Home

fulltext - MDH DiVA

Contents

1. 1 9 17 2533 41 49 57 65 73 81 pixel 1 120 60 100 H9OHO standard 50 H90H3 standard 2 80 2 40 0 0 E E g 20 H9OHO ee H90H3 40 blocked filmholes 20 blocked filmholes 20 142 10 142 0 0 1 9 17 2533 41 49 57 65 73 81 1 9 17 25 33 41 49 57 65 73 81 pixel pixel Standard VS Blocked filmholes 1 2 Standard VS Blocked filmholes 1 2 50 80 40 H90H5 standard f H90H8 standard E 30 50 5 5 40 20 H90H5 2 20 H90H8 blocked_filmholes T 20 blocked_filmholes 10 142 142 10 0 0 1 9 17 25 33 41 49 57 65 73 81 1 9 17 2533 41 49 57 65 73 81 pixel pixel Standard VS Blocked filmholes 3 4 Standard VS Blocked filmholes 3 4 120 50 100 H9OHO standard 40 H90H3 standard gt gt E 00 30 E 60 S H90H0 20 H90H3 40 blocked filmholes blocked filmholes 20 3 4 10 3 4 0 0 1 9 17 2533 41 49 57 65 73 81 1 9 17 25 33 41 49 57 65 73 81 pixel pixel Standard VS Blocked filmholes 3 4 Standard VS Blocked filmholes 3 4 50 70 40 H90H5 standard 60 H90H8 standard E 30 TI 2 2 407 g 20 H90H5 2 30 H9OHB Blocked filmholes 20 blocked filmholes _ 10 344 10 344 9 17 25 33 41 49 57 65 73 81 pix
2. 1000000 100000 H90H5 standard H90H5 obstacle plate 10000 1000 100 10 1 1 3 5 7 9 11 13 15 17 19 21 23 25 Intensity 47 Standard vs Obstacle plate PDF Standard vs Obstacle plate 1 8 15 22 29 36 43 50 57 64 71 78 85 pixel 70 10000000 MO 60 Kee H90H8 standard gt 90 T mou H90H8 obstacle plate E 40 H9OHB standard 2 10000 2 30 H90H8 obstacle plate 3 1000 20 100 0 1 1 9 17 25 33 41 49 57 65 73 81 1 3 5 7 9 11131517 19 21 23 25 Pixel Intensity 100 100000000 10000000 80 gt H9OHO standard H H9OHO standard en 5 100000 2 2 RHS H90H0 obstacle 2 40 H9OHO obstacle 5 cylinder 4 Z 1000 cylinder 20 100 Poy Dee A 1 3 5 7 9 11 13 15 17 19 21 23 25 Pixel Intensity Standard vs Obstacle cylinder PDF Standard vs Obstacle cylinder 50 10000000 Ro 40 1000000 100000 30 Poe ecard z H90H3 standard 0 2 10000 2 20 H90H3 obstacle 5 1000 H90H3 obstacle cylinder cylinder 10 100 10 1 8 15 22 29 36 43 50 57 64 71 78 85 1 3 5 7 9 1113 15 17 19 21 23 25 Pixel Intensity Standard vs Obstacle cylinder PDF Standard vs Obstacle cylinder
3. MatLab Fps Drv Low Emissions Nitrogen Oxides Siemens Industrial Turbo machinerv Revnolds Number Pressure Loss Coefficient Graphic User Interface Siemens Gas Turbine Parts per million Probabilitv Densitv Function Matrix Laboratorv Frames per second 6 7 Apparat E 2 BAC NTT PP 3 3 1 Siemens Industrial Turbomachinery AB sse eee eee 3 9 2 DG RE RRE ue et it a o e So 3 3 3 37 generation DLE burner oo ecesceccssessecsseceesssseceesecsecersecsrearsecersessecaesesseearenceees 4 IVT COMO ME 6 4 1 Development of evaluation strategy sees eee 6 ALI Rost processins t00l Mata i i ta 6 4 1 12 Testeyaluationy Tepeadtabilit yc EE 6 4 1 3 DLE plastic burner experiments with geometrical alterations 6 4 2 Water rig evaluation strategv sese T E PSUC DUM O aoee aAa A 8 4 4 Risk identification and user manual 8 AD sle HT 9 ee 10 4 7 Video Capture and collection QO VICA sese 10 e UD 11 eM UU UM E ME E ARE Er E VI i 11 A VO ANN e i 12 a d b T E a E 12 4 12 The Graphical User Interface eee eee 12 A VO Mean ATES IY TITTA OM CI EE 13 ATA POR 13 AD SIG e KR DIT itet a i a Ma tuta ti eee ees 13 A TO Alpi KE DE e ete TC mere A ea 15 SIE EIDE ee nara ean Te nee a taka 15 ek E Ee TEE 16 RESONS eee MP 17 Od eeel 17 5 1 1 Video clip edition Intensity by radius sss esse eee eee 17 DZ STS SS AC EEN 19 Fld QUIM OF E ee TTT 19 SAA ehre 20 alee e MOS De S E E EE 20
4. The reason of the geometrical alterations was to look over if better fuel air mixture could be obtained and accordingly thus to reduce hotspots in the burner and in that case reduce NOx emissions Sammanfattning I dagens industriella v rld ar kraven h ga ur milj perspektiv Siemens Industrial Turbomachinery AB SIT AB r ett f retag som r v ldigt m n om en god milj och l gger stor m da p att utveckla f rbr nningsprocesser i sina motorer som i sin tur reducerar NO utslappen kv veoxider De forskar ocks mycket i alternativa br nslen vilket r mer milj v nligt I avsikt att minska emissioner konstruerade SIT en vattenrigg f r att studera dynamiken p fl det 1 en DLE br nnare Dry Low Emission Ett analysprogram skapades f r att utv rdera olika experimentella tester av flodesdvnamiken i en 3 generation DLE brinnare tillh rande gasturbinen SGT 800 Malet med examensarbetet var att s kerst lla repeterbarheten och d rmed tillforlitligheten av de experimentella testresultaten f r fortsatt arbete 1 framtida projekt kring DLE br nnare N r repeterbarheten uppn ddes utf rdes olika geometriska ndringar i DLE br nnaren Detta gjordes avsiktligt f r att se om b ttre br nsleluftblandning kunde uppn s f r att reducera hotspots omr den med h g koncentration av br nsle i br nnaren och d rmed NO emissioner Abbreviations DLE NO SIT Re PLC GUI SGT PPM PDF
5. 12 4 13 Mean intensity imaging The existing function in MatLab post processing tool called mean intensity function calculates the mean intensity of each pixel in the entire movie sequence Further to this it displays how the fuel concentration is distributed as a mean value picture of the video clip The script summarizes the intensity values of a specific pixel from the green matrix frame by frame and then divides it with number of frames that the movie contains and afterwards takes on the next pixel for calculation and so on All the mean pixel values are stored in a parallel created matrix which eventually turns into a mean value picture for the entire movie Both GUI s use this calculation script for the mean value evaluation 4 14 PDF The definition of the Probability Density Function 1s In mathematics a probability density function pdf is a function that represents a probability distribution in terms of integrals Formally a probability distribution has density f if f is a non negative Lebesgue integrable function H E such that the probability of the interval a b 1s given by T na for any two numbers a and b This implies that the total integral of f must be 1 Conversely any non negative Lebesgue integrable function with total integral 1 is the probability density of a suitably defined probability distribution Intuitively if a probability distribution has density f x then the infinitesimal interval Ix x dx h
6. Pixel Comparison between the left and the right main gas cylinder HOH5 standard HOV5 standard 10 19 28 37 46 55 64 73 82 91 100 Pixel 46 Comparison between the left and the right main gas cylinder HOH6 standard HOV6 standard 1 10 19 28 37 46 55 64 73 82 91 100 Pixel Comparison between the left and the right main gas cylinder HOHB standard HOV8 standard 1 10 19 28 37 46 55 64 73 82 91 100 Pixel Comparison between the left and the right main gas cylinder HOH7 standard HOV 7 standard 1 10 19 28 37 46 55 64 73 82 91 100 Pixel Comparison between nave hole A and nave hole C HOnavA standard HOnavC standard 10 19 28 37 46 55 64 73 82 91 100 Pixel Intensity Standard vs Obstacles plate H90H3 standard H90H3 obstacle plate 1 8 15 22 29 36 43 50 57 64 71 78 Pixel Number 10000000 PDF Standard vs Obstacle plate 1000000 100000 10000 1000 100 10 1 H90H3 standard H90H3 obstacle plate 1 3 5 7 9 11 13 15 17 19 21 23 25 Intensitv Intensitv Standard vs Obstacles plate H90H5 standard H90H5 obstacle plate 1 8 15 22 29 36 43 50 57 64 71 78 85 Pixel Number 10000000 PDF Standard vs Obstacle plate
7. The simulation of air enters the water rig at the bottom The pre mixed fluorescent dye is stored in a tank close to the water rig A pump then provides the fuel water to the water rig where the fuel allocate through a apportion cylinder to the burner fuel hole that is in use To be able to perform experimental tests in the fluid dynamics laboratory at SIT it was necessary to make a risk identification and a flowchart for the water rig Later on a user manual for the water rig was created Outlet Camera Horizontal laser sheet Burner outlet Mixing tube Swirler Vertical laser sheet Camera Plate to redistribute let the flow Figure 4 2 1 Water rig combustion chamber replica 4 3 Plastic burner The plastic burner at SIT 19 a model of a real DLE burner shown in Figure 4 3 1 SIT designed the burner because they were interested to evaluate the flow field in the mixing tube The benefit of the opaque plastic burner 19 that it is possible to analyze the radial distribution of the air fuel mixing in different sheets in the burner The experiments were filmed horizontally at the burner outlet and 90 mm down in the mixing tube The plastic burner had only two functional main gas cylinders in comparison to the real one that has four In a real burner it is only possible to analyze the radial distribution at the burner outlet as the burner is made of steel The swirl cone has as earlier mentioned four main gas cylinders In order to
