 6.8 4.0 .1330 .1442 17.7 15.0 13.0 10.6 7.4 4.4 2.2 00668 .1265 50 47.0 a9.o 32.0 24.0 16.0 10.7 6.7 .2352 .2350 0 30 27.5 23.0 19.2 15.3 11.2 7.0 3.8 .1174 .1730 p 50 .2253 .3470 q 45 42.0 7.3 .1878 .2860 r 10 9.3 1.2 .0282 .0635 s 25 22.5 3.5 .0977 .1522 t 35 5.0 .1210 .2792 *u 35 5.0 .1645 .2204 V 35 4.5 .1300 .1740 w 35 4.7 .1565 .2005 X 35 5.2 .1365 .2762 10.0 5.0 6.7 37.0 30.0 24.0 '7 .4 19.0 16.7 18.0 12.2 4.5 13.3 20.5 10.0 6.0 * Results discarded for reason set forth in description of runs. Ratio o:f Pressure Gradients of Run No. 2 to Run No. l between various points (oorreoted to equal flows). l?ressu.re Pl- p2 P2- P3 p3... P4 p4 .... P5 P5- p6 P6- P7 P7- PS 20 1.685 .600 .708 .763 .988 1.245 1.345 25 1.030 .750 .698 1.000 .944 1.370 1.318 30 1.015 1.020 .563 .882 .890 1.255 1.252 40 .956 .970 .882 .956 1.090 1.108 1.500 45 .733 .700 1.052 .782 .902 •. 727 1.370 Average 1.082 .810 .780 .878 .962 1.14 1.38 RWl No. 3 - Sand No. 2 - No Gas. RWl pl Gal/M~·n p2 P3 P4 P5 p6 P7 Pa 18 11.3 6 0.3 .948 17.7 10.8 0.5 l.565 X 10 a 50 44 35 28 b 75 66 51 41 0 100 73 40 0.5 2.025 d 125 93 62 0.4 2.412 e 150 112 o.6 2.767 10 • Run No. 4 - Sand No. 2 - Saturated with CO2 at P1• p5 p6 p7 PS Gal/min. Gas/oil 27 15 11.2 0.7 .672 57 o.a 1.070 100 81 1.5 1.505 # # # *d 125 104 1.3 1.895 0.66 *e 125 104 0.7 1.582 1.136 *:e 150 124 1.0 2.178 1.010 g 150 1.0 2.09 1.117 h 125 0.7 1.785 0.922 *i 125 0.7 1.680 0.965 Run pl a 50 38 b 75 C p2 p3 p4 3 X 10 Rwi No. 5 - Sand No. 2 - Saturated with CO 2 at P8 • # RW'l pl Ps Gal/min. x 103 a 150 100 .753 b 100 50 .940 0 150 100 1.130 d 100 50 1.420 (73 F. temp.) No ga s-oil ratio was obtainable due to a slip in the gas meter a t such low rates. 11. ]),scription of Runs. The temperature of the room was kept at approximately 80° F for all of the runs. The screens used for sand No. 1 were IIBde up of slots of O.04 inch width and placed O.085 inch apart. The screens used for sand No. 2 were made up of slots of o.ooa inch width placed 0.042 inch apart. It was found necessary to put a thin layer of 65 mesh sand next to the exit slot to prevent sand from passing through. Two runs were made with eaoh sand. First the :now was measured for oil without dissolved gas and then the flow of the oil saturated with gas was · taken. Run No. 1 was made by using the CO 2 for pressure only by introducing it over the oil through B. Sub-runs (designated by letters in the tables) were ma.de with the pressure varying from 10 to 50 pounds per square inch. The time and weights were recorded. at the be~ ginning and end of the run. The gallons per minute were then calculated by the formula: Gallons/Minute ::; O.l 4lOW T where W is the weight in :pounds, Tis the time in minutes and 0.1410 a conversion factor from . pounds to gal loris. In Run No. 2 the oil was first saturated with gas by allowing the CO 2 to enter through C and bubble through the oil. The valve at D was opened slightly to allow a steady bubbling of the gas through the oil at the required pressure. Some difficulty was experienced in completely saturating the gas in this manner. The sub-runs were made in the same wa.y 12. as those for Run No. 1 with the addition that the initial and final readings of the gas meter were recorded. The gas- oil ratio was then computed by means of the formula: Gas/Oil = V - 0.01885 W 0.1410 W where V is the volume of gas passing through the meter in cubio feet, Wis the weight of oil in pounds and 0.01885 is a conversion factor from pounds to cubio feet. This formula takes into account the volume of the oil displacing gas. The gas-oil ratio ob- tained in this manner is in cubic feet per gallon. It is known that the solubility of a gas in an oil plotted against the pressure of saturation is very nearly a straight line. The runs were therefore plotted with the gas-oil ratio against the pressure of saturation. The