THE UNIVERSITY OF TULSA THE GRADUATE SCHOOL PARTICLE AND FLUID VELOCITY MEASUREMENTS FOR VISCOUS LIQUIDS IN A DIRECT IMPINGEMENT FLOW RESULTING IN MATERIAL EROSION by Stephen J. Miska A thesis submitted in partial fulfillment of the requirements for the degree of Masters of Science in the Discipline of Mechanical Engineering The Graduate School The University of Tulsa 2008THE UNIVERSITY OF TULSA THE GRADUATE SCHOOL. PARTICLE AND FLUID VELOCITY MEASUREMENTS FOR VISCOUS LIQUIDS IN A DIRECT IMPINGEMENT FLOW RESULTING IN MATERIAL EROSION by Stephen J. Miska i A THESIS APPROVED FOR THE DISCIPLINE OF MECHANICAL ENGINEERING By Thesis Committee » Chair Dr. Daniel Crunkleton iiABSTRACT Miska, Stephen J. (Master of Science in Mechanical Engineering) Particle and Fluid Velocity Measurements for Viscous Liquids in a Direct Impingement Flow Resulting in Material Erosion Directed by Professor Siamack A. Shirazi (210 pp., Chapter 6: Summary and Conchisions) (222 words) Velocity measurements by way of Laser Doppler Velocimetry (LDV} were made for particles entrained in a viscous liquid traveling in a submerged, direct impingement jet. In an attempt to obtain representative particle impact characteristics during material erosion, data was collected from the nozzle exit to the target surface in order to track icle sizes of 120 and fluid and particle velocities prior to impact with a wall. Average p: 550 jum were used to represent typical sand sizes, while much smaller particles with an average diameter of 3 jm were utilized in fluid velocity measurements. The carrier fluid viscosity was varied from | to 100 centiPoise, while the nozzle flow rate and fluid density were maintained constant, Changes in approach and estimated impingement velocity occurring due to fluid viscosity and particle size are then presented. For the same impingement geometry and flow situations, metal loss erosion ‘measurements have been made by way of an Electrical-Resistance (ER) probe. Oklahoma #1 sand particles with an average diameter of 150 um were suspended in a viscous carrier fluid at a measured sand concentration. The measured erosion rate and particle velocities iitat near target wall locations are then compared to observe the effect of viscosity on material erosion and impact speed. Particle tracking and erosion predictions made by ‘Computational Fluid Dynamics (CFD) can then be experimentally validated, ivACKNOWLEDGEMENTS ‘The author would like to thank Dr. Shirazi and Dr. McLaury for their support and guidance throughout his studies and graduate research work. Further, the author would like to thank his supportive family and friends, as well as other involved faculty, staff, and fellow graduate students for their continuous assistance during this endeavor.TABLE OF CONTENTS. Page ABSTRACT. iti ACKNOWLEDGEMENTS... v TABLE OF CONTENTS... ee a vi LIST OF TABLES... viii LIST OF FIGURES... xi CHAPTER |: INTRODUCTION 1 Erosion Prediction Overview.. 1 Research Goals and Approach 2 CHAPTER 2: LITERATURE REVIEW ...... 5 CHAPTER 3: EXPERIMENTAL FACILITY... i 12 General Description of Experimental Facilitis....... eee Flow Loop for Laser Doppler Velocimeter (LDV) Measurements 13 Laser Doppler Velocimeter Background. 17 LDV Signal Processing.s...+++ 19 Flow Loop for Erosion Rate Measurements. . 2 Electrical-Resistance (ER) Probe Background. ........cvvvvseesseeee 26 CHAPTER 4: EXPERIMENTAL FLUID AND PARTICLE VELOCITY RESULTS. 29 Introduction. #7 29 LDV Resutts....... 30 Particle Size Constant at 3 um ~ Effect of Viscosi 30 Particle Size Constant at 120 um — Effect of Viscosity . 50 Particle Size Constant at 550 ym — Effect of Viscosity . 66 Fluid Viscosity Constant at 100 CP - Effect of Particle Size...... 83 Fluid Viscosity Constant at 50 CP - Effect of Particle Siz 99 Fluid Viscosity Constant at 25 CP - Effect of Particle Siz us Fluid Viscosity Constant at 10 CP - Effect of Particle Siz 131 Fluid Viscosity Constant at 1 CP - Effect of Particle Siz@............. 146 Brief Note Regarding Measured Turbulent Intensities. 163 Brief Note Regarding LDV Measurement Uncertainty. 167 viCHAPTER 5: EXPERIMENTAL EROSION PRELIMINARY DATA. 168 ER Probe Results.. : ee : 168 Erosion Rate versus Viscosity 168 Comparison of Erosion Results to LDV Measurements .. 171 CHAPTER 6: SUMMARY AND CONCLUSIONS. .. 116 Recommended Future Work... 177 REFERENCES. 178 APPENDIX A: EXPERIMENTAL LDV DATA... 180 APPENDIX B: EXPERIMENTAL EROSION DATA.. 210 viiLIST OF TABLES Page 2-1 E/CRC Erosion Ratio Equation Empirical Constants. 3-1 Description of Viscosities for LDV Testin; 14 3.2 CMC or Glycerin Concentration Required for Viscosity Change... 25 41 Reynolds Number of Flow with Vatiable Viscosity... 42 Calculated Flow Rate from LDV Axial Vel. Profile 3 jm Measurements... 167 A-1_ Axial Vel. LDV Measurements, 3 im Aluminum, nozzle exit, x= 1 mm...... 180 A2 Axial Component LDV Measurements, 3 um Aluminum, z= 0 mm. 180 A:3_ Axial Component LDV Measurements, 3 jam Aluminum, 2= 1 meee 181 A-4 Axial Component LDV Measurements, 3 um Aluminum, z= 2mm... 181 A:S_ Axial Component LDV Measurements, 3 jam Aluminum, 2=3 mts 182 A-6 Axial Component LDV Measurements, 3 am Aluminum, 2= 4 mm... 182 A-7 Axial Component LDV Measurements, 3 jm Aluminum, z= 6 mm. 183 A-8 Axial Component LDV Measurements, 3 um Aluminum, z= 8 mm... 183 A-9 Axial Component LDV Measurements, 3 jm Aluminum, z O MM veces 184 A-10 Axial Component LDV Measurements, 3 pm Aluminum, z= 12 mm ssn A-L1 Radial Component LDV Measurements, 3 ym Aluminum, z= 0 mm... 185 A-12 Radial Component LDV Measurements, 3 am Aluminum, z= 1 mm... 185 ‘A-13 Radial Component LDV Measurements, 3 um Aluminum, 2= 2 MMos. 186 A-14 Radial Component LDV Measurements, 3 um Aluminum, 2=3 MMos 186 A-LS Radial Component LDV Measurements, 3 jum Aluminum, 2= 4 mm. 187 viiiAcl6 Radial Component LDV Measurements, 3 um Aluminum, z= 6 mm... A-17 Radial Component LDV Measurements, 3 pm Aluminum, 2= 8 mM... 188 A-18 Radial Component LDV Measurements, 3 jum Aluminum, z= 10 mm, 188 A-19 Radial Component LDV Measurements, 3 gm Aluminum, 2= 12 mm... 189 ‘A-20 Axial Vel. LDV Measurements, 120 ym Aluminum, nozzle exit, x= 1mm... 189 A-21 Axial Component LDV Measurements, 120 pm Aluminum, 2=0 mm... 190 ‘A-22 Axial Component LDV Measurements, 120 um Aluminum, z= 1 mm........ 190 A-23 Axial Component LDV Measurements, 120 um Aluminum, 2=2 mmo. 191 A-24 Axial Component LDV Measurements, 120 um Aluminum, z=3 mm........ 191 A-25 Axial Component LDV Measurements, 120 um Aluminum, 2=4 MMe 192 A-26 Axial Component LDV Measurements, 120 um Aluminum, 2= 6 mm... 192 ‘A-27 Axial Component LDV Measurements, 120 pm Aluminum, 2=8 mm ue 193 A.28 Axial Component LDV Measurements, 120 pm Aluminum, z= 10 mm ....... 193 A-29 Axial Component LDV Measurements, 120 pm Aluminum, z= 12mm... 194 A-30 Radial Component LDV Measurements, 120 pm Aluminum, 2=0 mM. 194 A-31 Radial Component LDV Measurements, 120 pm Aluminum, = 1 mm..oou 195 A:32 Radial Component LDV Measurements, 120 um Aluminum, 2=2 MM. 195 A-33 Radial Component LDV Measurements, 120 pm Aluminum, z=3 mm...... 196 A-34 Radial Component LDV Measurements, 120 um Aluminum, 2=4 mm... 196 ‘A-35 Radial Component [LDV Measurements, 120 jum Aluminum, z= 6mm... 197 A-36 Radial Component LDV Measurements, 120 jum Aluminum, z= 8mm... 197 A-37 Radial Component LDV Measurements, 120 um Aluminum, z= 10 mm....... 198 A-38 Radial Component LDV Measurements, 120 jm Aluminum, z= 12 mm....... 198 ixA-39 Axial Vel. LDV Measurements, 550 jum Aluminum, nozzle exit, x= | mm... 199 A-40 Axial Component LDV Measurements, 550 pm Aluminum, = 0mm... 199 A-41 Axial Component LDV Measurements, 550 pm Aluminum, z= 1 mm........ 200 A-42 Axial Component LDV Measurements, 550 pm Aluminum, z= 2 mm... 200 A-43 Axial Component LDV Measurements, 50 pm Aluminum, 2=3 mM se 201 A-44 Axial Component LDV Measurements, 550 pm Aluminum, = 4 mm eve: 201 A-45 Axial Component LDV Measurements, 550 pm Aluminum, 2 = 6 mm we... 202 A-46 Axial Component LDV Measurements, 550 im Aluminum, 2=8 mm... 202 A-47 Axial Component LDV Measurements, 550 pm Aluminum, z= 10mm... 203 A-48 Axial Component LDV Measurements, $50 im Aluminum, z= 12 mm ...... 203 A-49 Radial Component LDV Measurements, $50 wm Aluminum, 2 = 0 mm... 204 A-50 Radial Component LDV Measurements, 550 pm Aluminum, 2= 1mm... 204 A-SI Radial Component LDV Measurements, 550 jm Aluminum, 2=2 mm... 205 ‘A-52 Radial Component LDV Measurements, 550 wm Aluminum, = 3 mm... 205 A-53 Radial Component LDV Measurements, 550 um Aluminum, z= 4 mm........ 206 A-54 Radial Component LDV Measurements, 550 wm Aluminum, 2= 6 mit... 206 A-55 Radial Component LDV Measurements, 550 um Aluminum, z= 8 mm........ 207 A-56 Radial Component LDV Measurements, 550 um Aluminum, z= 10 mm....... 207 A-57 Radial Component LDV Measurements, 550 jum Aluminum, z= 12 mm....... 208 A-58-59 LDV Frequency Settings for Axial Velocity Measurements jose 208 B-I Measured Erosion Rate for OK #1 Sand (dyyg = 150 um) in CMC-Water 0... 210 B-2 Measured Erosion Rate for CA 60 Sand (dy. = 300 um) in CMC-Water ........ 210 B-3 Measured Erosion Rate for CA 60 Sand (djyg = 300 pm) in Glycerin-Water .... 210LIST OF FIGURES Page 1-1 Process to Predict Erosion Rates... 3 341 Schematic of LDV Flow Loop .... 13 3-2 Photograph of LDV Flow Loop... 13 3-3 Microscopic Image of 550 jm Aluminum Particles .. 1 1S 3-4 Microscopic Image of 120 pm Aluminum Particles ....... 15 3-5 Microscopic Image of 3 ym Aluminum Particles. Is 3-6 Map of LDV Velocity Measurement Locations... 3-7 LDV Measurement Process... 7 3-8 Cross-section of Intersecting Light Beams and Fringe Pattern. 19 3.9 Determination of the Half-Angle, x... . 20 3-10 Erosion Flow Loop Schematic . 3-11 Microscopic Photograph of Oklahoma #1 Sand....conussennnennnnnnren 23 3-12 Microscopic Photograph of California 60 Sand... 3-13 Oklahoma #1 Sand Size Distribution «ccc eee 23 3-14 Target Surface of Cormon ER Probe... 3-15 Sample Cormon Probé Software Output... 4-1 Axial Velocity Profile near Nozzle Exit for 3 um Particles. 4-2, Axial Velocity versus Distance from Nozzle Exit at 2-0 mm for 3 jim Particles. 32] 4-3. Axial Velocity versus Distance from Nozzle Exit at z=1 mm for 3 ym Particles... 1 33 xi4-4 Axial Velocity versus Distance from Nozzle Exit at for 3 ym Particles. mm 34 4-5 Axial Velocity versus Distance from Nozzle Exit at for 3 uum Particles. 34 4-6 Axial Velocity versus Distance from Nozzle Exit at for 3 um Particles. 4-7 Radial Velocity versus Distance from Nozzle Exit at 2-2 mm for 3 um Particles. 37 4-8 Radial Velocity versus Distance from Nozzle Exit at z=4 mm. for 3 ym Particles. 4-9 Radial Velocity versus Distance from Nozzle Exit at z=6 mm for 3 ym Particles. 4-10 Radial Velocity versus Distance from Nozzle Exit at 2-8 mm for 3 yum Particles... nse ee 40 4-11 Radial Velocity versus Distance from Nozzle Exit at 2-10 mm for 3 um Particles svnens 40 4-12 Radial Velocity versus Distance from Nozzle Exit at for3 um Particles. 