Pilot plant studies of a new hot water process for extraction of bitumen from Utah tar sands

This research project concentrates on the design and development of a pilot plant to extract bitumen from Utah tar sands with the minimum loss of water, taking into consideration the environmental parameters of this area through a modified hot water process. Although the hot water process is being used in Canada for the extraction of bitumen, the basic differences in the overall properties between the two tar sands, rule out the application of the Canadian hot water process to Utah tar sands, and warrant an appropriately modified hot water process technology specifically suitable to process Utah tar sands. Accordingly, a new process has been suggested to achieve the research goals and obviating the necessity of a tailings pond. To recover maximum water, solids are divided into two streams of +65 mesh and 65 mesh and dewatered separately. This modified process envisages the use of major equipment in the form of a new three-product, open-top hydrocyclone, a high-density thickener, and a spiral classifier. Three-product hydrocyclone is a new unit. An 11-inch plastic model of a new three-product, open-top hydrocyclone was conceived and constructed primarily to determine the flow pattern. This was tested with sand and water. The flow pattern was determined using a pitot tube. It was distinctly observed that unlike the two-product conventional hydrocyclone, where a free-vortex pattern exists, forced-vortex pattern for tangential component of velocity was confirmed. In the pilot plant, a 30-inch diameter three-product, open-top hydrocyclone was constructed and tested with a slurry containing bitumen, water, and sand. Bitumen has a specific gravity very close to one, it is very difficult to separate it from water. Since bitumen is known to have an affinity for attachment to air, fine bubbles of air were injected in the feed pipe of the three-product, open-top hydrocyclone to alter its specific gravity. The air thus injected attached to the bitumen and caused it to float in subsequent steps. More than 90% bitumen was recovered in the overflow and middling streams. Injection of air caused more bitumen to report at the overflow stream.

PILOT PLANT STUDIES OF A NEW HOT WATER PROCESS FOR EXTRACTION OF BITUMEN FROM UTAH TAR SANDS by Rajinder Kumar A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical and Fuels Engineering The University of Utah August 1995 Copyright © Rajinder Kumar 1995 All rights reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Rajinder Kumar This dissertation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. 7£f0 Iq'f~ M 31,1995 ~ Noel de Nevers );r7i 3~ !riJ~ Ma 31,> 1995 May 31, 1995 THE UNIVERSITY OF UTAH GRADUATE SCHOOL FIN AL READING APPROV AL To the Graduate Council of the University of Utah: I have read the thesis of Raj inder Kumar in its fmal fonn and have found that (1) its format, citations and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is ready for submission to The Graduate School . t2-~ct~ Donald A. Dahlstrom Clair, Supervisory Committee Approved for the Major Department Approved for the Graduate Council AnnW.Han Dean of The Graduare School ABSTRACT This research project concentrates on the design and development of a pilot plant to extract bitumen from Utah tar sands with the minimum loss of water, taking into consideration the environmental parameters of this area through a modified hot water process. Although the hot water process is being used in Canada for the extraction of bitumen, the basic differences in the overall properties between the two tar sands, rule out the application of the Canadian hot water process to Utah tar sands, and warrant an appropriately modUied hot water process technology specHically suitable to process Utah tar sands. Accordingly, a new process has been suggested to achieve the research goals and obviating the necessity of a tailings pond. To recover maximum water, solids are divided into two streams of +65 mesh and -65 mesh and dewatered separately. This modified process envisages the use of major equipment in the form of a new three-product, open-top hydrocyclone, a high­density thickener, and a spiral classifier. Three-product hydrocyclone is a new unit. An 11-inch plastic model of a new three-product, open-top hydrocyclone was conceived and constructed primarily to determine the flow pattern. This was tested with sand and water. The flow pattern was determined using a pitot tube. It was distinctly observed that unlike the two-product conventional hydrocyclone, where a free-vortex pattern exists, forced-vortex pattern for tangential component of velocity was confirmed. In the pilot plant, a 30-inch diameter three-product, open-top hydrocyclone was constructed and tested with a slurry containing bitumen, water, and sand. Bitumen has a specific gravity very close to one, it is very difficult to separate it from water. Since bitumen is known to have an affinity for attachment to air, fine bubbles of air were injected in the feed pipe of the three­product, open-top hydrocyclone to alter its speci'fic gravity. The air thus injected attached to the bitumen and caused it to float in subsequent steps. More than 90% bitumen was recovered in the overflow and middling streams. Injection of air caused more bitumen to report at the overflow stream. v CONTENTS ABSTRACT.............................................................................................................. iv USTOFTABLES .................................................................................................. ix LIST OF FIGURES ............................................................................ : ................... xi ACKNOWLEDGMENT......................................................... ........... ......... ..... ....... xiv CHAPTER 1 INTRODUCTION...... ........ ....... .... ..... .... ........ .... ... .... .............................. ..... 1 1.1 The U.S. Scenario.... ..... .............. ......... ....... .............. ............ ..... 1 1 .2 The Current Project.................................................................... 2 2 LITERATURE SURVEy ........................................................................... 4 2.1 Definition of Tar Sands ......................................................... 4 2.2 Geographical Distribution of Tar Sand Deposits ............. 4 2.3 Bitumen Properties: General, Physical and Chemical... 8 2.4 Selection of the Technology ................................................ 10 3 RESEARCH OBJECTIVES AND METHOD TO ACHIEVE THEM .............................................................................................................. 26 3.1 Research Objectives ............................................................. 26 3.2 Method to Achieve the above Mentioned Research Objectives .................................... : ..................... -..................... 27 4 SOLID-LIQUID SEPARATION THEORY AND TPOT HYDROCYCLONE.................................................................................... 34 4. 1 Settling of Solids................................................................... 34 4.2 Free and Forced Vortex ........................................................ 37 4.3 Definition of D50.... ................... ....... ........ ............ ................... 38 4.4 Pressure Drop in a Pipe ....................................................... 38 4.5 Viscosity of the Slurry ........................................................... 39 4.6 Density of the Slurry .............................................................. 39 4.7 TPOT Hydrocyclone .............................................................. 39 5 STUDY OF THE TPOT HYDROCYCLONE ..................................... 40 5.1 Flow Pattern in TPOT Hydrocyclone...... ..... ............ ........... 40 5.2 Variables Studied for the TPOT Hydrocyclone.......... ...... 53 5.3 Data Collection and Analysis .............................................. 54 5.4 Results and Discussions ...................................................... 56 6 BENCH SCALE THICKENING TESTS ON TAILINGS .............. 91 6.1 Flocculant and Flocculant Screening................................ 92 6.2 Bench Scale Thickening Tests ..................... : ..................... 94 6.3 Unit Area Calculations .......................................................... 96 7 DESIGN OF SPIRAL CLASSIFIER. ........ .......................................... 109 7.1 Sizing of Spiral Classifier.... ................................................ 1 09 7.2 Concentration of +65 Solids Obtained in a Beaker Test of Tailings........... ............................................................ 112 8 FINAL RESULTS AND DiSCUSSiONS .......................................... 113 8.1 Product Quality from TPOT Hydrocyclone........................ 113 8.2 Water Loss in the Process ................................................... 115 8.3 Comparison of Water Loss in New Process with the Water Loss in the Process Shown in Figure 2.7 .............. 116 8.4 Explanation of Increased Underflow Flow Rate above Critical Underflow Split......................................................... 11 7 8.5 Bitumen Loss to Underflow in TPOT Hydrocyclone........ 120 8.6 Final Recovery of Bitumen................................................... 120 9 SCALE UP OF THE TPOT HYDROCYCLONE ............................. 121 9.1 Determination of Diameter of TPOT Hydrocyclone ........ 121 9.2 Determination of Height of Overflow above the Middling Pipe......................................................................... 125 9.3 Diameter of Middling Pipe. ..... ................... ...... ... .......... ... .... 125 9.4 Determination of Diameter of Overflow Pipe.................... 125 1 0 SUMMARY AND CONCLUSiONS..................................................... 126 10.1 Recommendations..... .... ..... ...... .... .... .......... ..... ...... .... ....... .... 128 vii APPENDICES: A NOMENCLATURE ..................................................................................... 130 B EXPERIMENTAL DATA FOR TPOT HYDROCYCLONE.................................................................................... 135 -C DATA FROM BENCH SCALE THICKENING lESTS ........................................................................................................... 212 o FIGURES FOR FABRICATOR TO CONSTRUCT- 30-INCH TPOT HYDROCYCLONE.................................................. 227 E PI DIAGRAM OF BOILER AND CIRCUIT DIAGRAM FOR FLAME CONTROLLER ............................................ 244 F PICTURES OF THE TPOT HYDROCYCLONE AND BENCH SCALE EQUIPMENT USED FOR THICKENING TESTS .............................................................................. 247 REFERENCES ........................................................................................................ 250 viii LIST OF TABLES 2.1 World reserves of in-place bitumen........ ......... ....... ...... ... ...... ..... ... 5 2.2 U.S. reserves of in-place bitumen .................................................. 6 2.3 Extent of major Utah tar sand deposits .......................................... 8 2.4 Estimate of tar sand resources, Uinta Basin, northeast Utah .... 9 2.5 Estimate of tar sands resources, central-southeast region, Utah ...................................................................................................... 9 2.6 Fractional composition of Utah tar sand bitumens... ........... .... .... 11 2.7 Elemental analysis of extracted bitumens 'from selected Utah tar sands.... ... ... .......... ....... ....... ... ........ ............. ... ................................. 11 2.8 Physical properties of extracted bitumens from Utah tar sands ................................................................................................... 12 2.9 Chemical properties of extracted bitumen from Utah tar sands ................................................................................................... 13 5.1 Tangential velocity and radial acceleration at different distancesfrom the center of 11-inch TPOT hydrocyclone at 75 GPM (at port 1) and radial acceleration for two-product conventional hydrocyclone ....................................... 42 5.2 Analysis of the feed tar sand ................ : .......................................... 47 5.3 Critical volume split to underflow as function of various feed concentrations................... ................................................................. 71 5.4 Summary of variables and constants for Figures 5.7 to 5.27........................................................................................... 89 6.1 Polymers selected after screening ................................................. 95 6.2 Summary of results of thickening tests for solid concentration achieved .............................................................................................. 97 6.3 Calculation of unit area .................................................................... 100 7.1 Particle size distribution of underflow for experiment 3.............. 1 09 B.1. Summary of variables and constants in the experiment sets... 136 x LIST OF FIGURES 2.1 Location of major tar sands deposits in Utah ............................. 7 2.2 Effect of temperature on the viscosity of selected Utah bitumens... ... ......... ...................... ................ ... .................•.... .... ... ... ...... 14 2.3 A cO,mparison of viscosity of several Utah tar sands bitumens with Athabasca bitumen......... ............ ......... ... .... .... .... ...................... 15 2.4 Phase disengagement process ...................................................... 19 2.5 Canadian hot water process .... .... ... ....... ..... .......... ......... ..... .... ....... 20 2.6 Hot water process proposed in 1978 for Utah tar sand .............. 23 2.7 Schematic of moderate-temperature bitumen recovery process ................................................................................................ 25 3.1 Newly proposed flow sheet for hot water process for Utah tar sands ................................................................................................... 28 3.2 Terminal velocity plot ........................................................................ 29 5. 1 Tangential velocity and radial acceleration as a function of distance from center of 11-inch TPOT hydrocyclone.................. 43 5.2 Process flow diagram of tar sand pilot plant. ................................ 44 5.3 PI diagram of TPOT hydrocyclone ........ ~ .................... -..................... 45 5.4 PI diagram of steam header, tank, condensate pump and condensate tank ...................................................................... ~. 46 5.5 Particle size distribution of the feed tar sand ................................ 48 5.6 Dean - Stark unit ............................................................................... 51 5.7 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 3/4 inch) ......................................................................... 58 5.8 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 5/8 inch)......................................................................... 59 5.9 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 1/2 inch) ......................................................................... 60 5.10 Effect of air on percentage of feed bitumen reporting to the overflow and underflow at reduced feed flow rate (underflow diameter = 3/4 inch ) ................................ ~ ................... 