Evaluation of Steadiness and Drop Size Distribution in Sprays Generated by Different Twin-Fluid Atomizers

a Co esponding au ho : [email p o ec ed]me. u b .cz E alua ion o S eadiness and D op Size Dis ibu ion in Sp ays Gene a ed by Di e en Twin-Fluid A omize s Ma ouš Za emba1,a, Ma ek Mlk ik1 Milan Malý1, Jan Jedelský1 and Mi osla Jícha1 1B no Uni e si y o Technology, Facul y o Mechanical Enginee ing, Depa men o he he modynamics and en i onmen al enginee ing, Technická 2896/2, 616 69 B no, Czech Republic Abs ac . Twin- luid a omize s unde wen a signi ican de elopmen du ing he las ew decades. They a e common in many indus ial applica ions such as uel sp aying, mel a omiza ion and ood p ocessing. This pape is ocused on he e alua ion o ou di e en win- luid a omize s. The aim is o compa e he quali y o sp ays gene a ed by a ious a omize s wi h simila dimensions and in he same ope a ing egimes. A phase- Dopple anemome y (PDA) and pa icle image elocime y (PIV) we e used o measu e sp ay cha ac e is ics such as eloci y and size o he d ople s. Measu ed da a we e used o compa e d ople size dis ibu ion and o e alua e s eadiness o he sp ay. Visualisa ions we e made o suppo measu ed da a and o cla i y he p inciples o p ima y a omiza ion and i s in luence on he sp ay. 1 In oduc ion Twin- luid a omize s a e one o he mos uni e sal a omize s in indus y. They a e designed o a ious applica ions, o example: a omiza ion o highly iscous and was e uels, a omiza ion o mel s, ood p ocessing indus y, powde d ying and o he s. Due o he a ious ields o usage many di e en designs o win- luid a omize s we e de eloped. Gene ally, hese a omize s can be di ided in o wo ca ego ies: a omize s wi h in e nal o ex e nal mixing o phases. The ex e nal mixing a omize ope a es on he p inciple o gene a ing ela i ely high eloci y di e ence be ween gas and liquid. The eloci y di e ence helps o disin eg a e liquid [1]. A omize s wi h in e nal mixing a e cha ac e ized by he in e nal mixing chambe whe e usually wo phases a e mixed. Two-phase low is c ea ed inside he a omize and i is pushed downs eam h ough he a omize and disin eg a es a e passing he exi o i ice. Liquid is sha e ed by he apidly expanding gas [2-4]. This pape is ocused on ou di e en win- luid a omize s designed o sp aying o ligh hea ing oil. A omize s we e examined mos ly on he basis o inpu pa ame e s such as ope a ing p essu e, gas o liquid a io (GLR), dimensions o he nozzle, liquid physical p ope ies and ambien condi ions and i s in luence on he esul ing sp ay quali y [3-5]. Majo i y o s udies which deal wi h win luid a omize s a e ocused di ec ly on one ype o a omize . Howe e , a compa ison o mo e di e en designs o win- luid a omize s has no been su icien ly desc ibed ye . Compa ison can only be made unde he assump ion o using he same me hodology, he same liquids and he same ope a ing condi ions. We assume an applica ion o a omize s o uel sp aying o combus ion p ocesses. One o he mos impo an pa ame e s o hese sp ays is d ople size dis ibu ion. The size o d ople s in he sp ay is o en desc ibed using he Sau e mean diame e [2]. Ano he impo an aspec o he sp ay quali y is he s eadiness [6]. Uns eadiness has a nega i e impac on he noise le el o he combus ion. I aises he load o he combus ion chambe and i causes p oblems in he bu ne s a ing sequence [7]. Wo