Document MMEoxkQjZ6K7K9jYk6yEbG5g7

CARBIDE AID CARBON CHEMICALS CatFATT texaa cm PLANT AUGuar 8, 1956 To: Mr. Robert Dillon TEXAS CITY FLAM SUBJECT: Agitation Studies* grejact 817-1^5 cc: Mr. J. W. Biddle Mr. D. C. Burger/Mr. R. R. Rasmussen Mr. L. L. Dintlean Dr. A. W. Downes Mr. J. P. Ferrer Mr. G. K. Graeber Mr. G. J. Hanks Mr. J. R. Keman Dr. J. L. Marsh Mr. E. H. Mel 11 nyr Dr. T. R. Miller Mr. J. S. Nicholas Mr. B. R. Thompson Mr. C. A. Wales Mr. C. R. Welter Nr. T. A. Wilier Mr. J. C. Wright Attached Is your copy of "Agitation Studies" by E.K. Adrian and E. D. McCrary, done under tbe supervision of J. C. Wright. This work carers various factors involved 1a tbe design of tower type polymerization reaction vessels and makes a valuable contribution to the knowledge re quired to build successfully operating reactors of this type. Two primary factors were studied. These were: (l) Depth of 1apellers below surface to prevent vapor entrainment Into tbe liquid phase and (2) Methods of preventing recirculation of particles at the bottom of the reactor. We believe that the findings contained in this report can he ex trapolated to the 5-foot diameter continuous polymerlzer to permit more continuous operation. Mr. Adrian and Mr. McCrary are two of our summer hires and we be lleve their work will benefit them as well as the company. Mr. Adrian is a Junior student at Rice Institute majoring In Physics and Mr. McCrary Is a senior student at Notre Dame majoring In Chemical Engineering. WWM/jn UCC 040907 /* AGITATION STUDIES CONDUCTED AMD KEFQFSmJ BY EHLE K. ADRIAN E. D. McCROHY PROJECT NUMBER 817-1^5 PROCESS DEVELOPMENT DEPARTMENT CARBIDE AND CARBON CHEMICALS COMPANY A DIVISION OF UNION CARBIDE AND CARBON CORPORATION TEXAS CITY. TEXAS AUGUST 1?. 1956 v Ucc 0*0908 SUMMART TABLE OF COHTENTS Page 1 DmraDOCTioir 2 HtOCEDURE 2 DISCUSSION 2 A. Studies of Surface Effects and Mixing Characteristics of Various Impellers 2 1. The effect on AI of impeller depth 2 2. The effect of impeller diameter 3. The effect of increasing the numberof blades 3 5 B. Study of Bottom Agitation and Recirculation Control 5 1. Agitation in a conical bottom section 3 2. Agitation in a flat bottom vessel 3 3. Baffle designs 3 Studies of heat transfer 4 C, Effect of Varying Pellet Concentration 4 APPENDIX Figure 1-A, Effect of Impeller Depth (8 Blades) 5 6 Figure 1-B, Effect of Impeller Depth (6 Blades) 7 Figure 1-C, Effect of Impeller Depth (A- Blades) 8 Figure 2f Effect of Diameter 9 Figure 3# Effect of Blade Number 10 Figure Jf, Comparison of Baffle Design H Figure 5-A, Baffle Effect on Heat Transfer (30 Degree Baffles) 12 Figure 5-B, Baffle Effect on Heat Transfer Slanted Baffles (Vertical Section) 13 Figure 5-C, Baffle Effect on Heat Transfer (Disc and Baffle Section) Figure 5-D, Baffle Effect on Heat Transfer (Disc Alone) -5 Figure 5-E, Baffle Effect on Heat Tranafer Comparison of Interface Position ucc 040909 *6 Figure 6, Effect of Loading on AI Figure 7-A, 2^-Inch Diameter Glass Column Figure 7-B, 6-Inch Diameter Glass Column Figure 8, Slanted Baffle Sections ^7 18 19 20 ucc 040910 1 SUMttHY Agitation studies were made In the 2k-lnch and 6-inch diameter glass columns In order to determine the effect on agitator influence (AI) of the following; (1) Impeller depth below liquid surface (2) Impeller diameter (?) Humber of blades per Impeller (k) Concentration of agitated material Mixing properties, such as the point at which entrainment of the vapor phase first appears were noted In each case. It was found that the AI Is affected but little by impeller depth except when the impeller is shallow enough to cause entrainment. An optimum impeller diameter to column diameter ratio is found near the value 0.? for maximum AI. Increasing the concentration of solids lowers the AI to a minimum point beyound which there is no further change upon additional loading. Studies were made on baffle design in order to limit recirculation caused by a bottom Impeller. Slanted baffle sections placed immediately above the impeller were found quite satisfactory, and their relative ease of in stallation in plant scale equipment far outweighs the slight advantage gained by placing a disk on the agitator shaft above the impeller. Using the slanted baffles, studies in*column beat distribution were made in order to see if brine colls placed at the bottom of the column would cool the whole column under agitation. The baffles caused a definite hotcold interface to be set up with a large temperature drop over a slight in terval. Therefore, with these baffles it would be possible to place cooling colls in the bottom of an autoclave and kill the