4-13. Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from ‘Wall for 3 jim Particles 4-14 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 3 jum Particle : 4-15 Near-Wall Radial Velocities at 2=3 mm for 3 jum Particles... 4-16 Near-Wall Radial Velocities at 2=10 mm for 3 ym Particles... 4-17 Axial Velocity Contour in m/s at Fluid Viscosity = 100 CP for 3 ym Particles... 4-18 Radial Velocity Contour in m/s at Fluid Viscosity = = 100 CP for 3 pum Particles...cnseesnsnr eae 47 100 CP 4-19 Speed Contour in m/s at Fluid Viscosity for 3 um Particles, xii4-20 Speed Contour in m/s at Fluid Viscosity for 3 pm Particles..rnnn 21 Speed Contour in mvs at Fluid Viscosit for 3 yum Particles..vessennenintine 4-22 Speed Contour in m/s at Fluid Viscosity = 10 CP for 3 jum Particles...cesinmesineneninntnntnienteniesneeeseii 4-23 Speed Contour in m/s at Fluid Viscosity ~ 1 CP for 3 ym Particles: 4-24. Axial Velocity Profile near Nozzle Exit for 120 jum Particles... 4-25. Axial Velocity versus Distance from Nozzte Exit at 2-0 mm for 120 jum Particles. 4-26 Axial Velocity versus Distance from Nozzle Exit at z=] mm for 120 pm Particle 4-27 Axial Velocity versus Distance from Nozzle Exit at 2-2 mm for 120 um Particles... 4-28 Axial Velocity versus Distance from Nozzle Exit at 2-3 mm for 120 um Particles... 4-29 Axial Velocity versus Distance from Nozzle Exit at 2=4 mm for 120 jim Particles... . 35 4-30 Radial Velocity versus Distance from Nozzle Exit at 2=2 mm for 120 sm Particles. ae erate 56 4-31 Radial Velocity versus Distance from Nozzle Exit at 2=4 mm. for 120 um Particle: 4-32. Radial Velocity versus Distance from Nozzle Exit at z=6 mm. for 120 um Particle: 37 4-33. Radial Velocity versus Distance from Nozzle Exit at 28 mm. for 120 um Particles ....cscosecneee 58 4-34 Radial Velocity versus Distance from Nozzle Exit at z=10 mm for 120 um Particles . 58 4.35 Radial Velocity versus Distance from Nozzle Exit at z=12 mm for 120 um Particles... . 39 xiii436 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from Wall for 120 im Particles... 60 4.37 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 120 yim Particles... 61 438 Near-Wall Radial Velocities at 2=3 mm for 120 um Particles... 4.39 Near-Wall Radial Velocities at 2-10 mm for 120 pum Particles....sornen 63 4-40. Speed Contour in m/s at Fluid Viscosity = 100 CP for 120 wim Particles. .. 68 4.41 Speed Contour in m/s at Fluid Viscosity = 50 CP for 120 pm Particles... : 64 4-42 Speed Contour in m/s at Fluid Viscosity = 25 CP for 120 jum Particles ..cnnnees 443. Speed Contour in m/s at Fluid Viscosity = 10 CP for 120 pm Particle... 4444 Speed Contour nls at Fi Viscosity = 1 CP for 120 ym Particle... eu eee 65 445 Axial Velocity Profile near Nozzle Exit for 550 um Particles... 4-46 Axial Velocity versus Distance from Nozzle Exit at for 550 wm Particles... mm 4.47 Axial Velocity versus Distance from Nozzle Exit at z=1 mm. for 550 wm Particles... . 68 4-48 Axial Velocity versus Distance from Nozzle Exit at for 550 ym Particles... revaeaes 69) 4-49 Axial Velocity versus Distance from Nozzle Exit at for 550 wm Particles. 4-50 Axial Velocity versus Distance from Nozzle Exit at 2=4 mm for 550 pum Patticles.cesocnn 4-51 Radial Velocity versus Distance from Nozzle Exit at 2=2 mm for 550 pm Particles... seounscanee 72 xivI t i 4-52 Radial Velocity versus Distance from Nozzle Exit at 2-4 mm for 550 um Particles... 73 4-53 Radial Velocity versus Distance from Nozzle Exit at 2-6 mm for 550 um Particles.. ae 3 4-54 Radial Velocity versus Distance from Nozzle Exit at for 550 um Particles. 14 4.55 Radial Velocity versus Distance from Nozzle Exit at 2=10 mm for 550 um Particles.. 4-56 Radial Velocity versus Distance from Nozzle Exit at 2=12 mm, for 550 jum Particles 4.57 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from Wall for 550 jum Particles... 4-58 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 550 um Particles 7 4-59 Near-Wall Radial Velocities at z=3 mm for 550 jum Particles.....c.0:acnenne 78 4-60 Near-Wall Radial Velocities at 2=10 mm for 550 ym Particles..... 79 4-61 Speed Contour in m/s at Fluid Viscosity = 100 CP for 550 jim Particles. Sia enae 4-62 Speed Contour in m/s at Fluid Viscosity = 50 CP for 550 im Particles. retaeteeaeeea BL 4-63 Speed Contour in m/s at Fluid Viscosity for 550 um Particles. 5 CP 4-64 Speed Contour in m/s at Fluid Viscosity = 10 CP for 550 pm Particles. 4-65 Speed Contour in m/s at Fluid Viscosity =1¢P for $50 pm Particles nner ee 82 4-66 Axial Velocity Profile near Nozzle Exit for 100 CP Fluid. 4-67 Axial Velocity versus Distance from Nozzle Exit at z=0 mm for 100 CP Fluid... 84 cog4-68 Axial Velocity versus Distance from Nozzle Exit at z=1 mm for 100 CP Flt. see ctceeneneeneeee iseetectospeeeeeeeestseteateat 4.69 Axial Velocity versus Distance ftom Nozzle Exit at 2-2 mm, for 100 CP Fluid... 4-10 Axial Velocity versus Distance from Nozzle Exit at 2-3 mm for 100 CP Fluid... 4-11 Axial Velocity versus Distance from Nozzle Exit at for 100 CP Fluid. 88 4-12 Radial Velocity versus Distance from Nozzle Exit at ==2 mm for 100 CP Fluid... 4-13 Radial Velocity versus Distance from Nozzle Exit at z=4 mm for 100 CP Fluid... a1 4-14 Radial Velocity versus Distance from Nozzle Exit at 2=6 mm for 100 CP Fluid. 4-15 Radial Velocity versus Distance from Nozzle Exit at 2=8 mm for 100 CP Fluid...... 4-16 Radial Velocity versus Distance from Nozzle Exit at 2=10 mm for 100 CP Fluid... 4-77 Radial Velocity versus Distance from Nozzle Exit at 2=12 mm for 100 CP Fluid... aa 93 4-78 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from ‘Wall for 100 CP Fluid... 4-19 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from. ‘Wall for 100 CP Fluid... 4-80. Near-Wall Radial Velocities at z=3 mm for 100 CP Fluid 4-81 Near-Wall Radial Velocities at 2=10 mm for 100 CP Fluid. 98 4-82. Axial Velocity Profile near Nozzle Exit for 50 CP Fluid. -- 100 4-83. Axial Velocity versus Distance from Nozzle Exit at 2-0 mm for 50 CP Fluid .. 101 xvi4-84 Axial Velocity versus Distance from Nozzle Exit at for 50 CP Fluid .. 102 4-85 Axial Velocity versus Distance from Nozzle Exit at z=2 mm for 50 CP Fluid .. . 103 4-85 Axial Velocity versus Distance from Nozzle Exit at for 50 CP Fluid .. 103 4-87 Axial Velocity versus Distance from Nozzle Exit at for 50 CP Fluid ... . 104 4-88 Radial Velocity versus Distance from Nozzle Exit at 2=2 mm for 50 CP Fluid .. . 106 4-89 Radial Velocity versus Distance from Nozzle Exit at 2=4 mm for 50 CP Fluid ae . 106 4-90 Radial Velocity versus Distance from Nozzle Exit at 26 mm. for 50 CP Fluid . 108 4-91 Radial Velocity versus Distance from Nozzle Exit at 7=8 mm for 50 CP Fuid ... 108 4.92. Radial Velocity versus Distance from Nozzle Exit at z=10 mm. for 50 CP Fluid ... . 109 4-93 Radial Velocity versus Distance from Nozzle Exit at 2=12 mm for 50 CP Flu . 110 4-94 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from Wall for 50 CP Flui 1 4.95 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 50 CP Fluid..... 12 4-96 Near-Wall Radial Velocities at z=3 mm for 50 CP Fluid 113 4-97 Near-Wall Radial Velocities at 2=10 mm for 50 CP Fluid... eid 4-98 Axial Velocity Profile near Nozzle Exit for 25 CP Flui . 116 4-99 Axial Velocity versus Distance from Nozzle Bxit at 7-0 mm for 25 CP Fluid .... 117 xvii4-100 Axial Velocity versus Distance from Nozzle Exit at z=1 mm. for 25 CP Fluid .. ae fete 118 4-101 Axial Velocity versus Distance from Nozzle Exit at z~2 mun for 25 CP Fluid .... 7 oa avesistesteteceeseets EC) 4-102 Axial Velocity versus Distance from Nozzle Exit at 2-3 mm for 25 CP Fld ...ccccnsennsennnniennnnsintnennsteis 120 4-103. Axial Velocity versus Distance from Nozzle Exit at 4 mm for 25 CP Fltid wasssnsnesienenetnninninannninsniennesis 121 4-104 Radial Velocity versus Distance from Nozzle Exit at z=2 mm for 25 CP Fluid. .cccsiecennenneinstnentinenintntete eeesney) 4-105 Radial’ Velocity versus Distance from Nozzle Exit at z=4 mm for 25 CP Fluid ... 123 4-106 Radial Velocity versus Distance from Nozzle Exit at 2=6 mm for 25 CP Fluid .. . 124 4-107 Radial Velocity versus Distance from Nozzle Exit at z=8 mm for 25 CP Fluid ae eraser 25) 4-108 Radial Velocity versus Distance from Nozzle Exit at z=10 mm for 25 CP Fluid....... 7 essen . 125 4-109 Radial Velocity versus Distance from Nozzle Exit at 2=12 mm. for 25 CP Fluid . 126 4-110 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm fom, Wall for 25 CP Fluid.. . 127 4-I11 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm ftom Wall for 25 CP Fluid.. — . 128 4-112 Near-Wall Radial Velocities at z=3 mm for 25 CP Fluid parent 129) 4-113 Near-Wall Radial Velocities at 2=10 mm for 25 CP Fluid 130 4-114 Axial Velocity Profile near Nozzle Exit for 10 CP Fluid.. . 132 4-115 Axial Velocity versus Distance from Nozzle Exit at 2-0 mm for 10 CP Fluid 0. = . 133 xviii4-116 Axial Velocity versus Distance from Nozzle Exit at ==] mm. for 10 CP Fluid . iiiseaieseataiaeieauenaicteenentin 134 4-117 Axial Velocity versus Distance from Nozzle Exit at 2=2 mm for 10 CP Fluid ee 135 4-118 Axial Velocity versus Distance from Nozzle Exit at 2=3 mm for 10 CP Fluid ... : sevseusennene 136 4-19 Axial Velocity versus Distance from Nozzle Exit at z=4 mm for 10 CP Fluid sseunsmnessensneene 137 4-120 Radial Velocity versus Distance from Nozzle Exit at 2-2 mm for 10 CP Fluid .nsscsnsnnsiesisiennnnaisnecinensi one eesncnee 138 4-121 Radial Velocity versus Distance from Nozzle Exit at z=4 mm for 10 CP Fluid .. eententesnen . 139 4122 Radial Velocity versus Distance from Nozzle Exit at 2=6 mm for 10 CP Fluid..... . 140 4-123 Radial Velocity versus Distance from Nozzle Exit at 2=8 mm for 10 CP Fluid..... 140 4-124 Radial Velocity versus Distance from Nozzle Exit at for 10 CP Fluid . 7 7 ssn . 141 4-125 Radial Velocity versus Distance from Nozzle Exit at for 10 CP Fluid .. Sienna . 141 4-126 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm fiom Wall for 10 CP Fluid... . 143 4-127 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 10 CP Fluid. . 144 4-128 Near-Wall Radial Velocities at 2-3 mm for 10 CP Fluid . 145 4-129 Near-Wall Radial Velocities at 2=10 mm for 10 CP Fluid 146 4-130 Axial Velocity Profile near Nozzle Exit for 1 CP Fluid. . 147 4-131 Axial Velocity versus Distance from Nozzle Exit at 2=0 mm. for | CP Fluid . 148 xix4-132 Axial Velocity versus Distance from Nozzle Exit at 2=1 mm for | CP Fluid .. 149) 4.133 Axial Velocity versus Distance from Nozzle Exit at 7=2 mm_ for 1 CP Fluid 4-134 Axial Velocity versus Distance from Nozzle Exit at z=3 mm for 1 CP Fluid 151 4-135 Axial Velocity versus for 1 CP Fluid ... ee eee : tee oe 4-136 Radial Velocity versus Distance from Nozzle Exit at 2=: for 1 CP Fluid..... esseeensnseenn wee 153 4-137 Radial Velocity versus Distance from Nozzle Exit at =4 mm for 1 CP Fluid 154 4-138 Radial Velocity versus Distance from Nozzle Exit at 2=¢ for 1 CP Fluid .. . 