ke s Edwa ds and Ma x [8] ha e published a me hod designed o e e escen sp ays o e alua ing he s eadiness o he sp ay using inhomogeneous Poisson s a is ics and ime dependen poin measu emen s in sp ays. Se e al wo ke s used his me hod a e wa ds o sp ay examina ion, o ins ance: [7-11]. Mos ly o e alua ion o he in luence o ope a ional condi ions, physical p ope ies o he a omized liquid and dimensions o he a omize on he sp ay uns eadiness. The aim o his pape is o use hese cha ac e is ics o sp ays and compa e ou di e en win- luid a omize unde he same ope a ing condi ions and wi h he same me hodology. 2 Ma e ials and me hods All p esen ed da a was acqui ed by cold es ing o win luid a omize s in he Sp ay labo a o y a he Ene gy depa men , Facul y o Mechanical Enginee ing, B no Uni e si y o Technology. DOI: 10.1051/ C Owned by he au ho s, published by EDP Sciences, 2015 / 021 201epjcon EPJ Web o Con e ences , 021 59 92 2 (2015) 1 1                      !" 8 8 A icle a ailable a h p://www.epj-con e ences.o g o h p://dx.doi.o g/10.1051/epjcon /20159202118 2.1 Tes bench Examined a omize s we e ope a ed on a es bench equipped wi h uel and ai supplies. Ligh hea ing oil was used as an a omizing uel. Densi y o he oil is 874 kg·m-3 and dynamic iscosi y is 0.0185 kg·m-1·s-1 a 20 °C. Oil was pushed om a p essu e uel ank h ough a il e , egula ion al e and mass low me e in o an a omize . Comp essed ai was used as a omizing gas. I was aken om he cen al p essu ized line. Ai was pushed h ough a il e and egula ion al e in o he a omize . De ailed desc ip ion o he es bench can be ound in [12]. 2.2 Phase-Dopple Analyse Figu e 1 Schema ic layou o he PDA sys em, op iew. A 2D ib e based phase-Dopple anemome e (PDA) by Dan ec Dynamics was used o measu ing he d ople eloci y (in adial and axial di ec ion ela i e o he axis o he sp ay) and d ople size. The sys em measu es ime esol ed da a which enables us o make hei empo al analysis. Measu emen s we e pe o med in axial dis ance o 100 mm om he exi o i ice o an a omize in one adial p o ile. Schema ic layou can be seen in igu e 1. This pa icula dis ance was chosen due o he ac ha in his dis ance ully de eloped sp ay is expec ed [5]. This means ha we assume inished a omiza ion p ocesses: p ima y and seconda y a omiza ion. De ailed desc ip ion o he measu ing sys em can be ound in [12]. 2.3 Pa icle Image Velocime y Figu e 2 Schema ic layou o he PIV sys em, op iew. A PIV sys em was used o isualize he eme ging wo phase low in he egion nea o he exi o i ice o he a omize . Sp ay images we e cap u ed by a CCD PIV came a (TSI PIVCAM 13-8) wi h a ached Nikon 60 mm lens. The examina ion a ea o in e es is o dimensions 41 × 50 mm which gi es esolu ion o 40 μm/pixel. A se ies o images was cap u ed and a cha ac e is ic image o sp ay pa e n is displayed in igu e 4. De ailed desc ip ion o he PIV sys em can be ound in [13]. 