Design and Performance Evaluation of an Air-Blast Atomizer

Transcription

1 Design and Performance Evaluation of an Air-Blast Atomizer Dr. Olusegun Adefonabi Adefuye Department of Mecanical Engineering, Lagos State University, Lagos, Nigeria Bennet Ifenna Okoli Centre for Space Transport and Propulsion, National Space Researc and Development Agency, Abuja, Nigeria Abstract: Te present work describes te design, construction and experimental investigations of an air-blast atomizer for alf spray cone angle of 30 using stainless steel for foundry application. Outline detail of experimental setup to investigate effect of injection pressures on spray and flame lengts, te amount of fuel and te time taken to melt some selected materials was studied. An experimental study of air-blast atomization was conducted using te manufactured atomizer in wic te fuel (kerosene) flows under gravity at angle 45 o from te tank and was atomized by te oxygen stream flowing in a cylindrical cannel from a pressurized oxygen bottle (cylinder). Produced air-blast atomizer was experimentally investigated at different pressures ranging from 3 to 15 bars in te step of 3 bars [6]. As te injection pressure was increased from 3 to 15 bars, te spray and flame lengts increases. From 6 to 12 bars, visible increment was observed in te spray and flame lengts due to increase in injected pressure wic led to breaking-up of te liquid film into small droplets. As enoug pressure was provided from 12 to 15 bars, spray and flame lengts increased appreciably. 0.6 kg of aluminum melted in 13 minutes 43 seconds using 0.5 liter of kerosene; te volumetric flow rate and te mass flow rate obtained was 6.075*10-7 m 3 /sec and 4.921*10-4 kg/sec respectively. Similarly, 1.2 kg of brass melted in 17 minutes 13 seconds using 1 liter of kerosene; te volumetric flow rate and te mass flow rate obtained was 9.680*10-7 m 3 /sec and 7.841*10-4 kg/sec respectively. Te furnace efficiency of 2.2 % was calculated from te teoretical and actual energy used for melting te metal [6]. Keywords: Air-blast atomizer; stainless steel; kerosene; spray cone angle; furnace; foundry. 1. INTRODUCTION Atomization is a process wereby a volume of liquid is converted into a multiplicity of small drops [1]. Its principal aim is to produce a ig ratio of surface to mass in te liquid pase, resulting in very ig evaporation rates. Air-blast atomizers ave many advantages over pressure atomizers, especially in teir application to gas turbine engines of ig pressure ratio. Tey require lower fuel pressures and produce a finer spray. Moreover, because te air-blast atomization process ensures toroug mixing of air and fuel, te ensuing combustion process is caracterized by very low soot formation and a blue flame of low luminosity, resulting in relatively cool liner walls and a low exaust smoke. Te merits of te air-blast atomizer ave led in recent years to its installation in a wide range of industrial and aircraft engines. Most of te systems now in service are of te pre-filming type, in wic te fuel is first spread out into a tin, continuous seet and ten subjected to te atomizing action of ig velocity air. In oter designs, te fuel is injected into te ig-velocity airstream in te form of one or more discrete jest. In all cases te basic objective is te same, namely, to deploy te available air in te most effective manner to acieve te best possible level of atomization [1]. Madu [2] designed, constructed and tested a burner tat uses an admixture of used engine oil and kerosene for foundry application. Tests were carried out to determine te time taken to melt 1 kg eac of te selected engineering materials (copper, aluminum, brass, and lead). It took 49 minutes, 15 minutes, 22 minutes, and 7 minutes to melt copper, aluminum, brass, and lead respectively. Robert [3] studied te effect of atomization gas properties on droplet atomization in an air-assist atomizer were air and fuel mix witin te nozzle before exiting troug te outlet