1、NASA CONTRACTORREPORTNASA CR-2209ONcsiCASE FILECOPYDESIGN AND EXPERIMENTAL PERFORMANCEOF SHORT CURVED WALL DIFFUSERS WITHAXIAL SYMMETRY UTILIZING SLOT SUCTIONby Tah-teh Yang, William G. Hudson, and Carl D. NelsonPrepared byCLEMSON UNIVERSITYClemson, S.C.Jor Lewis Research CenterNATIONAL AERONAUTICS
2、AND SPACE ADMINISTRATION WASHINGTON, D. C. MARCH 1973Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-1. Report No. 2. Government Accession No.NASA CR-22094. Title and SubtitleDESIGN AND EXPERIMENTAL PERFORMANCE OF SHORTCURVED WALL DIFFUSERS WITH AXIA
3、L SYMMETRYUTILIZING SLOT SUCTION7. Author(s)Tah-teh Yang, William G. Hudson, and Carl D. Nelson9. Performing Organization Name and AddressClemson UniversityClemson, South Carolina12. Sponsoring Agency Name and AddressNational Aeronautics and Space AdministrationWashington, D.C. 205463. Recipients Ca
4、talog No.5. Report DateMarch 19736. Performing Organization Code8. Performing Organization Report No.None10. Work Unit No.11. Contract or Grant No.NGL 41-001-03113. Type of Report and Period CoveredContractor Report14. Sponsoring Agency Code15. Supplementary NotesProject Manager, Albert J. Juhasz, A
5、irbreathing Engines Division, NASA Lewis ResearchCenter, Cleveland, Ohio16. AbstractThe feasibility of designing short curved wall axially symmetrical subsonic diffusers utilizingsuction through slots in the diffuser walls to prevent flow separation was investigated. Apotential flow analysis was mad
6、e, and a digital computer program was written for determiningthe diffuser wall contour for prescribed boundary conditions. The flow field included branchflow so that the suction slot geometry could be a part of the diffuser design. One bell shapeddiffuser and three annular diffusers with area ratios
7、 of either 2. 5:1 or 3:1 were designed,fabricated, and tested. Minimum suction requirements for metastable operation ranged from6.3% to 12% when operating with inlet air velocities in the 100 to 250 ft/sec (30 to 76 m/sec)range. For stable operation suction rates from 10% to 22% were required. In al
8、l cases thediffuser effectiveness was above 95% based on the conventional definition, and from 81% to94% when the suction loss was accounted for. The exit velocity profiles were virtually flatwith no more than 9% variation over 95% of the exit area when operated with sufficient suctionto prevent flo
9、w separation.17. Key Words (Suggested by Author(s)Diffuser bleed Wall suction19. Security dassif. (of this report)Unclassified18. Distribution StatementUnclassified - unlimited20. Security Classif. (of this page) 21. No. of Pages 22. Price*Unclassified 113 $3_,QQ.-_ For sale by the National Technica
10、l Information Service, Springfield, Virginia 22151Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TABLE OF CONTENTSSection PageSUMMARY 1!LIST OF SYMBOLS
11、 . 31 INTRODUCTION 51 .1 Background1.2 Objective 71.3 Scope2 ANALYSIS AND DESIGN 82.1 Analysis for Determination of Diffuser Geometry 82.2 Digital Computer Design Program 132.3 Griffith Diffuser Geometry 143 INSTRUMENTATION AND APPARATUS 63.1 Test Faci1ity 163.2 Instrumentation and Measurements 17k
12、TEST CONDITIONS AND PROCEDURE 20lt.1 Test Conditionsk.2 Test Procedure5 DISCUSSION OF RESULTS 235.1 Minimum Suction Requirement 235.2 Velocity Distributions 65.3 Effect of Partial Blockage Downstream 325.A Effect of Non-Uniform Inlet Velocities 35.5 Effect of Unsteady Inlet Velocity 345.6 Diffuser P
13、erformance 36inProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-CONCLUSIONS 39REFERENCES 40APPENDIX A - PROGRAM DESCRIPTION AND LISTING 41INPUT DATA PREPARATION INSTRUCTIONS k2OUTPUT EXPLANATION 6PROGRAM LISTINGAPPENDIX B - DIFFUSER DESIGN 93SAMPLE IN
14、PUT DATA 7SAMPLE OUTPUT 99APPENDIX C - DEVELOPMENT OF PERFORMANCE EQUATIONS 108IVProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SUMMARYThis report presents results of an investigation concerning theanalysis, design, and testing of short curved wall
15、axially symmetricalsubsonic diffusers utilizing suction through slots in the diffuser wallsto prevent flow separation.A potential flow analysis was made and a digital computer programwas written for determining the diffuser wall geometry for prescribedboundary conditions. The prescribed boundary con
16、ditions included thefollowing for the fluid velocity at the diffuser walls (both inner andouter walls for the annular diffusers): a region of nearly constant highvelocity from the diffuser entrance to the slot; a region of rapid decelerationat the slot, and a region of nearly constant low velocity d
17、ownstream fromthe slot to the diffuser exit. Branch flow was included in the flow fielciand this allowed the suction slot geometry to be a part of the design.Four diffuser geometries were designed, fabricated and tested:(1) a 7-inch (18 cm) exit diameter bell shaped diffuser with an arearatio of 3:1