2.4 A omize s Fou di e en win- luid a omize s we e examined in his pape . They we e ope a ed a low p essu e 0.07 MPa and h ee egimes o GLR om 2.5% o 10%. The ope a ing condi ions can be ound in able 1. The p inciple o how ai o liquid is injec ed in o he mixing chambe will be desc ibed in he ollowing pa ag aphs. Table 1. Ope a ing condi ion a he p essu e o 0.07 MPa. A omi ze GLR (%) Mass low a e (kg/hou ) LHO Ai CFT 2.5 4.30 0.11 5 3.04 0.15 10 1.88 0.19 OUIG 2.5 4.08 0.10 5 2.65 0.13 10 1.70 0.17 OUIL 2.5 3.75 0.09 5 2.65 0.13 10 1.72 0.18 Y 2.5 3.49 0.09 5 2.54 0.13 10 1.68 0.17 2.4.1 CFT a omize This ype o a omize was de eloped in he Sp ay labo a o y a he Ene gy depa men , Facul y o Mechanical Enginee ing, B no Uni e si y o Technology as a p o o ype o a new design o win- luid a omize . The o iginal design was based on he indings by wo ke s Chin [14], Fe ei a [15], Tamaki [16]. The liquid is injec ed in o he mixing chambe pe pendicula ly o he main axis o he a omize . Two liquid s eams poin ed agains each o he in e ac wi h he ai s eam which is pushed h ough ou holes. These holes a e il ed unde a pa icula angle in angen ial and adial di ec ion. Swi ling mo ion o he ai imp o es mixing o he liquid and ai . The wo-phase low is pushed downs eam o he exi o i ice and i expands in o ambien ai a e wa ds. A schema ic layou can be ound in igu e 3d. EPJ Web o Con e ences 02118-p.2 2.4.2 Ou side-in gas e e escen a omize A s anda d ou side-in gas e e escen a omize (OUIG) was examined. This a omize wo ks on a p inciple o injec ing p essu ized ai in o a liquid s eam h ough a se o ae a o holes. A schema ic layou o he a omize is p esen ed in igu e 3b. Inside he a omize is a mixing chambe whe e uel is mixed wi h ai . Ideally small bubbles o ai a e c ea ed inside he liquid s eam, so called bubbly low [14]. This egime is ypical o alues lowe han 5% o GLR [3, 5]. Highe alues lead o change o in e nal low pa e n om bubbly low o an annula low. This pa e n is cha ac e ized by he liquid mass concen a ed on he walls o he a omize and ai co e in he cen e o he low. A wo phase low is hen pushed downs eam o he exi o i ice and i is disin eg a ed when i passes h ough he exi o i ice [17]. The cha ac e o he wo phase low depends mainly on he GLR and on he dimensions o he mixing chambe [3, 5, 18]. The e o e di e ences be ween examined ope a ing egimes a e expec ed. 2.4.3 Ou side-in liquid e e escen a omize The design o his a omize co esponds o he design o he p e iously desc ibed s anda d OUIG a omize . The only di e ence is ha uel and ai po en ies a e swi ched as i can be seen in igu e 3c. The liquid is injec ed in o gas s eam in o mixing chambe h ough a se o ae a o holes. 2.4.4 Y-Je a omize As he name sugges s, his a omize wo ks on he basis o injec ing uel in o he ai s eam unde a pa icula angle, see igu e 3a. Acco ding o [19-21] he in e nal low egime depends on liquid and gas mass lows, and GLR and dimensions o he mixing chambe . The a omiza ion p ocess is di ided in o h ee phases acco ding o [21]. Fi s ly, when he liquid s eam eaches he ai je inside he mixing chambe , i is pa ially dis up ed. This mode is called di ec collision. Secondly, he mix u e lows along he mixing chambe and annula liquid ilm is o med. The dis o ed ai s eam is s aigh ened up in he axial di ec ion, and some d ops a e gene a ed by he shea ing ai low. The in e nal low inside he mixing chambe has annula cha ac e o mos egimes. Thi dly, he mix u e eaches he exi o i ice and liquid ilm is disin eg a ed in o ligamen s and hen in o d ops by he expanding ai co e. Acco ding o [21], he d ople size dis ibu ion depends on he liquid ilm hickness inside he mixing chambe which is in luenced by he gas and liquid momen um a io. 2.5 S eadiness o he sp ay – measu emen me hod Acco ding o he heo e ical amewo k