8. BZ GUN VONICA eme 21 9 3 RE Te E He EE 22 e ES ODs L b e e RE R 22 92 le EE 23 SM PT ene ee ASA 24 Jok E 26 S05 Blocked Minar Dole ani a a a aa 21 Summary C OnClUSIONS E 29 Recommendations Future work sss ss sss esse esse esse KKK KN 30 SOT 31 APPEND E 32 Appendix 1 Risk identification ss sse eee 32 Appendix 2 Instruktioner for handhavande av vattenrigg esse eee eee eee 33 leie nelle EE ER ele wie E Te Me WEE 36 ele wie 9 ey T 37 Appendix 6 Experimental test schedule sees 38 ADDEN MESU EE 45 Appendix 8 User friendly manual Mat ab 55 Figure 3 1 1 The water rig at SIT AB fluid dynamics laboratory Lee 2 Fi Sure 2 2 1 SG I 800 Cnoine a da te nG d niece ohne hice aga dar e DI MODA ed 3 FiStiie 5 35 10 91d senerauon DLE DUDOI pase 4 Picute 3 5 2 31d eeneration DLE Dummer assesi sd a D Figure 3 3 3 SGT 800 combustion chamber T D Figure 4 2 1 Water rig combustion chamber replica 7 Meur 3 15 Plastic DUME EE 8 Facute 5 ATS ON EE 9 Figure 4 6 1 Fl resc in sodim EE 10 Figure 4 7 1 Toshiba JK L75M industrial probing camera sss sese eee eee eee 11 Figure 4 9 1 Calculations of MASS2100 DI 6 sensor sse eee 11 Figure 4 12 1 Graphic user interface window sss nene 12 Figure 4 15 1 Frame one of the movie sequence identifying one pixel eee na 14 Figure 4 15 2 The intensity change per frame for one pixel sees eee 14 Figure 4 15 3 An example of how a single pixel PDF may look ke 14 Figure 4 16
9. LEE Frame Li Show L Calculated a Lenger Celeulgic ine iss I Foot mean square X Mass Center ji L MaxMin Mean Calculation Graphs 10 1 Singel pixel PDF intensity i fuel mass Fuel Flow qisec 4 1 1 3D PDF radius ne intensity fuelmass one hole All pixel PDF 11 1 Click on three points as near but not on the plexiglas reflections as possible on the black amp white symmetry simulated picture 11 2 Press any key when the black amp white symmetry simulated picture shows where the center has been located to accept this center point if else close the window and try again 57 12 To start the graph analysis click he intensity mass flow compensation for 1 hole graph button The user has to go through the same procedure as in item 11 above 13 To start the mass center evaluation press the Mass Center button The user has to go through the same procedure as in item 11 above 14 To start the PDF analyzes for one pixel click on the Single pixel PDF button Afterwards the user has to click on a point on the mean value picture 15 To start the 3D PDF statistics press the 3D PDF radius button The user has to go through the same procedure as in item 11 above 16 To start the PDF statistics for all pixels press the All pixel PDF button The user has to go through the same procedure as in item 11 above GUI horizontal Horizontal video manager Loaded File HOH 1 avi
10. Mean valuepix image HS0H8_1 avi Load video Calculation Mean Calculation Root mean square Image 2 Max Min SSS Frame K 7 C Show Larger Calculated Images L Rooct mean square j Max Min Mean Calculation 15 Graphs 1 Singel pixel PDF 4 intensity I fuel mass Fuel Flow gisec 4 SD POFiradius 15 S Casa J16 o EE a me ln 58
11. be due to the changed inflow turbulence The mass center function revealed that the basket decreases the swirl Figure 5 3 12 and increases when the basket is being blocked upstream Figure 5 3 13 which verifies the basket influence on the swirl Basket blocked downstream Figure 5 3 14 decreases the swirl more than for non blocked basket Fuel distribution by radius es 240 e GER A S H90H5 bask et ANZI blocked upstream 1 H90H5 hask et 1 r 131925 31 3743 4955 61 67 73 79 blocked downstream Pixel Figure 5 3 10 Similar profile for basket and standard Probability density function 1 0000000 1000000 100000 10000 1000 100 10 1 H90H5 standard H90H5 hask et H30H5 bask et blocked upstream H90H5 bask et 1 4 7 10 13 16 19 22 25 blocked downstream Number of hits Intensity Figure 5 3 11 The PDF shows that basket causes slightly less mixed areas 25 p CR Aur j e 120 140 160 180 200 40 ew 80 100 12 140 160 18 200 Figure 5 3 14 Mass center comparison between standard left and basket blocked upstream right 5 3 4 C stage The idea was to look over if better fuel air mixture could be obtained in order to inject fuel in different positions near the burner Fuel hoes were mounted in to the arranged basket holes Figure 5 3 15 The C stage movies were later on analyzed and compared to the proximate burner main gas holes from the standard appli
12. fuel was injected from all 4 main gas cylinders The graph does not look different in shape for one hole Thus the periodicity interference gives a level curve The function takes the whole mean value picture and rotates three times and not the areas of interest of the mean value picture The areas of interest of the mean value picture are pixels within the burner circumference So depending on how the video clip is edited the user gets different results Figure5 1 1 illustrates 4 different video clip editing implemented by the user to check if the function which compensates for 4 holes has an effect on the test results When the burner circumference is centered in the frame everything is flawless As soon as the user has displaced the burner circumference in the frame during video clip editing conclusions can be drawn that the periodicity interference warps the mean value picture Centered Off center 1 Off center 2 Off center 3 Figure 5 1 1 Four different video clip editing 17 Improvement of the function was necessary to obtain reliable results Changes have been done and now the function takes only the areas of interest of the mean value picture and then performs the periodicity interference 1 e the mean intensity matrix is rotated around the burner center instead of the video clip center Figure 5 1 2 shows a comparison with the same video clip editing between the previous function and the improved function Figure 5 1 3 shows a comparis
13. 1 An example of how the user should click on the mean value picture 15 Figure 4 16 2 Probability Density Function for the entire radial fuel distribution logarithmic Se E A A A ET 15 Figure 4 17 1 An example of how a 3D PDF could look like enen nene nene 16 kieured 8 17 Mass center E e EE 16 Figure 5 1 1 Four different video clip edng sess eenennnnznznzenzennannennannannannnnnnn nanna 17 Figure 5 1 2 Same video clip editing with border adjustment add in to avoid dislocation of the UTNE COM Et ik g i hd 18 Figure 5 1 3 Comparison of the results for the border adjustment add in nn 18 Figure 5 1 4 Significant improvement of the repeatability due to video clip edition 19 Figure 5 1 5 Center search performed by five users sese eee eee eee 19 Figure 5 1 6 25 seconds seems sufficient to obtain reliable results aaa aaa nana nen neve vie venies 20 Figure 5 1 7 The shape is similar and the intensity level almost linear to the fuel mass flow 20 Figure 5 1 8 Effect of inconsistency in camera settings on the results 21 Figure 5 2 1 Mean Intensity picture of the vertical application left and with contour add on TLS Ni EE 21 eu 347 Th L s apedobstacle plate iii de 22 VI Figure 5 3 2 Comparison in radial fuel distribution between standard application and obstacle PLAL at ee a Ore tern Eee ere A reer ore eo mere reer errr 22 Figure 5 3 3 Comparison in PDF statistics between standard application and obstacle
14. 30H3 10s H30H3 25 Z H30H3 1021 H30H3 25 21 HS0H3 1022 5 HSOH3 25 32 HS0H3 1025 E HI0H3 25 25 HA0H3 1 0154 H90H3 25 s 1 7 1319 25 31 37 43 49 55 61 67 7379 1 7 131925 31 37 43 49 55 61 67 7379 pixel pixel H90H3 100sec 205 100 sec movie analyzed in 4 parts I H90H3_0 25sec HS01H3 1 00s er L H90H3 25 50sec H90H3 1 0021 L l H90H3 50 75sec H9OH3 1 01052 3 intensi H30H3 100sec 1 7 135192531 37 4349 5561 GF 7373 1 B 15 22 29 36 43 50 57 64 71 78 pixel pixel Figure 5 1 6 25 seconds seems sufficient to obtain reliable results 5 1 4 Fuel flow Several tests with various fuel flows were performed and analyzed in the MatLab program The standard amount of fuel entering the main gas hole number 3 is 2g s The range for the fuel simulation was set from 1 2 g s 2 8 g s The effect of these deviations can be observed in Figure 5 1 7 The shape of the fuel distribution curves are almost identical however as expected the intensity levels are almost linear to the flow Fuel flow variations k H90wv3 2 0g en H90 3 1 24 A A0 mA HAD 1 6 SS tanve 22 si i Mi Va H90w3 2 8g U 1 6 11 1621 2631 36 41 4651 56 bl bb 71 76 BI 56 Pixel Figure 5 1 7 The shape is similar and the intensity level almost linear to the fuel mass flow 5 1 5 Camera settings As earlier mentioned the camera was very light sensitive equipped w
15. 40 10000000 35 1000000 30 SS l i z SE H90H5 standard 00000 TTIR 5 20 ne H90H5 obstacle cylinder 45 H90HS obstacle 5 1000 T 40 cylinder 100 5 10 0 RES 1 1 3 5 7 9 11 13 15 17 19 21 23 25 Intensity 48 1 8 15 22 29 36 43 50 57 64 71 78 85 Pixel Standard vs Obstacle cylinder PDF Standard vs Obstacle cylinder 70 10000000 60 EE H90H8 standard gt 50 H90H8 standard 100000 S 5 40 F 10000 a Pe 2 30 H90H8 obstacle 5 1000 20 cylinder 100 10 ge a 1 1 8 15 22 29 36 43 50 57 64 71 78 85 I BS Ae EE Pixel Intensity Swirl Hole 5 4 Hole 8 200 150 a 2 100 a 50 0 Standard vs Blocked filmholes Standard vs Blocked filmholes 50 40 gt H9OHO standard gt H90H3 standard o 30 c c g H9OHO blocked 20 H90H3 blocked filmholes filmholes 10 0 1 8 15 2229 36 4350 57 6471 78 85 1 8 15 22 29 36 43 50 57 64 71 78 85 Pixel Pixel Standard vs Blocked filmholes Standard vs Blocked filmholes 70 60 gt H90H5 standard gt 50 H90H8 standard 5 8 40 2 H90H5 blocked 2 30 H90H8 blocked filmholes 20 filmholes 10 0 1 8 15 22 29 36 43 50 57 64 71 78 85 Pixel 49 Standard VS Blocked filmholes 1 2 Standard VS Blocked filmholes 1 2