61 5.11 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 3/4 inch, temperature = 140 deg F) ............................................................... 62 5.12 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 3/4 inch, temperature = 125 deg F) .............................................................. 63 5.13 Effect of underflow volume split on bitumen loss and sand recovery at underflow ....................................................................... 66 5.14 Effect of the feed flow rate on underflow solids concentration (wt.%), underflow flow rate, percentage of feed bitumen loss and sand recovery to underflow (underflow diameter = 3/4 inch) ..... ~ ................................................ 67 5.15 Effect of the feed flow rate on underflow solids concentration (wt.%), underflow flow rate, percentage of feed bitumen loss and sand recovery to underflow (underflow diameter = 5/8 inch) ...................................................... 68 5. 16 Effect of the feed flow rate on underflow solids concentration (wt.%), underflow flow rate, percentage of feed bitumen loss and sand recovery to underflow (underflow diameter = 1/2 inch)...................................................... 6 9 5. 17 Effect of feed concentration on critical volume split to underflow ............................................................................................ 72 5.18 Weir flow .............................................................................................. 73 5.19 Effect of underflow flow rate on percentage particle size reporting to underflow...................................................................... 76 xii 5.20 Effect of volume flow rate to underflow on 050 ............................. 77 5.21 Effect of feed flow rate on percentage particle size reporting to underflow (underflow diameter = 3/4 inch) .................................. 79 5.22 Effect of feed flow rate on 050 ( underflow diameter = 3/4 inch) ............................................................................................... 80 5.23 Effect of feed flow rate on percentage particle size reporting to underllow (underflow diameter = 1/2 inch) ................................... 81 5.24 Effect of feed flow rate on 050 ( underllow diameter = 1/2 inch) ............................................................................................... 82 5.25 Effect of sand concentration in feed on percentage particle size reporting to underllow (underflow diameter = 5/8 inch) ............ 84 5.26 Effect of sand concentration in feed on 050 .................................. 85 5.27 Effect of the air flow rate on percentage particle size reporting to underflow (underflow diameter = 3/4 inch) ................................... 87 6.1 Schematic of bench scale equipment for thickening tests ......... 93 6.2 Effect of feed concentration on the underflow concentration in the bench scale thickening tests .................................................... 98 6.3 A plot of height versus time for bench scale test no. 6.. .............. 102 6.4 A plot of height versus time on log-log scale for bench scale test no. 6 .............................................................................................. 103 6.5 Effect of interface solids concentration on interface settling velocity ................................................................................................. 1 ~4 6.6 Effect of underflow concentration on unit area ............................. 106 6.7 Effect of underllow concentration (wt.%) on unit area ................ 1 07 xiii ACKNOWLEDGMENTS In our Indian tradition there is a universal saying that "no knowledge is complete without a teacher and nobody can repay the debt of a teacher." In that spirit I would like to express my sincere thanks and gratitude to my supervisor, Professor Donald A. Dahlstrom, for his invaluable personal guidance, technical advice, unstinted support, encouragement, and above all the love and affection that I received from him in abundance, which provided me deep motivation and gave me indefatigable energy to complete this thesis research. I also acknowledge with a sense of humility the suggestions and guidance from members of my supervisory committee: Professor J.D. Seader, Professor Noel de Nevers, Professor A.L. Tyler, and Professor J.D. Miller. I would always cherish my association with them with nostalgia throughout my career wherever I may be. I also sincerely thank my colleagues in the Department of Chemical and Fuels Engineering and Dee Valentine and Lindy Charlatan for helping in the construction and running of the pilot plant and achieving the desired results from the experiments My special thanks are also addressed to my wife, Kamini, and little daughter, Prakriti, who have gracefully suffered with perseverance and loneliness during my long hours of absence from home to work on the pilot plant. Above all, I express my sincere thanks for the emotional support which I received from my parents, Mr. K.D. Khazanchi, my father, and Mrs. Lajya Wati, my mother, not only during this period but throughout my whole life. I will not be able to repay their debt. I feel overwhelmed that they have specially flown from India to be with me on this occasion, and I dedicate this thesis research to them. xv CHAPTER 1 INTRODUCTION This research project concentrates on the development .of a process for the extraction of bitumen by applying hot water technology for the separation of bitumen from the Utah tar sands. 1.1 The U.S. Scenario The U.S. by virtue of its being the most advanced country and the largest consumer of petroleum, accounting for about 300/0 of world consumption of crude oil, needs to develop new petroleum substitutes because of its decrease in petroleum production (predicted to be about 40% of its crude oil consumption by the year 2000). Alternative sources that can be usefully tapped in the U.S. may come from tar sands, oil shales, etc. There are very promising reserves of bitumen in the U.S. and some other countries. The extracted bitumen from these sources can be subsequently upgraded and treated in a \ conventional petroleum refinery to produce aviation fuel, gasoline, and other petroleum products. Accordingly, the University of Utah in collaboration with the U.S. Department of Energy has undertaken pioneering research to extract bitumen from the Utah tar sands. Efforts have been made to design and develop the most suitable, economical, and viable technology for this production. 2 1.2 The Current Project Hot water technology is being already applied and commercialized in the Canadian tar sands in the Athabasca region. Canada already has two commercial operating plants in this field, viz., Syncrude and Suncor, providing about 100/0 of the oil requirements of Canada. Tar sands, which are also referred to as "heavy oil sands" or "bituminous sands, II are a mixture of bitumen, sand, and fine solids down to less than 1 Jlm. Extraction of oil from tar sands can be a substitute for at least part of the U.S. oil imports. Depending upon the location of these reserves, Utah tar sands can contain up to about 130/0 bitumen.1 However, Utah tar sands bitumen content is appreciably lower than the Canadian tar sands, which makes mining and processing more difficult from an economic standpoint. Since bitumen is not in a liquid form in tar sands, technology for extracting bitumen from it may be costly as the tar sands must be mined and tailings must be disposed of after extraction. Therefore, it is essential that every effort be made to reduce costs of processing tar sands here in the U.S. Because over 90% of the U.S. tar sands are located in Utah and because they are lower in bitumen grade, as low as 60/0 as compared to Canadian tar sands that average 12 % bitumen, it is of the utmost importance to minimize the cost of mining and extraction to realize an economic and commercially viable product. The key to this plan is the reduction of the cost of tar sands processing and final disposal of the tailings. Utah is an arid land, with a state average of only 11 inches of precipitation per year (even less in the tar sands area). Water is at a premium. It is therefore imperative that the process to be developed should maximize recovery of water. This obviously means that the conventional 3 tailings ponds used in Canada, as reported in literature,2,3 cannot be employed in Utah because water loss by percolation and evaporation would be too high. Further it would be commercially exorbitant to design a tailing pond to meet EPA requirements. Therefore a method for tailings disposal must be developed whereby the solids in the tailings would be dewatered to higher concentration, which would permit transport by conveyors or high pressure pumping. A high­density thickener (a new development) could be applied here to maximize water recovery and minimize tailings disposal cost. This method would involve separation of solids at about 65 mesh, thereby creating two streams containing essentially solids coarser than 65 mesh and fine solids finer than 65 mesh. Streams containing much of the thickened +65 mesh coarse solids would be dispatched to spiral classifiers for maximum final dewatering. A high-density thickener, different from a conventional thickener, would be used to concentrate solids in the -65 mesh streams, hence enabling recovery of a maximum amount of water. All tailings solids would be finally disposed of in the old mining area where they will not rewet if placed properly. Thus the tailings pond would be eliminated. To achieve all the above mentioned objectives it is important to develop a process that can handle a large tonnage of tar sands at a minimum cost per unit of bitumen produced. The assumption and perception are that the process must recover 90+0/0 of the bitumen and at the same time be able to recover and recycle the maximum amount of water in the process. This approach may be cost effective and can result in an economically viable means of recovering bitumen from Utah tar sands. CHAPTER 2 LITERATURE SURVEY 2.1 Definition of Tar Sands The organic constituent of tar sand is bitumen, a very high molecular weight asphaltic solid which cannot be recovered by ordinary petroleum production methods.1 Tar sands contain bitumen in semisolid form at ambient temperature.4-10 The origin of the bitumen is not known. 11 The origin of tar sands is controversial as to whether they are the residue of conventional petroleum that migrated from a shale-like rock or whether organic debris was deposited in the sand and had matured over a period of time. However, whatever may have been the origin of formation of these tar sands, the fact remains that these are definitely "oil sands" or "bitumen sands" and are a potential source for future petroleum. 2.2 Geographical Distribution of Tar Sand Deposits Tar sand deposits are widely distributed throughout the world, with major deposits found in Canada, Venezuela, Africa, Europe, and the U.S. The world­wide distribution of tar sands is given in Table 2.1. In the U.S., 24 states are known to have occurrences of tar sands, but only 6 states have well-defined deposits that are considered to have hydrocarbon properties and are conceived to be .fit for commercial Table 2.1 World reserves of in-place bitumen 13 Deposits In-Place Bitumen (billion bbl) Canadian tar sands (Athabasca) Utah tar sands Other U.S. deposits (principally California Kentucky, and New Mexico) Venezuela Africa Europe 900 25 3 700 2 3 5 exploitation.12 These states and their known reserves are listed in Table 2.2. In the state of Utah, 55 tar sand deposits have been identified by H.R. Ritzma of the Utah Geological and Mineral Survey.1 Twenty-five of these deposits are located in the Uinta Basin in the northeast part of Utah, 24 lie in the central southeast portion of the state, and 6 lie in other areas. Although the total bitumen of these deposits accounts for over 95% of the total U-.S. bitumen Table 2.2 U.S. reserves of in-place bitumen13 State Alabama California Kentucky New Mexico Texas Utah In-Place Bitumen (billion barrels) 1.180 0.323 0.037 0.157 0.141 29.5 6 reserves, only a few of these deposits have been considered to be of commercial significance.1 Figure 2.1 illustrates the distribution of tar sands in the state of Utah. Table 2.3 lists these major deposits and their location. The tar-sand deposits in Utah have been mapped in a systematic manner since 1967. Based on these mapping studies and the sampling of the tar sands, estimates have been made of the quantity of bitumen (or oil) in each of the deposits. The criteria used is the same used to make estimates GREAT SALT LAKE N t BEAR LAKE ®S L C ASPHALT WHiTEROCKS .... RIDGE UTAHf\ ~ LAKEr! HILL CREEK . fit SUNNYSIDE " ~ P.R. SPRING LAKE POWELL Figure 2.1 Location of major tar sand deposits in Utah7 7 Table 2.3 Extent of major Utah tar sand deposits 1 In-Place Bitumen Deposit Name (billion barrels) 12.0- 16.0 Tar Sand Triangle 4.0- 4.5 P.R. Spring 3.5- 4.0 Sunnyside 1.3 Circle Cliffs 1.2 Hill Creek 1.0 Asphalt Ridge 0.6- 1.5 White Rocks Location Average Bitumen (Wt.%) SE, Utah 5.0 NE, Utah 12.2 NE, Utah 9.0 SE, Utah NE, Utah NE, Utah 13.1 NE, Utah 8 of oil fields and other hydrocarbons. The unit of evaluation employed is the standard 42 U.S. gallon petroleum barrel. Tables 2.4 and 2.51,14,15,16 summarize the estimated resources of bitumen in these sands, based on the Utah Geological and Survey (UGMS) Map 47 (1979), but with incorporation of the revisions carried out on the basis of recent published and unpublished sources.11 2.3 Bitumen Properties: General. Physical. and Chemical As in the case of all other hydrocarbons, it is of the utmost importance to have information of the general, physical, and chemical properties of the 9 Table 2.4 11 Estimate of tar sand resources, Uinta Basin, northeast Utah Giant or other Resources estimate (millions of barrels) major deposits Inferred or Total Measured Indicated conjectured . (All Categories) Asphalt Ridge 435 438 175 1,048 Asphalt Ridge, NW 2 3 95-120 100-125 Hill Creek 350 480 330 1,160 PR spring 2,500 1,200 550-1,100 4,250 Sunnyside 1,800 2,200 1,200-1,850 5,200-5,850 Whiterocks 50 40 35-50 125-140 Total(from above): 5,137 4,361 2,385-3,625 11,883-13,123 Total of 20 other deposits 50 60 112-202 227-317 Total-Uinta Basin 5,187 4,426 2,497 -3,827 12,110-13,440 Table 2.5 Estimate of tar sands resources, central-southeast region, Utah 11 Giant or other Resources estimate(millions of barrels) major deposits Measured Indicated Inferred or Total Conjectured (All Categories) Circle Cliffs 707 430 170 1,307 San Rafael Swell 35 55 355-455 455-545 Tar Sand triangle 2,500 3,600 3,400-7,900 9,500-14,000 Total(from above): 3,242 4,085 3,925-8,525 11,252-15-852 Total of 10 other deposits 10 6 7-10 23-26 Total-central-SE region 3,252 4,091 3,932-8,535 11.275-15,878 10 bitumen with a view to selecting the most viable technology for the recovery, upgrading and refining processes. Any optimization or development of technology is di rectly related to these properties. The tar sands of the Uinta Basin have similar properties, which differ significantly from those of the central, southeast region of Utah and the Canadian tar sands. Typical general properties, physical properties, and chemical properties of Utah tar sands in the various regions are given in Tables 2.6,2.7,2.8, and 2.9 respectively. Figure 2.2 contains viscosity17,18 data for bitumen present in Utah tar sands. Figure 2.3 compares the viscosity19 of Utah bitumen with Canadian Athabasca bitumen. It will be noted that the former is significantly higher. 