p esen ed by wo ke s Edwa ds and Ma x [8] i is possible o e alua e s eadiness o he sp ay using ime-based mul ipoin s a is ics o sp ays. S eady sp ay is de ined as he sp ay whose in e pa icle a i al ime dis ibu ion obeys inhomogeneous Poisson s a is ic. Consequen ly, uns eady sp ay is de ined as sp ay whose in e pa icle a i al ime dis ibu ion does no obey inhomogeneous Poisson s a is ics. De ailed desc ip ion o he me hod is p esen ed in [22]. The le el o uns eadiness is exp essed wi h alue  which ep esen s uns eadiness o he sp ay. This alue was calcula ed using ime esol ed da a om PDA sys em. 2. In eg al Sau e mean diame e As i was men ioned, examined a omize s we e o iginally designed o sp aying o highly iscous liquid uels. Rega ding applica ions in combus ion p ocesses i is impo an o know a d ople size dis ibu ion o he whole sp ay. The mos common pa ame e which desc ibes he d ople size dis ibu ion in sp ays is Sau e mean diame e (D32). This alue de e mines d ople dis ibu ion in one pa icula poin in he sp ay. Fo he desc ip ion o he whole sp ay using one speci ic alue wo ke s Jedelský e al. [5], deduc ed an in e gal Sau e mean diame e (ID32) which desc ibes global sp ay quali y wi h only one numbe . De ailed desc ip ion o he alue o ID32 can be ound in [5]. 3 Resul s and discussion Images o examined sp ays a e p esen ed in igu e 4. A di ec impac o he ope a ing egime on he sp ay Fi Schema ic la y ou o he PIV s y s em. EFM 2014 02118-p.3 6 gu e 3 de elopmen in nea -a omize a ea can be seen. A he lowes alue o GLR (2.5%) i is shown ha all a omize s p oduce la ge s uc u es – ligamen s. The eme ging wo phase low is ela i ely na ow and dense o all a omize s. Compa ing be ween a omize s i can be s a ed ha he Y-Je a omize p oduces he longes ligamen s which lead o la ge d ople s. This is in ag eemen wi h p e ious wo ks [19, 21]. Also he alue o ID32 in igu e 5 p o es his s a emen . Acco ding o [21] he amoun o uel inside he mixing chambe in luences he hickness o he annula liquid shee which has a di ec impac on he diame e o ligamen s and consequen ly on he d op size dis ibu ion. The CFT a omize pe o ms simila ly. A dis up ed liquid s eam can be seen o GLR = 2.5%. The majo i y o he mass is concen a ed in he cen e o he sp ay. The ligamen s a e p esen ed e en in he bo om o he igu e which poin s o he un inished b eakup p ocess. As expec ed, he CFT a omize has he second highes alue o ID32 in his egime. In con as , he OUIG and OUIL a omize s p oduce ela i ely well disin eg a ed sp ay e en in he egion nea he exi o i ice o he a omize s. E en hough, hey ha e he same diame e o he exi o i ice as he CFT. I should be no ed ha hese a omize s ope a e in he same ope a ing egime, mo eo e wi h simila liquid mass lows acco ding o able 1. Fo ha eason we assume ha he mixing p ocess has a di ec impac on he sp ay pe o mance especially in he low GLR egimes. The di e en sp ay pa e ns in luence he alue o ID32. The smalles d ople s a e p oduced by he OUIG a omize in his pa icula egime, see igu e 5. This co esponds o he p e ious wo ks on he e e escen a omize whe e i was p o ed ha a low GLR egimes, du ing which he a omize ope a es a bubbly in e nal low, he a omize is able o pe o m small d ople dis ibu ion hanks o he speci ic b eakup p ocess [3, 4]. When we inspec ed images om egimes o GLR = 5% we obse ed ha sp