16. 800 Data about the DLE burner rel area Main gas holes 36 Appendix 5 Formulas ed 1 Reynolds number Re jie u 2 The mass flow formula is give by m p c A 2 3 Ap PLC SC Pressure difference PLC is the pressure loss coefficient 4A fi 4 Hydraulic diameter 2 h 1 5 Momentum ratio SE P 70 7A c velocitv p density A area h height L I length Sw swirl number 6 Simplified velocity profiles USA U B B A R D D IO xU x rdr Ax BI pdr i i R 1 A R 2 RJU x rdr RJA xrar ii i Les A xx ER ow 3 Appendix 6 Experimental test schedule C stage Analys J mf ra PDF analys f r att lokalisera hotspots Radiella f rdelningen C stage position 9 positions on the basket in front of one slit in 3x3 pattern Massflow ai Laser strength HzH ti Name of videoclip sips clips Section position fuel OI Total nr of clips 38 Standard Analys J mf relse mellan v nster och h ger huvudgash len Rotationsanalys mha masscenterfunktionen mellan snitten Radiella f rdelningen J mf relse mellan standardfilmer och filmerna med geometriska ndringar 3 kg s Laser strength Max clip of clip nr pos fuel S B l l A S S 25 z S S S 2 S S d E i E e S e E i E ee oO SS E es im EE ee Ed ie ma EM IE ee ar gt EEEBEEEEEEEREEEEEBET 39 H9OnavA standard K H90navC_ standard ia ki HOHO st
17. Basket a pattern Mass flow ai Laser strength a Horizontal Nr of Hole Hole Blocked Name of video clip clips Section nr 00S part Mass flow fuel aj O 25 H90H5 Basket H90H8 Basket HOHO Basket 25 25 X H90HO_Basket_blocked_lower X H9OH3 Basket blocked lower X H90H5 Basket blocked lower H9OH8 Basket blocked lower HOHO Basket blocked lower 25 25 X H90H3 Basket blocked upper 1 Ee X H90H5 Basket blocked upper H90H8 Basket blocked upper HOHO Basket blocked upper 25 x HOH3 Basket blocked upper x HOH5 Basket blocked upper x HOH8 Basket blocked upper 1 1 q I I 42 Total nr of clips 26 Name of video clip of clip clips Section part fuel 205 Vertical rado Basket 2 g s 2 g s Vert rado Basket block up D x Vert rado Basket block low FE S 3 of clips Blockerade filmluftshal J mf relse av radiella fordelningenm mellan blockerade filmhal 1 2 3 4 1 2 3 4 Analys samt standard Blocked Horizontal Length film air Mass Name of video clip of clips hole row flow fuel H90H0 blocked upper H90H3 blocked upper H90H5 blocked upper 25 H90H8 blocked upper HOHO blocked upper 25 25 H90H8 Blocked lower 1 1 1 HOHO Blocked lower 25 25 H90H8 blocked filmholes HOHO_blocked_filmholes HOH3 blocked filmholes HOH5 blocked filmholes HOH8 blocked filmholes 1 25 25 25 1 Total nr 56 of
18. Ta 15 Intensity KO Lea PDF Standard vs C stage 9 11 13 15 Intensity H9OHS standard H90 posG 23 20 H90HO standard H9Q posl Mai 23 25 Number 100000000 10000000 1000000 100000 10000 1000 100 10 PDF Standard vs C stage HIOHO standard H90 posH B 11 13 185 AP Et lt 35 Intensity 54 Appendix 8 User friendly manual MatLab ae a To start access the MatLab Start menu lt Start toolboxes Visualization GUI Horizontal GUI gt Set the MatLab current directory to the folder that contains the video data Type the filename of the video in the filename input field Click the load button to load the file into MatLab Loaded video data can be previewed using the frame slider GU horizontal Horizontal video manager Loaded File Mean valuepix image 4 Load video Calculation Mean Calculation Root mean square Max Min K Show Larger Calculated Images Root mean square X 7 j ij t f Mar htin EA d Yi gi F wa errar ere ii jew 8 fikura buri N o Ia nel II H C Mean Calculation EE Graphs Singel pixel PDF intensity I fuel mass Fuel Flow qfsec 4 ID POF radius L jw interzitvifuelmazzjone hole All pixel POF1 55 6 Choose which calculation is to be performed more than one calculation can be chosen at the same time 7 Click the Calculation execution button 8 When a calculation is done the results ap
19. am The laser beam was positioned in two directions 1 Horizontally when filming the radial distribution of the fuel in the mixing tube or at the burner outlet 2 Vertically positioned when filming the flow expansion at the burner outlet Figure 4 5 1 Argon Laser 4 6 Tracer Fluorescent dye can be used to visualize the fluid dynamics of the particles For example it can be used to see the fuel air distribution in a DLE burner A fluorescent dye that was suitable for the argon laser 450 nm is the Fluorescein sodium salt C20HIONa2O5 It is an orange powder and when it s mixed with water it turns green Figure 4 6 1 The mixing concentration of the dye is 0 1 grams per 10 liters 100ppm of water The MatLab post processing tool as earlier mentioned requires that the green particles have great contrast to be able to visualize the pixels The test section in the water rig had to be jet black to reduce the background noise and to highlight the colors of the pixels o HO E ti 2 ZnCl COOH Phthalic B ss anhydride Resorcinol Fluorescein Figure 4 6 1 Fluorescein sodium salt 4 7 Video capture and collection device The camera that was used was a very sensitive powerful Toshiba JK L75M industrial probing camera set on a shutter speed of 25 fps Figure 4 7 1 with a D A Video converter The video clip editing software that was used to capture and store the movies called Adobe premiere 6 with an add on called Pinnacle st
20. amera 4 8 Pump A relatively large electric pump was used for the fuel simulation and because of the usage of small doses of fluorescent tracer the flow stability decreased With higher fuel mass flow the accuracy of the mass flow measurements were increased 4 9 Calibration Due to very low mass flow for the fuel simulation it was necessary to use a very sensitive mass flow sensor A frequent calibration was essential due to flow indicators carioles to get reliable and sufficient test results The mass flow indicator has an accuracy of 40 014 g s as calculated in Figure 4 9 1 Calculations on MASS2100 DI6 sensor gs bar rms Re of Flowrate 1 00 162 1 39 2 00 0 70 3 00 0 27 4 00 i 0 36 5 00 0 30 6 00 i 0 25 d UU O22 3 00 i 0 20 9 00 0 15 0 17 0 16 0 15 The following data are used forthe pressure drop calculation Liquid Mame water Density 998 kgim3 Minimum flows rate 1 ois Wiseosity 0 001307 M sim Maximum flow rate 12 gj Figure 4 9 1 Calculations of MASS2100 DI 6 sensor 11 4 10 Scaling The purpose of scaling was to get comparable test result with other methods that were used in the burner development The physical properties for a real burner under real engine operations had to be scaled down One SGT 800 burner consumes approximately 3 kg s of air and this corresponds to 100 kg s of water at equal Reynolds number The Reynolds number was calculated to be approximately 9038 in the water ri
21. ance of the laser sheet This procedure will enable good movie quality and therefore increased reliability can be achieved Test evaluation The 3D PDF function requires a high performance computer to perform Burner performance simulations The blocked film air holes effect on the swirl need to be further investigated Research of how to interpret the vertical averages with respect to the flow rotation Sources 1 Niklas Roos and Daniel Halling Experimental evaluation of the flow in a 3 generation dry low emissions burner for Siemens Industrial Turbo Machinery Finspong 2006 2 Private consultation with Daniel Lorstad 2007 3 Private consultation with Tomas Larsson 2007 4 CED results from Daniel L rstad 2007 Internet sites http www powergeneration siemens com 2007 06 13 31 Nr 10 11 12 Appendix Appendix 1 Risk identification Riskidentifiering enl What if Projekt Nummer Kiningmummer Anl ggning Vattenrigg 1 Sir mningslabb Datum 2007 02 03 Signatur Andersi Pevman Process Experimental tests AEV burners Sida 1 montera fast tl skada i slangarikablar Jpersonskada kabeldragning skyddskl der r tt personskada hantering mttl skada handhavande av vatteriggen pakerhetsventil jmtrl skada personskada Undvik all gonkontakt med lasern utbildning p utrustning anv av operat rsmanual personskada Buller personskada Ut pers
22. andard ed HOH1 standard HOH2 standard 25 HOH3 standard 25 HOH4 standard 25 HOH5 standard 25 HOH6_ standard 25 HOH7 standard 25 HOH8 standard x HOHO standard x HOV1_standard 25 HOV2_standard 25 HOV3_standard 25 HOV4_standard 25 HOV5_standard 25 HOV6_standard 25 HOV7 standard 25 F DS p clips 5 ka Laser strength Vertical Name of video Nr of Length of Section Hoe clips clips clip 40 Obstacles J mf ra med standard J mf ra PDF analys for att lokalisera hotspots Radiella fordelningen Analys Type1 L bent metal plate with 3x3mm height and with placed closely in front of each Obstacles maingashole angled towards the airflow Type2 2mm metal tread fastened closely in front of the main gas holes 0 8 Mass flow ai Laser strength Name of video clip clips clips E Obstacles FI H90H3 obstacles je ee eS A x H90HS_obstacles S d 5 x H90H3_Obstacles_type2 fs x H90H5_Obstacles_type2 25 5 x Loes Obstacles wpe2 25 65 ol ajo x H90V3 obstacles 28 5 H 0 OI O1 BS EM pe v once vet 1 MAN B Total nr of Extra clips OI Ht 41 Basket J mf relse mellan standard vre blockerad del samt undre blockerad del Rotationen mha masscenterfunktionen mellan snitten Radiella f rdelningen Analys The blocked part on the basket is 17 holes from the top and the 17 holes from the bottom side of total 41 holes