2.4 Selection of the Technology 2.4.1 Historical Background Traditionally, two types 11 of technologies have been adopted for recovery of hydrocarbon values from extensive tar sand deposits of North America. These are the in-situ or above-ground thermal production of bitumen or the hot water extraction of bitumen-derived liquid from surface mining of the deposit for onward processing of the mined ore. In any of these approaches it is essential that the bitumen be upgraded to provide a reasonably suitable feed stock for subsequent processing in a petroleum refinery. The in-situ thermal methods encompass steam soak, steam drive, in-situ corTibustion, RF heating, etc. All of these have been used to increase oil recovery of heavy · oils. The application of either of the approaches in this method is largely dependent on various parameters such as the viscosity and specific gravity of the bitumen, temperature, and pressure. Table 2.6 Fractional composition of Utah tar sand bitumens20 Analysis Whiterocks Asphalt Ridge Sunnyside Saturates, wt. % 29.2-30.2 32.4 24.9 Aromatics, wt.% 16.8-24.8 22.4 18.1 Resins, wt.% 40.1-43.1 37.6 30.0 Asphaltenes, wt. % 3.3 -4.1 7.3 23.7 Table 2.7 Elemental analysis of extracted bitumens from selected Utah tar sands21 Elemental Analysis Whiterocks Asphalt Ridge Sunnyside Carbon wt.% 86.40 84.50 84.90 Hydrogen, wt. % 11.50 11.00 10.30 Nitrogen, wt.% 1.11 1.10 0.90 Sulfur, wt. % 0.40 0.40 0.50 Oxygen, wt.% 0.50 3.00 3~40 Atomic HIC ratio 1.60 1.56 1.45 Molecular weight 500.00 490.00 588.00 ....L ....L Table 2.8 Physical properties of extracted bitumens from Utah tar sands11 Properties Tar Sand PR Spring PR Spring PR Spring Sunnyside Wh ite rocks Asphalt Triangle Rainbow I Rainbow II South Ridge Bitumen content, wt.% 4.50 14.10 8.50 6.50 8.50 8.00 10.90 Specific gravity 1.01 1.02 0.99 1.01 1.03 1.00 0.97 Gravity, API 8.60 7.80 11.80 8.80 5.50 10.30 14.40 Con radson carbon, wt.% 16.70 14.00 17.40 24.00 14.80 13.00 Ash, wt.% 0.20 3.30 1.40 1.90 2.40 0.80 0.04 Pour point, deg F 91.00 210.00 320.00 320.00 Viscosity, cps 42638(a) 8269(b) 2900(c) 7031 (c) 7373(f) 29245(d) 2015(e) Simulated distillation IBP, deg F 316.00 279.00 316.00 308.00 307.00 Volatility, wt. 0/0 34.40 39.90 22.80 14.30 32.40 22.10 IBP-400 deg F, wt.°la 0.70 1.30 0.50 0.70 0.90 0.90 400-650 deg F, wt. % 7.60 5.10 2.20 1.30 7.30 3.30 650-1000 deg F, wt.% 26.20 25.60 20.10 12.30 24.00 18.80 > 1000 deg F residue, wt. % 65.61 68.10 77.20 85.10 67.60 77.90 (a) Measured at 45 deg C(113 deg F) (b) Measured at 100 deg C(212 deg F) (c) Measured at 220 deg C(428 deg F) (d) Measured at 74 deg C(167 deg F) (e) Measured at 50 deg C(122 deg F) (f) Measured at 65 deg C(149 deg F) ...a-t\) Table 2.9 Chemical properties of extracted bitumen from Utah tar sands 11 Tar Sand PR Spring PR Spring PR Spring Sunnyside Whiterocks Asphalt Triangle Rainbow I Rainbow II South Ridge Elemental Analysis C, wt.% 84.30 84.70 81.41 81.70 83.30 85.00 85.20 H, wt.% 10.30 11.20 10.30 9.30 10.80 11.40 11.70 N, wt.% 0.40 1.30 1.40 1.40 0.70 1.30 1.00 S, wt.% 4.00 0.50 0.40 0.40 0.60 0.40 0.60 O,wt.% 1.00 1.80 6.30 7.20 4.40 1.60 1.10 Atomic HIC ratio 1.47 1.60 1.51 1 :36 1.56 1.61 1.65 Mr, glmol 571.00 702.00 1381.00 1561.00 1042.00 668.00 Gradient elution chromatography Saturates, wt.% 13.30 9.50 15.80 4.10 13.20 15.30 10.00 MNAIDNA oils, wt. % 9.70 10.20 3.50 5.30 21.00 8.50 11.40 PNA oils, wt. % 11 .70- 11.40 9.00 0.90 5.90 11.90 4.40 Soft resins, wt.% 25.90 13.90 5.80 4.00 13.90 16.70 18.40 Hard resins, wt. % 1.90 1.10 2.30 1.80 5.60 2.60 1.20 Polar resins, wt.% 3.50 2.00 3.60 1.10 1.70 2.70 3.70 asphaltenes, wt. % 30.60 31.30 35.90 55.70 29.80 31.20 39.90 Non-eluted asphaltenes, wt. % 3.50 20.60 24.10 27.10 8.90 11.10 11 .10 ....... U) 106r-------------------_ 105 I-- t! I .; .1 ;1/ / /I /f/ l· I ./ val· ./ /.J0 / . / / / • TDr 5Dnd TriDngle • P. R. Spring ~ Asphalt Ridge o Whiterocks 6 Sunnyside 102----I~ ----~I ----~1 ____~ I ____~ I ____L I __ _~ 2.7 2.8 2.9 3.0 3.1 3.2 111xl03 (11K) Figure 2.2 Effect of temperature on the viscosity of selected Utah bitumens17,18 14 10· G .e-n 0 Q. > -t- (/) 0 (J (/) > 101 0 Sunnyside ~ Asphalt Ridge <> P.R. Sprin 9 • Athabasca 0/ / ,// /ci /j; /0 / 0 o/ // / A// 6/ / d /0 /~o /~ 2.8 3.0 3.2 (1/T)X10 3 , °K-1 Figure 2.3 A comparison of viscosity of several Utah tar sands bitumens with Athabasca bitumen 19 15 16 The other method that has been applied involves surface mining. This involves surface mining of tar sands, transportation of the ore to the processing facility, separation of bitumen from sand substrate, disposal of spent sand, and up-grading of the resultant bitumen liquid. From surface mining, bitumen is recovered by the method of water-assisted separation, solvent extraction, solvent-assisted aqueous separation, or thermal processing .. However, these methods of surface recovery can be adopted only if they are cost effective. The disposal of spent sand poses a great environmental problem that has to be considered for all methods. 2.4.2 Canadian Tar Sands The main commercial thrust of Canadian tar sands was directed towards surface mining, recovery of bitumen using a water-assisted process, and research on recovery of hydrocarbon values of Athabasca tar sands of Canada, all initiated by Dr. K. A. Clark22 in the 1920s. Based on these studies, Dr. F. W. Camp23-26 and his co-workers suitably modified the Clark processing scheme during the course of their extensive research and were able to scale up these developmental studies for commercial exploitation. The result of this has been the establishment of commercial plants by Suncor and Syncrude of Canada,25,26,27 which are the only commercial plants in the Western Hemisphere for surface mining recovery of tar sands, providing about 100/0 of the total crude oil needs of Canada. 2.4.3 U.S. Tar Sands In the U.S. the method mostly used and actively supported by the Department of Energy has been the in-situ method of thermal-enhanced oil 17 recovery. The studies were conducted by Laramie Energy Technology Center and included laboratory combustion tube experiments and field tests involving both in-situ combustion and steam-derived techniques. Four technologies have been tested to recover hydrocarbon values of bitumen in the U.S. These are aqueous separation, solvent extraction, solvent­assisted aqueous separation, and thermal recovery. However, only modest efforts have been made for the application of these technologies. Due to the easy availability of petroleum crude at reasonable prices, most of these studies have now been abandoned. 2.4.4 Tar Sands Program of the University of Utah In the wake of the petroleum energy crisis of 1973, the University of Utah initiated a research and development program under the pioneering efforts of Professor Alex G. Oblad. The initial thrust focused on surface-mining recovery processes with greater emphasis on water-assisted and thermal technologies. The major objective of these studies was to develop the scientific and engineering base essential for commercial exploitation of Utah tar sands. However, there are a vast difference and variation in the properties of Canadian and Utah tar sands, the major difference being the viscosity of the two bitumens. In addition, Utah tar sand deposits are mostly consolidated, whereas the Canadian Tar sands are unconsolidated. Further, Utah tar sands have a lower bitumen content compared to Canadian tar sands. This vast difference in the composition and the properties of these two tar sands and the environmental differences in the two countries precluded the simple adoption of Canadian technology for recovery of bitumen from Utah tar sands. Consequently, it became essential to develop a commercially economical and 18 viable technology that should be distinctly applicable to Utah tar sands, taking into consideration the environmental conditions available and to meet the environmental regulations here in U.S. To meet these challenges the research team at the University of Utah has reached a stage where aqueous thermal recovery and subsequent processing and up-grading of bitumen have been undertaken with active support and funding from the Department of Energy, the State of Utah, and private industry. 2.4.5 Hot Water Process . As the name implies, the hot water process uses hot water for the recovery of bitumen from tar sands. The hot water decreases the bitumen viscosity and helps in the disengagement of the bitumen phase from the sand matrix when a high-shear force field is caused by stirring and adding certain chemicals in a digester. Figure 2.4 shows the phase disengagement28 process of sand and bitumen in the presence of chemical and mechanical forces. 2.4.6 Hot Water Process for Athabasca Tar Sands in Canada As shown in the flow sheet,26 Figure 2.5, the feed material is mixed with hot water, caustic (sodium hydroxide), and sodium carbonate to form pulp of 60-85% solids at a temperature of 82-930C in a rotating digestion drum. In this stage, phase disengagement of the bitumen from the tar sand takes place due to combined action of the drum motion, reaction of the caustic soda and sodium carbonate with the bitumen, and the sparging of steam into the drum. The system pH is maintained between 8.0 and 8.5 by addition of the alkali chemicals. .Chemical Forces , > Mechanical Forces ... _----- Figure 2.4 Phase disengagement process28 ~ <0 Tar Sand Feed Steam Water Digestion Condition Drum Gravity Separation Makeup water Viberating Screens Bitumen Middlings Froth Settler Naptha + t Air Flotation Cells Flotation Upgrading Centrifuges Bitumen Product Tailings .' .J ~ , . Figure 2.5 Canadian hot water process26 I\) o 21 After leaving the conditioning drum, the pulp is screened to remove large rocks and pebbles. The screened pulp (undersize) flows by gravity to the separation cell, which is a special high-torque thickener. The bitumen floats to the surface and flows over a weir to bitumen cleanup, while the sand settles down and is raked to the center for pump discharge to the tailings pond. A third stream, known as the middling stream, is removed by a peripheral bustle pipe, midway down the cell, with submerged orifices on the pipe bottom. The middling stream contains mostly water with suspended fine-tailings solids and bitumen. The middling stream is treated by conventional froth-flotation, if necessary, when bitumen recovery is less than 90% in the overflow. The final steps of the process involve further upgrading of the concentrate from the primary separation cell and the floatation cells, when necessary, by dilution with naphtha at a diluent ratio of 1 :0.6 and centrifugation to remove mineral matter and water. The final bitumen contains approximately 1 % by weight mineral matter and about 5% by weight water. Solids and liquid effluents from the extraction process are pumped to the tailings ponds at Suncor where the solids settle and water from the surface is reclaimed. Tailings ponds rise to a height of 350 feet before they are abandoned. At Syncrude, a 12-square-mile tailings pond is employed to settle out the fine solids. At both plants, availability of water is not a problem due to the Athabasca River in the area. In more recent work43 hydrocyclone-based process for rejecting solids from tar sand was carried out. Pilot tests demonstrated solids removal of greater than 90% while 85 to 92% of the bitumen was retained in the product stream. 22 2.4.7 Hot Water Process for Utah Tar Sand Unfortunately29 the hot water process as developed for the Athabasca tar sands cannot be applied directly to Utah tar sands due to differences in physical and chemical properties. In the case of high-grade Utah tar sands (Asphalt Ridge and P.R. Spring deposits) the sand grains are completely enveloped by the bitumen phase. This is in contrast with the Athabasca tar sands in which sand particles are reported to be largely separated from the bitumen phase by a film of water surrounding each sand particle.1,23,30,31 As a result , bitumen does not bond directly to the sand particles. Consequently, phase disengagement in Utah cannot be achieved easily. The absence of a water layer in Utah tar sands and strong bonding between sand and bitumen suggest that a more intense force field or some other agent is required in order to achieve phase disengagement. The hot water process designed for Utah tar sands involves high-temperature alkaline digestion in a high-shear force field as well as the possible use of a penetrating agent such as kerosene prior to digestion for breaking the bond between bitumen and solids. Another factor that affects the recovery of the bitumen is the viscosity. Viscosity of the Utah tar sands is about two orders of magnitude greater than the viscosity of the Athabasca bitumen. It has been found32,33-38 that addition of the penetrating agent, cited previously, also greatly facilitates the reduction of viscosity of the highly viscous Utah bitumen and hence increases the recovery of bitumen from tar sands. A flow sheet for Utah tar sand developed and tested in 1978 is shown in Figure 2.6.39,40 In contrast to the Canadian process flow sheet, Figure 2.5, after phase disengagement by digestion, the pulp is discharged directly into a CAUSTIC TAR SANDS ~~---SOLUTION + .1 DIGEsnON·1 (I D WATER l' I .. BITUMEN , (to refinery) ,SEPARATIONj AIR --+\ CELL / SCAVENGER ---, ,~ WATER (to refinery or ~ r.cycle) ~ I • TAILINGS Figure 2.6 Hot water process pn 'j)osed in 1978 for Utah tar sand39 N (J.) 24 standard flotation cell where the pulp is diluted with hot water to 11 % solids. Air bubble entrapment by the bitumen phase causes it to float to the top of the cell where it is removed. Sand particles settle to the bottom of the cell and are removed as tailings along with all the fine solids, down to less than 1 J.1m. These tailings, depending on the type of feed, are screened for removal of undigested feed particles or middling. These were then recycled back to. a digestion stage for further recovery of bitumen. However, the recovery of water from the tailings was not carried out and tailings were sent to a small tailings pond. In more recent work2,3 a new process was suggested. The impact of recycled tailings water from tailings pond was examined. The flow sheet for this process is shown in Figure 2.7. In all of the above mentioned processes, either for processing Canadian tar sand or Utah tar sand, a tailings pond is employed which causes water loss. rKerosene Na2C03 Tar sand pretreatment NaSP3010 --..,... .+.. +1 +, Digester (A) Bitumen concentrate Gravity separation cell (8) Tailings Air + Flotation cell (C) Recycle water Bitumen .-----4 •• concentrate Tailings Sedimentation tank (0) Figure 2.7 Schematic of moderate-temperature bitumen recovery process2 I\) 01 CHAPTER 3 RESEARCH OBJECTIVES AND METHOD TO ACHIEVE THEM 3.1 Research Objectives The basic and paramount objective of this research is to develop technology to recover the maximum amount of bitumen from Utah tar sands. As the production of petroleum within the U.S. is continuously decreasing, it is essential to develop new sources in this country. To this end this research is concentrated on developing an improved hot water process that will be more economical and meet the particular demands of the Utah environment, e.g., minimum water availability, and low-grade tar sands. Therefore, the general thrust of the research was as follows: 1. Water being at a premium, one of the thrusts of research should be the maximum recovery of water from the hot water process as applied to Utah tar sands. 2. Concern for water loss by percolation and evaporation calls for the elimination of the conventional tailings pond from the process. 