orifice. He used air, nitrogen, argon, and carbon dioxide as te atomizing gas to determine te effect of eac of tese gases on mean droplet size, number density, velocity and teir distributions in kerosene fuel sprays. Data were obtained wit tese atomizing gases using a base, air assisted case as a reference. Comparisons were made between te gases on a mass and momentum flux basis. Te results sow tat te presence of oxygen in te air atomized sprays assists in te combustion process, since it produces smaller and faster moving droplets, especially at locations near to te nozzle exit. Ligter gases suc as nitrogen more effectively atomized te fuel in comparison to te denser gases. Argon and carbon dioxide produced larger, slower moving droplets tan air and nitrogen assisted cases. Witold [10] studied kerosene atomization process under ig speed air stream. Experiments sowed tat in te case of stream-type injectors (atomizers), a large number of small injection oles in detonation camber sould be applied in relation to disintegration of injected fuel as well as spatial uniformity of created combustible mixture. Pipatpong [11] developed an air-assisted fuel atomizer for a continuous combustor. Tey summarized tat low pressure air atomization of refine palm oil fuel wit air pressure in te range of kpa can be used to develop air blaster or burners

2 3 Fuel stand ii. Polymer Mild steel ii. Cemically inert to oil. Strengt and ability to resist corrosion. Figure 1: Scematic of an air-blast atomizer [4]. 2. MATERIALS AND METHODOLOGY 2.1 Materials selection Te selection of materials to be used for te different components in tis design involves te following consideration: 2.3 Design considerations for te atomizer Te air passage sould be made aerodynamically smoot, wit minimum areas at te atomization edge to obtain maximum air velocities and to maintain tem during initial disintegration process. Te cylindrical part of te injector nozzle sould be sort; an increase of nozzle lengt is undesirable, since it leads to a decrease of te root angle of te spray. It is recommended tat te cone angle on te orizontal entrance be witin limits from 60 0 to [7]. It is advisable to calculate tese parameters on te basis of te teory of te spray, using te curves sown below: i. Cost and availability of te material. ii. Material property mecanical, pysical and cemical properties (tat is, its ability to resist corrosion due to prolonged usage). 2.2 Properties of fuel used Te kerosene fuel sample was collected from Nigerian National Petroleum Corporation (NNPC) approved gas station in Lagos Nigeria. Te specification of te fuel is presented below. Table 1: Properties of fuel used [5]. Property Unit Kerosene Cemical formula - C 10 H 22 Calorific value kj/kg Self-ignition temperature o C 640 Final boiling point o C 249 Ignition delay period S Flame propagation rate cm/s 11.8 Flame temperature o C 1782 Kinematic 39 o C m 2 /s 2.2 Specific 15.6/15.6 o C Colour - Colourless Sulpur content wt % 0.04 Table 2: Oter materials used and teir properties S/N Parts Material Selected 1 Fuel Mild steel tank Material Property Strengt and ability to resist corrosion. 2 Hose i. Galvanized i. Ability to resist corrosion. Figure 2: Dependence of discarge coefficient, nozzle space factor and spray angle on te geometric caracteristic of te injector A - experimental points [7]. Determining te dimension of te air-blast atomizer from Figure 2; te spray angle was cosen to be Diameter of te nozzle orifice, d c was calculated using equations 1 and 2 below: 4G dc Cd 2P (1) Were: G = mass flow rate C d = 0.27 = discarge coefficient (from Figure 2) ρ = 810 kg/m 3 = density of kerosene P 1 = 3 bars = 300,000 Pa = injected pressure P atm. = 101,300 Pa = atmosperic pressure ΔP = P 1 P atm. = 198,700 Pa = pressure differential But, 365