23. application show that the air flow direction into the swirler has a major importance for the swirl 7 Recommendations Future work Water flow rig already under construction Robust camera and laser mount with distinct positions to simplify for the user for upcoming projects Drain hole enlargement so that emptying of water goes much faster The now existing drain hole for the water rig is too small and takes approximately 30 minutes to empty the water flow rig Current water supply through the water rig is 3 kg s The water rig has a capacity to handle mass flows around 8 kg s of water Increased main mass flow would obtain results closer to reality Water flow rig The valve for the fuel flow setting is very sensitive Little changes by the user can cause major oscillations of the fuel mass flow The pump for delivering trace fluid is to powerful and too big for this kind of experiments since very low mass flows are in use Change the current overpowered pump to a more appropriate one for example a highly mounted tank which provides a small amount of fuel with help of the gravitation 30 Camera with distinct aperture and focus so more repeatable studies can be obtained Light exclusion for the water rig to reduce the background noise and to highlight the contrast of the pixels Replacement of laser optics to obtain an even laser sheet with no disturbance Plexiglas scratches have to be removed to avoid disturb
24. as probability f x dx Informally a probability density function can be seen as a smoothed out version of a histogram if one empirically samples enough values of a continuous random variable producing a histogram depicting relative frequencies of output ranges then this histogram will resemble the random variable s probability density assuming that the output ranges are sufficiently narrow It is used for this project to allow the division of the area into a number of intervals This PDF statistics tool was to be implemented in the MatLab post processing tool 4 15 Single pixel PDF This function allows the user to select a point for example where the greatest fluctuation is situated Subsequently the function use the chosen pixel Figure 4 15 1 and examines the intensity changes throughout the entire movie sequence see Figure 4 15 2 The black line displays the intensity change per frame and the red line shows the mean value of the intensity throughout the whole movie The green line indicates the mean value change per frame which in technical term is called the dynamic mean value Since 250 frames are used for the mean value the dynamic mean value is invariant with time at this number of frames However the small variation in dynamic mean value indicates that 250 frames are sufficient for mean averages http en wikipedia org 13 Figure 4 15 1 Frame one of the movie sequence identifying one pixel Single pixel PDF functio
25. cation The best achievement from this study was position F and position G as shown in Figure 5 3 16 26 Number 10000000 1000000 100000 10000 1000 100 10 S HIR Figure 5 3 15 C s age fuel inlet positions PDF Standard vs C stage PDF Standard vs C stage 10000000 1000000 100000 10000 1000 100 10 H90 posG zz A 1 3 5 7 9 11 13 15 17 19 21 23 25 1 3 5 7 9 11 13 15 17 19 21 23 25 Intensity Intensity Figure 5 3 16 PDF graph for C stage 5 3 5 Blocked film air holes As earlier mentioned the 3 generation DLE burner is equipped with film air holes positioned around the mixing tube Its main purpose is to make sure there is no fuel along the wall and to minimize the risk of flashback that could occur at the wall The experimental tests were performed to see if blocking all or some of the air holes had any impact on the radial fuel distribution Tests were conducted by blocking the two downstream rows the two upstream rows and all of the film air holes on the mixing tube Figure 5 3 17 The conclusions of the analysis from this experiment were 1 The radial profiles seem to be moving outwards with blocked film air holes 2 By blocking all the holes the PDF curve is similar to the standard but the mean values are smaller 3 Blocking upstream or downstream film air holes causes less mixed areas The mass center analysis indicated that blocking the upstream film air holes increases the swirl whil
26. clips 44 Appendix 7 Results Comparison between the left and the Comparison between the left and the right main gas cylinder right main gas cylinder 150 100 E H9OHO standard E H90H1 standard 2 H9OVO standard 2 H90V1 standard E 50 0 1 8 1522 29 36 43 50 57 64 71 78 85 1 8 1522 29 36 43 50 57 64 71 78 Pixel Pixel Comparison between the left and the Comparison between the left and the right main gas cylinder right main gas cylinder 100 60 50 s 40 a 60 H90H2 standard E 30 H90H3 standard S D T 40 H90V2 standard 20 H90V3 standard 20 10 0 0 1 8 1522 29 36 43 50 57 64 71 78 85 1 8 15 22 29 36 43 50 57 64 71 78 Pixel Pixel Comparison between the left and the Comparison between the left and the right main gas cylinder right main gas cylinder 50 50 a 40 gt 40 6 30 H90H4 standard 9 30 H90HS standard 2 20 H90V4 standard 20 H90V5 standard 10 410 0 0 1 8 15 22 29 36 43 50 57 6471 78 1 8 15 22 29 36 43 50 57 64 71 78 Pixel Pixel Comparison between the left and the Comparison between the left and the right main gas cylinder right main gas cylinder 50 60 40 50 40 E 30 H90H6 standard E an H90H 7 standard D 8 D S
27. e blocking downstream and blocking downstream upstream do not Figure 5 3 18 21 1 2 downstream 3 4 upstream Figure 5 3 17 Blocked film air holes Fueldistribution by radius Probability density function HUH standard HHUHS standard Less alr through 90 section 50 10000000 ii a 1000000 gt H90 5 blocked film 100000 HS0HS5 blacked a air holes 1 0000 film air holes 20 HS0H5 blocked E Qa HS0HS5 blocked Mi 10 upstream film air 100 upstream film air More air through holes 10 holes 0 HU section HI0HS blocked 1 HHHS blacked 1 9 17 2533 41 49 57 65 73 81 downstream film air 1 4 7 10 13 16 19 22 25 raed film alf FDIES i holes l Pixel Intensity Figure 5 3 18 Comparison between blocked film air holes and standard Figure 5 3 19 shows a summary of mass center rotation for the different geometrical alterations It shows that the obstacle plate causes the largest swirl increase and basket blocked downstream causes the largest decrease in swirl e Hole 5 a Hole 8 Angle Figure 5 3 19 Summary of mass center rotation for geometrical alternations 28 6 Summary Conclusions Repeatability analysis The MatLab post processing tool with its functions is a very useful aid when test repeatability is secured Several different error factors were investigated and eliminated in this project regarding movie length camera settings clip editing burner center searc
28. e cap and hole 8 proximate to the mixing tube The main function of the swirl cone is to blend the compressed air with the injected fuel from the lance space cap and the main gas cylinders The swirl cone is very important at the burner inlet since it provides the most of the air fuel mixture as earlier mentioned The mixing tubes purpose is to mix all of the injected fuel with compressed air as evenly as possible The mixing tube also adds a slight of compressed air through the film air holes that is positioned around it The extra added air is to prevent the mixing tube wall from the possibility of flame propagation backwards flash back along the boundary layer where the velocity is small At the burner outlet the ignition of the fuel air mixture takes place and then leads to the combustion chamber The burner outlet is mounted to the wall that separates the compressed air from the combustion chamber Figure 3 3 3 The pilot holes are positioned at the burner outlet The pilot holes are able to inject oil as well as gas The purpose of the pilot holes is to add a small amount of fuel in order to retain the stability of the main flame which will result in greater stability in the combustion chamber The pilot influence the NO values in a negative way due to added unmixed fuel which contributes to higher local flame temperature As a result of the locally higher flame temperature the pilot holes influence the NO values in a negative way Dual
29. el 50 Swirl Hole 5 a Hole 8 Standard vs Basket PDF Standard vs Basket H90HO0 standard 1 7 1319 25 31 37 43 49 55 61 67 73 79 Pixel blocked downstream 100000000 H90H0 standard 10000000 1000000 Si 100000 H9OHO basket 10000 Z H90HO basket blocked 2 1000 upstream 100 H9OHO basket downstream 1 1 8 15 22 29 36 43 50 57 64 71 78 85 l 4 ae I e ocke Pixel ner downstream Standard vs Basket PDF Standard vs Basket H9OHB3 standard 1000000 gt I i C H90H3 basket Z 10000 g 3 1000 H90H3 basket H90H3 basket Se blocked upstream blocked upstream HOOH3 basket 1 3 5 7 9 111315171921 2325 blocked 1 9 1725 33 41 49 57 65 73 81 H90H3 basket Erer Pixel blocked Intensity downstream Standard vs Basket PDF Standard vs Basket 10000000 H90H5 standard H90H5 standard 1000000 alia gt H90HS5 basket 5 100000 H90H5 basket 8 g 10000 D H90H5 basket 3 1000 H90H5 basket blocked upstream 400 blocked upstream H90H5 basket 10 H90H5 basket 1 blocked downstream 10 13 16 19 22 25 Intensitv 1 4 7 91 Standard vs Basket H90H8 standard H90H8 basket Intensity H90H8 basket 1 9 17 25 33 41 49 57 65 73 81 H90H8 bas