155 4-139 Radial Velocity versus Distance from Nozzle Exit at 2=% for 1 CP Fluid .. eerie 156 4-140 Radial Velocity versus Distance from Nozzle Exit at 2-10 mm for | CP Fluid...... sees 156 4-141 Radial Velocity versus Distance from Nozzle Exit at 2=1 for | CP Fluid... is7 4-142 Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from Wall for 1 CP Fluid. 158 4-143 Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 1 CP Fluid. 159 4-144. Near-Wall Radial Velocities at 2=3 mm for 1 CP Fluid ...cccnenennennees 161 4-145 Near-Wall Radial Velocities at 2=10 mm for { CP Fluid .. . 162 4-146 Axial Turbulent Intensity versus Distance from Nozzle Exit at z=0 mm. ee feeders 164 4-147 Radial Turbulent Intensity versus Distance from No: at 0.3 mm from Wall. . 1654-148 Radial Turbulent Intensity versus Distance from Nozzle Centerline at 0.3 mm from Wall for 25 CP Fluid. 166 5-1 Erosion Rate for 300 and 150 um Particles for Different Fluid Viscosities....... 169 5-2 Brosion Rate for 300 jum Particles using CMC and Glycerin to Change Fluid Viscosity... - 170 5-3. Average Radial Velocity 0.3 mm from Wall and Erosion Rate Erosion Rate versus VisCOSity ..ccsnsnnnnsnnensnnennnsn 5-4 Average Radial Velocity 0.6 mm from Wall and Erosion Rate Erosion Rate versus Viscosity :CHAPTER | INTRODUCTION Erosion Prediction Overview In the production and pipeline transport of various fluids, such as oil and natural gas, solid particles may be entrained in the fluid. These particles, commonly consisting ‘of numerous types and sizes of sand, can travel apart from the streamlines of the fluid and impact the surface of the pipe. With time, enough particles may impinge a pipe wall at a sensitive location, such as an elbow or tee, to result in a measurable wall thickness loss. ‘This may ultimately lead to severe erosion damage causing a leak in a pipeline, a dangerous and costly problem. As a result, a pipeline’s service life may often depend on the rate at which a pipe wall is eroded The erosion rate, or amount of material loss over a certain time period, depends ona large number of factors. The target material, or material experiencing « thickness Joss, such as a pipe wall, influences the rate at which damage occurs. Its density, hardness, yield strength, and microstructure combine to present a certain resistance toward erosion occurring from solid particle impact. Furthermore, the solid particle’s diameter, sharpness, and shape will influence its trajectory, speed, and momentum transfer into the target, thereby requiring the analysis of various particle types in predicting erosion. Finally, the carrier fluid being transported through a pipeline will further affect the solid particle’s movement as it approaches the target. As a result, the fluid’s density and viscosity must be carefully considered in particle tracking and erosionanalysis. By considering the aforementioned properties of the target, solid particles, and carrier fluid, it is desirable to be able to predict the erosion rate from a single erosion equation, Other factors depending on these properties may be found in this expression, such as particle impact speed and impingement angle at the target. An experimental study of erosion may be performed by measuring actual target ‘material losses over time for a given geometry and flow situation, In addition, particle ‘tacking, or measuring a particle's velocity, angle, and location as it approaches and impinges a target, is also possible by experiment. At the same time, computational fluid dynamics (CFD) can use transport phenomena theory and computer-based numerical techniques to predict the particle trajectory, including particle impact velocity, for many geometries and flow cases. By incorporating current erosion equations based on ‘empirical testing, CFD can then also provide erosion predictions for the geometry of interest. Research Goals and Approach Research work at the Erosion/Corrosion Research Center (E/CRC) at The University of Tulsa has been directed at improving an erosion ratio equation by way of a combination of experimental and computational work. More specifically, the direct jet impingement of solid particles released from a fluid submerged nozzle has been carefully studied, Such a flow situation allows for solid particles to travel through a carrier fluid moving with a controlled flow rate prior to being released through a nozzle. At the nozzle exit, solid particles will travel through a certain length prior to impacting a target. When properly studied and evaluated, this flow scenario can be related to erosion in anothercommon geometry, such as a pipe elbow or tee. Figure 1-1 below shows the process used to predict the erosion rate in CFD for a given geometry and the steps needing experimental validation, Create Geometry, Mesh, and Define Flow Flow Solution Validate |* Particle Tracking *——{ Validate |* 4 * Erosion Equations Evaluate i Predict Erosion Rate Figure 1-1 — Process to Predict Erosion Rates As scen above, once the desired geometry is meshed and flow is defined in CFD, the computational simulation will provide a flow solution from which particle tracking of the solid particles is predicted. As mentioned previously, particle tracking involves predicting a solid particle’s speed and angle at any given location in the stagnation length ofa certain geometry. At the end of this distance is located the target where impact speeds and impingement angles are found. CFD utilizes this information along with erosion equations incorporated into the CFD code to predict the erosion rate of the target. An experimental facility with an identical geometry and flow can then be designed. Particle velocities are then measured in the stagnation length between a nozzle and target in order to examine particle behavior before and at impact. The particle tracking predicted by CFD can then be compared to experimentally measured values inan attempt to validate and improve CED results. At the same time, actual metal losses resulting from solid particle erosion in the same geometry and flow situation can be experimentally determined in an attempt to evaluate and make adjustments to erosion equations. Once CFD particle tracking is fully validated and the erosion equations are experimentally evaluated, erosion rate predictions for any geometry can be accurately made. The research to be discussed involves measuring the effect of the carrier fuid viscosity on solid particle tracking and erosion rate for the direct impingement geometry escribed thus far. As a result, particle velocity, including near wall velocity, has been, measured and can be compared to CFD predictions. Different fluid viscosities are used and Laser Doppler Velocimetry (LDV) is utilized to measure particle and fluid speeds ‘The velocity of particles near the wall is examined for different carrier fluids in order to explain erosion rates that occur as a result of particle impact.CHAPTER 2 LITERATURE REVIEW Erosion resulting fiom solid particle impact on a material has been studied by numerous researchers using Laser Doppler Velocimetry (LDV) particle tracking, direct ‘metal loss measurements with erosion probes, and theory-driven computational fluid dynamics (CED). A summary of some of these studies will be presented in this literature review. Meng and Ludema (1995) investigated various erosion models and equations Which have been used to make erosion predictions in the past. A conclusion from this research indicated that a single comprehensive model cannot be practically established for general use due to the complexity of the factors influencing erosion. Since over one hundred various parameters and factors have been suggested to affect material erosion and many different erosion mechanisms have been proposed, a single mode! explicitly using all necessary factors and variables has yet to be produced. Common factors, however, such as particle velocity, particle size, particle shape, target material hardness, target yield strength, carrier flow velocity, carrier fluid density, and carrier duid viscosity, have consistently been cited as key variables dictating erosion. Bitter (1963) suggested that erosion occurred simultaneously by two separate ‘material wear mechanisms. ‘These included deformation wear, which occurs due to ‘material cracking instigated by the normal velocity component of impinging particles, and cutting wear, which occurs when particles impact at angles parallel to the surface,thus causing scraping of the material, It was further proposed that erosion occurs as a sum. of these two wear mechanisms, thus requiring separate equations to predict deformation ercsion and cutting erosion. This called for the need to track particles in two mutually- nomal directions, ‘The erosion equations that are used by the E/CRC (Erosion-Corrosion Research Center) at The University of Tulsa were developed for many different materials including carbon steel (Ahlert, 1994) and aluminum (Mclaury, 1996). These equations were largely based on direct impingement testing in air and liquid and are a function of particle impact velocity and angle as well as hardness of the target material. The general form of this erosion ratio equation is shown below in Equation 2-1, ER=A*R*V"*FQ) 1) where ER is the erosion ratio (material mass loss/mass of impacting particles), A and n are empirical constants, V is the particle impact velocity, Fs is a particle shape factor, and F(@) isa function of the impingement angle which is presented in Equation 2-2, ro)-{ a0? +bo (@s6,) 2 x cos’@ sin(w 6) + ysin70+2z, (0>6,) where 0 is the particle impingement angle, @,, a, b, w, x, y, z are empirical constants and depend on the material of the target. The empirical constants above were determined for carbon stee!] and aluminum and are can be seen in Table 2-1 on the following page.Table 2-1 - E/CRC Erosion Ratio Equation Empirical Constants, Empirical ‘Target Material Constants Carbon Steel ‘Aluminum = 10 %BE? ; A fn tase ite euedaeenti a 15° 1° a -38.4 34.79 eae eae aceeence007 123 w 1 5.205 bl 3.147 0.147 y 0.3609 0.745 z 2.532 1 n 1.73 1.73 It is of value to note that carrier fluid properties, such as fluid density or viscosity, are not explicitly present in the F/CRC erosion ratio equation, Rather, itis assumed that factors such as viscosity may affect the particle impact velocity and its exponent. Furthermore, it is possible that the empirical factors affecting the particle impact function are dependent on the carrier fluid properties and particle size. Another commonly used erosion equation presented by Oka et al. (2005) is founded on the mechanical property of material hardness and is generally more comprehensive than the E/CRC erosion equation. The erosion ratio equation developed by Oka et al. is given