ay is mo e dispe sed. Signi ican di e ence o he sp ay pa e n can be seen o he Y-Je a omize which again p oduces a na ow sp ay wi h la ge ligamen s. In compa ison wi h p e ious egime, he ligamen s a e smalle . This co esponds o he alue o ID32 as can be seen in igu e 5. The d op o ID32 alue is p esen o all a omize s as was expec ed. The amoun o a omizing ai was inc eased as well as he po en ial ene gy a ailable o he a omiza ion. The a omizing ai educes he hickness o liquid s uc u es and inc eases i s eloci y. The di e ence o he alue o he ID32 om 5% o 10% o GLR is conside ably lowe han di e ence om 2.5% o 5%. When we inspec a classic e e escen a omize (OUIG) he ID32 di e ence be ween 2.5% and 5% is caused by he change o he in e nal low pa e n. F om bubbly o slug o annula [1, 23, 24]. Which also means he e is a change in he b eakup mechanism [25]. The change om 5% o 10% means ha he in e nal low pa e ns migh emain he same. Bu he amoun o he ai mass low is inc eased. Thus, only ai mass low is esponsible o he change o ID32 alue. The in e nal wo-phase low pa e ns in CFT and OUIL a omize s we e no examined p ope ly be o e. So i can be only assumed ha he abo e men ioned p inciple ela ed o he OUIG a omize is p esen also in he o he wo a omize s. Figu e 4 Typical images o he sp ay pa e ns in a ea nea o he a omize . Regimes a GLR = 10% a e cha ac e ized by well dispe sed sp ay in whole image. E en he Y-Je a omize p oduces ela i ely well dispe sed sp ay bu compa ed o he o he a omize s, ligamen s and o he liquid s uc u es a e s ill p esen . As i was men ioned in [21] he Y-Je a omize wo ks on he p inciple o he annula liquid shee disin eg a ion. The liquid is accele a ed inside he mixing chambe by he high speed lowing ai . The i s p essu e d op o he liquid is a e en e ing he mixing chambe , see igu e 3. The second p essu e d op comes a e mix u e passes h ough he exi o i ice. The con e sion o an ini ial po en ial ene gy o he kine ic ene gy is hus di ided in o wo s eps. This leads o he lowe p essu e d op a he exi o he a omize . Tha is why, liquid s uc u es and ligamen s a e s ill p esen in he egimes o GLR = 10%. Again, he alue o ID32 is he highes o his a omize in con as o he o he h ee a omize s. The e alua ion o he s eadiness o examined sp ays show demons able dependency on he ope a ing egime o all a omize s. Bu no simple end was obse ed. EPJ Web o Con e ences 02118-p.4 Figu e 5 ID32 and  ela ionship a ying GLR alue o each a omize . The CFT a omize p oduces he mos uns able sp ay a he highes GLR alue. Wi h dec easing GLR o 5% he s eadiness o he sp ay inc eases and wi h u he dec ease in GLR, sp ay s eadiness dec eases. Thus we can obse e local minimum o he sp ay uns eadiness a 5% GLR. The alue o  o a omize s OUIG and OUIL inc eases as he GLR dec eases as expec ed [9]. Fo OUIG a omize pe iodical pulsa ions we e obse ed as can be seen in igu e 6 which suppo s he esul s om alues o . In his igu e sepa a ed clouds o do s can be obse ed. Each cloud ep esen s one pulse o liquid. I was de ec ed as se e al d ople s wi h eloci y a ying om abou 2 m/s o 16 m/s. The pe iod o he pulsa ing egime is app oxima ely 110 ms. This beha iou migh be a ibu ed o he in e nal low pa e n. As i was men ioned in [7, 8, 9], e e escen sp ays a e inhe en ly uns able and in e nal wo-phase low pa e ns migh lead o he highly uns