reaction in this region with out endangering the reaction going on above. ucc 040Si 1 2 INTRODUCTION: In order to supplement studies already made on agitation In a tower type vessel (1) and to do special work on the problem of recirculation of particles In a liquid-solid system, Investigations were made using glass columns filled with water or water-acetone solutions and polyethylene pellets. The mixing properties of various Impellers and baffle systems were observed and described, using as a standard measurement, agitator Influence (AI), the distance from the center of the Impeller to the nominal boundary of the agi tated material. Results were compiled and are presented in this report In graphical form. PROCEDUREt A. Twenty-Four Inch Column, All experiments carried out In the 24inch column were done using polyethylene pellets (8p.gr, .913) in water. AI was measured from the center of the impeller downward to the lowest point of good pellet movement. Impeller speeds were recorded in revolutions per min ute and converted to feet per minute. B. Bottom Agitation Studies. For bottom agitation studies the 6inch diameter glass column was used. A conical section was put on the bottom. The tests were carried out using polyethylene pellets (Sp.gr, .943) in an acetone-water solution (Sp.gr, .9). AI in this case was defined as the dis tance from the lower Impeller upward to the highest point of pellet agitation. C. Heat Transfer Studies. The conical bottom was replaced by a flat bottom and heat transfer studies were made. Brine colls were placed In the bottom two inches of the 6-inch column. Hot water (60 degrees, C.) was allowed to flow into the top of the column, and the flow rate out the bottom was ad justed to keep the water level constant. Thermometers were placed at various intervals in the column in order to record the temperature drop down the column. An Impeller was placed 5-inches off the bottom of the column to pro vide agitation. The equipment used for these tests is illustrated in Figures T-A, 7-B, and 8. DISCUSSION! A. STUDIES OF SURFACE EFFECTS AND MIXING CHARACTERISTICS OF VARIOUS IMPELLERS (24-inch diameter column) 1. The effect on AI of impeller depth below the liquid surface at various peripheral speeds for the 10-inch Mlxco Impellers containing eight, six, and four blades Is shown In Figures 1-A, 1-B, and 1~C. These graphs show that at depths greater than two Inches and above peripheral speed 12? feet per minute the difference in AI achieved by changing impeller depth at constant speed is within the range of experimental error. Entrainment of the vapor phase is responsible for the odd curve produced at the two-inch depth. To avoid vapor entrainment the top impeller should be located at least four inches below the liquid level. (lj Brittain. B. F., Agitation Studies in a Tower Type Vessel. Carbide and Carbon Chemicals Company, Texas City, Texas, 1933. UCC 040912 3 2. The effect of Impeller diameter 1* shown in Figure 2. Mixeo impeller blade* (2-1/2" X 2W) were mounted on disk*, forming impeller* rang ing from six and one-half inches to fourteen and one-half Inches in diameter. The results show that there is an optimum value for maximum AI of the ratio, impeller diameter to column diameter. The high efficiency of the eleven and one-half inch diameter Impeller indicates that the ratio lies in the neighbor hood of 0.^. 3. Increasing the number of blades on the 10-inch Mlxco impeller resulted in increasing the AI proportionally. Figure 3 shows an average of the values from graphs 1-A, 1-B, and 1-C with the exception of the data taken at the two-inch depth. At the 2-inch depth the surface is less turbulent with a greater number of blade*) thus vapor entrainment Is diminished as the blade number is Increased. B. STUDY OF BOTTOM AGITATION AND RECIRCULATION CCWEROL (6-inch column) 1. Agitation in a conical bottom section was accomplished by placing small impellers within a conical glass section fastened to the bottom of the 6-inch column. A series of small disks was horizontally mounted on the rotating shaft within the conical section. This arrangement, however, was unsatisfactory due to poor mixing upon heavy loading. Observations were then made using a three-bladed, four-inch diameter turbine Impeller in the conical section. Vlth the conical section fully baffled, dead spots occurred upon loading. With no baffles there was a swirl ing effect which kept all particles moving but tended to keep Individual par ticles in the same horizontal plane. A single disk of slightly leaa diameter than that of the column was horizontally mounted on the rotating shaft from four to six inches above the impeller in the hope of limiting recirculation. Particles settled past the disk with little trouble, and recirculation was greatly reduced. The effects of the baffles were the same aa above. 