3 G Q Cd A 2P Te effect of viscosity on te atomizer (nozzle) is given by te Reynolds number at te inlet of te atomizer: Re Were: 4G id c ϑ te contraction coefficient (assumed from te range of ) [9]. d c is te diameter of te nozzle. After calculation of d c, oter geometrical sizes of te injector: te nozzle lengt l c, lengt of entrance port l oz, diameter of inlet port d oz, was selected based on [8] approac. (2) (3) Table 3: Design parameters for te air-blast atomizer Design Data Nozzle alf spray angle 30 0 Diameter of nozzle, d c 0.89 mm Number of inlet ports, n 2 Diameter of inlet port, d oz Lengt of entrance port, l oz Nozzle Lengt, l c 3 mm 40 mm 50 mm mg (4) If te pipe is attaced at te bottom, fuel flows along tis pipe out of te tank to a level 2. A mass, m as flowed from te top of te reservoir to te nozzle and it as gained a velocity, v. So, from Torricelli s equation: v 2g( 1 2 ) We now ave an expression for te velocity of te fuel as it flows from te fuel tank to te atomizer at a eigt ( 1 2 ) below te surface of te reservoir. Were: 1 = 2.1 m = eigt of fuel tank stand + eigt of fuel level in te tank 2 = 0 = eigt of atomizer (ground level) g = 9.81 m/s 2 = acceleration due to gravity = 810 kg/m 3 = density of kerosene Fuel velocity, v v 2g( 1 2 ) v 2*9.81* 2.1 = 6.42 m/s (5) (6) Pressure inside te fuel pipe, P P = ρg (7) = 810 * 9.81 * 2.1 = Pa Figure 3: Manufacturing diagram of te atomizer [6]. 2.4 Flow of kerosene from te fuel tank to te atomizer Using te energy conservation concept to determine te velocity of flow along a pipe from a reservoir, we considered te ideal reservoir in Figure 4 below. Figure 4: An ideal reservoir. Te level of te fuel in te reservoir is 1. Considering te energy situation tere is no movement of te fuel, so kinetic energy is zero but te gravitational energy is: 2.5 Separation losses in pipe flow Tese are losses wic occur as a result of various pipes fittings suc as bends, valves, and also sudden enlargement and contraction of te pipe. For losses due to friction, using Darcy -Weisbac equation: 2 l v f f d 2g But, Reynolds number for a pipe is given by: vd Re u vd u Were: v = 6.4 m/s = fuel velocity u = kinematic viscosity (1 cst = 10-6 m 2 /s) = 2.2 * 10-6 m 2 /s d = ydraulic diameter d 4 r r 2 2 o i 2ro 2r o 2ri r i (8) (9) (10) d o = m = external diameter of fuel pipe, r o = 5.1*10-3 m = external radius of fuel pipe. d i = 0.01 m = internal diameter of fuel pipe, r i = 5*10-3 m = internal radius of fuel pipe

4 d = 2* 10-4 m and vd Re u Were: d i = 0.01 m = internal diameter of fuel pipe f = ead loss due to friction in te pipe f = friction coefficient v = 6.42 m/s = fuel velocity l = 2.97 m = lengt of fuel pipe Re = Pa vd u Determination of flame lengts at different injected pressures Te pressurized oxygen and kerosene were well mixed before exiting te atomizer. Visible atomization process was observed as te mixture of oxygen and kerosene was seen discarging troug te atomizer (nozzle) orifice. Te atomizer sprays te mixture and it was ignited using a ligter. Wit te elp of te pressure regulator attaced to te oxygen bottle; at different injected pressures ranging from 3 to 15 bars, te flame lengts were captured using a camera [6]. Since te flow is laminar as Reynolds number is less tan 2000, 64 f = 0.11 Re 2 l v f = Pa f d 2g 2.6 Test facility A scematic of te test facility is sown in Figure 6, te atomizer was mounted on a tick plate for support and te fuel and oxygen pipes were fixed to te adaptors. Te oxygen for te test was provided from oxygen refrigerant cylinder of 150 bars pressure capacity. A pressure regulator was used to keep te pressures constant at any pressure injection feed. Te different parts of te atomizer were assembled togeter in te following order: te oxygen inlet and fuel inlet adapter was bolted to te atomizer and finally, te fuel ose and te oxygen ose was connected to te fuel inlet and to te oxygen inlet adapter respectively. Te burning device operates on te principle of combustion in wic oxygen (air) is required or supplied to enance burning [6]. 