eady beha iou o he sp ay. Fo ins ance, he slug low is one o he mos uns able pa e ns. The obse ed pulsa ing egime is a special case. Fo o he a omize s simila beha iou no ye been obse ed. The Y-Je a omize pe o ms di e en ly. The mos uns able egime is o GLR = 5%. Then he alue o  dec eases o bo h GLRs; 2.5% and 10%. No pulsa ing egime was obse ed as i was men ioned o he OUIG a omize . Acco ding o igu e 4 o Y-Je a omize , we migh assume ha sp ay o 2.5% is a special case whe e he eme ging liquid looks like dis u bed liquid s eam which p oduces la ge liquid s uc u es. Thus he Poisson s a is ic migh no be a p ope ool o e alua ion o he sp ay s abili y. Mo eo e , he da a a e o his pa icula egime was nea ly hal o he da a a e o he o he h ee a omize s (da a a e o he OUIL a omize was a ound 1 kHz and o he Y-Je a omize i was 470 Hz). When we inspec he mos uns able sp ay (Y-Je a omize a 5% o GLR) we can obse e ha he d ople axial eloci y has bimodal dis ibu ion. Figu e 6 Time dependency o d ople s axial eloci y o OUIG a omize a 2.5% o GLR in he cen e o he sp ay. Figu e 7 Time dependency o d ople s axial eloci y o Y-Je a omize a 5% o GLR in he cen e o he sp ay. This end migh be explained as andom pulsa ing egime. Majo i y o d ople s a e p esen ed in low eloci y ield (up o 5 m/s) and he es o measu ed d ople s in he highe eloci y ield ( om 6 o 11 m/s). This bimodal dis ibu ion shows ha he e migh be wo di e en mechanisms how d ople s a e o med. E en hough he OUIG a omize has pe iodical pulsa ing egime Y-Je a omize is mo e uns able due o he bimodal eloci y dis ibu ion. EFM 2014 02118-p.5 4 Conclusions The aim o his pape was o e alua e he quali y o he sp ay conside ing he size o d ople s and s eadiness. These wo ac o s a e impo an o se e al easons especially o he applica ions in indus ial p ocesses. Gene ally, i can be s a ed ha he Y-Je a omize p oduces he la ges d ople s in all egimes. One assump ion o his is ha liquid in Y-Je a omize expe iences wo p essu e d ops and hus he inal p essu e d op is no as high as o o he a omize s which esul s in la ge d ops. Mo eo e , he s abili y o he Y-Je a omize s is poo which suppo s esul s o he ID32 alue. Unusual beha iou can be seen o GLR = 2.5% which was a ibu ed o he unusual sp ay pa e n and low da a a e. A omize s OUIG and OUIL pe o m simila ly. Thei s abili y inc eases wi h he alue o GLR and he alue o ID32 inc eases as alue o GLR dec eases as expec ed. The main di e ence be ween hese a omize s is in he s abili y. The OUIG is mo e uns eady a egime o GLR = 2.5% whe e pe iodical pulsa ing sp ay was obse ed. Regimes o 5% and 10% GLR a e mo e s able o he OUIL a omize . I should be poin ed ou ha hese wo a omize s ha e he same dimensions bu he ai and liquid po s a e in e sely connec ed. The newly de eloped CFT a omize beha es in a simila way as he OUIL and OUIG a omize o he ID32 alues. The main di e ence can be ound in s abili y. The local minimum is o GLR = 5% whe eas he maximum alue o  is o 10% o GLR. P esen ed esul s show ha o he classic cons uc ion o e e escen a omize s (OUIG and OUIL) expec ed beha iou was ound. The o he wo a omize s pe o m di e en ly. This migh be a ibu ed o he ac ha hey ope a e wi h di e en liquid b eakup mechanisms. The b eakup mechanism was examined be o e only o he OUIG and Y-Je a omize bu CFT and OUIL a omize s a e s ill in de