30. f 20 H90V6 standard 20 H90V7 standard 10 10 0 0 1 8 15 22 29 36 43 50 57 64 71 78 1 9 1725 33 41 49 57 65 73 81 89 Pixel Pixel 45 80 60 40 20 0 Intensity Comparison between the left and the right main gas cylinder H90H8 standard H90V8 standard l 1 8 15 22 29 36 43 50 57 64 71 78 Pixel Intensitv Comparison between nave hole A and nave hole C 100 80 60 H90navA 40 H90navC 20 0 1 7 13 19 25 31 37 43 49 55 61 67 73 79 Pixel Comparison between the left and the right main gas cylinder HOHO standard HOVO standard 10 19 28 37 46 55 64 73 82 91 100 Pixel Comparison between the left and the right main gas cylinder HOH1 standard HOV 1 standard 1 10 19 28 37 46 55 64 73 82 91 100 Pixel Comparison between the left and the right main gas cylinder HOH2 standard HOV2 standard Pixel Comparison between the left and the right main gas cylinder HOHB standard HOV3 standard Pixel Comparison between the left and the right main gas cylinder HOH4 standard HOV 4 standard 100 19 28 37 46 55 64 73 82 91 1 10
31. fect the results 4 1 3 DLE plastic burner experiments with geometrical alterations The purpose of all geometrical modifications was to investigate if better fuel air mixture could be gained to reduce the NO emissions The different geometrical modification that has to be implemented is the following Obstacle plate Obstacle cylinder Basket Blocked film air holes C stage 4 2 Water rig evaluation strategy To study the flow dynamics in a 3 generation DLE burner and consequently evaluate why differences in NO is obtained from burners in the same configuration SIT designed a water burner test rig The advantage of a water rig is that tests can be performed with a very low cost The water rig at SIT replicates one real burner segment Figure 4 2 1 of SGT 800 with the dimensions of the outer hull 580x660x2250 mm to get the same flow geometry as a real SIT combustion chamber The rig has two visualization windows and it s made of 20 mm thick Plexiglas The water rig is constructed with water drainage at the top and the only way the water can exit is through the inner test section with the dimensions 270x240x700 mm where the burner is mounted The water rig has a capacity to handle a mass flow of 8 kg s but the existing water flow that simulates the air flow supplies maximum 3 kg s The water rig has two water inputs one that simulates the air entering the burner and one that is representing the fuel using water mixed with fluorescent dye
32. fuel capability dr e Fuel transfer section A s Turbulator section B s Mixing tube section C Pilot gas injection 1 e Main gas injection 2 s Pilot liquid injection 3 e Main liquid injection 4 Figure 3 3 3 SGT 800 combustion chamber 4 Method 4 1 Development of evaluation strategy 4 1 1 Post processing tool MatLab To investigate the mixture of instant data PDF statistics of transient fuel concentration had to be evaluated Automatic search of burner circumference and burner radius in movies was essential in order to calculate the geometry In order to estimate swirl the mass center of the injected fuel into the burner has to be calculated Highlight contours of the expansion shape in MatLab vertical application in order to ease visual analyze for the user Secure reliability of the experimental tests by using movies e How long movie is required to obtain averages of sufficient accuracy averages PDF Evaluate statistical accuracy e How much may the result differ if the same experiment is repeated averages PDF 4 1 2 Test evaluation repeatability Research in movie lengths Decide the quantity of frames vyhich is required to obtain a secure reliability Determine the magnitude of fuel flow variations that may influence on the test results Video clip editing Verify the importance of the burner circumference position in the frame Decide if camera settings aperture focus may af
33. g using a water flow of 3 kg s while in a real SGT 800 it is 452000 It is a big difference around a factor 50 in Reynolds number but sufficient enough to have turbulent flow in the water rig For the complete spread sheet see appendix 4 4 11 MatLab The MatLab program that was used was 7 3 0 267 R2006b with a picture analysis add on called image post processing tool Without this toolbox it would not be possible to analyze movies since MatLab in its basic form is not capable to do so MatLab use RGB system to restore analyses of a picture MatLab loads the movie and converts it into frames This implicate that one picture or one frame from a recorded movie stores in three matrices that 19 red green and blue matrices and they have exact the same dimensions as the picture The matrices consist of property information of the picture and by that the intensity values of each pixel is stored The intensity of each pixel is described numerical between 0 and 255 where 255 19 the maximum intensity Each intensity value has a coordinate position in the matrices 4 12 The Graphical User Interface MatLab has a user friendly GUI Graphical User Interface shown in Figure 4 12 1 which allows the user in a simple way to program scripts linked to functions that applies a standard call back syntax The good thing about the GUI is its simplicity therefore with a little effort the user can have total control Figure 4 12 1 Graphic user interface window
34. ght 5 3 3 Basket The purpose of the basket investigation was to evaluate how much the fuel distribution is influenced The smaller SGT 700 engine is equipped with burners with baskets SIT s usage of the basket for the SGT 800 is only for test purposes The main purpose of the basket is to increase the pressure drop over the burner which stabilizes the air flow through it and reduces the risk of Low Frequency Pulsations LFP The basket also changes the inflow direction through the swirler which influences the swirl and fuel distribution Also the inflow turbulence is affected which influences the fuel mixing The tests were performed in order to quantify these effects There was no designed basket for the plastic burner A design of a basket with 41 rows of holes that had the exact accurate fitting as the plastic burner was made Earlier tests have been performed using blocked parts of the basket on the SGT 700 engine and simulated tests have been performed in the water rig The test categories included basket with 17rows blocked upstream or downstream Figure 5 3 9 24 PE Blocked downstream 17 rows Sa Figure 5 3 9 The designed basket for the plastic burner The tests in the water rig was captured and evaluated in the MatLab program From the fuel distribution graph Figure 5 3 10 similar profile for basket and standard is obtained The PDF Figure 5 3 11 shows that the basket causes slightly less mixed areas which may
35. h and fuel flow The length of the movie clips should be at least 25 seconds for the radial fuel distribution graph and 100 second for the mass center evaluation Identical laser camera and rig settings with strictly even fuel flow are absolutely essential for any comparison between tests performed at different occasions The archived repeatability is now sufficient to detect flow changes due to geometrical modifications parameter study Increased repeatability may be needed to detect differences between similar burners Some equipment requires improvements to analyze smaller variations than performed in these parameter studies 29 MatLab post processing tool The existing MatLab post processing tool for concentration field measurements was further enhanced and developed Several PDF functions have been created for concentration field measurements A mass center function was added to investigate the fuel flow rotation The vertical expansion application was enhanced to simplify the evaluation for the user Geometrical alterations As the overall goal was and still is to reduce exhaust NO levels it was important to find ways to obtain as even distribution of the fuel air mixture as possible Several different geometrical alternations on the burner were tested and analyzed Among those modifications the obstacle cylinder was the most promising and also relatively simple to manufacture and mount on real burners The results from the basket
