1、N 7 3“ 122 9 4NASA CR 121024ASE FILECOPYAN IMPROVED DESIGN METHOD AND EXPERIMENTALPERFORMANCE OF TWO DIMENSIONAL CURVED WALLDIFFUSERSBYTAH-TEH YANG , W.G.HUDSONAND ALI M.EL-NASHARCLEM SON UNIVERSITYCLEMSON , SOUTH CAROLINAPREPARED FORNATIONAL AERONAUTICAL AND SPACE ADMINISTRATIONNASA LEWIS RESEARCH
2、CENTERGRANT NGR 41-001-031A. JUHAS2, PROJECT MANAGERProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-1. Report No.CR-12102*2. Government Accession No. 3. Recipients Catalog No.4. Title and SubtitleAn Improved Design Method and Experimental Performance
3、 of TwoDimensional Curved Wall Diffusers5. Report Date6. Performing Organization Code7. Author(s)Tah-teh Yang, W.G. Hudson, and Ali. M. El-Nashar8. Performing Organization Report No.10. Work Unit No.9. Performing Organization Name and AddressMechanical Engineering DepartmentClemson UniversityClemson
4、, South Carolina11. Contract or Grant No.NASA Grant NCR-* 1-001-03112. Sponsoring Agency Name and AddressNational Aeronautics and Space AdministrationWashington, D. C. 205613. Type of Report and Period CoveredContractor Report14. Sponsoring Agency Code15. Supplementary NotesProject Manager, Albert J
5、. JuhaszAir Breathing Engines Division, NASA Lewis Research CenterCleveland, Ohio16. AbstractA computer design program was developed to incorporate the suction slots insolving the potential flow equations with prescribed boundary conditions. Using the contourgenerated from this program, two Griffith
6、 diffusers were tested having area ratios AR = 3 and4. The inlet Reynolds number ranged from 6 x 1CP to 7 x 106. It was found that the slotsuction required for meta-stable operation depends on the side-wall suction applied Valuesof slot suction of 81 of the inlet flow rate was required for AR = 4 wi
7、th meta-stable con-dition, provided that enough side-wall suction was applied. For AR = 3, the values of slotsuction was about 25% lower than those required for AR = k. For nearly all unseparated testruns the effectiveness n was 100% and the exit flow was uniform. In addition to theGriffith diffuser
8、s, dump and cusp diffusers of comparable area ratios were built and testedThe results obtained from these diffusers were compared with those of the Griffith diffusersFlow separation occurred in all test runs with the dump and cusp diffusers17. Key Words (Suggested by Author(s)Contour wall diffuserBo
9、undary layer controlDiffuser performanceDi ffuser des ign18. Distribution StatementUnclassified - unlimited19. Security Classif. (of this report)Unclass i fi ed20. Security Classif. (of this page)Unclassi f ied21. No. of Pages7222. Price* For sale by the National Technical Information Service, Sprin
10、gfield, Virginia 22151Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-FOREWORDThe research described herein, which was conducted by the MechanicalEngineering Department of Clemson University, was performed under NASAGrant NGR-41-001-031. Project Mana
11、ger was Mr. Albert Juhasz of theAirbreathing Engines Division, NASA-Lewis Research Center.i iProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TABLE OF CONTENTSSection PageFOREWORD i i iTABLE OF CONTENTS ivSUMMARY 1I INTRODUCTION 31 . 1 Background 41.2
12、 Objectives1.3 ScopeI I SYMBOLS 6III ANALYSIS AND DESIGN 93.1 Two-Dimensional Griffith Diffuser 93.1.1 Design Concept 93.1.2 Analysis 113-1.3 Incorporation of Suction Slot in the Design. . 153.1.4 Computer Design Program 173-1.5 Computer Analysis Program.- 73.2 Cusp Diffuser 183.2.1 Background 83.2.