2. Agitation in a flat bottom vessel was on the whole more successful and practical than that in the cone. The effect of a single rotating disk hori zontally mounted four to six inches above the bottom impeller was much the same as in the cone. When vertical baffles were placed above the rotating disk, swirling was eliminated above the disk, but pellets were driven up into the column. The impeller in the bottom of the column gave good agitation, however, even when loaded with pellets. 3. Besides the standard vertical baffles we studied other baffle designs which greatly reduced recirculation. A short vertical baffle section reduced the channeling effect caused by the long vertical baffles. Placing of a disk of slightly less diameter than that of the column below the baffle section and Immediately above the impeller reduced recirculation. It was found, however, that the extra recirculation prevention given by the disk was outweighed by the mechanical difficulties involved in installing it in plant scale equipment. Best results were obtained from a slanted baffle Beetion placed Just above the impeller. Figure 4 shows the comparison be tween the slanted and vertical baffles. The slanted baffles provided a UCC 040913 k definite limit on the AI and much more efficiently controlled the upward motion of the particle*. In a reaction Teaael having a limited settling area, a slanted baffle section would decrease chances of recirculation of solid particles. if. Studies of beat transfer in the 6-inch column showed that with adequate baffling, cooling colls in the bottom of the column would not cool the entire column even with violent bottom agitation. Figures 5-A, 5-B, 5-C, 5-D, 5-E show that for slanted baffles a temperature drop of from 25 to 30 degrees occurred over a range of only two inches. This temperature drop in dicated that a definite hot-cold interface is set up very similar to that observed as agitator Influence. A change in flow rate through the column did not affect the position of the hot-cold interface but only changed the outlet temperature depending on the amount of contact the water had with the cooling colls. The setting up of this Interface shows that it is possible to have a quench zone to kill the reaction in the bottom of a reactor and yet not affect the reaction occurring in the rest of the column. The slanted baffles deter mine the limits of this cold zone and also prevent the recirculation of re actants. C. EFFECT GF VARYING PELLET COSCEHTRATICW (6-inch column) Polyethylene pellets were continually added to a column agitated by one 3-bladed, 4-inch diameter turbine impeller. The speed and total volume were kept constant. It was found that the AI dropped off as the column was loaded. This happened probably because of frictional losses of kinetic energy which occurred at higher concentrations. Figure 6 shows the magnitude of these results. UCC 0409'H APPENDIX ucc 040915 EFFECT OF IKPELLER DEPTH OB >QITATCR mUflOBl 1ft - TTCW MTTTQ XMPEUJB 8 BUCT8 PTBCT OF IMPELLER DEPTH <M AGITATOR IHTUaaCB io meg mco ihpbxp 6 BLADES T. 50 E7PECT OF PggTfTgR DEPTH OH ADITAICR HtFUJEHCl io ncH Mnco imbbub k BLADES fl yiOURB 1 -c ucc 040918 A.I. V. PKRlfHJtKAL SPEED FOR VARIOUS DUMBER IMPELLERS urn ar vmABao warn car bums 610-001 unco nmnin. k* , 8 BUBW 16. cammiBgr of baffu gang 11. hattim Emci car heat giAHam . mTOW VW TOCKRAJUKE (30* hums) ,12 > I g 0 R 1 a -A ucc 040922 mmi etbct m nig Tmnsran ynrrmw va. mgmuaURM {SlNCttA MfLw With Turtle*! Seetle*) aj. DISTAKCB ABOVS XMPELLBH (jSCHBfl) ML^una 3 -E UCC 040923 uurrij mcT m bus traksto ', : 1 ngcm TB BMHBUKHW8 (Slak * uta Bootle# - 8 XnAoo Aon T^fV<r) DISTANCE ABOVE IMPELLER (INCHES) TIQUR8 5-0 ucc 040924 BJtfTU MT1CT (Bt MAI TRAaarfR amm va mcBUCTORB Dirt * 2 Inohet JUxsw !bgll*r (Bo Baffloa) X5. M0UR1 3-D ucc 040925 nwnrjfli Kgrriar vs ibupbbul mbbp r .-..I N tooHer Alooo (No Boftlae at Sl*fc) V Vertical + Dlik 8 Inahaa Abow Impeller (No Baffle*) A 30* Baffles a&d Sisk 8 luehe* Above Impeller 30* Baffles 8 laches Above lapeller O Slanted Baffle with Vertical Sootlon f I OUM M ucc 040926 PTBCI CJ LOADIBO OT ifllTATOW fflELUiriCl MMcnmiATiiar or teutb vh. a<i? XT. ^cc 040927 34-INCH DIAMETER GLASS nm IIMH ttmwue Aim* c2 non* 35 'M BArrLMj (twin) MCo IMAmuMl -Aueee* Qasmts GUaSS Sections DRAIN / '*%&'/&///// SUfTOAT FIGURE ?-A ucc 040928 S-t~rC fc-iJCH DIAMETER &LAS3 CO^JMu 19 comcal AerrroM nauRE 7~B ucc 040929 hLAt cU BAFh-k SEC 11C 'S 20