36. he fuel has less blended zones more high intensity points which may increase NO emissions Figure 4 16 1 An example of how the user should click on the mean value picture PDF 1000000 100000 10000 1000 100 10 i 13 5 7 0 1113 15 17 19 21 25 25 Intensity Humber of hits Figure 4 16 2 Probability Density Function for the entire radial fuel distribution logarithmic scale 4 17 3D PDF radius To investigate the quality of fuel air mixing in the DLE burner the need for a function that could locate unmixed areas arose The only difference between this function and all pixel PDF as earlier mentioned is that this function checks intensity changes for all pixels within a certain radius interval through the whole movie sequence and then plots a 3D curve Figure 4 17 1 for every radius interval The user has to go through the same procedure to find the radius of the burner as earlier described Any peaks that descend and rise in the 3D graph can indicate that there is poor fuel air distribution that can result in hotspots in the DLE burner causing increased NO emissions 15 Number of pixels 150 200 0 Radius Intensity Figure 4 17 1 An example of how a 3D PDF could look like 4 18 Masscenter The main function of the swirl cone is to define the flow rotation and initiate the fuel air mixing The center of mass 1s a function of the positions and masses of the particles that is included in the system In MatLab the mas
37. ith focus and aperture set manually by the user The main purpose of the experiment was to examine the change in error factor when the camera settings were modified and then turned back to their original settings and position to simulate different test occasions 20 Figure 5 1 8 shows that major error factor was discovered during these studies since the brightness of each pixel changes in the MatLab post processing tool To avoid and eliminate this error factor the user had to apply the same settings for all the experimental tests so comparable test results could be achieved It was necessary to adjust the camera settings once and use exactly equal settings for upcoming experiments H 90ra focus aperture H 9035 focus aperture 1 H903 focus aperture 2 H 903 focus aperture_3 1 7 1519 25 31 37 43 49 55 61 67 Fa 7955 Pixel Figure 5 1 8 Effect of inconsistency in camera settings on the results 5 2 GUI vertical GUI Vertical is similar to GUI Horizontal but it analyzes the expansion form of the fuel exiting at the burner outlet Therefore the laser sheet is positioned vertically when filming The expansion of the fuel is very important to analyze By filming the expansion of the fuel at the burner outlet the user can see where the recirculation zones created by vortices are located Figure 5 2 1 By knowing the recirculation zones and the angle of the expansion conclusions can be drawn if the stagnation point of the fla
38. ket Pixel blocked downstream blocked upstream 10000000 1000000 100000 e 10000 1000 100 10 Swirl POF Standard vs Basket H90H8 ba sket blocked upstream God basket l blocked 10 13 16 dovwnetream Intensity Hole 5 a Hole Vertical expansion shape Standard Basket blocked upstream Basket Basket blocked downstream 52 Summary of swirl Hole 5 Hole 8 PDF Standard vs C stage PDF Standard vs C stage 100000000 10000000 10000000 1000000 1000000 100000 100000 H90H0 standard H90H8 standard 10000 HID POSA toang H90 posB 1000 1000 100 100 10 10 Number Number KW A gt es A L all 19 15 17 19 21 23 25 Intensity Intensity PDF Standard vs C stage PDF Standard vs C stage 10000000 10000000 1000000 1000000 100000 100000 H90H8 standard H90H8 standard 10000 H90 posC kent H90 posD 1000 1000 100 100 10 10 1 FACT l T lt E A Ue EC 11 13 19 17 19 21 23 25 Intensity Intensity Number PDF Standard vs C stage PDF Standard vs C stage 10000000 10000000 1000000 1000000 TAREE E H90V3 standard 2 6g s posF 1000 1000 100 100 10 1 10 E 9 SE ee ato 1910 19 39 20 35 9 11 13 15 17 19 21 23 25 Intensity Intensity 53 Number Number 10000000 1000000 100000 10000 1000 100 10 100000000 10000000 1000000 100000 10000 1000 100 10 PDF Standard vs C stage tio
39. m f r br nsle simulering med k rl BB005 Kontrollera att e ventilerna AA001 4A002 4 A008 r 1 ppet l ge e ventilerna AA003 A004 AA005 AA006 AA007 r 1 st ngt l ge e pumpen AP005 r aktiverad e Alternera fl det efter behov med ventilerna AA010 och eller AA001 e Fl det indikeras pa massfl des m tare CF005 3 2 2 Vattensystem f r br nsle simulering med k rl BB010 Kontrollera att e ventilerna IAA004 A A005 A A008 r 1 ppet l ge e ventilerna AA001 AA002 AA003 AA006 AA007 r 1 st ngt l ge e pumpen AP005 r aktiverad e Alternera fl det efter behov med ventilerna AA010 och eller AAOOSI e Fl det indikeras pa massfl des m tare CF005 3 3 Vattenriggen e Ventilernas AA100 AA118 l ge best ms av provansvarig beroende pa ndam l 4 terst llning av vattenrigg efter prov 4 1 Vattensystem for br nsle simulering Se punkt 1 1 4 2 Vattenriggen Se punkt 1 2 4 3 Vattensystem for luft simulering e Kontrollera att ventilerna AA050 AA055 AA059 AA060 r 1 st ngt lage e Tomvattenriggen genom ppning av ventilen AAQ58 e F r trvekutjimning skall ventilen AA061 vara i ppet lage under t mning av vattenriggen 34 Appendix 3 Flow Sheet AAO6O AA061 r AAIDI 1 AA100 gt lt gt lt AADSO VAAIIO AAII4 CF010 AA115 AA118 AAD55 BBO15 AAQS9 X AAQS8 X GL lt 35 Appendix 4 Scaling Inlet air in real life one burner SGT
40. me is positioned upstream or downstream the burner outlet For an upstream positioned flame the larger risks there is to get flashback Flashback is a phenomenon that forces the flame down the burner outlet and it could rapidly destroy the burner outlet The mean calculation function evaluates the mean intensity of each pixel through the entire movie sequence The function searches for the contrast to find the edges of the expansion shape and therefore it 1s very sensitive to camera disturbance The analyze function evaluates the expansion angles of the fuel by using a MatLab function called max The user has to click at the most vivid point on the mean value picture to allow MatLab to find the largest number in an array 1 6 columns MatLab then plots all the coordinates of its largest number as the pixel that was chosen initially by the user If there is more than one pixel in the current column that has the same maximum all points will be plotted to avoid loss of data Due to camera disturbance as earlier mentioned the function did not work properly therefore a plot contour Figure 5 2 1 of the mean value picture was required so the user could easily visualize the expansion angles of the fuel The red color shows that there are high intensity more fuel values in the middle of the expansion Disturbance Figure 5 2 1 Mean Intensity picture of the vertical application left and with contour add on right 21 5 3 Geometrical alte
41. n was created to evaluate how long movie that is required to get sufficient and reliable analyses results in MatLab 250 i i MW l m 50 Figure 4 15 2 The intensity change per frame for one pixel The PDF function seen in Figure 4 15 3 shows how many times the pixel has a certain intensity through the entire movie sequence The x axis shows the intensity and the y axis shows number of times it hits certain intensity The blue curve is the PDF for one pixel and the green curve 1s the ideal PDF with same mean value but less high intensity hits 1 e less spots of high fuel content The oval circle indicates that there are high intensity values for this certain pixel Number of hits Intensity Figure 4 15 3 An example of how a single pixel PDF may look like 14 4 16 All pixel PDF To investigate the quality of fuel air mixing in the burner and locate hot spots a function was created to detect the fuel distribution over the entire flow field To evaluate the areas of interest the user has to click on three points as shown in Figure 4 16 1 nearby the burner circumference on the mean value picture to calculate the radius and the center point of the burner The function investigates the intensity changes for all pixels within the radius measured in pixel of the burner and then plots a curve as shown in Figure 4 16 2 The amplitude of the curve suggests the amount of hot spots in the burner Hot spots can occur if t
42. on of the differences in radial distribution graphs between the previous function and the improved function Centered Off center 2 Off center 1 Off center 3 Figure 5 1 2 Same video clip editing with border adjustment add in to avoid dislocation of the burner centre clip editing Horizontal 50HO0 25sec centered ane hole 100 centered comp holes E 0 centered com holes improved 1 6 15 22 29 56 43 5057 64 71 78 Fi pixel clip editing Horizontal 30H0 25sec offcenter one hole offcenter2 comp 4holesi offcenterer2 compl holes improved function 1 9 17 25 33 41 49 57 65 73 81 88 pixel clip editing Horizontal 430H0 25sec offcenterd ane hole offcenterd comp sholesi tentererl compli holes improved 1 9 17 2533 41 49 57 65 73 81 89 dara a pixel clip editing Honzontal 90HO 25sec offcenterS one hole offcenters comp sholesi officenteri comped holes improved 1 9 1725 3341 49576573681 89 function pixel Figure 5 1 3 Comparison of the results for the border adjustment add in 18 Figure 5 1 4 illustrates a comparison of all four video clips editing with the previous function and the new improved function Previous function Improved function alikaklippringar 1 olikaklippringar 1 olikaklippningar_2 olikaklippringar_ olikaklippningar_3 olikaklipp