3. For economic reasons, ore as well as tailings must be either conveyed or pumped as truck transport is too expensive. Canada employs very large tailings ponds since they have no area problem and the Athabasca River can easily furnish all water required. 27 4. Utah has low-grade tar sands (lesser bitumen content) compared to Canada. Therefore, the thrust of the research calls for recovering over 90% of the bitumen. In summary, the aim of this research is to develop a technology for extraction of bitumen from Utah tar sands which will pave the way for establishment of an economic and commercially viable pJant in the future to help meet the demand for petroleum and petroleum products. 3.2 Method to Achieve the above Mentioned Research Objectives 3.2.1 Newly Proposed Process for Utah Tar Sands A new hot water process is proposed for recovery of bitumen from Utah tar sands. The process to be applied is considered eminently suitable to Utah IS environmental conditions. The new modified hot water process is shown in Figure 3.1. To achieve the objective of maximum dewatering of the tailings, separation of the solids at about 65 mesh is necessary because all solids (coarse and fines) cannot be dewatered to the maximum content together in a thickener or in a spiral classi'fier. A separation of solids at about 65 mesh (212~Lm) is required since it is well known from thickener operations that +65 solids may settle in the floor of the thickener, thereby enormously increasing the torque requirements. All solids cannot be dewatered in a spiral classifier since it will need an enormous pool area because of low settling velocities of the fines. Further, 'fines will plug the pores of the coarse sand, thereby resisting the water drainage. Settling velocity plots are given in Figure 3.2. Settling velocities for solid particles of density 2.7 g/cm3 in water, at 70 0 F, and particle Tar Sand ~ Storage 1-2hr Crusher Penetrating Agent (Kerosene) Steam Condensate Q) ~ c .Q ~ i5 "0 CJ) 1ij CJ) J: Condensate & Makeup Water • Boiler Oversize to waste by conveyor Air Compressor Screens , Steam Dilution Water To Bitumen Clarifier To High Density Thickener Makeup Water + Sodium Hydroxide Bitumen Removed asWeirAow Bitumen Removed as Weir Flow from TPOT Hydrocyclone r-------,~ from a Special Clarifier Middlings Three­Product, Open-Top, Hydro cyclone Bitumen Removed as Weir Flow from a Special Clarifier r----------p,--~. ~. ~~ AA Recycled Process Water High Density Thickener Tailings To be Pumped to Final Disposal Figure 3.1 Newly proposed flow sheet for hot water process for Utah tar sands I\) (X) 10 -~- Equivalent standard Tyler screen mesh Theoretical screen mesh Il) 0 0 Il) 0 0 N N ~ ~ N m", Il) Il) It') --0 -0 ------rNt)ilN)'-" -- co ~ ~ rt) N - - rti N ' 0 0 8 0 00 00 0 1l)CX) Il) ~~ 21! 0 T Q. Il) Il) \0 vl~ Q "1. N. N ~I~ Oil) Q \O~ rt) T 1 J 0 Il) N \0 N- :r I I I ~"" ~ L ",..,.. " ./~ 1""":..,... ~ ~ V..ofI' V" ~I'" L~ ~~ ~~ ~ ".... ~ fL 'tr J"// L~ ..... 1""'" ~ " I/ j ) L ~ j'fl/ 'iJ If rL: "~'" I~-,.'. ."" 1 I r ~ ~ t.l'L Jf 'L I [j 11'1.1' ./ ~ ,... L '{J :.J. ~ / / j~ lJ L / ~"" ilL ~ lL ~o .,Vj It '1/ ~/ v~ ~~ I/l~ 01.(. c; O~.L.L..L 0 I /I' A:..Ill. LL tl; I I I I ... AJ III J -, ~~~~"" r-r ~~~tr$~Q ~~ .~8~ ...... 0. 0 'V 0 ....... "="" ~ ~ 0 ·1::1 FI "'i .... 1 t--i ~,~ 'V~' ~~ r~v ....... f=~ 01 fII ....... r-ct ~~ v,(/).L 1 I ~ / / 11 '£ IJ r'" 1 'l' Ii , L_ '1 1/ IJ 'I I 11~ I/L I V~ ~ V11 ~ '11 L If II ~ V .L11L- L1L..1L..' l.L LL ...J. Notes LI I IJ I.J. J If I. Numbers on curves represent If/ 'I J Ii III 1/ Ii.. 'I true (not bulk or apparen t ) VL '( ~Ii rfj J If specific gravity of particles J 11 referred to water at 4 0 C. 2. Stokes-Cunningham corr- 1.. L ILL 1 ection factor is included for 1 1L L Ii. 1 1 fine particles settling in air. LL 1 L ILl 3. Physical properties used IJ 'j I I il Fluid Temp. Viscosity Density ...:L centi20ise I b./cu.ft. 1 1 1 L Air 70 0.0181 0.0749 "/1 . J / Water 70 0.981 62.3 lL ~ / I 111111111 I I 1111111 1 1111 ill tiL 1 , lL 111!1 Lill 11111 ] 10 100 1000 10,000 Particle diameter, J.lm Figure 3.2 Terminal velocity plot41 29 30 diameters ranging from 1 to 212J.1m fall in range of 3x10-6 to 1 x1 0-1 tVs whereas settling velocities for particles of diameter ranging from 212 to 1 0000 ~lm fall in the range of 1 x1 0-1 to 2 tVs. To increase the settling velocities of -65 mesh particles, they should be 'flocculated and treated in a high-density thickener for maximum dewatering. A three-product, open-top hydrocyclone (hereafter ref.erred to as TPOT hydrocyclone) was developed specifically for this purpose to make the bulk separation of solids at about 65 mesh. This unit was also tested (in a small diameter unit)42 on sand and plastic particles for liquid-solid separation potential. Other alternatives to the hydrocyclone were a two-product pressurized cyclone followed by a bitumen-water gravity separator and a conventional separator of the type being used in Canada. The former was tested on a small scale by Syncrude which employed a three-stage, hydrocyclone process with various recycle streams basically in a counter­current process of solids and water. Syncrude indicated that it needed further research.43 As Utah bitumen is usually close to a specific gravity of 1.0 (same as water) separation would be even more difficult. The conventional separator used in Canada (referred to in Figure 2.5) does not make a separation of coarse and fine solids at about 65 mesh size and therefore pumps the wide range of all solids to a tailings pond at a dilute solids concentration. In the proposed modified process, the tar sand is first crushed to minus 2 inches and the penetrating agent (e.g., kerosene) is added (if necessary) before sending the tar sand to storage. The primary aim of adding the penetrating agent5,7,8,32-38 is to decrease the viscosity of the bitumen and break the bond 31 between bitumen and sand. Second, the agent also greatly helps to unconsolidate 'the tar sand. After storage for 1-2 h, the tar sand is sent to a second crusher. Since the tar sand is already soaked with kerosene, the energy requirement at this crushing stage is drastically reduced. The tar sand is then sent to the digester where it is mixed with hot water and a small amount of alkali chemicals. The digester is a rotating tubular sloped r~actor with internal lifters to create the required force field to separate bitumen from the surface of the sand. The digester is heated by applying direct steam counter--currently through a stationary valve at the discharge end. The residence time of the slurry in the digester should be such that at least 95-97% minimum of the bitumen is separated from the surface of the tar sand. The digester is similar to that employed in Canada and is expected to work just as well on Utah tar sands. The slurry from the digester is sent to a double-decker screen( 1/2 inch and 1/8 inch decks). The oversize 'from the two screens is combined with the other +65 mesh tailings and conveyed to final disposal in the old mined area. The screen underflow is pumped to the three-product, open-top hydrocyclone to begin the bitumen recovery and separate the solids into two streams at about 65 mesh. Since bitumen has a definite affinity to attach to air,44,45 air is also injected into the feed line of TPOT hydrocyclone to reduce the specific gravity of the bitumen. An air compressor is used to pump the air through a small-diameter orifice to improve contact with the bitumen. The TPOT hydrocyclone yields three streams: 1. Overflow bitumen rich stream 2. Middling stream containing mostly fine sand, water, and some bitumen 3. Underflow stream containing mostly coarser sand (65 mesh) and \ 32 water with some bitumen. The overflow stream and middling streams are sent to separate clarifiers. These clarifiers had been developed in the last two decades and can produce a weir flow of concentrated bitumen with some water and fine solids and a second overflow containing most of the fine solids and water. These two streams are concentrated in the high-density thickener. The underflow f~om the clarifiers is then returned to a spiral classifier, where the +65 mesh solids report to the spiral product whereas the classifier overflow contains primarily the -65 mesh. The overflow from the spiral classifier is sent to the high-density thickener. The solids are flocculated with a polyacrylamide type flocculant. The solids concentration upon flocculation is crucial and must first be determined in order to concentrate to a very high underflow solids content. Underflow from the high­density thickener is the -65 mesh dense tailings (60-65 wt.% solids or higher) to be pumped for final disposal in the old mined area. The overflow from the high-density thickener is relatively clear water, which can be recycled back to the process for reuse. The underflow should be so concentrated that it does not seek its own level but instead exhibits a IIslump.1I The underflow from the TPOT hydrocyclone is also sent to a spiral classifier where an appropriate amount of diluent water is added prior to the feed point, as required. Three streams are produced in the spiral classifier. The +65 mesh stream from the spiral discharge is sent to the old mined areas by a belt conveyer. The overflow stream containing fine solids (-65 mesh) is also sent to the high-density thickener. Bitumen from the spiral classifier is removed as weir flow. 33 Specifically, research was carried out to study the application of a new TPOT hydrocyclone to hot water process, bench scale thickening tests on the fine tailings, and bench scale tests on the spiral classifier. CHAPTER 4 SOLID-LIQUID SEPARATION THEORY AND TPOT HYDROCYCLONE 4.1 Settling of Solids 4.1.1 Gravity Settling of Solids When a particle of diameter Dp and density Pp settles in stagnant fluid it experiences the gravitational force acting downwards and buoyant and drag forces acting upwards. These forces are given by the following expressions: Gravitational force (F g) = Buoyant force (Fb) = wt. of fluid displaced = -611 tDP3 P P g (4.1) Drag force (Fd) =.! V 2 1tD/ C (4.3) 2 PI s 4 0 where Co is given by the following expressions46: -Creeping flow regime ( Stokes law, Rep < 0.1) C = 24 o Re P -Intermediate regime (0.1 < Rep < 500 ) (4.4) C _ 18.5 o - Re 0.6 p - Turbulent regime ( Newton's law 2x1 05 >Rep > 500 ) Co = 0.44 (4.5) (4.6) Rep is particle Reynolds nurrlber and is given by following expression (4.7) 35 At the terminal velocity of the particle when it is no more accelerating, the following equation must be satisfied (4.8) Substituting the expression for various forces in .above equation, the following expression was obtained: (4.9) The above expression can be solved for the settling velocity V s: 36 (4.10) 4.1.2 Settling of Solids in Swirling Flow In the swirling motion which is typical of the hydrocyclone all three components Vz, Vr and Vt exist. In a swirling motion, depef'ding on geometry and density of the particle, the particle will lag the fluid. The slip velocities, Vsr, V sz. and V st, of a particle in a swirling motion are given by the following expressions. 4.1.2.1 Radial Slip Velocity (Vsr) In a swirling motion a particle will experience a radial force acting outwards due to the acceleration generated by the tangential velocity. The radial acceleration is given by following expression: V2 a =_t r r (4.11 ) Substituting ar in equation 4.10 following expression was obtained for the radial slip velocity: Vsr = (4.12) 37 4.1.2.2 Vertical Slip Velocity (V sz) In the vertical direction the particle experiences the downward gravitational force opposed by the upward buoyant force and vertical drag force. Equation 4.10 applies to determine the vertical slip velocity. 4.1.2.3 Tangential Slip Velocity (Vst) Since no significant force acts in the tangential directron, the particle will move a.long with the fluid in that direction. 4.2 Free and Forced Vortex A perfect 'free-vortex is a potential flow, and there is complete conservation of angular momentum. Tangential velocity in perfect free-vortex by definition obeys the following relationship: V tr = constant (4.13) and there is inward radial velocity whose magnitude decreases with decreasing radius. In case of a two-product conventional hydrocyclone a similar kind of relationship exists as in equation 4.13 for tangential velocity component which is as follows: V tr" ::! constant (4.14) where n normally has values between 0.5 to 1. 38 On the other hand in case of forced-vortex the following relationship exists: V trn= constant (4.15) In perfect forced-vortex the fluid moves as a solid body and the radial component of velocity is zero. For perfect forced-vortex the value of n in equation 4.15 is 1. In a swirling motion existence of forced- and free-vortex may depend on the other two velocity components viz. radial and axial components of velocities. 4.3 Definition of D.5jl In the context of the TPOT hydrocyclone Dso is defined as the particle size which has equal probability of reporting to underflow, and overflow and middling combined. 4.4 Pressure Drop in a Pipe Pressure drop through a pipe of length L is given by following expression (4.16) where f is friction factor and is given by the following equations: Re < 2.1x10-3 (4.17) f = O.O~91 2.1x103 < Re < 1x10s for smooth pipes (4.18) Re~iPe 4.5 Viscosity of the Slurry Viscosity of the slurry was calculated by the following correlation:47 Jlm = 1.0+2.5 </)y +10.05 cp/ + (0.00273)exp(16.6 </)y) (4.19) J.Lo 4.6 Density of the Slurry Density of the slurry was calculated using the following equation: (4.20) 4.7 TPOT Hydrocyclone 39 Detailed geometry of the 30-inch TPOT hydrocyclone is given in Appendix B. The feed to the TPOT hydrocyclone enters tangential creating the swirling motion. However, in the upper part of the TPOT hydrocyclone, above the middling pipe, two vortex breakers are installed. The vortex breakers are cross-shaped having a height of 6-inch each. These vortex-breakers break the vortex in the upper section of the TPOT hydrocyclone. Only the axial component of the velocity exists in the upper section above the vortex breakers. In the upper section of the TPOT hydrocyclone, settling is by gravity against the upward current of the fluid, and in the lower section the particles experience an additional radial force generated due to the swirling motion. The TPOT hydrocyclone differs in geometry from the two-product, conventional hydrocyclone. In contrast to the two-product hydrocyclone, there is no upper wall in the TPOT hydrocyclone as a result of which fluid moves up and a third product can be withdrawn from the upper part of the TPOT hydrocyclone. CHAPTER 5 STUDY OF THE TPOT HYDROCYCLONE The experimental studies conducted on the new TPOT hydrocyclone in a pilot plant are presented in this chapter. 