below in Equation 2-3. e m= anst"-(¥) -(3) *p*T(0) ¢ where ER is the erosion ratio (material mass loss/mass of impacting particles), Hv is the material Vickers hardness, and 0 is the material density (kg/m’), Vis the particle impact velocity (nv/s), D is the particle diameter (1m), and V’ and D’ are the impact velocity andparticle diameter used during the experiments (104 m/s and 326 ym, respectively). K, kl, and 43 are a constant and exponents which depend on the properties of the impacting particles, and &2 is an exponent found to be dependant on both the material hardness and particle properties. ¥(@) is again an impingement angle function which is seen in Equation 2-4, F@)= (Sin )"1+ Hol -sin ay" (2-4) where Hv is the material kers hardness, and n/ and n2 are empirical exponents which ‘were found to be dependant on the hardness of the material and properties of the impacting particles. As with the E/CRC erosion ratio equation, fluid properties such as viscosity are assumed to affect the particle impact velocity and angle, and must be found experimentally or computationally. Zhang et al. (2007) presented Laser-Doppler Velocimeter (LDV) data for the measured velocities of particles entrained in water approaching a target in a direct impingement geometry. Furthermore, CFD simulations of the experiments were performed, with computational predictions matching experimental particle tracking results very well. Erosion rate measurements were also made for this geometry by way of anelectrical-resistance (ER) probe during various flow conditions using air or water as the carrier fluid. Again, good agreement between the erosion data and CFD predictions ‘was noted when a newly modified erosion equation incorporating E/CRC’s and Oka’s models was utilized. Clark (1992) suggested that experimental data on erosion rates for a given material can be predicted only if representative particle impact velocities and trajectories are known. Liquid viscosity and density, as well as particle impingement velocity, angle ofimpact, and size range, were listed as key variables in laboratory testing, It was further concluded that forces acting on a particle close to a target surface must be further investigated by way of using analytical models of particle trajectories combined with ‘experimental erosion test methods. Earlier, Clark (1991) performed experiments to measure the impact rate and energy of particles in a slurry pot erosion tester. For these tests, erosion was estimated by the size of impact craters caused by glass bead particles on a copper rod terget for a water- glycerin carrier fluid of varying viscosity. Using this method, the calculateé particle impact speed decreased significantly with increasing carrier liquid viscosity. It was found that increasing viscosity and decreasing particle size resulted in a reduction in erosion rates due to the flow interaction of the fluid, particles, and eroding surface. Wong (1993) presented an analytical model to quantitatively describe particle trajectories and impact velocities on a cylindrical erosion target. This model considered cfiets influenced by the carrier fluid properties, such as drag, inertia and pressure variation acting on a particle approaching the target surface. When compared to and ‘experimental data gathered by Clark (1991), the effects of change of liquid vise particle size were predicted accurately by the model. At the same time, the model predicted the formation of a sliding bed of particles along the target surface for lower impact velocities, thus suggesting the potential for erosion caused by abrasion or deformation wear. ‘Niu (2001) presented a numerical approach for the time-averaged Navier-Stokes equations incorporating a turbulence model in order to evaluate particle trajectories and near-wall interactions in an erosion situation, The study and resulting computationsconsidered flow turbulence intensity, particle size, and impact velocity effects on erosion, Numerical predictions of erosion ultimately validated experimental data. It was concluded that erosion decreases under higher inlet turbulent intensities and the location of maximum erosion significantly depends on the amount of turbulence in the fluid. Furthermore, it was displayed that erosion was most pronounced for particles having high inertia and that large erosion rates were exhibited for larger abrasives. In a related numerical study of erosion by particle impact, Niu (2000) showed by way of computational results that turbulent intensity, particle size, inlet flow velocity, and temperature all have a significant influence on erosion. It was displayed tha: the location of maximum erosion moved closer toward the stagnation point, which is located at the center of an impinging jet, as turbulent intensities in the flow were increased. At the ssame time, it was found that this maximum point of erosion was displaced further away from the stagnation point as inlet jet velocity was decreased. Carrier fluid properties, such as viscosity or density, are known to potentially affect inlet jet velocities and turbulent intensities, Similar studies performed earlier by Dosanjh (1985) revealed comparable erosion predictions but expressed mode! limitations caused by the lack of suitable experimental data needed for validation. Laser-Doppler Velocimetry (LDV) has successfully been used in past experimental studiés to measure fluid and particle velocities. Barata (2004) used LDV to measure the flow field resulting, from the direct impingement of an axisymmetric jet on a solid target after encountering a confined cross flow. During this research, particles as small as 1-4 im were used to accurately measure fluid flow in turbulent gas jets. Sasaki (1980) used LDV to measure the falling velocity and size of large rain droplets ranging in size from 100.1 to 1.0 mm. Ancimer (1999) presented LDV measurements of seeded particles in a spark ignition engine and was able to obtain valid velocity measurements with sufficient frequency shifts and high data rates. Phares (2000) provided a method to theoretically determine wall shear stress at the target of a normally impinging, submerged jet carrying an incompressible fluid, During this study, theoretical predictions were compared to available wall shear stress ‘measurements made within the impingement region. ‘The resulting analysis showed that the most accurate indirect measurements along the impact surface were found using an electrochemical method. Further experimental studies involving direct metaods of velocity measurement for an impinging jet containing a carrier liquid of variable viscosity have not been found. uCHAPTER 3 EXPERIMENTAL FACILITIES General Description of Experimental Facilities ‘Two experimental facilities have been constructed to measure solid particle velocities and erosion rates, Both facilities are flow loops designed to te-circulate a liquid of variable viscosity at or below 8.0 gallons per minute, resulting in a maximum fluid velocity of approximately 10.5 m/s. Particles can be readily added and suspended in the liquid, which is pumped at the desired flow rate through a nozzle positioned directly ata target. Nozzle inner diameters in both facilities are 8 mm or 5/16 inch. For all tests, the nozzle exit is submerged in the liquid and located 12.7 mm or 0.5 inch eway from the solid target. This distance scrves as the stagnation length during which particle tracking can be observed. One flow loop is used primarily for particle velocity measurements with Laser Doppler Velocimetry (LDV). The other experimental facility is utilized for erosion rate or metal loss measurements with an electrical-resistance (ER) probe Preliminary erosion measurements were performed by Okita (2008) using this flow loop. ‘These two experiments will ultimately make it possible to relate particle impact speed to erosion rate for various fluid viscosities and particle sizes used in the facilities, The following sections describe both flow loops in greater detail, as well as provide background on ER probes and LDV. 2Flow Loop for Laser Doppler Velcimeter (LDV) Measurements In order to obtain solid particle velocity measurements using Laser Doppler Velocimetry (LDV), a flow loop such as the one seen in the schematic of Figures 3-1 and photograph of Figure 3-2 was constructed. | Bypsss valve Nome Tenge (LOW rmess;) Shay mise Vices \ rue Figure 3-1 — Schematic of LDV Flow Loop Figure 3-2 — Photograph of LDV Flow Loop 13The entire system is filled with at least 70-75 liters of a fluid having a desired viscosity achieved by mixing appropriate amounts of water and glycerin. The tested viscosities and comesponding mixtures at room temperature are described in Table 3-1 (Nomenclature: centiPoise — CP). Table 3-1 — Description of Viscosities for LDV Testing Fluid Viscosity (CP) ‘Amount of Water (%) Amount of Glycerin (%) 1 100 0 10 50 50 25 33 7 30 22 78 100 13 87 ‘The upper tank (or testing tank) contains around 45 liters of the liquid so as to maintain a water level which ensures a submerged jet af the nozzle exit. This testing tank is made of transparent acrylic so LDV measurements may be taken between the nozzle exit and the solid target. About 50 grams of highly reflective aluminum particles with an average diameter of 3, 120, or 550 jm are added to the liquid in the reservoir tank and suspended in the fluid by way of slurry mixer. These particles have a density of approximately 2,650 kg/m? (about the same as the density of OK #1 sand commonly used in industry). Microscopic photographs of these particles can be seen in Figures 3-3 Uhrough 3-5 (diameter of photograph’s viewing circle is approximately 2 mm). 