elopmen . The s abili y o he Y-Je and CFT a omize s has no been examined be o e as well. Thus compa ison wi h p e ious esul s is no possible. Deepe knowledge o he mixing and b eakup p ocesses should be ob ained in o de o be able o explain such a complex p ocess as i is in he case o hese a omize s. Acknowledgemen s Au ho s acknowledge inancial suppo om p ojec No. P101/11/1264 unded by he Czech Science Founda ion and NETME Cen e, egional R&D cen e buil wi h he inancial suppo om he Ope a ional P og amme “Resea ch and De elopmen o Inno a ions” wi hin he p ojec “NETME Cen e (New Technologies o Mechanical Enginee ing)”, Reg. No. CZ.1.05/2.1.00/01.0002 and, in he ollow-up sus ainabili y s age, suppo ed h ough NETME CENTRE PLUS (LO1202) by inancial means om he Minis y o Educa ion, You h and Spo s unde he “Na ional Sus ainabili y P og amme I”. Re e ences 1. LI, Z. H., Y. X. Wu, C. R. Cai, H. Zhang, e al., Fuel, 97, 306-314, (2012). 2. Le eb e, A. A omiza ion and Sp ays. Ed ion ed.: Taylo & F ancis, (1988), ISBN 9780891166030. 3. So ani, S. D., P. E. Sojka AND A. H. Le eb e E e escen a omiza ion. P og ess in Ene gy and Combus ion Science, 27, 483-521, (2001). 4. Kons an ino , D., R. Ma sh, P. Bowen And A. C ay o d, A omiza ion and Sp ays, 20, 525-552, (2010) 5. Jedelský, J., M. Jícha, J. Sláma and J. O áhal, Ene gy & Fuels, 23, 6121-6130, (2009). 6. Le eb e, A. H. GAS Tu bine Combus ion, Second Edi ion. Ed ion ed.: Taylo & F ancis, (1998). ISBN 9781560326731. 7. Jedelský, J. and M. Jícha, A omiza ion and Sp ays, 18, 49-83, (2008). 8. Edwa ds, C. F. and K. D. Ma x A omiza ion and Sp ays, 5, 435-455, (1995). 9. Luong, J. T. K. and P. E. Sojka, A omiza ion and Sp ays, 9, 87-109, (1999). 10. Liu, M., Y. Duan and T. Zhang, Expe imen al The mal and Fluid Science, , 34, 657-665, (2010). 11. Liu, M., Y. Duan, T. Zhang and Y. Xu Expe imen al The mal and Fluid Science, 35, 190-198, (2011). 12. Jedelský, J., M. Za emba, M. Malý and M. Jícha, EPJ Web o Con e ences. 67, (2014). 13. Za emba, M., M. Malý, J. Jedelský and M. Jícha, AIP Con e ence P oceedings, 1608, 272-275, (2014). 14. Chin, J. S. and A. H. Le eb e, Jou nal o Enginee ing o Gas Tu bines and Powe , 117, 266- 271, (1995). 15. Fe ei a, G., J. Aan onio Ga cia, F. Ba e as, A. Lozano, e al., Fuel P ocessing Technology, 90, 270- 278, (2009). 16. Tamaki, N., M. Shimizu & H. Hi oyasu, P oc. ILASS-Eu ope, 412–417, (2004). 17. Chin, J. S. and A. H. Le eb e, A omiza ion and Sp ays, 3, 463-475, (1993). 18. Sen, D., M. A. Balzan, D. S. Nobes and B. A. Fleck, Jou nal o Visualiza ion, 17, 113-122, (2014). 19. Mullinge .PJ and N. Chigie , Jou nal o he Ins i u e o Fuel, 47, 251-261, (1974). 20. Paci ico, A. L. and J. I. Yanagiha a, Jou nal o he B azilian Socie y o Mechanical Sciences and Enginee ing, 36, 13-22, (2014). 21. Song, S. and S. Lee, A omiza ion and Sp ays, 6, 193- 209, (1996). 22. Jedelský, J., Jícha, M.; Beins ein, Z., In P oceedings o 11 h T iennial In e na ional Con e ence on Liquid A omiza ion and Sp ay Sys ems.1. Den e , CO: Uni e si y o Den e , s. 1-6, (2009). 23. Huang, X., X. Wang and G. Liao, Jou nal o Visualiza ion, 11, 299-308, (2008). 24. Le eb e, A. H. and J. S. Chin, A omiza ion and Sp ays, 3, 463-475, (1993). 25. San angelo, P. J. and P. E. Sojka, A omiza ion and Sp ays, 5, 137-155, (1995). EPJ Web o Con e ences 02118-p.6