13、2 Contour Generation 19IV APPARATUS AND INSTRUMENTATION 224.1 Test Facility 224.2 Griffith Diffuser 24.3 Dump Diffuser 3i vProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Section Page4.4 Cusp Diffuser 234.5 Instrumentation and Measurements 24V TEST C
14、ONDITIONS, PROCEDURE AND DATA REDUCTION 325.1 Test Conditions 325.2 Test Procedure and Data Reduction 33VI DISCUSSION OF RESULTS 366.1 Griffith Diffuser 66.2 Dump Diffuser 446.3 Cusp Diffuser 52VII CONCLUSIONS 47.1 Griffith Diffuser 547.2 Dump Diffuser 47.3 Cusp Diffuser 4VIM REFERENCES 55APPENDIX A
15、 - Outline of Surface - Source - DistributionMethod in Solving Channel Flow Problems UsingCascade Theory 56APPENDIX B - Derivation of the Expressrion for theDiffuser Effectiveness 64DISTRIBUTION LIST 65Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-
16、SUMMARYThis report presents the results of an investigation concerningthe design and testing of two-dimensional subsonic short diffusersemploying boundary layer control using suction slots. The main goalwas to design and test a two-dimensional Griffith diffuser using animproved computer design progr
17、am. This program incorporates the suctionslots in the potential flow solution. The diffuser contour was arrivedat by prescribing the potential flow velocity along the diffuser wallto have a region of constant high velocity at the inlet, a region ofconstant low velocity at the exit, and a region of c
18、oncentrateddeceleration connecting the inlet and exit regions. The suction slotwas located at the deceleration zone to control flow separation withinthis region. Side-wall suction was applied through porous side wallsin order to maintain two-dimensional flow. Two diffusers, with arearatios, AR, of 3
19、 and k, were tested. The inlet and diffuser lengths were1.0 in (2.5* cm) and 6.2 in (15-7 cm) respectively. Slot suction ratesaround 8% of the inlet flow rate were required for the diffuser withAR = 4 under metastable operating conditions, provided enough side-wallsuction was applied. Slot suction r
20、equirements for the diffuser withAR = 3 were about 25% lower when a constant side-wall suction of 15% ofthe inlet flow was applied. For stable operation the required slot suctionranged from 28 to 30 percent for the diffuser with AR = 3 and 30 to 34percent for the diffuser with AR = A. The values of
21、the diffuser effective-ness n for the unseparated flow were very close to 100%. The flow emergingfrom the exit plane was uniform throughout most of the exit area. Goodcorrelation was obtained between predicted and experimental values of wallpressure distribution and center line velocity distribution
22、.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-The secondary goal was to compare the performance of a dump (a dumpdiffuser is a sudden enlargement) and a cusp diffuser having variable arearatios between approximately 3 and k and having the same inl
23、et and diffuserlengths as the Griffith diffuser. These diffusers were each fabricatedwith two different suction slot locations. For the dump diffuser, slotswere provided at the inlet and exit corners; for the cusp diffuser theslots were located downstream of the cusp.For the dump diffuser (AR = 3 an
24、d k), the location of the suctionslot proved to be critical regarding the diffuser effectiveness. Withthe suction slot located at the inlet corner, much higher values ofeffectiveness n were obtained than with the slot located at the exitcorner. However, flow separation occurred in all the test runs.
25、For the cusp diffuser (AR = 2.7 and 3-7), it was not possible toachieve unseparated flow using the available suction capability.Accordingly, no standing vortices were observed. The flow behaved asa jet emerging from the diffuser inlet section and no pressure recoverywas obtained.Provided by IHSNot f
26、or ResaleNo reproduction or networking permitted without license from IHS-,-,-SECTION IINTRODUCTION1.1 BackgroundResearch on curved wall short two-dimensional diffusers employingboundary layer suction has been going on since 1969 at Clemson University1 . The diffuser contour was arrived at by prescr
27、ibing the potentialflow velocity along the diffuser wall to have a region of constant highvelocity at the inlet, a region of constant low velocity at the exit,and a region of concentrated deceleration connecting the inlet and exitregions. This contour is similar to that used on experimental high-lif
28、tlaminar airfoils, originally suggested by A. A. Griffith 2 and ishenceforth referred to as the Griffith diffuser. A suction slot waslocated at the deceleration zone in order to control flow separation ofthe retarded fluid. Suction through the diffuser side-walls wasnecessary in order to achieve two
29、-dimensional flow. Test results indicatedthat a high diffuser effectiveness (98%) and a uniform exit velocity distri-bution were achieved. Suction rates several times higher than those pre-dicted theoretically were necessary to achieve these results. It waspointed out in l that the incorporation of
30、the suction slot in thepotential flow solution was needed. This could result in a diffuserrequiring a lower slot suction rate than the one previously designed.A different approach towards designing an unseparated diffuser wastaken by Ringleb 3. He suggested a method of preventing separation bymeans