2.7 Experimental investigations Te fuel (kerosene) was stored in a tank of 0.8 m lengt and 0.4 m diameter mild steel of 27 litres capacity, te liquid flows under gravity and injected troug a small diameter orifice at te centerline of te atomizer. A calibrated valve was used to regulate te flow of kerosene from te fuel tank. Pressurized oxygen (air) was injected troug a pipe to te fuel pipe and te fuel drops troug te pipe duct at an angle of 45 o to form te liquid streams. Te nozzle increases te velocity of te fluid [6] Determination of time taken to melt metal carge Te volume of kerosene in te fuel tank was noted. Te crucible pot and te metal carge to be melted was prepared and weiged using digital weiging macine. Te crucible pot wit te metal inside was placed in a pit furnace and te taps tat controls te fuel and te oxygen line was opened. Once visible atomization process was observed te fuel was ignited. As te fuel was ignited, a stop watc was used to ceck ow long it took to melt a particular metal. Before and after eac test (melting of te metals), te volume of kerosene was noted and te difference between te initial and te final volume gives te amount of kerosene used to melt te metal [6]. 3. RESULTS AND DISCUSSIONS Figure 5: Designed atomizer [6]. Te experimental setup was developed for te measurement of spray caracteristics like spray lengts and flame lengts at different pressures varying from 3 to 15 bars. Potograps were taken by ig speed camera to capture te spray lengts and flame lengts at different injected pressures. Also te time taken to melt te selected materials was also obtained Determination of spray lengts at different injected pressures Te rate of flow of fuel from te tank was kept constant for all te experiments by controlling te valve. Te pressurized oxygen and kerosene were well mixed before exiting te atomizer. Visible atomization process was observed as te mixture of oxygen and kerosene was seen discarging troug te atomizer (nozzle) orifice. Using a pressure regulator attaced to te oxygen bottle; at different injected pressures ranging from 3 to 15 bars, te spray lengts were captured using a camera [6]

5 Figure 8 sows tat entrainment of secondary air starts beyond 700 mm. As velocity of te spray decreases as it leaves te nozzle, te spreading of te spray becomes more pronounced because of te decrease in te velocity of te spray. Figure 8: Spray for α = 60 and ΔP = 6 bars. Figure 6: Complete assembly of te system [6]. Table 4: Spray lengts at different injection pressures (Measured values) Figure 9 sows a jet wic was produced wen te fluid (oxygen and kerosene) was discarged troug te nozzle. Te spray was linear up to 600 mm because te velocity at te tip of te nozzle is more wic extends until was affected by te air from te atmospere. Beyond point 600 mm, entrainment of secondary air occurs. Te free jet was produced wen te fluids was discarged in te surrounding wit no confinement. S/N Pressures Spray lengts (mm) (Bars) Figures 7 to 11 sows te sprays from te produced atomizer aving alf spray angle of 30 at different injected pressures. Figure 7 sows tat entrainment of secondary air starts beyond 400 mm. As velocity of te spray decreases as it leaves te nozzle, te spreading of te spray becomes more pronounced because of te decrease in te velocity of te spray. Figure 9: Spray for α = 60 and ΔP = 9 bars. Te spray as observed in Figure 10 sows tat entrainment of secondary air starts around 700 mm; before tis point, te spray was seen to be linear because te velocity of te spray was greater at tat region. Te entrainment of te surrounding in te jet increases te mass of te jet but decreases te velocity of te jet as it sprays. Figure 7: Spray for α = 60 and ΔP = 3 bars. Figure 10: Spray for α = 60 and ΔP = 12 bars