5.1 Flow Pattern in TPOT Hydrocyclone Before running pilot-plant tests on bitumen, an 11-inch diameter "plastic model" of the TPOT hydrocyclone was constructed to determine the type of flo 'w pattern which is necessary for understanding the classification principle of the separation. This plastic unit was only tested with sand and water. The flow pattern was determined using a pitot tube.42 Although a pitot tube does not yield the precise flow pattern for the three­dimensional flow typical of this type hydrocyclone, it did provide reasonable estimates of the tangential velocity within the unit, which is the predominant velocity component. This was observed in the plastic model and confirmed the existence of a forced-vortex pattern rather than the free-vortex pattern in the conventional two-product, pressurized hydrocyclone. Following is the expression48 for radial acceleration for conventional, two-product hydrocyclone where a. = 0.45, n = 0.8. Vinlet is inlet velocity and Rc is radius of hydrocyclone: (5.1) 41 Table 5.1 compares the measured values of radial acceleration for TPOT hydrocyclone and calculated values of the radial acceleration for the two­product, conventional hydrocyclone at 75 GPM inlet flow rate with inlet velocity of 4.15 m/s. It can be concluded from Table 5.1 that TPOT hydrocyclone is a relatively low centri'fugal force unit that is more appropriately applied in the processing of tar sands because a solid separation at 65 mesh is required. Figure 5.1 plots the tangential velocity and radial acceleration for the TPOT hydrocyclone at 75 GPM inlet flow rate as a function of distance from the center. Based on the data obtained and the results achieved with the plastic model, the next step was to construct and test the TPOT hydrocyclone in a pilot plant. This resulted in the construction of a 30-inch diameter TPOT hydrocyclone to carry out the tests with a slurry containing bitumen, water, and sand from Utah tar sands. 5.1.1 Experimental Procedure to Test the 3Q-lnch Diameter Hydrocyclone at the Pilot Plant with a Slurry of Bitumen, Water, and Sand Detailed dimensions of the 30-inch TPOT hydrocyclone are shown in Appendix C. A process flow diagram of the equipment used to test the TPOT hydrocyclone is shown in Figure 5.2 and PI diagrams are given in Figures 5.3 and 5.4. Tar sand was digested in tank T-1. The particle-size distribution of the digested tar sand is shown in Table 5.2, and Figure 5.5 is a cumulative plot of particle size distribution. An agitator was employed to create a shear force field to facilitate the detachment of bitumen from the solids. A slurry of solids was 42 Table 5.1 Tangential velocity and radial acceleration at different distances from the center of 11-inch diameter TPOT hydrocyclone at 75 GPM (at port 1 )42 and radial acceleration for two-product conventional hydrocyclone Distance From Tangential Vel. "Radial Accel. Center (mm) Vt(TPOT) (m/s) ar(TPOT) (m/s2) (m/s2) 0 0 0 10.92 0 0 21.08 0 0 3409.5 30.99 0.2256 1.642 1251.9 40.89 0.3170 2.458 608.9 51.05 0.3901 2.981 341.9 60.96 0.4511 3.338 215.6 71.12 0.5517 4.290 114.4 81.03 0.6370 5.008 102.9 90.93 0.8412 7.782 76.2 100.33 0.9540 9.091 58.3 110.99 1.1918 10.738 · 45.3 120.90 1.3106 14.217 36.3 * Radial acceleration for 11-inch two-product conventional hydrocyclone using equation 5.1 <> Tangentia.l velocity (m/s) b. Radial acceleration (m/s2) 1.4-----------------------------------.16 1.2 14 12 .-.. .-.. 1 N J! J! E E ~ ~... ~ 10 ... > ca ~ f).8 C 0 (,) '; 0 a; > 1i '; c CD C) c ca I-. c.a. 8 CD a; (,) (,) 0.6 <C :! 6 'tJ ca a: 0.4 4 0.2 2 O~~~~~~~~~~~~~~~~~~O o 20 40 60 80100 120 140 Distance From Center, r, (mm) Figure 5.1 Tangential velocity and radial acceleration as a function of distance from center of 11-inch TPOT hydrocyclone 43 P-2 T-3 Return Pump Return Tank T-2 Flow Measuring Tank Manual Flow Control Valve T-1 Digestion tank Overflow T-2 Scale P-2 TPOT 1 P-3 T-4 Three Product Condensate Condensate Open Top Pump pump surge Hydrocyclone tank ~I T-5 T-4 T-1 Figure 5.2 Process flow diagram of tar sand pilot plant T-5 P-4 B-1 C-1 Condensate Condensate Boiler Air Compressor Tank Pump L '\;-4 ~I B-1 Steam 1 .5 mm Converging Nozzle Air ~ ~ Manual Control On-Off Valve Middling On-Off Valve Overflow 2" Feed 2" Tangential Entry Manual Control Valve Feed By- Pass to Tank 2" 45 On-Off Valve Undeflow Slurry Pump From Tank 2" Figure 5.3 PI diagram of TPOT hydrocyclone s-......... ___ On-Off Valve Low Condensate ---~Pump 46 Pressure Reduction Valve 2" Steam Header On-Off Valve Condensate to Condensate pump surge tank Condensate Tank To Atmosphere Wrap around heat exchanger ight Glass .. Condensate to boiler Figure 5.4 PI diagram for steam header, tank, condensate pump and condensate tank 47 Table 5.2 Analysis of the feed tar sand Wt. % Bitumen = 12.76 Wt. % Moisture = 1.44 Wt.% Sand = 87.24 Particle Size Distribution Mesh Size Particle Size Cum. wt.% Mesh Range Average Particle 0/0 (mm) Finer Than Size(mm) bywt. 14 1.168 98.487 14+ 1.3340 1.514 20 0.850 98.158 14x20 1.0090 0.329 24 0.707 97.817 20x24 0.n85 0.341 28 0.600 96.146 24x28 0.6535 1.671 32 0.500 92.583 28x32 0.5500 3.563 35 0.425 84.922 32x35 0.4625 7.661 50 0.295 52.549 35x48 0.3600 32.373 65 0.212 29.646 48x65 0.2535 22.903 100 0.150 15.889 65x100 0.1810 13.757 150 0.106 11.261 100x140 0.1280 4.628 200 0.075 7.000 140x200 0.0905 4.261 270 0.053 5.074 200x270 0.0640 1.926 325 0.043 5.051 270x325 0.0480 • 0.023 400 0.038 4.044 325x400 0.0405 1.007 400- 0.0190 4.044 48 <> Cumulative % (0/0 finer than) 100,--------------------------------- 90 80 - 70 c as ..J...:.!. cCI) 60 ;;:: ::!!. -0 t/. 50 i CI) > ~ 40 :; E :::s 0 30 20 10 o~--~~~~--~~~~~~~~~~ ·0.01 0.1 1 10 Average Particle Size(mm) Figure 5.5 Particle size distribution of the feed tar sand .. 49 prepared in tank T-1 to a temperature of about 1700F with the use of steam provided by boiler 8-1 to the jacketed heat exchanger of tank T -1. Feed slurry was pumped to the TPOT hydrocylone using pump P-1. Compressor C-1 was used to produce compressed air which was injected in the feed pipe through a nozzle of 1.5 mm diameter. Feed entered the TPOT hydrocylone tangential, and, due to the centrifugal force action, coarse solids moved towards the wall and, due to gravity, downwards and reported primarily at the underflow. The fine solids reported primarily to the overilow and middling. Since the bitumen has a de'finite affinity to attach to fine bubbles of air, bitumen tends to move towards the overflow and middling streams. The pilot plant was constructed to determine the bitumen separation and recovery efficiency of the TPOT hydrocyclone. Therefore all three streams, viz., overflow, middling, and underflow, were continuously recycled back to tank T -1. New tar sands were periodically added in tank T-1 to replace old tar sand; thus tar sand quality was maintained. Condensate from the jacket of tank T-1 was pumped back to the condensate tank T-5, using condensate pump P-3, and finally sent back to boiler 8-1 whenever need for water arose in the boiler. Overflow, middling, and underflow were diverted to the flow measuring tank T-2 to measure their flow rate. Flow measuring tank rested on a scale. The flow was diverted to the scale using flexible pipes for a timed pe!iod. From the initial and final readings of the weigh scale and time, flow rates of all three streams were thus calculated. After measuring the flow rate, slurry from flow measuring tank T -2 was transferred to tank T -3 by gravity where an agitator 50 was employed to keep the slurry well mixed. The slurry was then pumped back to the main tank T -1 using pump P-2. Samples of all three streams were taken and analyzed for bitumen, water, and sand content to determine the composition of overflow, middling, and underflow streams. From the composition and flow rates of the individual streams, flow rates of bitumen, water, and sand were determined for all three product streams, and by material balance the feed stream was determined. Tables containing results for different operating conditions are in Appendix A. 5.1.2 Analysis of Samples Containing Bitumen. Water. and Sand Several Dean-Stark20 units were used to analyze the samples containing bitumen, water, and sand. Figure 5.6 shows the Dean-Stark unit. The apparatus consists of a glass extraction flask, a trap, and a reflux condenser, all connected by means of glass joints. The flask has four indentations spaced evenly about the neck of flask to support a round-bottom extraction thimble containing the sample to be analyzed. The following procedure was adopted to analyze samples for bitumen, water, and sand. A sample, which may be overflow, middling, or underflow from a typical pilot plant run, was collected in a sample bottle and then allowed to settle for about 48 h before starting its analysis. Free water from the sample bottle was decanted, and the bottle was weighed before and after decanting; the resultant difference in weight indicated decanted water. The flask of the Dean-Stark extraction unit was filled with solvent toluene to extract bitumen out of the sample. A cellulose thimble wrapped with fine filter paper was placed in the Water ~Condenser Indentations- ~ ........... Boilino Ch'ps Figure 5.6 Dean-Stark unit39 51 Flask Material: Pyrex Glass • 52 flask on indentations in order to block the passage of solids th rough it. The contents of the sample bottle were thereafter mixed with toluene and the same poured into the thimble using a funnel. The empty bottle was rewashed with toluene to remove every trace of bitumen, water, and sand from the sample bottle and poured into the thimble. The sample bottle was then placed under the hood so that it could dry completely. The Dean-Stark unit was thereafter secured with the condenser and water trap in place. The flask of the Dean-Stark unit rests on a hot plate. The hot plate was switched on and water to the condenser started to condense the vapors of water and toluene. Heating of the flask caused toluene to vaporize, and when vapors of toluene passed through the thimble, part of toluene vapors condensed and caused vaporization of water in the thimble. Bitumen, being a very high boiling point hydrocarbon, does not evaporate at these temperatures. Water and toluene vapors condensed, and the resultant condensate fell into the trap. Water, being denser and immiscible in toluene, formed a lower, separate layer in the trap and was removed from time to time. Toluene, being lighter, falls back into the flask through the thimble. Bitumen, being soluble in toluene, was extracted and passed through the thimble into the flask of the Dean-Stark unit, leaving behind only the sand in the thimble. The Dean-Stark unit was run until the time when no more water and bitumen were left in the thimble. This was insured by observing the trap for any increase in the water level and by closely observing the color of toluene falling til in the flask from the thimble. Colorless falling toluene from the thimble was an indication that all the bitumen from the sample poured into the thimble had been extracted and were in the flask. The heater was switched off, and the unit was allowed to cool down. All residual water from the trap was removed, and the 53 total amount of water thus collected was weighed. The thimble, which was drenched by toluene, was removed from the flask and placed to dry up until all toluene evaporated. Typically it took about 24-48 h to dry the thimble. This dried thimble contained only sand, and by weighing and subtracting the weight of the thimble the amount of sand in the sample was determined. Once the amount of sand and water was established, the amount of bitumen was calculated by the difference. 5.1.3 Exact Determination of the Amount of Bitumen in a Sample The above method of analysis for bitumen by difference is not fool proof, although it gives a sufficient indication of the amount of bitumen, water, and sand. Since the proportion of bitumen in the sample, especially in the case of the underflow sample, is much less than water and sand, even a small experimental error can greatly affect the calculated amount of the bitumen in the sample. Therefore it is essential that all experimental steps be carried out very carefully in order to keep good control. 5.2 Variables Studied for the TPOT Hydrocyclone Based on the above experimental procedures, studies were carried out under actual pilot plant operating conditions to determine the influence of the following variables: -Effect of air quantity on the recovery of bitumen to overflow, middling, and underflow. -Effect of temperature on recovery of bitumen to overflow, middling, and underflow II 54 -Effect of feed flow rate on 0 50 -Effect of underflow volume flow rate on 0 50 -Effect of feed concentration on 050 -Effect of air flow rate on 050 5.3 Data Collection and Analysis The following data sheet was used for the collection of data 5.3.1 Flow Rates The following data were collected to determine flow rates of overflow, middling, and underflow - Final reading of scale (Ibm) - Initial reading of scale (Ibm) -Time Flow rates for overflow, middling, and underflow were determined using the following formula. FI t (Ib I) Final reading of scale (Ibm) -Initial reading of scale (Ibm) owra ems = ' . () Time s 5.3.2 Collection of Data for the Determination of Compositions of a Sample • (5.2) The following data sheet was used for data collection to determine the composition of a typical sample of overflow, middling, or underflow. - Wt. of sample bottle + sample - Wt. of sample bottle after decanting water - Wt. of decanted water - Wt of Thimble wrapped with fine filter paper - Wt of dry sample bottle after putting sample into the thimble - Wt. of sample poured into the thimble - Wt. of water collected from Dean-Stark unit - Wt. of dry thimble + sand - Wt. of sand in the sample - Wt. of water in the sample 55 = C1 =C2 = C2-C1 ·=C3 =C4 = C1-C2-C4 =C5 =C6 = C6-C3 = C5 + C2-C1 -wt. of bitumen in the sample = (C1-C4)-(C6-C3)- (C5+C2-C1) The sand collected in a typical sample was wet screened to determine the particle size distribution. 5.3.3 Analysis of the Data Collected in the above Section • The above data were used to carry out the following analysis to determine different parameters for all streams. Composition and flow rate of the feed stream were back calculated knowing the flow rate and composition of all 56 output streams, viz., overflow, middling, and underflow from the TPOT hydrocyclone. The following parameters were calculated for the three different streams: 1. Volumetric flow rate ( m3/s) 2. Flow rate of sand (gls ) 3. Flow rate of water (gls) 4. Flow rate of bitumen (gls) 5. WI. bitumenIWt. water (gig) 6. Sand wt.% 7. Water wt.% 8. Bitumen wt.% 9. WI. % of feed sand 10. Wt.% of feed water 11. WI. % of feed bitumen 12. Wt.% of different particle sizes reporting in different streams. Appendix A contains the information on the variables studied and the individual analysis for each stream for all the experiments with the TPOT hydrocyclone. 5.4 Results and Discussions 5.4.1 Study of the Effect of Air on Bitumen Recovery • As mentioned earlier, bitumen has a definite affinity to attach to air. Accordingly the effect of air on the recovery of bitumen to the three streams was 57 also studied. Air was injected into the TPOT hydrocyclone in the feed pipe well before the TPOT hydrocyclone. It was desirable to provide for enough length of the feed pipe before the entrance of the feed to the TPOT hydrocyclone so that enough time was available for air to attach to bitumen, thereby making it lighter than water. Tables in Appendix A summarize the effect of air injected into the TPOT hydrocyclone at various experimental conditions. The effect of air flow rate is plotted in Figures 5.7 through 5.10. 5.4.1.1 Observations As can be seen from these plots, the effect of air flow, although substantial in bitumen recovery, to the overflow points towards a negligible, or it has little effect on bitumen loss in the underflow stream. In the case of overflow, as is evident from the plots, bitumen recovery increased with the increase in the air flow. However it reached a plateau for a certain flow rate of the air, and any amount of further injection of air did not help in any increased bitumen recovery at the overflow. 5.4.2 Study of the Effect of Temperature on Bitumen Recovery Since bitumen viscosity increases exponentially with the decrease in the temperature (refer to Figure 2.2 ), the effect of temperature on bitumen recovery to overflow, middling, and loss to underflow was studied. Figure 5.11 and 5.12 .. plots the effect of air flow at temperatures of 140 OF and 125 OF respectively. <> % Water reporting to overflow A % Bitumen reporting to overflow o % Bitumen reporting to underflow !