31、of two standing vortices located near the boundaries of the flowBracketed numbers refer to references in Section 8Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-passage. This was done by designing the contour of the flow passage insuch a way as to a
32、llow the formation of two standing vortices with otherflow control methods like boundary layer suction. A cusp provides theproper device for the standing vortices. Perkins and Hazen 4 havereported a successful attempt in using a short cusp diffuser.1.2 ObjectivesThe primary objectives of this resear
33、ch effort described in thisreport were: (l) to re-examine experimentally the slot suction require-ment in two-dimensional Griffith diffuser with diffuser contour determinedfrom the improved design program. The improved design program treats thesuction slot geometry as an integral part of the diffuse
34、r design. A slotsuction requirement less than 10% was chosen as a limit of practicalinterest for an area ratio of to 1. (2) to examine experimentally theinteraction between the side-wall suction and the slot suction requirements.The secondary objectives were: (1) to compare the performance,effective
35、ness and exit velocity distribution, of a dump diffuser (suddenenlargement) having the same area ratio and length to inlet width ratio asthe Griffith diffuser at comparable suction rates. (2) to establish unsep-arated flow with standing vortices in a cusp diffuser having an arearatio comparable to t
36、he Griffith diffuser and compare its performancewith that of the Griffith diffuser.1.3 Scope(1) Improve the existing design computer program which was used in thedesign of flow channels with prescribed boundary conditions (inlet, exit,and wall velocity distribution) to incorporate the suction slot i
37、n theProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-potential flow solution. Use the modified computer program to generatethe contour of a Griffith diffuser.(2) Use the existing analysis computer program 6 to analyze theflow field in the diffuser ge
38、ometry generated from the improved designprogram.(3) Use conformal mapping techniques described in Section 3-2.2 toarrive at the contour of a cusp diffuser having an area ratio of k to 1.(k) Fabricate a Griffith, dump and cusp diffusers. Allowance wasmade in the design of the dump diffuser to have t
39、he suction slot locationinterchangeable between the inlet corner and the exit corner of the diffuser.Two suction slots were provided in the cusp diffuser, both downstream ofthe cusp. Only one suction slot was used at one time, the other wasblocked.(5) Carry out experimental tests to obtain the sucti
40、on requirementfor stable and meta-stable operations, effectiveness, wall velocitydistribution, centerline velocity distribution and exit velocity dis-tribution for each diffuser. Compare the measured and the analyticallypredicted velocity distributions. The range of the inlet velocity usedwas from 3
41、0 ft/sec (9 m/sec) to 260 ft/sec (79 m/sec).Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SECTION I ISYMBOLSA. cross sectional area at diffuser inletAR area ratio, exit area to inlet areaA coefficient in equation A-6pqA constant in Kings lawB coeff
42、icient in equation A7pqC, C complex constants in equation 28F complex potential, $ + ifyi . used in complex numbers, defined as v-+i unit vector along x-axis-j unit vector along y-axis-k unit vector along z-axis-n outward normal unit vectorN number of segments, see equation A-6P static pressure at a
43、ny location along the diffuser wallP stagnation pressure at end of air ductPJ static pressure at inlet sectionP static pressure at cusp line, see Figure 38Pfc static pressure at vena contracta, see Figure 38P atmospheric pressureo L mP static pressure at exit sectionp any point either on the body su
44、rface or in the fluidq any point or segment on the body surfacer distance between two points p and q, see Figure A-(4A./peri meter) U.Re Reynolds number, -7-:. , , . . : :7 fluid kinematic viscosityProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SPs5
45、uuu.maxVVVnWmaxpqX.Y.ZZPZ1cascade spacingrefers to body surface; also to area on body surface; alsolength along segment in two-dimensional or length alongstream!inelength along streamlinevelocityfree stream velocityfree stream velocity vectorcomponent of fluid velocity in x-directioncomponent of flu
46、id velocity in y-directioncomponent of fluid velocity normal to body surfacecomponent of fluid velocity tangential to body surfaceaverage velocity in the diffusers inlet planeaverage velocity in the diffusers exit planeaverage velocity at cusp lineaverage throat velocity at vena contractawall veloci
47、ty upstream of suction slotwall velocity downstream of suction slotmaximum velocityvelocity inside the boundary layer upstream of suction slotD.C. voltage of hot wire or hot film probeD.C. voltage of hot wire or hot film probe at U = 0maximum D.C. voltcomplex velocity at point p due to the segment q
48、cartezian coorindatescomplex coordinates of any point in the argand plane, z = x + iycomplex constants in equations 30a and 30breal number in equation 27Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Y,y complex constants in equation 289,A anglesc; complex variable defined as + in5,n real and imaginary parts of ?potential functiontfi stream functionp fluid dens i tyn diffuser effectiveness, defined by eq. (31)Provided by IHSNot for ResaleNo reproduction or netw