14Figure 3-4 — Microscopic Image of 120 jum Aluminum Particles Figure 3-5 — Microscopic Image of 3 pm Aluminum Particles 15The particle-liquid mixture is pumped through a Hydra-cell Pump and past a bypass valve which controls the velocity at the submerged nozzle exit located in the acrylic tank. Regardless of the fluid viscosity or particle diameter, the pump always displaces a flow rate of about 8.0 gallons per minute. Local solid particle velocity measurements with the LDV are to be made at several different locations at and away from the nozzle centerline and exit, For the exact measurement locations, refer to Figure 3-6 below. Note that z= 0 mm refers to the nozzle centerline, while z= 4 mm refers to the distance at the OD of the nozzle, and 12 mm is 12 mm away from the centerline. In addition, the target is located at a distance of 12.7 mm from the nozzle exit. It is also necessary to observe that due to the geometry of the facility and LDV probe which emits laser beams, the closest near-wall axial velocity measurements can be made 0.6 to 0.7 mm from the target. On the ether hand, it is possible to make radial velocity measurements as close as 0.1 mm from the target wall. e Target | _42.7 mm] Figure 3-6 Map of LDV Velocity Measurement LocationsLaser Doppler Velocimetry Background The LDV measurement process is briefly described in Figure 3-7 below. ‘Measurements @ Computer Displays Velocity HistogrAms Figure 3-7 — LDV Measurement Process In short, a single laser beam is generated and travels into a color burst, which separates the laser into four individual beams, When the probe is positioned as seen ; above, two green beams measure the axial velocities of the particles, while two blue beams measure the radial velocities (these velocities can then be used to find the magnitude or speed as well as the impingement angle at the target). These beams then | travel through a fiber optic probe, arc emitted toward the submerged nozzle, and are arranged so that the beams converge at the location where local measurements are to be made, ‘The passage of solid particles changes the reflected fringe pattern of the emitted ; beams. This information is then sent through a signal processing system, and velocity measurements are presented by way of a histogram so that an average velocity and any 7fluctuations may be recorded in FIND software, Both radial and axial particle velocity components are recorded. A detailed step-by-step process is listed below: 1 A laser produces a single beam of light that is directed into the Color Burst. ‘The laser used for this study is an Innova 70-C 5-Watt argon-ion laser which produces a 1.6 mm diameter beam (actual laser power in this research was approximately 2 Watts). ‘The Color Burst splits the single beam into four beams (two are g:een beams for axial velocity and two are blue beams for radial velocity), ‘The laser beamn enters the Color Burst and convenes on an Acousto-Optic Bragg Cell which splits the beam into two separate beams. One beam is shifted by 40 MEZz and the other beam remains unshifted. Next, a prism separates the beams into green (A = $14.5 ‘nm) and blue (A = 488.0 nm) beams. The four beams are directed into fiber optic, couplers located at the top of the Color Burst. These couplers focus and direct each beam into a transmitting fiber optic network. Each beam travels through the fiber optic network to a focusing lens within the fiber optic probe. The lens directs each beam to intersect and create a ‘measurement or control volume seen in Figure 3-8 (about 0.2 mm wide by 8 mm Jong). A lens with a focal length of 749.2 mm and half-angle, ¢ of 1.881 is used. Solid aluminum particles pass through the fringe spacing of the messurement, volume. A receiving lens within the fiber optic probe collects and directs the reflected light into a receiving fiber. The light travels through the receiving fiber to the signal processing stage. 185. A Color Link Plus Receiver collects and separates the reflected light into the two beam colors. The light is then converted to electrical signals that are sent to an Intelligent Flow Analyzer (IFA) 655 in bursts. The IFA 655 separates noise fiom signals and the signals are filtered and output to a computer. 6. Flow Information Display (FIND) software receives the digital signals and displays real-time velocity distribution histograms as well as data distribution statistics (mean velocity and turbulent intensity % based off velocity fluctuations for both velocity components). LDV Signal Processing LDV measurement is based on the principle that intersecting beams of light with the same wavelength but different frequencies create an alternating light and dark imerference fringe pattern. Figure 3-8 below shows the cross-section of two intersecting beams with different frequencies and the created light and dark fringe pattem, Alternating Light/Dark Fringes 19Tie fringe spacing distance, dy, s a function of the angle at which the beams intersect and their wavelength, It ean be found by using Equation 3-1 below. a 2sin 34 Note that dy is the fringe spacing (in m), A is the wavelength of the beams (in m), and ic is the half-angle between the two intersecting beams as seen in the next Figire 3-9. “Transmitting Lens Measurement Location ‘Transmitted Laser Beams Figure 3-9 - Determination of the HalfAngle, « ‘When a particle passes through the fringe pattern, light from the laser beams is reflected off of it, The velocity of the particle can then be calculated by finding the product of the fringe spacing and Doppler frequency of the reflected light. This is shown, in Equation 3-2 vadf 32 where v is the calculated particle velocity (in mvs), dyis the fringe spacing (in m), and fis the Doppler frequency of the reflected light in (in Hz). 20‘Another necessary component of the LDV system is a two-axis traverse, on which the fiber optic probe is mounted. The traverse moves the probe in the X-direetion (up and down) and Z-direction (left and right) between particle velocity measurements. This traverse, which is also controlled by the FIND software, is accurate to 0.01 mm and has a range of 600 mm for both axes. As a result, the velocity in the region of interest (between the nozzle exit and target, as well as up to 12 mm from the nozzle centerline) can be very accurately mapped out within a 2-D plane. It is important to note that due to the geometry of the target and LDV beams, the closest near-wall radial velocity is about 0.1 to 0.3 mm from the wall while the closest axial velocity measurement can be made approximately 0.6 to 0.7 mm from the target. Flow Loop for Erosion Rate Measurements ‘A separate flow loop was constructed to measure the erosion rate resulting from direct impingement of a submerged jet in the same geometry as the one used in the LDV flow loop. Figure 3-10 shows the schematic of this experimental set-up used by Okita (2008) which consists of a centrifugal pump with a variable speed motor, a single testing tank, and a mixer (o better suspend solid particles in the fluid. A carrier fluid with a certain viscosity flows from the testing tank (which holds about 50 liters of the liquid) to the pump from a valve at the bottom of the tank, at which point the pump displaces the {fluid through the straight nozzle back to the testing tank. The fluid exits the end of the nozzle and flows toward an electrical resistance probe that serves as the target. Once again, the distance between the nozzle exit and the target is 12.7 mm. Sand is mixed with liquid prior to the experiments so that the mixture of solid particles and fluid is constantlystimed by the mixer. This prevents excess sand from sinking to the bottom of the tank and constraining the pump inlet, Recall that the straight nozzle (normal to the target) is completely submerged in the liquid. As in the LDV flow loop, the nozzle inner diameter is about 8 mm, The nozzle exit velocity of the mixture can be controlled by changing the frequency of the variable speed motor. For a given fluid viscosity and sand corcentration, the frequency can be adjusted to produce a desired flow rate and average nozzle exit velocity. Mixer ‘Testing Tank Frequency Dial Connon, Probe: Bypass Valve Nozzle Centrifugal F Pump mee Motor Figure 3-10 — Erosion Flow Loop Schematic ‘Two different sand particles were used for experiments with this flow loop: Oklahoma #1 sand with an average size of 150 microns and California 60 mesh sand with an average particle size of 300 microns. The SEM pictures of Oklahoma sand and California sand are shown in Figures 3-11 and 3-12. Density of the sand is sbout 2,650 kg/m?. Note the increased particle sharpness observed in the California 60 sand. Furthermore, the sand size distribution for the OK #1 sand is seen in Figure 3-13. Itis, clear that although the average size is about 150 um, some sand particles may be as small a 80 im or as large as 350 pm,‘Sines Ee niet Figure 3-11 - Microscopie Photograph of Oklahoma #1 Sand Figure 3-12 - Microscopic Photograph of California 60 Sand ‘Oklahoma #1 Sand Size Distribution Volume % Particle Diameter (um) Figure 3-13 — Oklahoma #1 Sand Size Distribution 23In order to vary the viscosity of the liquid, either CMC (Carboxymelhyl Cellulose) or glycerin was mixed in with water. Since both CMC-water and glycerin- water mixtures have densities similar to that of water, viscosity was increased without significantly changing the density of the carrier fluid, Recall that glycerin was used to make the carrier fluid in the LDV tests since the glycerin solution possesses a more transparent appearance and allows LDV beams to penetrate through the mixture and acrurately make particle velocity measurements, At the same time, glycerin’s costliness makes it an unpractical fluid to be used in erosion testing where the appearance of the {uid is irrelevant, In order to run numerous experiments efficiently and economically, the use of CMC for erosion measurement experiments was necessary. It is important to note that CMC is made of long chain polymers which when mixed with water might act as a non-Newtonian fluid (for higher viscosities) whereas the glycerin solution is a Newtonian fluid. ‘Tests were performed to investigate the similarity of a mixture made with CMC dissolved in water and a mixture consisting of water and glycerin. Erosion measurements ‘were made for viscosities ranging from 1 to 50 CP. The following procedure was used by Okita for gathering erosion rate measurements: 1, Water is mixed with CMC to obtain the desired viscosity based on a pre- determined relation, Table 3-2 shows the CMC concentration and the corresponding viseosity that was obtained. ‘The CMC is added slowly and the mixture is stirred by an electric mixer until solid lumps are fully dissolved. A 1 CP test consisting of only water was also performed. 24Table 3-2 - CMC or Glycerin Concentration Required for Viscosity Change Viscosity | CMC required | 99% Glycerin Total Volume (cP) (kg) |__ concentration (2%) _ liters) 10 0301 | 50 50 | 25 0773 | 67 30 30 1076 | B 30 2. After the appropriate viscosity is made by mixing water/CMC or water/glycerin, the flow rates of the liquid are measured. This is done by attaching a long plastic tube to the end of the nozzle and collecting the fluid in a measuring container. The time required to pump 10 liters into the container is then recorded, Based off the cross-sectional area of the nozzle, an average nozzle exit velocity of 10 m/s corresponds to a filling time of 20 seconds, while 40 seconds are required for an average velocity of 5 nvs. The frequency of the motor is appropriately changed for each viscosity case in order to get the exact desired flow rates corresponding to these exit velocities. 3. The nozzle is then directed toward the electric resistance probe, so that the probe face is normal to the initial flow direction of the jet so that normal/direct impingement will occur. The ER probe (provided by Cormon LTD) is connected to the appropriate CEION software on a computer and sand particles are slowly added to the fluid. The erosion rate is measured in mils per year (MPY) for 10 m/s. and 5 m/s nozzle exit velocities. Multiple tests at the same condition are performed to obtain a solid average. 