(II % Water reporting to underflow A 0.0002 0.0004 0.0006 ' 0.0008 0.001 0.0012 Air Flow Rate (Std. m3/s) Figure 5.7 Effect of air on % of feed bitumen reporting to the It overflow & underflow (underflow diameter = 3/4") 58 <> % Water reporting to overflow A. % Bitumen reporting to overflow o % Bitumen reporting to underflow -=- % Water reporting to underflow 35 ~ 0 ;: Q) 30 '1::J C :;:) oa ~ 25 ;: Q) > 0 .o.. 20 C) c t: &'15 Q) a:: ~ .Q..) ~ 10 oa c Q) .E:.::.J 5 iii ~0 0 0 ...- C\j (\') .q- LO (0 ...... co 0) ...- 0 0 0 0 0 0 0 0 0 0 00 00 00 00 00 00 00 00 0~ 0~ ci ci ci ci 0 ci 0 0 0 Air Flow Rate (Std. m3/s) .. Figure 5.8 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 5/8 inch) 59 <> % Water reporting to overflow 11 % Bitumen reporting to overflow a % Bitumen reporting to underflow <J. % Water reporting to underflow 45~----------------------------------------~ ~ o 't: 40 Q) 'C C :l 35 ~ ~ o E 30 ~ o S 25~ __ --~~----~~------~------~----__ ~~ C) c t: &. 20 aQ..:.) oS 15 ~ ~ 10 Q) E :1 =m= 5 '#. o ,.... o o o ci C\I (") ~ o 0 0 o 0 0 o 0 q ci ci 0 o . lO co r--- co en ,.... o 0 0 0 0 0 o 0 0 0 0 0 q 0 0 q q ci o ci ci 0 0 Air Flow Rate (Std. m3/s) Figure 5.9 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 1/2 inch) 60 • ~ o i: ~ % Water reporting to overflow A % Bitumen reporting to overflow o % Bitumen reporting to underflow 4- % Water reporting to underflow 60--~==========================~~ ~ 50~--------------~~~----~-- c __ A-----~ :::» ~ ~ t 40 CI) o> -o C) .5 30 1: o Q. CI) a..:. .! 20 ~ ~ c Q) -§ 10 m 'i/. O~~~TT~~~~~~~-r~~~~~~~T4 o 0.0001 ·0.0002 0.0003 ·0.00040.0005 0.0006 0.0007 Air Flow Rate (Std. m3/s) Figure 5.1 o Effect of air on percentage of feed bitumen reporting to the overflow and underflow at reduced feed flow rate (underflow diameter = 3/4 inch) 61 II ¢ % Water reporting to overflow 11 % Bitumen reporting to overflow o % Bitumen reporting to underflow .. % Water reporting to underflow ~45 __ ---=============================--~ o ;: ~ 40 c: :::) olJ 35 ~ o i 30 o> .2 25 C) c: ~ 20 Q. Q) a: -Q; 15 ~ olJ 10 c: Q) o § 5~---------~~----------~-----V-­!:: m ~ 04-__ ~~~-r~~~~~~~~~~~~-r~~ o 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 Air Flow Rate (Std. rn3/s) Figure 5.11 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 3/4 inch, temperature = 140 deg F) • 62 <> % Water reporting to overflow /l. % Bitumen reporting to overflow o % Bitumen reporting to underflow -=- % Water reporting to underflow 45~~==========================~--~ ~ o t: 40 CI) 'C c: ;:) ~ 35 ~ o :e 30 ~ o £! 25 C) c: 1: 8. 20 a.Q.:.) .s 15 ~ ~c: 10~--------~~--~~----~ Q) E ::::s ~ 5~ __________ 6-----~------~----->--~~ o~~~~~~~~~~~~~~~~~~~~ o O~0002 0;0004 0:0006 0~0008 0;001 0.0012 Air Flow Rate (Std. m3/s) Figure 5.12 Effect of air on percentage of feed bitumen reporting to the overflow and underflow (underflow diameter = 3/4 inch, temperature = 125 deg F) 63 64 5.4.2.1 Observations As is evident from these plots the trend of bitumen recovery to different streams was similar to the trend for the experiments carried out at 170 of, but the plateau value of the bitumen recovery to the overflow decreased with the corresponding decrease in the temperature. Since bitumen recovery in the overflow decreased with the decrease in the temperature, it is imperative that the temperature be kept high within economic considerations. 5.4.3 Study of the Effect of Volumetric Split to the Underflow on the Bitumen Loss and Sand Recovery to the Underflow Study of the effect of volume split to the underflow was carried out by varying the underflow diameter and feed flow rate for each underflow diameter of the TPOT hydrocyclone. 5.4.3.1 Critical Underflow Diameterl Critical Volume Split It was found that at a certain feed flow rate and feed concentration of sand in the feed, there existed a critical underflow diameter, hence critical volume split to underflow, which also yielded a concentrated underflow of about 70 wt. % sand with less than 10% loss of bitumen. Further increase in underflow diameter resulted in a dilute and drastically increased flow rate and II bitumen loss to the underflow. Critical volume split to the underflow is conceived to be a function of following parameters: 1 . Sand concentration in the feed 2. Feed flow rate 3. Diameter and geometry of the TPOT hydrocyclone Dc = f(sand concentration in feed, feed flow rate, TPOT hydrocyclone diameter and geometry) 65 There are several control systems employed today for regulating underflow concentration and flowrate in the pressurized cyclones that could probably be utilized on the TPOT hydrocyclone to maintain operation in the desirable region of solid concentration and volume split. The control mechanism consists of a flexible elastomer-sleeve which is specially molded so that controlled air pressure can be applied to change the underflow diameter as required. This could be detected by measurement of the underflow solids concentration. 5.4.3.2 Observations Figure 5. 13 plots the percentage of bitumen loss, and sand recovery to the underflow as a function of the feed volume split to the underflow at a constant feed flow rate. As may be observed from the plot, with the increase in the feed volume split, percentage of bitumen loss increased very gradually while sand recovery to underflow increased very rapidly. Based on this plot an underflow nozzle can be selected which controls bitumen loss to 10% or less. This could be achieved automatically by proper controls. As evrdenced, 0.75 inch diameter underflow gave the desired results. With this diameter underflow only around 7.5% of the bitumen was lost in the underflow, and around 70% of the feed sand was also discarded. Figures 5.14 through Figure 5.16 show that at excessively lowered feed flow rate, bitumen loss to underflow and underflow /). % Bitumen reporting to underflow <> % Sand reporting to underflow 90------------------------------------~ o== 80 ;: CD ~ 70 :» ti ~60 >o () CD a: 50 ~ C tU en ~ 40 c tU en en .3 30 c CI) .Ea 20 in '0 ' fI. 10 O~~~~~~~~~~~~~~~~~~~ o 2 4 6 8 10 12 14 16 010 of Feed Volume Flow Rate to Underflow Figure 5.13 Effect of underflow volume split on bitumen loss and sand recovery at underflow • 66 [] 010 Situ men lost ~ Underflow flow rate <> % of feed sand to underflow • Underflow concentration 0.0018 100 ~ 0 :e 0.0016 90 CI) 'C C ~ 80 0 -0.0014 -'CC (,) 70 as CI) UJ ~0.0012 'C CI) -E 60 t1? ~ 0.001 0~ a: 50 o-lS == (I) .20.0008 0 u. 40 -J == C CI) :0e O.0006 E 30 :J CI) !: 'cC m ~0.0004 ~ 20 0 .-:: 'C c as 0.0002 10 (I) -'ifl ~ 0 0 -d <.0 <.0 <.0 <.0 <.0 <.0 <.0 <.0 <.0 <.0 C <.0 <.0 <.0 <.0 <.0 <.0 <.0 <.0 <.0 <.0 0 0 ~ C\I C') ~ LO <.0 " 0 0 0 0 o · 0 0 0 e0x > e0 n 0 0 0 ~ . 0 0 ~ ~ ~ ~ ~ ci ci 0 ci ci 0 0 0 0 0 Total Flow Rate, F, (m3/s) Figure 5.14 Effect of the feed flow rate "on underflow solids concentration (wt.%), underflow flow rate, percentage of feed bitumen loss and sand recovery to underflow (underflow diameter = 3/4 inch) 67 .. ..-.. U) """"'- CI) E "-" ... ~ u. .a...) as a: ~ 0 u::: ~ 0 ;: "- CD "C c ::) [] % Situ men lost It. Underflow flow rate <> % of feed sand to underflow -0. Underflow concentration 0.0009 100 ~ 0 t 0.0008 90 Q) "C C ;:) 80 0 0.0007 ... "C c 70 as en 0.0006 "C Q) 60 Q) II. 0.0005 0~ 50 ~... U) 0.0004 0 ..J 40 c: Q) 0.0003 E 30 .:.::.J iii 0.0002 ~ 20 0 .....:: "C c: 0.0001 as 10 en ~ i 0 0 -d (!) (!) (!) (!) (!) (!) (!) (!) (!) (!) c (!) (!) (!) (!) (!) (!) (!) (!) (!) (!) 0 0 ,.... C\I ('t') ~ LO (!) . - ~ co 0') 0 0 0 0 0 0 0 0 0 0 ~ 0 0 ~ 0 ~ 0 . ~ ~ ~ 0 ci ci 0 ci 0 ci a n 0 Total Flow Rate, F, (m3/s) Figure 5.15 Effect of the feed flow rate on underflow solids concentration (wt.%), underflow flow rate, percentage of feed bitumen loss and sand recovery to underflow (underflow diameter = 5/8 inch) 68 • 69 [] % Situ men lost A Underflow flow rate ~ % of feed sand to underflow • Underflow concentration 0.00035 100 ~ 0 't: 90 Q) ..- "C 0.0003 c ..(..I..). ::» ('I) 80 0 E - ~ "C ... C ::s 0.00025 70 as u. en "C ar Q) 60 Q) .. as II.. a: 0.0002 0~ ~ 50 -~ 0 U) iL 0.00015 0 ...J ~ 40 c Q) ;0; E ~ 30 ::J Q) 0.0001 - " iii c ~ ::l 20 0 ~ "C 0.00005 c:: as 10 U) 0~- ~ 0 0 -u (0 (0 (0 (0 (0 (0 (0 (0 (0 (0 c (0 (0 (0 (0 (0 (0 (0 (0 (0 (0 0 0 ,..... C\I C") ..q- LO (0 r-.... co 0) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C! 0 ~ 0 ~ 0 ~ ci ci ci 0 0 0 0 ('""'\ ,., 0 Total Flow Rate, F, (m3/s) • Figure 5.16 Effect of the feed flow rate on underflow solids concentration (wt.O/o) , underflow flow rate, percentage of feed bitumen loss and sand recovery to underflow (underflow diameter = 1/2 inch) 70 flow rate, increased dramatically, and underflow solids concentration substantially decreased. For this flow rate and sand concentration in the feed, the corresponding underflow diameter becomes too large and is greater than critical diameter. 5.4.3.3 Determination of Critical Volume Split The following procedure is outlined to determine the approximate critical volume split to underflow. Table 5.3 was prepared for 5, 10, 16.81 and 20wt.% feed solids concentrations. These calculations were made from results of experiment 49, which was a critical point as plotted in Figure 5.14, wherein about 1 0% or less bitumen was lost in the underflow and percentage of feed sand reporting also to underflow was satisfactory. Sand had a specific gravity of 2.7 whereas water and bitumen are essentially 1.0. Accordingly feed volume split to the underflow can be easily determined. These volume split are listed in Table 5.3 and can be applied to Utah tar sands as long as a 65 mesh classification split is to be achieved. A plot between critical volume split to underflow and wt. % solids in feed is shown in Figure 5.17. 5.4.4 Weir Flow and Bitumen Recovery 5.4.4.1 Definition of Weir Flow If the overflow or the middling stream in the TPOT hy'tjrocyclone is restricted, the slurry level in the overflow will increase and the level in the TOPT hydrocyclone may touch its top such that only bitumen may start flowing down from the top. This is essentially a fourth stream, termed weir flow, coming out of the TOPT hydrocyclone along with overflow, middling, and underflow. Since air Table 5.3 Critical volume split to underflow as function of various feed concentrations wt. % solids in feed Total flow rate (m3/s) Critical volume split ( volume % of feed) 5 10 16.81 20 19.797x10-3 9.584x10-3 5.435x10-3 4.466x10-3 3.27 6.57 11.91 14.50 71 injected into the feed is attached to the bitumen, part of the bitumen floats up to the top and falls down as the fourth stream. If the above flow restriction in the slurry is applied, a mechanism, such as a skimmer, may have to be employed to remove relatively pure bitumen at this level (weir flow) of TPOT hydrocyclone. Figure 5.18 illustrates the weir flow for a typical run of the TPOT hydrocyclone. 5.4.4.2 Observations Tables in Appendix A illustrate the runs with and without weir flow while the amount of air injected remained constant. In run 23, overflow in TPOT • hydrocyclone was adjusted such that the level in TPOT hydrocyclone reached the top and pure bitumen began flowing down from the top of the TPOT hydrocyclone. Since no means of measuring as to how much weir flow bitumen was reporting, this was calculated by difference from the amount of bitumen in 16~----------------------------------~ -"D .m 14 '0 'if!. CD 12 E ::::s -~ ~ 10 o =e CD "D :5 8 .o. := Q. U) 6 Co) "C Q) E ::::s 4 ~ a; Co) = ".: 2 (.) o~~~~~~~~~~~~~~~~~~~ o 2 4 6 8 10 12 14 16 18 20 Concentration of Solids in the Feed, Cf, (wt.%) Figure 5. 17 Effect of feed concentration on critical volume split to underflow 72 73 74 the feed. From runs (refer to Appendix A, experiments 24, 25), it was seen that roughly 300/0 of the bitumen must have been recovered by weir flow. 5.4.5 Determination of Os and Proposal of a New Correlation for D5Jl 5.4.5.1 Primary Parameters Studied which Affect the 050 Parameter 1 = Underflow flow rate (m3/s) Parameter 2 = Feed flow rate (m3/s) Parameter 3= Feed concentration (sand wt.%) Parameter 4 = Air flow rate (Std. m3/s) Experiments were designed such that only one of the above parameters was varied in a set of experiment keeping the other three constant during that set of experiments to derive the relation of effect of the above mentioned parameters on 050. The following power law relation was assumed to exist between these parameters and the 050 050 = C(Parameter 1)x1(Parameter 2)x2(Parameter 3)x3(Parameter 4)x4 (5.3) where C is a constant and x1, x2, x3, x4 are the exponents of the parameters. During one set of the experiment, three parameters of the four were kept constant and incorporated in constant C. Hence 0 50 may be ~xpressed as follows for that set of the experiment. 050= C1 (Parameter 1) x1 (5.4) 75 050= C2 (Parameter 2) x2 (5.5) 050= C3 (Parameter 3) x3 (5.6) 050= C4 (Parameter 3) x4 (5.7) where, C1, C2, C3 and C4 are given by following expressions: C1= C(Parameter 2)x2(Parameter 3)X3(Parameter 4)x4 (5.8) C2= C(Parameter 1 )x1 (Parameter 3)x3(Parameter 4)X4 (5.9) C3= C(Parameter 1)x1(Parameter 2)x2(Parameter 4)x4 (5.10) C4= C(Parameter 1 )x1 (Parameter 2)x2(Parameter 3)x3 (5.11) 5.4.5.2 Effect of Underflow Volumetric Flow Rate on 050 Figure 5.19 plots the effect of underflow flow rate on the percentage of particle size reporting to the underflow. Based on Figure 5.19, 050 particle size is traced for each underflow size and is plotted in Figure 5.20 on a log-log scale. The following relation is found to exist between D50 (mm) and underflow flow rate Fu (m3/s). .. -6 -\38 0 50 = 7.834x10 Fu (5.12) at Average Feed flow rate = 8.07x10-3 m3/s ~ 0 :;: ~ Q) "C C :;:) .0.. . t» C E 0 a. Q) a: Q) N (jj Q) c:; 1: as D. f!. <> % of feed solids to underflow (Underflow flow rate = 1.207E-4 m3/s) /j. % of feed solids to underflow (Underflow flow rate = 3.183E-4 m3/s) o % of feed solids to underflow (Underflow flow rate = 3.273E-4 m3/s) o % of feed solids to underflow (Underflow flow rate = 6.654E-4 m3/s) o()o % of feed solids to underflow (Underflow flow rate = 1.1 03E-3 m3/s) 100 90 80 70 60 50 40 30 20 10 0 0 0.1 0.2 0.3 0.4 0.5 0.6 • Particle Size (mm) Figure 5. 19 Effect of underflow flow rate on percentage particle size reporting to underflow 76 -E ..E... Q It) Q 77 1----~~------------------------_. D =7.834x10-6 F -1.38 50 u R2 = 0.9797 0.1 -'-------,---r-~--,-,.--r-r':>__--~___,___r~~~ 0.0001 0.001 0.01 Underflow Flow Rate, Fw (rn3/s) II Figure 5.20 Effect of volume flow rate to the underflow on D50 Temperature =170 of Average Solids Concentration =14.52 wt.% solids Air Flow Rate = nil The exponent of volume flow rate to underflow is -1.38 hence x1 =-1.38 5.4.5.3 Effect of Feed Flow Rate on 050 78 Figures 5.21 and 5.23 plots the effect of feed flow rate on the percentage particle size reporting to the underflow. Experimental conditions of the plot are given in Table 5.3. Based on Figures 5.21 and 5.23, 050 particle size is traced for each feed flow rate and is plotted in Figures 5.22 and 5.24 on a log-log scale. The following relations were found to exist between 050 (mm) and feed flow rate F (m3/s) for 0.75 and 0.5 inch underflow diameters. at at 0 50 = 165.8Ft419 Underflow flow rate Average solids concentration Average air flow rate Temperature Underflow diameter Average solids concentration (5.13) =0.75 inch =16.24 wt.% solids = nil =170 of and (5.14) =0.5 inch =13.002 wt.