6 Te entrainment of te surrounding as observed (beyond 900 mm) in Figure 11 was due to increase in mass of jet wic depends on te difference in te momentum flux witin te jet and tat of te surrounding (note tat as te jet was discarged into a still surrounding, te surrounding was set in motion). Te entrainment of te surrounding will continue as long as te difference in te momentum flux exists. Due to entrainment of te surrounding as observed in Figure 13, te axial velocity of te jet decreases making te flame to start spreading beyond 600 mm. Because of te increased velocity of te jet, te flame as seen in te Figure was linear witout any interference from te surrounding air. Figure 13: Flame for α = 60 and ΔP = 6 bars. Figure 11: Spray for α = 60 and ΔP = 15 bars. Figures 12 to 16 sows te flames from te produced atomizer aving alf spray angle of 30 for different injected pressures. Table 5: Flame lengts at different pressures (Measured values) In Figure 14, te caracteristic feature of te flame as it spreads was due to te difference in te density of te jet and te surrounding. A ot jet in a cold surrounding spreads faster tan a cold jet in te same surrounding. Spreading of te flame wic started to occur beyond 1200 mm was due to entrainment of te surrounding. S/N Pressures (Bars) Flame lengts (mm) Te free unconfined jet spreads in te surrounding and te entrainment of te secondary air occurred around 400 mm as observed in Figure 12. Te spreading of te flame was due to entrainment of te surrounding. Te free jet as no confinement and ence can spread till te difference between te momentum flux (mass of te jet * velocity of te jet) of te jet and te surrounding becomes zero. Entrainment of surrounding depends on mass flow rate and jet velocity. Figure 14: Flame for α = 60 and ΔP = 9 bars. Figure 15 sows tat te flame starts spreading as a result of te surrounding air beyond point 1300 mm. Figure 15: Flame for α = 60 and ΔP = 12 bars. Figure 12: Flame for α = 60 and ΔP = 3 bars. Figure 16 sows tat te spreading of te ot flame as a result of te surrounding air started beyond point 1500 mm. It was 369

7 Spray Lengts(mm) Spray Lengts(mm) Flame Lengts(mm) International Journal of Science and Engineering Applications observed tat tere was a witis flame formed very close to te tip of te atomizer; te flame was at te region were te velocity of te jet is ig witout entrainment of te secondary air. Te witis flame indicates tat tere was complete combustion process of te fuel. Effect of injection pressures on flame lengts (Regression values) Pressures(Bars) Figure 18: Effect of injected pressures on flame lengts. Figure 16: Flame for α = 60 and ΔP = 15 bars. Te grap in Figure 17 was gotten from te regression analysis of te two sample items, te injection pressure and te spray lengts. After te calculations, a relationsip between te pressures and te spray lengts was establised, tus te grap. Te equation of best fitting line tat described all te points was establised and te straigt line grap was plotted. At any point on y or x axis te oter corresponding values can be obtained from te oter axis. It was also proven tat te two sample items (te pressure and spray lengt or flame lengt as te case may be) as a linear relationsip tat exists between tem. Effect of injection pressures on spray lengts (Regression values) 1800 Figure 19 is te grap of measured spray lengts versus injected pressures and regression values of spray lengts versus injected pressures plotted to compare ow close te points (measured spray lengts and regression values of spray lengts) are. Te grap of injection pressures versus te regression values of spray lengts is a linear grap but tat is not te case of te grap for injection pressures versus te measured spray values; te discrepancies in te profile of te measured spray lengts was from errors wile taking te readings Effect of injection pressures on spray lengts Regression values Measured values Pressures(Bars) Figure 17: Effect of injected pressures on spray lengts. Te grap in Figure 18 was gotten from te regression analysis of te two sample items, te injection pressure and te flame lengts. After te calculations, a relationsip between te pressures and te flame lengts was establised, tus te grap. Te equation of best fitting line