254, The sand concentration of the fluid flowing through the straight nozzle is also recorded after ER probe measurements are obtained. At this time, the vinyl tube is again placed around the end of the nozzle and the mixture is collected into a bucket. Appropriate volumes of the fluid with sand are then collected at the calibrated velocities (10 mvs and 5 mvs). The sand particles are separated from the mixture using a sieve and all liquid is left to evaporate, Sand concentration can then be calculated by dividing the volume of sand by the volume of the gathered mixture. 5. Using the determined sand concentration, the erosion rate is converted from MPY to mils per pounds (mils/Ib). This allows for metal loss resulting from erosion to be presented as a function of the amount of sand or solid particles contained in the fluid. Equation (3-3) shows this conversion where the sand rate is given in Ib/s. or( ==Rqwpy)*#—_—__*¥_+_l b 365days* 24hr*60min* 60sec tit. «045488 GA 0.454 3) Electrical-Resistance (ER) Probe Background Electrical-resistance (ER) probes are used by many oil, gas, and chemical/refinery companies as an instrument to detect sand and to monitor pipe wall loss resulting from erosion and/or corrosion. In these experiments, a Cormon Ltd ER probe (seen in Figure 3-14) is used to measure the erosion or metal loss on a target wall. The probe is composed of two electrically conductive elements, both of which are made of Inconel 625. A coil-shaped sample element is exposed to the flow and a reference clement that sits behind the sample element is protected from the flow by the black epoxy coating. 26‘The resistance between the two elements is converted to a metal loss or erosion rate and is presented in the following equation: 34 where R ~ element clectrical resistance, length, A = element cross-sectional area, Figure 3-14 Target Surface of Cormon ER Probe As solid particles impact the sample element, the element experiences metal loss from erosion. Since the sample element undergoes a change in volume and the reference element does not, this results in a change in resistance seen in the above equation, From the difference in the resistances of the exposed sample clement and the constant resistance of the epoxy-protected reference element, the amount of metal loss can be calculated with Cormon CEION software (a sample output is seen on the next page). At the same time, the instantaneous temperature is measured by the probe since this property effects electrical resistance. The temperature change results in error that is accounted for 7and corrected by electronics that receive the output signal fiom the probe, The processed data containing temperature-corrected metal losses over time based on resistance changes is received by a computer and is plotted in real-time using software supplied by Cormon, A sample output screen is presented in Figure 3-15. Figure 3-15 - Sample Cormon Probe Software Output ‘The erosion rate is calculated from the position of two markers that are placed by the user (usually these are placed at the point the pump is tuned on and ofl), ‘The slope of the line between the two markers in Figure 3-15 represents the metal thickness-loss rate (in an/min). The metal loss rate is converted to mils/yr (MPY) and is presented in the lower right-hand comer of the display. Again, the erosion rate can be converted to mil/pound of sand by using Equation 3-3. 28CHAPTER 4 EXPERIMENTAL FLUID AND PARTICLE VELOCITY RESULTS Introduction Axial and radial velocities of solid aluminum particles traveling from a submerged nozzle normally directed toward a target, as seen in the test setup displayed in Figure 3-1, were recorded with Laser Doppler Velocimetry at the locations seen in Figure 3-6. Consequently, particle speeds and trajectories were measured as they exited the nozzle and approached the solid wall. While the fluid flow rate was held at a constant 8.0 gallons per minute, the fluid viscosities used were 1, 10, 25, 50 and 100 CP for three particle sizes of 3, 120, and 550 microns. The data in this chapter is presented in the following manner L) With particle size held constant at 3, 120, or 550 microns and viscosity changed between 1, 10, 25, 50 and 100 CP 2.) With viscosity held constant at 1, 10, 25, 50, and 100 CP and particle size changed between 3, 120, and 550 microns ‘The measured data, axial and radial velocities, arc then used to construct speed contours based on the magnitude of the measured velocities at all of the locations of Figure 3-6. Velocity fluctuations and estimated variation are also included in select discussion. 29Particle Size Constant at 3 Microns — Effect of Viscosity LDV Results When aluminum particles with an average diameter of 3 pum are seeded in the carrier fluid, minimal to no-slip between the particles and the fluid can be assumed, At the same time, viscosities are varied to observe the change in pa a5 flow approaches the solid-wall target. Table 4-1 below shows the estimaied Reynolds number for pipe flow through the nozzle. Since diameter, approximate fluid density, and average nozzle exit speed (flow rate) are held constant, viscosity is the only variable causing the change in Reynolds number. Table 4-1 — Reynolds Numbers of Flow with Variable Viscosity Fluid Viscosity (CP) | Reynolds Number (Ren) Flow Regime (Pipe) T $5,000 Fully turbulent 10 8,500 Fully tarbatent 25 3,400 “Transitional/Turbulent 30 1,700 ‘Transitional to laminar 100 350 Fully Laminar ‘These values for Reynolds Number were approximated by using Equation 4-1 below, where D is the nozzle diameter, p is the fluid density (about 1,000 kg/m), V is the average fluid velocity through the nozzle, and is the dynamic viscosity. Re, =pVD/qt 30 (4-1) le and Muid trajectoryFigure 4-1 seen below shows the axial velocity profile generated from LDV ‘measurements at 1 mm from the nozzle exit. Typical pipe flow is observed within the 8 mn nozzle diameter. Again, note that the flow rate is held constant at 8.0 gallons per minute even as fluid viscosity is altered. Integrating each of the curves of Figure 4-1 over the cross-sectional area of the nozzle will yield an approximate flow rate of 8.0 GPM for cach case. Furthermore, the average velocity of each curve results in a value of about 10.5 mis. ‘Axtal Velocity Profile (1 mm from Nozzle Exit) ‘2 mieron Aluminum, 8.0 GPM Flowrate I Pze4 mm a= Oman Sawer Tanna e006 scr seice ‘tal Volocty (mi) ea ete a teeteateeeeeeeeOteaty eaeeigtteeee a eceect ate Distance trom Nozzle Centertine mm) Figure 4-1 - Axial Velocity Profile near Nozzle Exit for 3 ym Particles It is clear that the flow predictions based on calculated Reynolds number seen in Table 4~ 1 are accurately displayed in the velocity profiles of Figure 4-1. The 100 CP viscosity case evidently possesses a laminar velocity profile which matches a theoretical laminar curve fit based on flow within a pipe. As expected, the 50 CP and 25 CP cases begin to show a transition from laminar to turbulent flow. The lower viscosity cases of 10 CP and 311 CP exemplify typical turbulent flow as seen in the constant axial velocity profiles followed by a sudden drop at the nozzle diameter caused due to wall shear stress. ‘The change in the axial velocity component (direction parallel to nozzle wall) versus distance from the nozzle exit along the nozzle centerline (defined as 2-0 mm) is seen in Figure 4-2 on the following page. Recall that the target is located 12.7 mm from the nozzle exit and the final axial velocity measurement with LDV ean be made approximately 0.6 mm from the wall due to limitations imposed by the target geometry. InFigure 4-2 it is seen that the axial velocity is retained or held constant for each viscosity case until the particles travel about 8 mm from the nozzle exit. At this point, there is sudden drop in axial velocity with an apparent drop to nearly zero at the wall (the last axial velocity measurements are below 2 m/s). Axial Velocity, z= 0 mm ‘S micron Aluminum, 8.0 GPM Flowrate ° 4 a8 = = 1 Tol 444 a a . = a : a” re oi ie & no . 3 2 ° 7 a er a Distance fom Nezzlo Exit (mm) Figure 4-2 - Axial Velocity versus Distance from Nozzle Exit at 2-0 mm for 3 ym Particles 32It is furthermore displayed that the higher viscosity cases of 100 CP and 50 CP possess higher initial axial velocities at the centerline due to the laminar flow transition. The lower viscosities of 25, 10, and 1 CP (which are closer to turbulent flow) possess very similar initial axial velocities. Ultimately, however, it appears that the axial velocity at the wall decays to below 2 mi for all of the viscosity cases at this location. Figures 4-3 through 4-5 show the change in axial velocity versus distance from the nozzle exit for radial distances away from the nozzle centerline (2=1, 2, and 3 mm). It is again seen that velocities are constant until 8 mm from the nozzle exit at which point values rapidly drop to below 2 m/s by the time particles reach the wall. Furthermore, initial velocities for all viscosity cases become more similar away from the centerline due to wall shear stress effects. ‘Axial Velocity, z= 1 mm ‘Smicron Aluminum, 6.0 GPM Flowrate «1 ee . . 