% solids ~ 0 :e Q) "C C ::) -0 C) c t: 0 c. Q) a: Q) N en Q) Cl 1: as D. ~0 <> % of feed particle size to underflow (feed flow = 7.76E-3 m3/s) A % of feed particle size to underflow (feed flow = 7.12E-3 m3/s) o % of feed particle size to underflow (feed flow = 5.44E-3 m3/s) c % of feed particle fize to underflow (feed flow = 4.70E-3 m3/s) <> % of feed particle size to underflow (feed flow = 3.36E-3 m3/s) 90 80 70 60 50 40 30 20 10 o ~ o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Paricle Size (mm) Figure 5.21 Effect of feed flow rate on percentage particle size reporting to underflow (underflow diameter = 3/4 inch) 79 , -E 1~------------------------~------~ D50 = 1.658x102F1.419 R2 = 0.9887 E - 0.1 o an Q 0.'01 ~--------.,..---~---,---,....--,-.....,.-.....,.-~ 0.001 Feed Flow Rate, F, (m3/s) Figure 5.22 Effect of feed flow rate on D50 (underflow diameter = 3/4 inch) 0.01 • 80 <> % of feed particles to underflow (feed flow = 7.S36E-3 m3/s) A % of feed particles to underflow (feed flow = 6.234E-3 m3/s) o % of feed particles to underflow (feed flow = 5.467E-3 m3/s) [J % of feed particles to underflow (feed flow = 4.702E-3 m3/s) c % of feed particles to underflow (feed flow = 2.43E-3 m3/s) 100,-------~~--~--~~r_------~ 90 [J SO ~ 0 1: 70 "cCD ::J 60 .0.. C) c ;..:. 50 0 c. CD a: 40 CD c:; 1: a.m,.. 30 20 10 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Particle Size (mm) Figure 5.23 Effect of feed flow rate on percentage particle size· reporting to underflow (underflow diameter = 1/2 inch) 81 o &I) C 1----------------------------~------_, D50 = 3.814x1 02F1.432 R2 = 0.904 0.01 +-----:..--r-----,-----r---,----r----,.---,.--T""-1 0.001 Feed Flow Rate, F, (m3/s) Figure 5.24 Effect of feed flow rate on D50 (underflow diameter = 1/2 inch) 0.01 82 • Average air flow rate Temperature = nil =170 of 83 From equations 5. 12 and 5.13 an average value of the exponent of feed flow rate was taken. Accordingly 1. 419 + 1. 432 Exponent of feed flow rate = = 1. 426 2 hence x2 = 1.426 5.4.5.4 Effect of Feed Concentration on 050 . Figure 5.25 plots the effect of feed concentration on % particle size reporting to the underflow. Experimental conditions of the plot are given in Table 5.4. Based on Figure 5.25 D50 particle size is traced for each feed solids concentration and is plotted in Figure 5.26 on a log-log scale. Following relation is found to exist between D50 (mm) and feed solids concentration Cf (wt.%). (5.15) at Underflow diameter = 0.625 inch Average feed flow rate Average ai r flow rate = 7 .83x1 0-3 m3/sec = nil • Temperature =170 of Exponent of solids concentration in feed = 2.183 hence x3 =2.183 100 90 ~ 80 0 ;: "QC) 70 C :l 0 60 ... ac 1: 0Q . 50 Q) m: Q) N 40 (jj Q) Q 1: 30 as Q. ~0 20 10 0 0 0/0 of feed particles to underflow (sand <> concentration in feed = 15.35 wt.%) 0/0 of feed particles to underflow (sand f1 concentration in feed = 7.58 wt.°la) 0.2 0.4 0.6 0.8 1 Particle Size (mm) 1.2 1.4 • Figure 5.25 Effect of sand concentration in feed on percentage particle size reporting to underflow (underflow diameter = 5/8 inch) 84 E -E o It) Q <> 050 (mm) 1,-----------------------------~--__ 0 50 = 1.441x10-3 C t 2.183 R2 = 1 0.1~----~~--~------~~------~--~ 1 10 30 Sand Concentration In the Feed (wt.%) Figure 5.26 Effect of sand concentration in feed on D50 .. 85 86 5.4.5.5 Effect of Air Flow Rate on 050 As is illustrated in Figure 5.27, air flow rate does not appreciably affect the 050. With the increase in the air flow rate some coarser solids may report to the overflow, but air flow rate has very little effect on 050 particle size. Hence exponent of the air flow rate is assumed to be zero. Thus x4 = 0.0. Finally, the following relationship for 050 particle size is obtained: CF1.426C 2.183 o - f so - F 1.38 u (5.16) It can be shown by simple calculations that the value of constant C is about 2.88x10-5; hence for 050 (mm) the equation is as follows: 2. 88x1 0-5 F1.426C f 2.183 Oso = F 1.38 u (5.17) If the above equation for 050 is compared with the following 050 equation for the two-product conventional hydrocyclone in the literature.34 KO .10°·60 0.8 •• 0.5 o - c i ° f""1 so - FO.5 (ps _ PI )0.5 (ti.18) It may be observed that in the TPOT hydrocyclone, 050 is proportional to the feed flow rate to the power 1.426, whereas for the two-product, conventional <> % of feed particles to underflow (air flow = 1.0824E-3 Std. m3/s) d % of feed particles to underflow (air flow = 0.9789E-3 Std. m3/s) o % of feed particles to underflow (air flow = 0.772E-3 Std. m3/s) [] % of feed particles to underflow (air flow = 0.565E-3 Std. m3/s) c 010 of feed particles to underflow (air flow = 0.358E-3 Std. m3/s) v % of feed particles to underflow (air flow = 0.0 Std. m3/s) 100 90 =0= 1: 80 CD "C :CJ 70 .0.. . Cc ) 60 t: 0 50 Q. CD a: CD 40 N CiS cC:D; 30 t: [] Das. 20 ~0 10 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Particle Size (mm) .. Figure 5.27 Effect of air flow rate on percentage particle size reporting to underflow (underflow diameter = 3/4 inch) 87 88 hydrocyclone D50 is inversely proportional to the feed flow rate to the power 0.5. This opposite dependence is a direct consequence of the flow pattern. In the TPOT hydrocyclone, flow pattem is a forced-vortex whereas in the two-product, conventional hydrocyclone the flow pattern is a free-vortex. Table 5.4 contains summary of variables and constants for figures 5.7 to 5.27. .. Figure No. 5.7 5.8 5.9 5.10 5.11 5.12 5.13 5.14 5.15 5.16 Table 5.4 Sum-mary of variables and constants for Figures 5.7 to 5.27 Experiment Variable Constants no. 1-6 Air Flow Rate Feed Flow Rate Underflow Diameter Temperature Solids Concentration 12-16 Air Flow Rate Feed Flow Rate Underflow Diameter Temperature Solids Concentration 7-11 Air Flow Rate Feed Flow Rate Underflow Diameter Temperature Solids Concentration 17-22 Air Flow Rate Feed Flow Rate Underflow Diameter Temperature Solids Concentration 25-30 Air Flow Rate Feed Flow'Rate Underflow Diameter Temperature Solids Concentration 31-36 Air Flow Rate Feed Row Rate Underflow Diameter Temperature Solids Concentration 37-41 Undeflow Feed Flow Rate Diameter Air Flow Rate Temperature Solids Concentration 47-51 Feed Flow Rate Underflow Diameter Air Flow Rate Temperature Solids Concentration 42-46 Feed Flow Rate Underflow Diameter Air Flow Rate Temperature Solids Concentration 52-56 Feed Flow Rate Underflow Diameter Air Flow Rate Temperature Solids Concentration 89 Average Value Of The Constants 7.89E-03 m3/sec 0.75 inch 170 deg F 14.57 wfO/o 6.06E-03 m3lsec 0.625 inch 170 deg F 11.31 wfO/o 6.03E-03 m3/sec 0.5 inch 170 deg F 10.84 wfOk 5.18E-03 m3lsec 0.75 inch 170 deg F 15.71 wfO/o 7.94E-03 m3/sec 0.75 inch 140 deg F 13.675 wfO/o 7.53E-03 m3/sec 0.75 inch 125 deg F 14.52 wfOk 8.28E-03 m3/sec nil - 170 deg F 14.51 wfO/o 0.75 inch nil - 170 deg F 16.24 wfOk .0. .625 inch nil - 170 deg F 15.36 wfO/o 0.5 inch nil - 170 deg F 13.002 wfO/o 90 Table 5.3 (Continued) Figure No. Experiment Variable Constants Average Value Of no. The Constants 5.19-5.20 37-39 Underflow Feed Flow Rate 8.28E-03 m3lsec 41 and 47 Diameter Air Flow Rate nil - Temperature 170 deg F Solids Concentration 14.51 wtO~ 5.21-5.22 47-51 Feed Flow Underflow Diameter 0.75 m3/sec Rate Air Flow Rate nil - Temperature 170 deg F Solids Concentration 15.48 wtO/o 5.23-5.24 52-56 Feed Flow Underflow Diameter 0.5 m3/sec Rate Air Flow Rate nil - Temperature 170 deg F Solids Concentration 13.002 wtO~ 5.25-5.26 43 and 69 Feed Solids Underflow Diameter 0.625 inch Concentration Air Flow Rate nil - Temperature 170 deg F Feed Flow Rate 7.83E-03 m3/sec 5.27 1-6 Air Flow Rate Feed Flow Rate 7.89E-03 inch Concentration Underflow Diameter 0.75 - Temperature ..1 70 deg F Solids Concentration 14.57 m3/sec CHAPTER 6 BENCH SCALE THICKENING TESTS ON TAILINGS Thickening of tailings is of utmost importance in the process of extraction of bitumen from the hot-water process as it aids in the recovery of water from the slurry. The tailings to be thickened are basically fine solids (finer than -65 mesh) and water. These are the residuals primarily of overflow and middling streams after the removal of bitumen from these two streams in the clarifiers. The high-density thickener is a very recent development and may be employed for thickening of fine solids (-65 mesh) from tar sands tailing. The high-density thickener was developed in Australia51 where they process bauxite to manufacture alumina (AI203) for -the electrolytic production of aluminum. The bauxite is leached in a caustic solution under high temperature and pressure to solubilize the alumina-trihydrate (AI203.3H20). The bauxite contains iron compounds, and it generates very fine solids ( generally less than 10 J.1m) which have a specific gravity of 3.3 to 3.5 and are termed "red mud. II The bauxite processors have found that if they flocculated these solids at a dilute concentration (2 to 5 wt.% solids) a different and better flocculi structure resulted which, when the solids height was maintained above the thickener .f. loor at 15 to 20 feet or more, resulted in a very high underflow solids concentration. This is caused by the high weight of solids above the floor and the flocculi structure that exhibited this compressibility. In the case of red mud, underflow solids concentration increased from 35 wt. % solids in a conventional thickener to 53 92 wt. % solids in the high-density thickener. This is for the -10 Jlm solids, a dramatic 50% more water elimination from the solids. The very thick slurry is pumped by a high pressure piston pump about 3.5 km to a final disposal area. Studies on tar sand tailings were confined to bench scale tests. Accordingly, typical overflow and middling tailings were collected in barrels to conduct bench scale thickening tests after bitumen was skimmed off from the ~ top of the barrels. Special batch thickening tests were performed with equipment used to develop the red mud application pictured in Figure 6.1. There is a slowly rotating rake which acts to furnish channels in the thickened solids so that fluid can more easily escape. The slurries that were tested had a size distribution that was about 900/0 -65 mesh. 6.1 Flocculant and Flocculant Screening In order to concentrate the fine tailings, a flocculant is used to agglomerate the solids with a view to increase their underflow solids concentration and settling rates. In order to concentrate these solids the flocculi structure is very important. High underflow solids concentration is achieved by the weight of solids above in order to compress them. A number of flocculants were tried in order to assess their effectiveness in producing the desired optimal flocculation of the tailings. The screening tests were conducted using a small sample of the tailings, about 200 ml in beakers, .. and using different polymer flocculant solutions having concentration of 0.2 gm/l of dry substance. These screening tests were necessary to discern the specific polymer which helps to produce the optimum flocculi structure. Polymer solution was added into the beaker containing tailings and agitated with a 93 f.- 22.2 ~ Motor Rake 49.5 1 76 * All dimensions in Centimeters .. Figure 6. 1 Schematic of bench scale equipment for thickening tests 94 spatula until the time the flocculi in the beaker appeared. The amount of the polymer solution to produce flocculation was noted along with other parameters such as clarity of the supernatant, settling rates, and amount of flocculi produced. Since amount of solids in the beakers were almost the same in all the cases, the lesser amount of floes implied that these floes were more dense. These were the basic requirements of the tests as the main aim was to ascertain the polymer which could produce high solids concentrations in the underflow of a high-density thickener. Based on the preliminary beaker tests some flocculants were identified for the bench scale tests of the solids. Characteristics of the polymers selected for bench scale tests are given in Table 6.1. 6.2 Bench Scale Thickening Tests Bench scale thickening tests were conducted using facilities at EIMCO Process Equipment Company. The schematic diagram of the bench scale equipment for thickening tests is shown in Figure 6.1. A measured amount of slurry was taken in the test equipment to conduct bench scale thickening tests. Polymer and the amount of polymer required to flocculate, as calculated from the beaker tests, were taken in a big syringe at a concentration of 0.2 g/I of polymer concentration. The polymer was injected in the slurry through a distributor to disperse equally the polymer in the slurry and was gently agitated to produce a good mixing of the polymer in the slurry. As a result of the injection of polymer and gentle agitation, the solids in the slurry were flocculated and settling was initiated. The rake mechanism was installed, and the rake motor started. The level of the slurry was noted down as a function of time. 95 Table 6.1 Polymers selected after screening S.N. Polymer Characteristics 1 Percol 156 High molecular weight, anionic polyacrylamide 2 Percol406 Liquid grade, high charge, cationic coagulant 3 Percol611 High molecular weight, anionic poyacrylate flocculant 4 Percol435 Cationic flocculant 5 Percol402 Liquid grade, high charge, cationic coagulant 6 E10 Very high molecular weight, low anionic, polyacrylamide flocculant 7 Percol592 Cationic flocculant 8 Percol727 Very high molecular weight, anionic polyacrylamide flocculant • 96 The results of these tests are in Appendix B. Table 6.2 summarizes the underflow concentration of the solids achieved in the bench scale tests. As observed from the results in Table 6.2, tests 1 to 8, the best thickening results exhibiting high solids concentration were obtained by using polymer Percol-156 and Percol-406 together. Percol-156 is the predominant polymer, and Percol 406 represents a very small proportion in the ratio of 95:5 required to produce high-density solids in the underflow. Table 6.2 represents the tests conducted by using the selected polymer, viz., Percol-156 and Percol-406, with a varying percentage of concentration of solids in the feed. These tests indicate that the increase in the percentage of solids in the feed increased the corresponding percentage of solids in the underflow. This may be partly because at high feed solids concentration, there was a greater depth of flocculated solids to produce more compression. Figure 6.2 illustrates the solids concentrations achieved in the batch thickening tests at various ·starting concentrations. Tables in Appendix B contain the time versus depth data required to achieve the unit area for the thickener. These tables also contain results for underflow concentration achieved and unit volume etc. 6.3 Unit Area Calculations Unit Area(UA) calculations were carried out in cases where feed concentration was 6.49 wt. % solids. II 6.3.1 Definition of Unit Area (UA) The required thickener area is decided by the layer requiring the maximum area to pass unit solids through it, which is also called the rate controlling layer. Unit area is generally expressed in units of m2/tonne/day. 97 Table 6.2 Summary of results of thickening tests for solids concentration achieved Test no. Polymers used Percentage Solids in Percentage Solids in feed underflow 1 Percol 406, 156 0.62 36.41 2 Percol 611 ,435 0.62 36.81 3 Percol 402, 156 0.62 46.79 4 Percol 156 406 0.62 49.55 5 Percol 592,406 1.77 37.66 6 Percol 727,406 1.77 39.55 7 Percol E 10,406 1.77 46.40 8 Percol 156,406 1.78 57.76 9 Percol 156, 406 1.77 50.55 10 Percol 156, 406 1.77 50.80 11 Percol 156,406 4.55 57.97 12 Percol 156,406 6.49 61.10 13 Percol 156,406 8.36 60.38, 14 Percol 156,406 10.41 64.05 • 65~---------------------------------. 