tat described all te points was establised and te straigt line grap was plotted. At any point on y or x axis te oter corresponding values can be obtained from te oter axis. It was also proven tat te two sample items (te pressure and flame lengt) as a linear relationsip tat exists between tem Pressures(Bars) Figure 19: Effect of injected pressures on spray lengts (measured and regression values). Figure 20 is te grap of measured flame lengts versus injected pressures and regression values of flame lengts versus injected pressures plotted to compare ow close te points (measured flame lengts and regression values of flame lengts) are. From te grap it can be deduced tat te readings gotten for te measured flame lengts and te corresponding regression values of flame lengts are almost te same. Te error margin can be said to be negligible

8 Flame Lengts(mm) International Journal of Science and Engineering Applications Effect of injection pressures on flame lengts Pressures(Bars) Figure 20: Effect of injected pressures on flame lengts (measured and regression values). Materials Regression values Measured values Table 6: Time taken to melt metals [6] Quantity (kg) DESCRIPTIONS Time taken to melt material (seconds) Volume of kerosene used (litres) Aluminum Brass 1.2 1, CONCLUSION Te design of te air-blast atomizer was carried out for maximum injection pressure of 15 bars. Te experiments of volumetric flow rate, spray lengts, and flame lengts were carried out wit te injection pressure ranging from 3 to 15 bars. Experiments sowed tat at 3 bars, te spray and flame lengts are small and also tat at 3 bars; liquid film was not breaking into small droplets. As injection pressure increases from 6 to 12 bars, spray lengts and flame lengts increases due to breaking of liquid film into small droplets. Te effect of entrainment of te surrounding air was more at ig pressures. Te design and construction of tis air-blast atomizer like any oter atomizer fabricated locally requires little data and is very easy to construct. Te atomizer produced te desired result for wic it was designed for, in melting te selected metals. It can be concluded tat te airblast atomizer produced can be used for bot surface and pit furnaces using kerosene as fuel [6]. 5. REFERENCES [1] Lefebvre, A. H Air-blast Atomization: Prog. Energy science, vol. 6, pp [2] M. J. Madu, I. S. Aji, B. Martins Design, Construction and Testing of a Burner tat uses an admixture of used engine oil and Kerosene for foundry application. International Journal of Innovative Researc in Science, Engineering and Tecnology, vol. 3, issue 9. [3] Robert Aftel Effect of atomization gas properties on droplet atomization in an air-assist atomizer. MSc. Tesis at Virginia Polytecnic Institute and State University, Virginia. [4] Lefebvre, A. H., and Ballal, D. R. Gas Turbine Combustion Alternative Fuels and Emissions, tird edition. CRC Press. [5] Nigerian National Petroleum Corporation (NNPC) Warri Refining and Petrocemical Company Ltd, Tecnical Report, 4:87. [6] Bennet Ifenna Okoli Design and Performance Evaluation of an Air-blast Atomizer, MSc. Tesis, Department of Mecanical Engineering, Lagos State University, Lagos, Nigeria. [7] Vasiliov, A.P., Koderaftsov, B.M., Korbatinkov, B.D., Ablintsky, A.M., Polyayov, B.M. Palvian, B.Y Principles of Teory and Calculations of Liquid Fuel Jet, Moscow. [8] V. A. Borodin, Yu. F. Dityakin, L. A. Klyacko, V. I. Yayodkin Atomization of liquids, Foreign Tecnology Division. [9] Bayvel, L Liquid Atomization. Hemispere Publising Corporation, Taylor & Francis. [10] Witold Perkowski, Andrzej Irzycki, Krzysztof Snopkiewicz Lukasz Grudzien, and Mical Kawalec Study on kerosene atomization process under ig speed air stream. Journal of Kones Power train and Transport, vol. 18, no 4. [11] Pipatpong Watanawanyoo, Sumpun Caitep and Hiroyuki Hiraara Development of an air assisted fuel atomizer (liquid sipon type) for a continuous combustor. American Journal of Applied Sciences 6 (3):