4 So fi 4 a = 3 rp © & 8 x 2 8 ° I x i, = j 3 * Be ‘ 1 2 ° Distance trem Novae Ext (mm)Figure 4-3 - Axial Velocity versus Distance from Nozzle Exit at z=1 mm for 3 um Particles Aviat Velocity, 2= 2 mm ‘3 mleron aluminum, 0 GPM Floweate Distance tam Noctl Ext mn) Fe eee le220mm —F ul drei es x x g j e¢— “ee : : Bed 7 : Figure 4-4 - Axial Velocity versus Distance from Nozzle Exit at 2=2 mm for 3 ym 27 Particles ‘Axial Volocity,2=3 mm ‘moron Aluminum, 8.0 GPM Flowrate 10 Asta vetocky (ms) a 4 4 i a k i a Rx + . + . x +s 4 ° 2 4 3 a 0 2 “ Distancefron Norae Ext jm) Figure 4-5- Axial Velocity versus Distance from Nozzle Exit at z=3 mm for 3 um Particles 34Figure 4-6 on the following page shows the axial velocities at 4 mm from the nozzle centerline, which is the location of the nozzle outer diameter. As a result, the trend displayed by these curves varies significantly from those of Figures 4-2 through 4-5. Initial velocities are reduced to nearly 0 m/s due to the presence of the nozzle wall Around 8 mm from the nozzle exit, there is a rise in velocities followed by a minor peak below 5 mis at 1-2 mm from the wall, finally preceded by a drop in velocities to below 2 nys.at the wall. Note that the curves for 10, 25, $0, and 100 CP appear to follow a very similar trend, whereas the curve for 1 CP (the most turbulent case) possesses velocity values which are consistently slightly higher. This may occur due to an increased. ‘turbulence resulting from low viscosity or may display a slight change in location attributed to LDV positioning error. Since this is the location where direct nozzle flow terminates, a slight upward or downward shift from the centerline can create a large change in the velocities recorded The axial velocity profiles for z=6, 8, 10, and 12 mm from the centerline are not shown since this velocity component is nearly negligible (well below 1.5 m/s) for all locations outside of the nozzle diameter. The recorded values however are included in the Appendix A. 35Axial Velocity, 2 = 4 cam ‘micron Aluminum, 8.0 GPM Flowrate is fa = # ae . Sance Pas en ae a 7 x x % mince yw xP 3 t >. z a 3 oe 4" Distance from Hoazlo Ext (mm) Figure 4-6 - Axial Velocity versus Distance from Nozzle Exit at 2=4 mm for 3 yam Particles At the same time, the radial velocity component, which travels parallel to the target wall, was recorded for the same radial distances, As seen in Figure 4-7, which shows the radial velocity versus distance from the nozzle exit at 2 mm from the centerline (this component is negligible at the centerline), the radial velocity component does not, become apparent until approximately 8 mm from the nozzle exit, at which point there is @ significant increase in its magnitude. A peak in this component occurs about | mm from the wall which is then followed by a sudden drop at the wall for the 10, 25,50 and 100 CP viscosity cases. The | CP case does not display this sudden drop at the wall. At the same time, the radial velocity peak values seen at | mm from the wall increase with increasing viscosity. This is most likely attributed to the higher initial axial velocities occurring near the centerline for the more viscous cases due to the turbulent to laminar 36transition evidenced in Figure 4-1. These axial velocities are transferred to radial velocities as particles tend to move outward away from the nozzle exit and closer towards the solid wall. This trend is further observed in Figures 4-8 through 4-12 as the radial distances are increased from 4 mm to 12 mm (outside the of nozzle diameter). Radial Velocity, z= 2mm ‘Srieron Aluminum, 8.0 GPM Flowrate oo a 80 : 30 i fel = ca 7 eee i a | nase a eae 3 eK ner i: ath 2 20 a x 10 ; g ee co | —@ g K 7 ceantea ’ ° 7 2 @ om Distance from Nozze Exit (rm) Figure 4-7 - Radial Velocity versus Distance from Nozzle Exit at 2=2 mm for 3 um Particles 37Radial Velocity,z= 4mm ‘Qmicron Aluminum, 8.0 GPM Flowrate eo - 40 120 g 100 —| [ear mince ju see i0ee i a 10P 40 2 x ac oo} — 2 A ° 2 4 6 a 0 2 “ Distance trom Nazzp Exit (mm) Figure 4-8 - Radial Velocity versus Distance from Nozzle Exit at 2=4 mm for 3 um Particles By observing the repeating trend of Figures 4-7 through 4-12 and noting the variation in maximum radial velocities of the particles at 2 mm from the centerline to the nozzle diameter (at 4 mm from the center of the nozzle) and then finally to 12 mm outward (8 mm outside the nozzle diameter), a rising maximum radial velocity is observed followed by a gradual decrease. By the time particles reach the z=6-8 mm position at a distance Jess than 1 mm from the wall, the radial velocity component reaches a maximum of slightly over 14 m/s for the 100 CP case, Maximum values for all viscosity cases are ‘generally seen in the 6-8 mm radial region, with increasing viscosity normally resulting in higher radial velocity values. There are a few exceptions, notably the appearance of the 50 CP case possessing slightly higher radial velocities than the 100 CP case irom 2-8 mm and onward. This may be attributed to velocity profile irregularities caused by a 38transition from laminar to turbulent flow (which occurs soon after 50 CP). Measurement Uncertainties due to turbulent fluctuations may also contribute to this phenomena. Finally, it is important to note again that for all near-wall radial velocities, the 1 CP or water case does not experience as significant of'a decrease as the more viscous cases of 10, 25, 50, and 100 CP. This occurrence will be discussed in more detail in upcoming figures. Radial Velocity, z= 6 mm ‘3 micron Aluminum, 8.0 GPM Flowrate Joo! Lae ic sa |e i: a [ | pee 20 8 wet? oo} ga 2 a 2 2 if See eee aa ee cone oie tonsa) Figure 4-9 - Radial Velocity versus Distance from Nozzle Exit at z=6 mm for 3 pm Particles 39uo eo Loco oo 2 a neve i wo tee pet — [iter 40 yo ® 20 3 es order dete tafe EEE TEC Distance from Nozale Extn) Figure 4-10 - Radial Velocity versus Distance from Nozzle Exit at 2=8 mm for 3 um. Particles Seefs ep caesar eee eases eases 720 a0 a0 ¥ ed i t ¢ oat veloc (me) 40 | 20 eeceantecee 00 8} A, ° 2 4 6 5 +0 @ + Distance rom Nozale Ext mm) Figure 4-11 - Radial Velocity versus Distance from Nozzle Exit at 2=10 mm for 3 um Particles 40Radial Velocity, z= 12 mm ‘2 micron Aluminum, 8.0 GPM Flowrate 204 100 80 4 sasoce x Radial velacty (ms) 40 % so a pees 3 aaa ee re.) a 20 ~e ~ & 00 +22 at st #— ih ° 2 ‘ 8 a 0 2 1“ Distance rom Nozzle Ext (mm) Figure 4-12 - Radial Velocity versus Distance from Nozzle Exit at 2=12 mm for 3 ym Particles Itis clear from Figures 4-7 through 4-12 that the measured radial velocity undergoes significant changes in magnitude in the last | mm to the wall. However, itis more difficult to observe the effect of viscosity on the radial velocity at certain fixed distances from the wall. Figure 4-13 displays the radial velocity as a funetion of radial distance from the nozzle centerline (2) at a fixed distance of 0.3 mm from the wall. This is where the closest accurate radial velocity measurements could be made with LDV, ‘most likely because the thickness of the control volume (volume of measurement) is slightly more than 0.2 mm. Further note that these measurements were taken at a distance from the peak in velocity observed in Figures 4-7 to 4-12. Itis evident that the peak in radial velocity occurs between 4-8 mm for all viscosity cases. Furthermore, the radial velocity is clearly and consistently highest for the 1 CP fluid at this near wall Location, 41For the more viscous cases of 10, 25, 50, and 100 CP, radial velocity appears to increase slowly with increasing viscosity. However, all of these values are still significantly less than one-half the radial velocities measured for the patticles seeded in the 1 CP case. Further sections of this chapter compare this near-wall velocity versus measured erosion rates, Radial Volocity (0.8 mm from wall) 3 micron Aluminum, 8.06PM Flowrate 20 20 ¥ 10 x x 00 = Fo noe ae Te| : sic 30 +. % ; . 20 |g eg | = = = a = a 10 yg ee og RB | Distance from Nozzle Centetne (nm) Figure 4-13 - Radial Velocity versus Distance from Nozzle Centerline at 0.3 mm from Wall for 3 jum Particles Figure 4-14, located on the following page, again shows the radial velocity versus radial distance from the nozzle centerline, However, these values were measured at a fixed distance of 0.6 mm from the wall, or the approximate location of the peak observed. in Figures 4-7 to 4-12. Again, itis evident that the maximum radial velocity occurs 4 to 8 ‘mm from the centerline and outside of the nozzle diameter. It is immediately evident that the velocity values increased significantly from those recorded at 0.3 mm closer to the 42wall (by as much as 4 times or more) for the viscous fluids of 10, 25, 50, and 100 CP. ‘The water case of 1 CP appears to have mostly unchanged velocity values between 0.3, ‘mm and 0,6 mm from the wall. At the same time, an increase in viscosity appears to usually increase the radial velocity at this location. This most likely occurs due to the increased axial velocities at the nozzle exit which occur as a result of the laminar ‘transition previously addressed. However, there is the notable exception of the 100 CP case possessing the lowest radial velocities outside of the nozzle diameter of z=: mm. Once above a given viscosity, the fluid or seeded particles may struggle to retain radial velocities as the fluid is pushed further outward and away from the nozzle exit. Within the nozzle diameter, the radial velocities continue to increase with increasing viscosity. Radial Velocity (0.6 mm from wall) ‘meron Aluminum, 8.0 GPM Plowrate nop > siesta ‘20 * = 7. & * S 100 a = fe x Fa see i. + 4 etal i a3 . xioce « lk ¥ soa — f 20 2 Gry imines seeteesiataeetaeeatesteeestsieesiaeaetaeal ° 7 7 3 aieanraecninaeairiiea Distance rom Nazze Ceterine tin) Figure 4-14 - Radial Velocity versus Distance from Nozzle Centerline at 0.6 mm from Wall for 3 um Particles In order to better display the change in radial velocities in the near wall region, this measured value is plotted versus distance from the nozzle centerline at a set radial 43distance with a focus on the 1 mm region closest to the wall. Figure 4-15 shows this plot ata distance of 3 mm from the centerline, while Figure 4-16 contains the plot at a radial distance of 10 mm. It is important to note that the final two data points closest to the wall for each of these graphs lie inside of the 0.3 mm accurate measurement region and were ct fully used in further analysis, Nevertheless, it is clear that for the viscous cases there isa significant drop in radial velocity between 0.6 mm and 0.4 mm from the wall. A. mach less apparent drop occurs for the non-viscous 1 CP water case. This may imply that a boundary layer approximately 0.5 mm thick forms as a result of increased viscosity for the 10, 25, 50, and 100 CP carrier fluids. At the same time, this may suggest post-impact speeds are being mensured at the near-wall regions, thus resulting in the significant velocity readings. If this were true, then particles in the 1 CP fluid make lesser contact or have less of a momentum transfer as they make contact with the wall. Recall these results are for very small particles with minimal velocity slip between it and the fluid, Larger particle sizes closer to typical sand will be investigated in upcoming sections. 