first few runs did not agree at all and it was finally found necessary to bubble the gas through the oil for about one and one half hours to insure saturation. The cooling caused by the evaporation of the co 2 increased the solubility of the gas to such an extent that it was found necessary to install the water bath R in order to heat the gas as it came from the cylinder. After the correct solu.bili ty curve was obtained all runs were discarded which did not agree with the curve. The formation of bubbles could be seen very clearly through the glass sides of the flow tube. They were found to form at a point midway in the sand compartment. In Rwis No. 3 and No. 4 a ve1~ fine sand from 150-200 mesh was used. In Run No. 3 the gas was again used only as a 13. source of pressure. It was fowid that the fine sand plugged up the pressure tubes after a short time, and for this reason few of the intermediate pressures could be read. Run No. 4 was similar to Run No. 2. In this case it was found that the gas formed at a distance of about 2/5 of the length of the sand compartment from the entrance. The for- mation took place at slightly different distances in the top, center and bottom of the f'low tube. The curve below illus- trates the line of appearance of the bubbles in the sand. The arrow indicates the direction of the flow~ The bubbles in this case had about the same size as those formed in Run No. 3. However, they were very greatly deformed while passing through the sand, while in No. 3 they were able to pass through with great ease while retaining their spherical shape. The bubbles did not increase in size more than the amount that would be predicted by their expansion due to the decrease in pressure. In other words, the bubbles did not combine with the newly formed ones as they flowed through the • chamber. The results of these runs were plotted as in the curves that follow. In order to reproduce graphically the effect of the gas dissolved in the oil, the ratio of the pressure gradients at different points of Run No. 2 to those of Run No. l 14. were calculated as already shown in a preceding table. The average of this ratio was then found for each one of the points and the results plotted~ The resulting averages were found to rm.ke a smooth curve except for the average ratio of the pressure gradient between the first screen and the point P2 • The explanation for this discrepancy is that during the course of the runs there was at times a distance of one half inch to one inch in the upper corner in vmich there was no sand, d.ue to compression. This fact made the pressure gradient between the first screen and - the point :P 2 unreliable. Run No. 5 was made by saturating the oil with gas at a given pressure and then putting a gas pressure over the oil of 50 pounds per square inch greater then the saturation pressure. By manipulating the valve at H, a back pressure was maintained equal to the pressure of saturation. As heat was not available sub-run (a) was made at 7-t' F and consequently gave a lower rate of flow. The balance of the sub-runs were ma.de at 80° F. This series of runs do not check. The reason for discrepancies may be the difficulty in maintaining the temperature of the room at a constant value. However, since the flow of oil with- out dissolved gas at 50 pounds differential p~essure is 0.00948 gallons per minute (by Run No. 3), the results seem in general to indicate an increased flow caused by the decrease in viscosi~ ty due to the dissolved gas. Due to the limited time available, a check of this run was impossible. .. ., I•• . ., -r---• • -: • 1_ I;;! .. 4...... _• t • ~ - -.. ,. ...- ,. r;- • 15. Conclusions. 