49 47 454-~~~~~~~~~~~~~~~~~ o 2 4 6 8 10 12 wt. % Solids in the Feed Figure 6.2 Effect of feed concentration on the underflow concentration in the bench scale thickening tests 98 99 Hence unit area is area (m2) of thickener required to process the 1 tonne of solids in a day to achieve the indicated desired concentration. As is known,52 the unit area (UA) of the thickener can be obtained from a batch test and can be calculated using the Wilhelm-Naide method. Following is Wilhelm-Naide equation for the calculation of the unit area. (b -1 )(b-1) C~b-1) UA = ---.:b:...-__ _ ab (6.1) Cu is underflow concentration in the continuously operating thickener and is equivalent to the average concentration (Ca) in the batch test and may be calculated as follows from the batch test: C = CoHo U H U (6.2) where Co. and Ho are the initial concentration and initial height in the batch test and Hu is the height of interface. Constants a, and b appear in the hindered settling velocity as follows: (6.3) • where Vi is settling velocity of layer of concentration Ci. Table 6.3 summarizes the unit area calculations in the case of data in in Appendix B for thickening test 12 with percentage solids in feed 6.49 wt.O/o. Time Height Velocity {h) {m) {mId) 0.000 0.760 0.025 0.620 57.615 0.033 0.600 41.968 0.042 0.575 32.823 0.050 0.565 26.851 0.067 0.550 19.559 0.083 0.530 15.297 0.167 0.490 7.129 0.250 0.465 4.561 0.333 0.435 3 .322 0.500 0.425 2.126 1.000 0.405 0.991 1.500 0.390 0.634 2.000 0.330 0.462 • Table 6.3 Calculation of unit area Intercept Interfacial '-Cone. of tangent{m) tonne/m3 0.560 0.585 0.5,42 0.605 0.518 0.633 0.509 0.644 0.496 0.661 0.477 0.687 0.440 0.744 0.417 0.785 0.389 0.843 0.381 0.861 0.364 0.901 0.350 0.935 0.292 1.124 UA m2/tonne/d 1.70E-03 2.16E-03 2.99E-03 3.39E-03 4.12E-03 5.45E-03 9.69E-03 1.43E-02 2.39E-02 2.79E-02 3.89E-02 5.10E-02 1.94E-01 Underflow Solids tonne/m3 42.763 43.807 45.240 45.807 46.680 47.964 50.662 52.531 55.058 55.821 57.485 58.862 65.821 -'" o o 6.3.2 Steps in Calculation of Unit Area 6.3.2.1 Determination of Parameters a and b 6.3.2.1.1 Step 1 101 Height H (m) and time t (h) were plotted in Figure 6.3 and on log-log scale in Figure 6.4. The following equation for the height as a function of time was obtained. H = 4. 067x1 0-1 t-O·1015 (6.4) 6.3.2.1.2 Step 2 Equation 6.4 was differentiated with respect to time to obtain interface velocity V (m/h) as a function of the time t (h). Hence, following equation was obtained. V = -4.127x1 0-2 t-O·1015 (6.5) 6.3.2.1.3 Step 3 To obtain the interface concentration at different times the tangents were drawn in Figure 6.3 (not shown) and the intercept of that tangent to the height axis (V-axis) was used to calculate the interface concentration. V.. elocity was plotted against the interface concentration in Figure 6.5, and the following equation for interfacial velocity Vi (mId) and interfacial concentration ex (tonne/m3) was obtained. 0.8-------------------, 0.7 0.6 .-.. E '-"0.5 :t: .: .c C) 0.4 "ii l: 0.3 0.2 0.1 H = 4.067x10-1 rO.1015 R2 =0.9723 o 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Time, t, (h) Figure 6.3 A plot of height versus time for bench scale test no. 6 .. 102 --E 1~~----------------------------~ H = 4.067x10-1 rO.1015 R2= 0.9723 0.11-~~~~~nmr-~~~~~~~~~ 0.0001 0.001 0.01 0.1 1 10 Time, t, (h) Figure 6.4 A plot of height versus time on log-log scale .. for bench scale test no. 6 103 .. 100~----~--------------------------1 Vi = 6.S17x1 0.1 CX·8.255 R2 = 0.9691 :> 10 .ie:-; o "ii > en .5 i en CD () ~ ! .5 1 0.1~--~--~--~~~~r---~--------~ 0.4 1 2 Interface Concentration, Cx, (tonne/m3) .. Figure 6.5 Effect of interface solids concentration on interface settling velocity 104 Hence a = 6.517x1 0-1, b = 8.255 6.3.2.2 Calculation of Unit Area Using Wilhelm-Naide Equation 6.3.2.2.1 Step 4 105 (6.6) Wilhelm-Naide equation (equation 6.1) was used for unit area (UA) calculations in m2/tonne/day. The following equation was obtained after substituting values of a and b. UA = 7.284x10-2C7.255 u (6.7) where Cu = underflow concentration, tonne/m3. Figures 6.6 and 6.7.plot the unit area against underflow concentration in tonne/rn3 and wt.% solids in the underflow respectively. From the bench scale tests conducted and the above discussions of the experiments, it may be observed that Percol-156 and Percol-406 in little quantity have produced the best results. Figure 6.2 illustrates the percent solids in feed versus the concentration achieved in the thickening tests. It is further observed that high concentration of feed in the tailings leads to high concentration in 'he thickener underflow. This drives the conclusion that in continuous operation, the thickener should be operated at maximum feed concentration. This conclusion is opposite to results with the high-density thickener. Generally speaking, the feed solids concentration is usually diluted (by thickener surface clean water) to 106 1-------------------------r----------------~ -... "C as c c 0 0.1 ~ E --- <i ;:) fi .Q..) C... "2 ;:) ... Q) c 0.01 ~ (,) :c I-J .00 1 -4---~-r--...,.-.,..........,.~~r___-_...,-__,__,.__r__r_....,.._,._r_1 0.1 1 10 Underflow Concentration, Cu, (tonne/m3) • Figure 6.6 Effect of underflow concentration on unit area 107 1~---------------------------r--~ ~ ~ c c 0 ~ 0.1 E ~ ~ ::l af .C.D. c..(, ·2 :.:..l CD C ~ :(c) 0.01 I- 0.001 -+----~--...,......-_rl_.....,____r_'_r""_r_~ 10 100 Underflow Concentration, Cu, (wt.% solids) • Figure 6.7 Effect of underflow concentration (wt.%) on unit area 108 somewhere between 2 to 8 wt.% solids as a different floc structure is formed. As the compression height experienced in these tests, particularly with dilute concentrations, did not result in high enough compression pressures, the concentrations were much lower at low feed solids concentrations. Accordingly more tests should be performed in the future. The unit area calculations discussed in this chapter may be used for the design of the thickener for continuous operation. • CHAPTER 7 DESIGN OF SPIRAL CLASSIFIER The spiral classifier dewaters the coarse solids, +65 mesh solids, separately from the -65 mesh solids. Minimum moisture contents can be achieved for both solids, thereby maximizing water recovery and recycling. 7.1 Sizing of Spiral Classifier Table 7.1 contains analysis of the underflow solids for a typical run with 3/4 inch underflow nozzle (results of experiment 3). Table 7.1 Particle size distribution of underflow for experiment 3 Mesh size Percentage solids 28+ 3.57 28x32 5.31 32x35 9.51 35x48 39.74 48x65 22.09 65x100 14.25 100x140 3.55 140x200 1.28 200- 0.7 110 In an actual running plant this will be the feed to the spiral classifier. As is observed there are 80.22% of +65 solids and only 19.78% of -65 solids. Following were the flow rates: Solids flow rate = 778.77 gls Water flow rate = 296.52 gls Wt.% solids = 72.42 which implies Flow rate of +65 solids = 624.73 g/s Flow rate of -65 solids = 154.04 g/s Water flow rate = 296.52 g/s Following are the calculations for the sizing of spiral classifier when the feed to spiral classifier is diluted to 5%, 100/0, and 150/0 sand by weight. 7.1.1 Sizing of Spiral Classifier when Feed Is Diluted to 5 wt.% Solids If x g/s amount of dilution water is added then the following expression should be satisfied: 778.77 = 0.52 x+296.52 which gives x=15278.88 gm/sec (7.1) Since -65 mesh solids are to be removed from the spiral crassifier as overflow, its concentration from the spiral classifier can be calculated as follows: Overflow flow rate = 15278.88+296.52+154.04 gls . . . . Flow rate of -65 mesh solids Overflow concentration from spiral classifier = -----------­Total flow rate Volumetric flow of overflow (7.2) =-------1-5-4-.-04- ------- 15278.88 + 296.52 + 154.04 = 0.9793 wt. % solids 15278.88 + 296.52 + 154.04 = _______- -=2... ....7 .:..- 1000 (7.3) =15.632 I/s =247.81 gpm I/s Pool area equation from the Denver Spiral Classifier manual is as follows: 111 P I (ft2) Overflow (gpm) 00 area = P I 't (Settling rate at overflow concentration (wt. 0/0)) 00 area capaci yx . . . Settling rate at 5 wt. % solids concentration (7.4) Pool area capacity for 65 mesh size is 15 gpm/ft2, settling velocity for 65 mesh size particle at 0.9793 wt. % solids is 1.0 inch/sec and settling velocity for 65 mesh size particle at 50/0 solids concentration is 0.8 inch/sec. Hence pool area is as follows: • P I A (ft2) 247.81 00 rea = 1 0 = 13.22 15x(-' ) 0.8 (7.5) 112 Solids to be raked(+65 mesh)=624. 73x3600/(1 000x1 000) = 2.249 TPH Similarly it can be shown that for 10 wt. % feed and 15 wt. % feed pool area is 6.982 ft2 and 4.93 ft2 respectively with 2.249 TPH of solids to be raked. The solids to the spiral classifier come from the underflow of the TPOT hydrocyclone, and since TPOT hydrocyclone can be operated in a regime where relatively little fines report to underflow (below critical underflow split). Thus it is not expected that dilution of the feed to the spiral classifier below 15 wt. % solids is required. However above calculations show that with less dilution, the pool area decreased substantially. This is advantageous, since a smaller unit will serve the purpose. Basically, for economic reasons, a spiral classifier should be selected with minimum pool area, and should give the required raking capacity of solids. The spiral classifier is usually selected based on the limiting parameter, minimum pool area, or the raking capability. If the selection is on the basis of raking 'capability, it is necessary that for such a model the desired dilution of feed should be calculated based on the pool area. 7.2 Concentration of +65 Solids Obtained in a Beaker Test of Tailings The +65 mesh from the underflow mentioned above were separated and diluted with water and raked up in an inclined plain with the help of a spatula. Solids concentration obtained was about 82 wt. %. At 18 wt. % moisture, the solids can be transported by a belt conveyer back to old mine area"'Nith a water loss of 0.146 tonne per tonne of dry solids. It was also analyzed for bitumen content, which was found to be less than 20/0 of feed bitumen. Thus additional bitumen should be collected in spiral classifiers by weir overflow to recover better than 90% of original bitumen. CHAPTER 8 FINAL RESULTS AND DISCUSSIONS 8.1 Product Quality from TPOT Hydrocyclone During processing of the tar sand slurry in the TPOT hydrocyclone most of the coarse sand reports at the underflow and· most of the fine sand reports at the middling and overflow. Bitumen in the feed is concentrated to a higher value with respect to per unit weight of the dry sand in overflow and middling. Following is the analysis for a typical case (Appendix A, experiment 3) 8.1.1 Quality of the Feed Wt. % bitumen in the slurry Wt. % sand in the slurry Wt. 010 water in the slurry = = = 1.18 14.99 83.83 Wt. % bitumen in the slurry = 1.18 = 0.0787 Wt. % sand in the slurry 14.99 8.1.2 Quality of the Underflow Wt. % bitumen in the slurry = 1.01 Wt.% sand in the slurry = 71.70 Wt. % water in the slurry = 27.30 0/0 of feed bitumen = 10.59 0/0 of feed sand = 59.61 0/0 of feed water = 4.06 (8.1) • 114 Wt. % bitumen in the slurry = ~ = 0 0141 Wt. % sand in the slurry 71. 70 . (8.2) 8.1.3 Quality of the Middling Wt. % bitumen in the slurry = 0.99 Wt.% sand in the slurry = 7.54 Wt. % water in the slurry = 91.47 % of feed bitumen = 42.15 0/0 of feed sand = 25.31 % of feed water = 54.92 Wt. % bitumen in the slurry = 0.99 = O. 1313 Wt. % sand in the slurry 7.54 (8.3) 8.1.4 Quality of the Overflow Wt.% bitumen in the slurry = 1.5 Wt.% sand in the slurry = 6.07 Wt.% water in the slurry = 92.43 0/0 of feed bitumen = 47.26 % of feed sand = 15.07 0/0 of feed water = 41.02 • Wt. % bitumen in the slurry = ~ = 0.2471 Wt. % sand in the slurry 6.07 (8.4) It can be observed from the above analysis that bitumen is concentrated in the overflow and middling. About 60% of the feed sand is discarded at 115 underflow with about 10% of the feed bitumen reporting there. (The percentage of discarded sand at underflow can be increased to about 75-80% if the unit operates at the critical underflow split.) 8.2 Water Loss in the Process Some water in the process is lost in the concentrated tailings. Tailings from the spiral classifier are +65 mesh sand and from high density thickener are -65 mesh sand. As mentioned before +65 mesh sand can be concentrated up to 82 wt. % sand and -65 mesh sand can be concentrated up to 65-70 wt. % sand. The following analysis gives the water loss per tonne of tar free sand. Weight of the tar free sand = Weight of the +65 mesh per tonne of tar free sand = Weight of the -65 mesh per tonne of tar free sand = 1000 kg 703.54 kg 296.46 kg Water loss with +65 mesh solids Water loss with -65 mesh solids = (18/82)(703.54) = 154.44 kg = (30170)(296.46) = 127.05 kg Total water loss per tonne of tar free sand 154.44+127.05 = 281.49 kg Water loss in this process may be expressed as follows for per tonne of tar free sand. 18 % +65 30 % +65 Water loss = (82)X1000X 100 +(70)X1000X(1.0- 100 ) (8.5) • 8.3 Comparison of Water Loss in New Process with the Water Loss in the Process Shown in Figure 2.7 116 Following is the comparison of water loss with per tonne of tar free sand for the new process and process mentioned in the literature.25,38 In the Iiterature38 the water cycle is given for the process in Figure 2.7. There were 250/0 of +65 solids in the feed tar sand. It was mentioned that 16% of the water was lost in coarse sand, and fine sand and from evaporation. The basis of this was that the concentration of sand at flotation was 20 wt.%. Hence total water lost in the process was as follows: 16 1000 Water loss (kg/tonne of tar free sand) = 99 x2()"x80 = 646.46 (8.6) Total water loss using equation 8.5 for new process is as follows: 18 25 30 25 Water loss (kg / tonne of tar free sand) = (-)x1 OOOx - + (-)x1 000x(1--) 82 100 70 100 = 376.30 (8.7) From equations 8.6 and 8.7, it can be seen that water loss was reduced by about 41.76 % by the new process. 8.4 Explanation of Increased Underflow Flow Rate above Critical Underflow Split 117 As evidenced from Figures 5.14 and 5.15, underflow flow rate increased dramatically and its concentration dropped substantially below the critical underflow split. Above the critical feed flow rate, underflow concentration and flow rate were about constant. However, below critical flow rate, not enough sand went into the TPOT hydrocyclone to maintain underflow concentration; hence the underflow concentration dropped and the underflow flow rate increased. This phenomenon can be explained on the basis of high underflow viscosity and density above critical flow rate. Analysis was carried out for graph 5.14. Data for this graph in respect of below and above the critica.l point (experiments 49 and 51.) 8.4.1 Analysis for Experiment 49 and 51 Assuming that pressure differential across the underflow pipe remains constant below and above the critical point, the expression for pressure drop is as follows: where L\P _ 32f1P1Q1 2 L - 1[20 5 f _ 0.0791 1 - 1 Re~iPe (8.8) (8.9) • 118 8.4.1.1 Calculations for Experiment 49 (8.10) Q=Fu1 (8.11 ) Du = 3/4" = (25.4)(0.75)/1000 = 0.01905 m (8.12) Vu Q = (8.13) A A = 1tD~'Pe (8.14) 4 Q V = 2 u (1tDpiPe) 4 (8.15) Vu = 6.475x10-4/((3.1415)(0.019052/4)) =2.272 m/sec Equation 4.15 was used to calculate the viscosity of the slurry. 845.95 41v = Volume fraction of solids = (845.95 + ~2~. 32 + 5.91) = 0.4856 2.7 1 (8.16) • J.lm = 1.0+(2.5)(0.4856)+(10.05)(0.4856)2 + (0.00273)exp((16.6)(0.4856)) Jlo = 13.23 hence J.1m = 13. 23J.1o J.1o = Viscosity of water at the operating temperature = 0.34x1 0-3 Pa.s Expression 4.16 was used to calculate slurry density. 100 kg Pm = 71.68 100-71.68 = 1823 m3 (2700 + 1000 ) Re = (1823)(0.01905)(2.272) = 17539 (13.23)(0. 34x1 0-3) aP _ (32)(.0791)(1823)QU12 => T - 1t2D5(17539~) Similarly, the above expression