43. onskada utbildad i mtrl skada Brand Brand ing personskada Belastningsskador komponenter personskada Genomg ende utbildning f r extern tillf llig personal Kommunik D lig kommunikation mtrl skada och eller handledning av ation EET Pet J E 32 Appendix 2 Instruktioner for handhavande av vattenrigg 1 Fore efter prov 1 1 Vattensystem for br nsle simulering Kontrollera att e ventilerna AA001 AA008 r 1 st ngt lage e ventilerna AA001 AA002 eller AA004 AAOOS r 1 ppet l ge e pumpen AP005 r str ml s 1 2 Vattenriggen Kontrollera att e ventilerna AA100 AA118 r 1 st ngt l ge 1 3 Vattensystem for luft simulering Kontrollera att e ventilerna AA050 AA055 AA058 AA059 AA060 AAO61 r 1 st ngt lage 2 Uppstart av vattenrigg 2 1 Vattensystem for luft simulering Fyll vattenriggen genom att e stilla ventilerna AA050 A A059 1 ppet lage e stilla ventilerna AA055 AA058 AA060 1 st ngt lage e stilla trvekutjimningsventilen AA061 1 ppet lage 2 2 Vattenriggen Se punkt 1 2 2 3 Vattensystem for brinsle simulering Se punkt 1 1 F rhindrar att pumpen gar torr vid oavsiktlig aktiveringjstr msattning 3 Under prov 3 1 Vattensystem for luft simulering Kontrollera att e ventilerna AA050 AA059 ar 1 ppet lage e ventilerna AA055 AA058 AA060 AA061 r 1 st ngt l ge e Flodetindikeras pa massfl des m tare CFO10 3 2 1 Vattensyste
44. pear in any of the two result windows and will be loaded into memory for further analysis 9 When the calculations are finished the operator can view the results again by using the image recall panel The intensities from the RMS result are often very weak hence the intensity enhancement box to greaten the contrasts GUlhorizontal Horizontal video manager Loaded File Mean valuepix image Load video Calculation Mean Calculation Root mean square Max Min 6 Ces 17 Frame Show Larger Calculated Images i LI Ee Max Min t L Mean Calculation Graphs intensity I fuel mass Fuel Flow qisec 4 S intensity fuelmazzjone holej U U Extra Hint The Batch button is used to calculate an entire set of recordings To use the batch command just click the check boxes of the calculations that are to be performed and start the calculations by clicking the batch button This will automatically process all videos in the current MatLab directory with selected calculations 06 10 To get a correct value in the mass flow graph use the fuel mass flow input field Object 11 11 To start the graph analysis click the intensitv mass flow compensation for 4 holes graph button object 10 GUlhorizontal Horizontal video manager Loaded File HS0H8_1 avi Mean valuepix image HOH 1 avi Load video Calculation Mean Calculation Root mean square Image 2 Max Min
45. plate 22 Figure 5 3 4 Mass center comparison between standard left and obstacle plate right 23 Facute 9 24 ODStac lec e 23 Figure 5 3 6 Comparison in radial fuel distribution between standard application and obstacle VRAS Gji pish 23 Figure 3 3 7 Comparison in PDE statistics si a i h E 24 Figure 5 3 8 Mass center comparison between standard left and obstacle cylinder right 24 Figure 5 3 9 The designed basket for the plastic burner en eee eae 20 Figure 5 3 10 Similar profile for basket and standard 25 Figure 5 3 11 The PDF shows that basket causes slightly less mixed areag nn 25 Figure 5 3 12 Mass center comparison between standard left and basket right 26 Figure 5 3 13 Mass center comparison between standard left and basket blocked downstream GUS nan hh 26 Figure 5 3 14 Mass center comparison between standard left and basket blocked upstream di E T 26 Fieur e E Stace RT E Position sortens da na Pa ae Pa dead 21 Fiere 5 5 16 N DI orapl for Stag tined je e Bad 21 Ficute 5 5 17 Bloeked film dir NOES i re eve G 28 Figure 5 3 18 Comparison between blocked film air holes and standard nana aaa aa ana nn 28 Figure 5 3 19 Summary of mass center rotation for geometrical alternations 28 VII 1 Aim The purpose of the graduate project was to implement water rig testing of gas turbine burners in order to study their flow dynamics The results will be useful fo
46. r comparison of flow dynamics and fuel concentration between individual burners Water rig test procedure development to secure reliability of the water rig tests by Making improvements of the post processing tool in the MatLab program and to find a way to secure repeatability of the experimental test results Evaluation and recommendations are also vital information for the rig improvements for future experimental testing Research and find a substitute for the corresponding laser or a fluorescent tracer that is suitable for the present laser Evaluate the burner performance and study the results that geometrical modifications alterations may generate on the air fuel mixture regarding Radial fuel distribution Expansion shape Flow rotation Probability density function statistics LA Apparatus Water flow rig 2 water systems one for air and one for fuel simulation Argon laser 450nm 750mW Toshiba JK L75M video device camera 25fps D A Video converter and a video clip editing software PC Fluorescein sodium salt 100ppm MASS2100 DI 6 mass flow measurement equipment Figure 3 1 1 The water rig at SIT AB fluid dynamics laboratory 3 Background 3 1 Siemens Industrial Turbomachinery AB In today s industrial world there are high demands on the environmental aspects This is something that every company has to contribute to and Siemens Industrial Turbomachinery AB is no exception SIT AB i
47. rations 5 3 1 Obstacle plate Based on CFD calculations an L shaped plate with the surface 3x3 mm was manufactured The main purpose of the experiment was to create local increased turbulence to investigate if better fuel air distribution could be achieved when the L shaped plate was fastened right in front of the main fuel holes Figure 5 3 1 On the fuel distribution graph illustrated in Figure 5 3 2 the mean value of the obstacle plate increases but the shape is similar to the standard application The PDF graph Figure 5 3 3 shows that less blended zones more high intensity points are achieved with the obstacle plate which may be due to of the higher mean value in comparison to the standard application The obstacles are reducing the slot area which results in a higher swirl as seen in Figure 5 3 4 Figure 5 3 1 The L shaped obstacle plate Standard vs Obstacle plate im OI q oO HUEN standard ti Ja CH HS0H5 obstacle Intensity bh CH D 1 6 15 22 29 36 43 50 57 64 71 fo 85 Pixel Figure 5 3 2 Comparison in radial fuel distribution between standard application and obstacle plate PDF Standard vs Obstacle plate 10000000 mm 1000000 H90H5 standard H90H5 obstacle plate 100000 10000 Number 1000 10 10 CH 1 3 5 7 9 11 18 15 17 19 21 23 25 Intensity Figure 5 3 3 Comparison in PDF statistics between standard application and obs
48. ringar_3 nlikaklippringar 4 mlikaklippringar 4 1 6 1176 21 26 31 3 4 4 51 56 61 BB 71 76 81 BB 15 9 307 21 2529 3557 414540 5357 6165 Ba ra 77 Bi pixel pixel Figure 5 1 4 Significant improvement of the repeatability due to video clip edition 5 1 2 Center search An additional research of the center search function was performed to look over the reliability of the manual location of the burner circumference executed by the user This was considered primarily that various users could cause various errors Figure 5 1 5 demonstrates the level of influence can be disregarded when five different users performing the center search on the same video clip Flourescein 90 H3 intensit 1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 at pixel Figure 5 1 5 Center search performed by five users 5 1 3 Quantity of frames To improve the quality and to achieve repeatability several tests were made on three different movie lengths 10 25 and 100 seconds 25 fps Figure 6 1 6 shows the movie length results from main gas hole nr 3 The graph shows clear improvement with 25 seconds or longer but regarding time and cost the amount of frames corresponding to the sequence length of 25 seconds was considered sufficient The 100 second movie was also analyzed in four equal parts and compared to the full clip to investigate if any changes occurred during the sequence 19 HS0HS 10sec 2g s HS0HS 25sec Zoe H
49. s a company that is keen about the environment and therefore spends a lot of effort in developing combustion processes in order to reduce NOx emissions on their engine products They are also researching in optional fuels which are more environment friendly 3 2 SGT 800 engine The SGT 800 engine Figure 3 2 1 was developed under the late 90 s for industrial applications such as generate electricity heat propulsion and for marine purposes It has been a very successful and popular engine among many companies worldwide due to the good attributes that it has The SGT 800 has an electric power output of 45 MW and an efficiency of 37 in simple cycle The main purpose is to use this engine in combined cycle which means that the engines exhaust gas goes into a heat recovery steam generator for maximum efficiency and minimal heat losses Figure 3 2 1 SGT 800 engine 3 3 3 generation DLE burner Due to the complex geometry of the DLE burner many parts are manufactured manually This implicate that the burners don t have identical geometry and dimensions This can cause for example that one burner can have different NO emissions than the other and this can result in different NO emissions between individual gas turbines Generally a swirl burner Figure 3 3 1 injects fuel axially into airflow with a certain tangential momentum This contributes mainly to a more efficient air fuel mixture and therefore a better combustion process Toge