1. It was found, as would be expected, that the pressure varies directly as the distance from the exit screen when oil without gas flows through sand. 2. When gas is :present in the oil, the pressure gradient increases as the gas is expelled in the fonn of bubbles. This is caused first because the viscosity of the oil increases when dissolved gas is expelled and secondly because the bubbles, when flowing through the sand, must be deformed thus causing a resistance due to the increase in surface of the gas-oil interface. The latter is what has been considered in this paper as the Jamin effect. 3. The Jamin effect is not pronounced when the sand is coarse because the interstices in the sand are of such dimensions that the deformation of the bubbles is slight. How- ever, in a fine sand this effect is very pronounced. 4. Equilibrium is not reached immediately between the oil and its dissolved gas when flowing through the sand. The oil retains the gas in a supersaturated condition for acertain distance. Although this distance is less in a. case of the fine sand where the flow was slow than in the coarse sand where the flow was rapid, the time of supersaturation was greater in the fine sand. This is probably due to the fact that the turbulence is a great deal less. 5. The bubbles did not combine or flow in separate channels from the oil. They were evenly distributed from to1> to bottom of the flow chamber, increasing in number rather 16. than in size as the pr essure decreased. 6. A run was made to determine how much the flo w would be i n creased by using a. back pressure e qual to the :p ressure of sa turation and using a 50 pound per s quare inch greater driving :pressure, but due to limited time the run could not be sufficiently ohe clced. Acknowledgements. Apprecia. tion should be expressed to C" J. Coberly of Kobe, Inc., who very kindly built and loaned the a pparatus used and who assisted with many helpful suggestions. C. F. Hansen and E. Messrs. s. Croasdale, members of his staff, also helped in the design of the apparatus and gave other valuable assistance. The flow tube itself was designed entirely by Mr. Coberly and his staff. Appreciation 1s due Dr. w. N. Lacey, under whose direction this work was done, for many constructive ideas, his interest and co-operation, and his critical perusal of this thesis. 17. B I BL I OGRAP HY "Movement of' Fluids in Porous Solids." Nutting, Oil and Gas Jour. Dec. 23, 1926. p. 26. "Textures of Oil Sands with Relation to the Production o~ Oil." Melcher, Bur •.A.In. Ass. Petr. Geel. VLII (1924) ttEffect of Pressure on the Migration and Aooumula tion of Petroleum." Van Tuyl and Beckstrom. B.A.A.P.G. X (1926) "Measuring 1ratural Forces that Hold Back Flow of Oil in Ground." L. c. Uren. Natl. Petr. News. Vol. 19 No. 6. Feb. 9, 1927. PP• 67-76. "Oil Sands and Production Relations." H. C. George and W. F. Cloud. Okl. Geol. Sur. Bull. no. 43 Aug. 1927. 140 p. "Use of the Acetylene Tetrachloride Method of Porosity." c. E. Sutton, u. S. Bureau of Mines. Reports of Investigations. Ser. no. 2876, June, 1928. 10 pp. rrpetr. Recovery by the Soda Process." P. G. Nutting, Oil and Gas Journal. vol. 27, no. 22. Oct. 18, 1928. pp. 146 and 238. "Capillary Phenomena as Related to Oil Production." F. s. Tickell, .Amer. Inst. of Mining and Met • . Eng, Tech Pub. 138, Oct., 1928. 12 pp. ttRelative Propulsion Efficiencies of Air and Natural Gas in Pressure Drive Operation. 11 H. R. Power, .Amer. Inst . .Mining and Met. Eng. Tech Pub. 148, Oct.; 1928. 18 pp. "Volume Controllers Aid in Giving Uniform "Push" in Repressuring Sands." c. E. Mason, Natl. Pet. News. vol. 20, no. 3, Jan. 18, 1928. PP• 87-88. ffQil Recovery Investigations of the Petroleum Experiment Station of the u. s. Bureau of Mines." R. v. A. Mills, J. Chalmers, J. s. Desmond, Amer. I nst. Mining and Met. ~ng., Teah Pub. l44, Oct, 192.8. 36 pages. "Roak Pressure -- Subsurface Formations. 0 L. L. Brundred. Oil Bu.l. vol. 14, nos. 3 and 4. Mar. and .Apr. 1928. pp. 293, 299, 301, 40l, 403, 405, 407.