18、 and a length to inlet radius ratio of 3.5:1-(2) a 9-inch (23 cm) exit diameter annular diffuser with an area ratioof 3:1 and a length to inlet width ratio of 5-0:1.(3) a 9-inch (23 cm) exit diameter annular diffuser with an area ratioof 2.5:1 and a length to inlet width ratio of 3-9:1-(4) an 18-inc
19、h (kB cm) exit diameter annular diffuser with an area ratioof 3:1 and a length to inlet width ratio of 3-6:1-For inlet air velocities in the 200-250 ft/sec (6l-?6 m/sec) range,metastable operation required 6.5% suction for the bell shaped diffuser whileProvided by IHSNot for ResaleNo reproduction or
20、 networking permitted without license from IHS-,-,-stable operation required 12% suction. Corresponding suction rates for the9-inch (23 cm) annular diffuser (3:1 area ratio) were 8.5% and 16.0% formetastable and stable operation respectively. For the 9-inch (23 cm) annulardiffuser with 2.5:1 area ra
21、tio, 6.3% and 12% suction rates were required. Forthe 18-inch (46 cm) annular diffuser, only 7% suction was required for meta-stable operation and 10% for stable operation.Suction rate requirements were found to increase as inlet velocity wasdecreased. Thus for stable operation of the 18-inch (46 cm
22、) diffuser therequired suction rate was about 20% at 100 ft/sec (30 m/sec) inlet velocity.In all cases the diffuser effectiveness was above 95% based on theconventional definition, and from 81% to 94% when the suction loss wasaccounted for. The exit velocity profiles were virtually flat with nomore
23、than - 9% variation over 95% of the exit area when operated withsufficient suction to prevent separation.Non-uniform inlet tests and tests with periodic shedding of wakes upstreamof the diffuser inlet were conducted on the 9“inch (23 cm) annular diffuserwith AR=3:1* Non-uniform inlet conditions caus
24、ed no significant increasein the required suction rates. However, periodic shedding of wakes upstreamof the diffuser inlet caused the suction requirement for stable operationto be increased from 16.0% to 18.0%.A downstream obstruction in the form of a perforated metal sheet atthe diffuser exit sligh
25、tly increased the minimum suction rate requirement atlow inlet velocities and had no significant effect upon the suction raterequirement at inlet velocities above 150 ft/sec (46 m/sec).Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-LIST OF SYMBOLSAA
26、RCPDFSKEmnPAreaArea ratio, exit to inletPressure recovery coefficient,P - Ps.eDiameterFraction of inlet flow removed by suctionKinetic energy per unit mass flowingMass flow rateNatural coordinate normal to streamlinesPressure (either static, dynamic, or total depending uponsubscript)Local to upstrea
27、m velocity vector magnitude ratioRadial coordinate of axially symmetrical systemA /perimeter)(velocity)Re No Reynolds number TC*Acs/PelI fluid kiiflui kinematic viscositys Natural coordinate along a streamlineU Fluid velocity outside the boundary layerU Mean velocity at a sectionW Width of diffuser
28、inlet passageX Axial coordinate of axially symmetrical systema Velocity vector angle measured from horizontal6 Boundary layer displacement thickness- P .)inletDiffuser effectiveness,Diffuser effectiveness,s.em. (P - P .). . ,i s,e s,i ideal(Ps,e s t, Js,e s,i idealProvided by IHSNot for ResaleNo rep
29、roduction or networking permitted without license from IHS-,-,-i r i u K Kinetic energy coefficient, jr I Hr dAA$ Maximum value of the velocity potential functionVelocity potential function Maximum value of the stream functionijj Stream functionp Fluid densitySubscriptscs Cross sectionale Diffuser e
30、xitd,i Dynamic inleti Diffuser inletmax Maximums,e Static exi ts, i Static i nlett,e Total exitt,i Total inlet1 Upstream2 DownstreamProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SECTION IINTRODUCTION1.1 BackgroundIn an internal flow system, a diffu
31、ser is a transitional sectionwhich connects a flow passage having a smaller cross-sectional area toa flow passage of larger area. Two desirable characteristics of adiffuser are that it will provide a nearly uniform velocity distributionat its exit plane and that it will transform most of the kinetic
32、 energyof the fluid at entrance into potential energy in the form of a higherstatic pressure at the exit plane.A previous report by the authors 1 concerned the feasibility ofdesigning short two-dimensional diffusers utilizing suction through thediffuser wall slots. The design philosophy was to presc
33、ribe the diffuserwall velocity so as to have essentially constant high velocity upstreamof a slot, constant low velocity downstream of the slot, and thus haveall the deceleration occur across the slot. By removing through the slotthose fluid particles having Insufficient kinetic energy to get across