50. s correspond to pixel location and the amount of mass corresponds to the intensity values of each pixel A function was created to evaluate how much the fuel mass center rotates between filming of different sheets and geometrical alterations in the plastic burner or at the outlet on a real burner The function locates the mass center Figure 4 18 1 of the fuel and an angle on the mean value picture With the information of the mass center angle and the fuel injected angle swirl numbers can be estimated The formulas that were in use can be found in Appendix 5 For example it s possible to evaluate the differences between different burners to look over how much manufacturing variations deviate Center of the burner Mass center Fuel injected angle 20 Mean value picture Figure 4 18 1 Mass center evaluation 16 5 Results 5 1 Repeatability After several experiments that weren t satisfying investigation of error factors that affected the experimental test results started To analyze reliability of the equipments that were in use for the water rig during the experiments several tests were necessary to review the results before going further 5 1 1 Video clip edition Intensity by radius The function called intensity by radius in MatLab GUI was not reliable enough To create periodicity for the graph the mean intensity matrix rotates the image three times to compensate for 4 main gas holes to see how it would look like if the
51. study the downstream fuel distribution that originates from each individual fuel nozzle The fuel was injected from one nozzle at a time This significant procedure increases the knowledge of the contribution of each nozzle to the total fuel distribution However there are uncertainties in studying the total fuel distribution using the plastic burner If the sum of the distribution from all individual holes is used the error due to background light is multiplied If the total fuel distribution is measured there are uncertainties of the mass flow through each nozzle since only the total fuel mass flow 19 measured Figure 4 3 1 Plastic burner 4 4 Risk identification and user manual The purpose of risk identification was to eliminate and decrease the risks that could occur in the water rig and around the rig during normal operations The complete sheet can be found in Appendix 1 Also a user manual was created to ensure safe operation of the water rig Appendix 2 4 5 Argon laser The fluid dynamics laboratory at SIT uses an argon laser Figure 4 5 1 of 750 mW power and a wavelength of 450 nm The laser is used to visualize the air fuel mixture in the burner The laser is a class four laser and therefore one of the most powerful and dangerous lasers on the market The laser may damage the eyes immediately The areas of interest was to film a thin sheet in order to get a two dimensional environment which could be evaluated by the MatLab progr
52. tacle plate 22 Figure 5 3 4 Mass center comparison between standard left and obstacle plate right 5 3 2 Obstacle cylinder A 2mm in diameter steel wire was mounted on the main fuel cylinder in front of the main fuel holes Figure 5 3 5 to create local turbulence and to achieve better fuel air distribution On the fuel distribution graph as it can be seen in Figure 5 3 6 that the obstacle cylinder has an identical profile to the standard application The PDF graph Figure 5 3 7 shows that more effective air fuel mixture can be achieved with less high intensity points compared to the standard application This type of obstacle also reduces the slot area and increases the swirl However the swirl increases less than for obstacle plate Figure 5 3 8 Figure 5 3 5 Obstacle cylinder Standard vs Obstacle cylinder H90H5 standard H90H5 Obstacle cylinder intensity 1 8 15 22 29 36 43 50 57 64 71 78 85 pixel Figure 5 3 6 Comparison in radial fuel distribution between standard application and obstacle cylinder 23 PDF Standard vs Obstacle cylinder 10000000 1000000 100000 10000 1000 100 10 1 1 3 5 7 9 11 13 15 17 19 21 23 25 Intensity H90H5 standard H90H5 obstacle cylinder Number Figure 5 3 7 Comparison in PDF statistics Figure 5 3 8 Mass center comparison between standard left and obstacle cylinder ri
53. ther with the usually divergent geometry of the burner mouth it also creates recirculation zones at the burner outlet which traps hot combustion products that stabilizes the flame and also acts as a permanent ignition source Mixing Swirl EA PO tube cone Main liquid fuel injection Ge fuel an Liquid fuel Pilot gas fuel njection Pilot liquid fuel injection Man gas fuel injection Figure 3 3 1 3rd generation DLE burner The 3 generation low emission burner is an important part of the low NO combustion process 1 Low NO values are achieved when good fuel air mixture is obtained 2 When the fuel burns at low temperatures A negative attribute with the low NO emission process is the impairment of the flame stability and combustion Insufficient stability in the combustion process can cause flame pulsation and vibrations that can transmit between burners due to the acoustic nodes in the flow system that communicates with the unstable flame In the middle of the space cap the lance is positioned The main function of the lance is to regulate the engine by changing its length and anchoring the main flame to prevent it from pulsations The space cap consists of four fuel injection holes A B C and D see Figure 3 3 2 The space cap provides a touch of compressed air mixed with injected fuel into the swirl cone The Swirl cone has four fuel injection cylinders Each cylinder has 9 injection holes with hole 0 proximate to the spac
54. udio 10 The movies were edited in Adobe premiere 6 and it was necessary to set the properties so MatLab could read the movies It is very important that the user clips the area of interest of the video clip and excludes unnecessary data since MatLab takes long time to process a large video clip The recommended frame resolution is 300x300 pixels The horizontal radial fuel distribution experiments that was captured and analyzed was the comparison between the left V for left in video clip name and the right CH for right in video clip name main gas cylinder The obstacle plate the obstacle cylinder and the blocked film air holes were all individually compared to the standard application using their respective gas holes The C stage positions were compared to the nearest gas hole on the standard application The vertical experiments were captured at the burner outlet to analyze the expansion shape of the flame The gas holes that were in use during the vertical experiments were gas holes number 0 from respective gas cylinder at the same time The file naming system is named after lt laser sheet position gt lt main gas cylinder gt and lt gas hole number gt 10 The translation of the video clip names is for example H9OHS5 standard avi which signifies horizontal laser sheet 90 mm down in the mixing tube hole number 5 on the right main gas cylinder and standard application Figure 4 7 1 Toshiba JK L75M industrial probing c
55. za am UNIVERSITY School of Innovation Design and Engineering gt ea IN AERONAUTICAL ENGINEERING 15 CREDITS BASIC LEVEL 300 DLE burner water rig Simulations W A 4 od E N a ki ch K E ki ot K ko p ei Pr 2 gat Authors Peyman Mohammadi and Anders Arato SI E M E N S Report code MDH IDT FLYG 0187 2008 GN300 15HP E Abstract In today s industrial world there are high demands on the environmental aspects Siemens Industrial Turbomachinery AB SIT AB is a company that is keen about the environment and therefore spends a lot of effort in developing combustion processes in order to reduce NO nitrogen oxides emissions on their engine products They are also researching in optional fuels which are more environment friendly In order to provide lower emissions the SIT designed a water rig to study the flow dynamics ina DLE Dry Low Emission burner An analyze program GUI horizontal was developed with new functions and the existing functions were improved The program s function was to evaluate different experimental tests of the flow dynamics in the 3 generation DLE burners of the SGT 800 gas turbine engine The aim was to ensure repeatability to enhance reliability of the experimental test results for further comparison for upcoming projects concerning future DLE burners When repeatability was achieved implementations of different geometrical modifications were performed in the 3 generation DLE burner

Download Pdf Manuals

Related Search

Related Contents

  Proffessional Quadraphonic DJ Mixer  Handbuch_B-Speech_DestinatorSP_Symbian(deu).  Atlus Persona Q: Shadow of the Labyrinth  PENETRANT SPRAY    Black & Decker FS1202BN Instruction Manual  Sputnik - User Guide - v1.0 - Bax  TP-LINK TL-SC3230N surveillance camera    

Copyright © All rights reserved. Failed to retrieve file