44‘Near Wall Radial Velocitias, 2= 3 mm ‘Smieron Aluminum, 8.0 GPM Flowrata ~~" i a y2=0mm 1207 ne=4 num . ¢ 40.0 $$ SD yss tan ae = i ® gz 80 z oO E i meee Qe wicca + m9cP ? rsoce 60 a25cP 3 x * * xioce i x ice @ 40 oo 6 6 a e & i is ite aA neeeee Areca 00 “9 0 RA kote Distance rom Nozzle Exit(mm) Figure 4-15 Near Wall Radial Velocities at z=3 mm for 3 pm Particles NoarWall Radial Velocities, 2= 10 mm ‘meron Aluminum, 8.0 GPM Flowrate 407 oop Seer reeeee x 100 z eee x x an = x ; . j 60 x 40 * & 20 peeeiccetccet a oo us 2 «24 B28 kB Distance rom Nozzle Exit (rm) Figure 4-16 —Near Wall Radial Velocities at 0 mm for 3 um Particles 45In order to better visualize the axial and radial velocity measurements made by LDV in the numerous locations of the testing region, contour plots such as those seen in Figures 4-17 and 4-18 were created. Figure 4-17, which uses data from the 100 CP case, clearly shows the high axial velocities within the nozzle exit region and the decrease in this velocity component as measurements were made closer to the wall. The laminar velocity Profile close to the nozzle exit (on the far left-hand side of the contour) is also clearly visible, At the same time, Figure 4-18, which is also based on the 100 CP viscosity uid, ro x eo + 7 7 wsoce . x & asc 60 7 xoce i 2 xtcp 40 * 7 + 7 # 29 pF? 5 op} ____—_— Ne @ 21 22 29 we 25 126 127 128 Distance trom Nore Ext (mm) Figure 4-39 ~ Near Wall Radial Velocities at z=10 mm for 120 pm Particles Particle Speed in m/s - 8.0 GPM Flowrate - 120 micron Aluminum - 100 CP 10] Distance from Nozzle Centerline (mm) 6 8 10 Distance from Nozzle Exit (rm) Figure 4-40 — Speed Contour in mvs at Fluid Viscosity = 100 CP for 120 um Particles 6Particle Speed in m/s - 8.0 GPM Flowrate - 120 micron Aluminum - 50 CP i] 10} 14 Distance fom Nozale Centetfne (rm) 6 8 Distance from Nozzle Ext (mm) Figure 4-41 — Speed Contour in m/s at Fluid Viscosity = 50 CP for 120 yxm Particles ‘Speed in ms - 8.0 GPM Flowrate - 120 micron Aluminum - 25 CP 10] 2 Distance fom Nozzle Centering (rm) 4 10 6 8 Distance from Nozale Ext (mm) Figure 4-42 — Speed Contour in m/s at Fluid Viscosity = 25 CP for 120 um Particles 64‘Speed in mis - 8.0 GPM Flowrate - 120 micron Aluminum - 10 CP 10) i i & 7 4 6 7 is : , i , E b 6 8 Distance from Nozzle Est (mma) Figure 4-43 ~ Speed Contour in m/s at Fluid Viscosity }0 CP for 120 ym Particles ‘Speed in mvs - 8.0 GPM Fiowrate - 120 micron Aluminum - 1 CP 19] Distance fom Nozale Centerine (rm) 4 6 a Distance from Nozzle Ex (mm) Figure 4-44 Speed Contour in m/s at Fluid Viscosity ~ 1 CP for 120 ym Particles 65Particle Size Constant at 550 Microns — Effect of Viscosity ‘The third and largest particle size which was seeded in fluids of variable viscosity hed an average diameter of 550 um. Such particles were used to exaggerate the size of seme of the largest sands which might be encountered in industrial pipelines. As in previous testing, axial and radial velocity measurements were recorded at rumerous locations to observe particle movement in a direct impingement jet. All other testing conditions, namely fluid flow rate, were maintained constant. Axial Velocity Profile (1 mm from Nozzle Exit) {650 micron Aluminum, 8.0 GPM Flowrate Sk: aN -e10cP 2 500P 2508 e100P eer 20mm, dal Volocty (mvs) pee ea ae ore ar eons teem tol Distance rom Nazzle Centering (wm) Figure 4-45 - Axial Velocity Profile near Nozzle Exit for 550 pm Particles ‘Ata distance of | mm from the nozzle exit, the axial velocity profiles seen in Figure 4-45 were recorded. It is thus seen that even with these largest particles, the 100 CP case maintained a profile resembling a theoretical laminar fit of the form seen below in Equation 4-2. Note that R equals 4 mm and is the radius of the nozzle, while r= 0 corresponds to the centerline location:U=Ugall=(r/ RY] (42) The 50 CP case still mainly displays laminar behavior, while the 25, 10, and 1 CP cases produce results which indicate a transition to turbulent flow within the nozzle. Wall shear stress is again responsible for the drastic reduction in axial velocity near to the nozzle diameter. Axial Velocity, = 0 mam 50 micron Aluminum, 8.0 GPM Flowrate . . ° i = = ioe Foto 2 2 asco meee * 1a250P i. $2 fae 3 % x10 26 & id a4 jest ° ° 2 4 ° 2 “0 2 “ Distance from Nozze Ext (mm) Figure 4-46 - Axial Velocity versus Distance from Nozzle Exit at 2-0 mm for 550 ym Particles ‘As seen above in Figure 4-46, centerline axial velocities for these largest particles were again maintained until about 8-9 mm from the nozzle exit. Here, there is a drop to below 2 m/s for all tested viscosities. Initial velocities increase with increasing viscosity, especially for the near-laminar cases of 100 and 50 CP. The transition to turbulence is clearly visible by the results of the lower viscosities. Overall, itis seen that regardless of viscosity, initial velocity, or type of flow at the nozzle exit, particle velocities always ordecayed to below 20 percent of their initial value at the nozzle by the final measurement location about 0.6 mm from the wall. Figures 4-47 to 4-49 show the change in axial velocity versus distance from the nozzle exit for radial distances of 1, 2, and 3 mm from the nozzle centerline. A trend very similar to Figure 4-46 is again observed. As with the smaller particles in the previous sections, initial axial velocities visibly decrease for the more viscous fluids as focus is shifted away from the centerline, largely due to the effects of the nozzle wall located at 4 mm from the centerline, At3 mm from the centerline, the initial axial velocities actually become larger for the 1, 10, and 25 CP cases than for the 100 and 50 CP fluids as a result of a decreased shear region in the less viscous cases. Axial Velocity, z= 4 mm {650 micron Aluminum, 8.0 GPM Flowrate age Omm z=4mm 2 ee eeeeiiiee: ‘| 4 ‘Adal Yolocky (mi) bx oe 7 for 8 4 sy ; 5 ° sats aes gaara ale Distance rom Nozze Exit (nm) Figure 4-47 - Axial Velocity versus Distance from Nozzle Exit at 2=1 mm for 550 um ParticlesAxial Velocity, z=2 mm ‘550 micron Aluminum, €.9 GPM Flowrate iex=omm, ae sence ace ‘ll elocty ie) o 2 a 6 a 0 2 4 Distores rom Novae Exon) Figure 4-48 - Axial Velocity versus Distance from Nozzle Exit at 2-2 mm for 550 ym Particles Axial Velocity,2= 9 mm 560 micronAluminum, 8.0 GPM Flowrate 2 i’ = }- hn ae i. i ee aK ; ay ‘ Distance from Nozzle Exe (rim) Figure 4-49 - Axial Velocity versus Distance from Nozzle Exit at 2=3 mm for 550 ym Particles 69‘The results displayed in Figure 4-50 show the axial velocities along a z=4 mm line which extends from the nozzle wall. It is seen that the trend displayed by particles along this line varies significantly from the trend observed within the nozzle exit region. Furthermore, it is clear that interference from the nozzle wall causes initia! axial velocities to drop well below 2 mis atz=4 mm. As particles move beyond 10 mm from the nozzle exit, they experience a slight peak in axial velocity followed by a rapid decay at the wall. This peak is highest for the least viscous fluid which may suggest that inereased viscosity may impede axial flow in this outer region. It is again important 10 note that a slight radial shift towards the centerline will ultimately result in a transition of the observed trend to one which resembles the plot of Figure 4-49. For distances outside of 2=4 mm, any peaks in axial velocity become negligible as, the radial velocity component again becomes dominant. As a result, an analysis outside of the nozzle diameter only includes the radial velocity component. In this outer region, axial velocities become nearly fully converted into a radial direction parallel to the wall 4s particles tend to move outward from the nozzle centerline. 10‘Axtal Velocity, z= 4 mm ‘550 micron Aluminum, 8.0 GPM Flowrate 40 as 20 x 22 9 100CP sae F 20 “ base i x ioe Bas x x xICP, a so * + * x é . 4 os} * e = a Maca oo 0 2 : ‘ a o 2 4 Distance trom Nozzle Exit (mm) Figure 4-50 - Axial Velocity versus Distance from Nozzle Exit at z=4 mm for 550 um Particles The following pages contain Figures 4-51 through 4-56 which plot the radial velocity of the 550 micron particles versus distance form the nozzle exit for incremental distances from the nozzle centerline. As in previous figures, all five tested carrier fluid viscosities are presented. As would be expected, axial velocities are very small at distances close to the nozzle exit. It is seen, however, that the radial velocities sharply peak about 1 mm from the wall. In many instances, such as at z=6 mm, this near-wall peak radial velocity is almost as large as the initial axial velocity at the nozzle centerline. This would imply that most of the initial particle speed is converted into a radial direction along the target. Itis again important to note that the largest peak in radial velocity occurs between 4-8 mm from the centerline. nUpon examining the last 1 mm to the wall in Figures 4-51 to 4-56, one can sec the peak in radial velocity followed by a significant reduction by the time particles travel to a location 0.3 mm from the wall, This drop in velocity, seen best in the viscous cases, may be attributed to the formation of a boundary layer or some other interference resulting from the solid wall. Upcoming results and discussion focus on the changes which occur in the near-wall region and the potential implications of such particle behavior. Radial Velocity, 2= 2mm 660 micron Aluminum, 8.0 GPM Flowrate Some $0. o-timm ao + 10 a = zoo 2. i ea $08 ps0 aja a j in 260 so saoee i ICP a0 a 7 we *| i 4 19 ar y % 8 oo| 8 9 4 _ __§ ° 2 ‘ 6 8 %0 ra 1" Distance rom Nozzle Exit mm) Figure 4-51 - Radial Velocity versus Distance from Nozzle Exit at 2-2 mm for 550 um Particles 2Radial Velocity. 2= 4mm ‘850 micron Aluminum, 8.0:GPM Flowrate 20 7 ‘ on 0 Bie A a i ? es * a ; i i il wg ' ool He _@ —» a 5 8 age oe Hea aOR one STE Distance trom Noeae Ext (rm) Figure 4-52 - Radial Velocity versus Distance from Nozzle Exit at 2=4 mm for 550 um_ Particles. Radial Velocity, = 6 mm {880 micronAluminum, #.0 GPM Flowrate 07 3 Vs ‘ 0 * 9 1006P Ban % axcP nia volo dis) & 2 0 Bt i = y 1 + Q eo hw ¥ # a \ Distance tom Nozze Est) Figure 4-53 - Radial Velocity versus Distance from Nozzle Exit at z=6 mm for 550 yum Particles ( 3Radial Velocity, z= 8mm ‘580 mieron Aluminum, 8.0 GPM Flowrats he: 140 >— oe ——_—— Yassin 20 = § i a i ry x4 | | ® only as) e a. Beit 4 3 ° 0 2 u Distancotom Home Ext yum) Figure 4-54 - Radial Velocity versus Distance from Nozzle Exit at 2=8 mm for 550 um Particles Radial Velocity,2= 10 mm ‘550 micron Aluminum, 8.0 GPM Flowrate oo} 4+ i. 33 Distance from Nozzle Exit (mm) Figure 4-55 - Radial Velocity versus Distance from Nozzle Exit at 2-10 mm for 350 um. Particles 74,Radial Velocity, z= 12mm 650 micron Aluminum, 8.0 GPH Flowrate 207 \ ‘00 : 4 76 t—| i | [emeer > ‘a | | mso0° so xo 260 : i s x] | | x0° i Bel | lee \ 40 a | en 20 v2 ; oo +B a * a 4 a 2 4 6 8 10 % 1“ Distance trom Nozze Exit (mm) \ Figure 4-56 - Radial Velocity versus Distance from Nozzle Exit at 2-12 mm for 550 ym \ Particles In order to better visualize the change in radial velocity versus radial distance from the nozzle centerline, the plots of Figures 4-57 and 4-58 were generated. At the same time, these plots display the change in radial velocity at two different locations away form the wall in the near-wall region. Figure 4-57 shows the velocities at a distance i f 0.3 mm from the wall, which is located well after the peak seen in Figures 4-51 to 4- j 56. ‘The plot of Figure 4-58 shows these same measurements at 0.6 mm from the wall, or the approximate location of the peak. \ From these results, it seems that radial velocity decreases as viscosity increases at ‘ 03 mm from the wall. The lower viscosity cases of | and 10 CP appear to consistently possess the highest velocities here, which may imply the formation of thinner boundary layers, The results of Figure 4-58 (located an additional 0.3 mm away from the wall) 15