34、the adverse pressure gradient in the region of the slot, flow separationcould be avoided. The concept of having deceleration occur abruptly andapplying slot suction to this narrow region was originally suggested byA. A. Griffith as reported by Lachmann 2 and is henceforth referred to asthe Griffith
35、diffuser.The two-dimensional Griffith diffusers designed and tested were onlypartially successful. When sufficient suction was applied to preventNumbers in brackets refer to the references at the end of this report.Provided by IHSNot for ResaleNo reproduction or networking permitted without license
36、from IHS-,-,-separation, the diffuser exit plane velocity distribution was nearly uniformand very low total pressure losses were achieved. However, the suction ratesrequired for stable operation varied from 16% for 2:1 area ratio tokk% for A:l area ratio. Problems were encountered in designing the o
37、ptimumslot geometry and also in controlling the sidewall boundary layer.In 1952, L. R. Manoni 3 designed several bell channel diffusersbased upon the Griffith concept. He used an electrical analogy tank todetermine the wall geometry required for the specified wall velocity dis-tribution. While his d
38、esign method was only approximate and somewhatcumbersome, he was able to demonstrate that such a concept would enable thedesign and operation of a relatively short diffuser without separationprovided that sufficient slot suction could be applied.In order to design planar flow channel geometries, J.
39、D. Stanitz kin 1953 performed a double transformation of the describing equations intoa coordinate system where the geometry did not need to be known for theequations to be solved. After the equations were solved on the transformedplane, the inverse transformation yielded the desired geometry.In 197
40、1, C. D. Nelson 5 extended the work of Stanitz by developing amethod for transforming the flow equations of an axially symmetricalcoordinate system. Nelson also devised an approximation to account for thebranch flow due to the slot by solving the channel design problem with amore complicated set of
41、boundary conditions.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-1.2 ObjectiveThe objective of the research effort described in this report was toexamine the concept and the practicality of a short axially symmetricalsubsonic curved wall diffuser
42、using suction through slots in the diffuserwalls to prevent flow separation.,3 ScopeIn order to achieve the objective a research program was carried out to(1) Develop a digital computer program which will determine .the wallshape of an axially symmetrical or a two-dimensional diffuser for prescribed
43、boundary conditions.(2) Utilize the program to design one bell-shaped diffuser and threeannular diffusers with area ratios of about 3:1.(3) Fabricate and test each of the above diffusers to determine(a) The minimum suction requirement for unseparated flow as afunction of inlet air velocity;(b) Perfo
44、rmance characteristics such as diffuser effectiveness,total pressure loss, and exit plane velocity distribution;(c) Effect of non-uniform inlet velocity distributions;(d) Effect of periodic shedding of wakes upstream of the diffuserinlet;(e) Effect of imbalance between inner and outer wall suction r
45、atesper unit slot length;(f) Effect of downstream blockage in the form of perforated steelplates.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-SECTION 2ANALYSIS AND DESIGN2.1 Analysis for Determination of Diffuser GeometryThe describing equations o
46、f incompressible irrotational motion andcontinuity in the axial ly symmetrical coordinate system are9R 3Xand(D(2)9X 8RThe geometry of the channel being designed might appear as shown inFigure 1.Figure 1. Possible flow channel geometry on thephysical plane.Transforming equations (l) and (2) into natu
47、ral coordinates by usingds = cos a dX + s.in a dR (3)dn = -sin a dX + cos a dR (k)and introducing the stream function and velocity potential in natural coordinatesdijj = R Q dnd(J = Q ds (5)(6)Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-results i
48、n the equations9 In Q 9a ._9n 9s and9 In9s . . . . = 0R 9s 9n(7)(8)The channel geometry on the natural coordinate plane might then appearas shown in Figure 2.=0 =$Figure 2. Possible flow channel geometry on naturalcoordinate system.Now using equations (5) and (6) to transform from natural coordinatesto a (jc coordinate system results in the following equations of irrotationalmotion and continuity. 9 In Q 9ctand1 JW2 9dR 9IThe flow field on the - plane might appear as shown in Fig
