REG NACA-TN-4080-1958 Some effects of vanes and of turbulence in two-dimensional wide-angle subsonic diffusers.pdf

1、Washington June 1958 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 4080 Page SUMMARY 1 INTRODUCTION . 2 SYMBOLS 4 DIFFUSER GEOMETRY AND ENTRANCE VELOCITY AND PRESSURE PROFILES . 8 DIFFUSER GEOI4EZRY 8 FZUXPLOTS g ENTRANCE VELOCITY AND F%.ES

2、SURE PROFaES 9 . Wall Curvature 9 Potential-Flow Velocity Profiles 11 PARAMETERS INVOLVED IN PROBLEM 11 MATHENATICAL SOLUTIONS 11 . DISCUSSIONOFPARAMETERS 12 WATER-TABLE TESTS 13 GEIJERAL DESCRIPTION 14 . Water Table 14 PlacementofBluingJets 14 Water-Table Diffusers 15 Aspect Ratio 15 . Boundary-Lay

3、er Trip 16 . Secondary Flow 16 Turbulence of Entering Flow . 17 Dynamic-Sinilarity Considerations 18 Method of Procedure in Photographic Studies . 19 WATER-TABLE DIFFUSERS WITHOW VANES 19 . Large Water-Table Diffuser 19 General description 19 Separations resulting from increasing angle in vaneless d

4、iffusers 21 Effect of changes in Reynolds number on separation 25 Variation of angles of separation with change in L/W . . 26 . Effect of turbulence 26 . Small Water-Table Diffuser 28 Reasons for building small diffuser 28 . Small-diffuser tests at small angles 28 Small diffuser at very large angles

5、 29 . Superimposed Plots of Large- and Small-Diffuser Results 30 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 4080 INSERTION OF VANES IN LARGE WATER-TABLE DIFFUSER .* 30 Tests With Vane Having an Airfoil Profile 31 . Tests With Flat-Plate

6、Vanes 31 . Vanes used 31 Short vanes at an L/W ratio of 8 32 Rationalization of findings on optimwn vane placement . 33 . Long vanes at an L/W ratio of 8 36 Vanes at different L/W ratios 38 General Conclusions From Tests With Vanes Inserted INSERTION OF RODS IN LARGE WATER-TABLE DUFFUSER . First Exp

7、eriences With Rod Insertions Sizes and Placement of Rods Used in Present Work . Rod diameters . Placement . Results and Discussion of Tests With Rod Insertion Upstream rods , Downstream rods Comparison of rods with vanes . AIR-DIFFUSER TESTS. APPARATUS OBSERVATIONAL TECHNIQUES AND COMPIRlATION OF PP

8、;RAMETERS Location of Separations Visualization of flow using smoke Visualization of flow using tufts Determination of Recovery and Efficiency . Location of static-pressure orifices Measuring and recording static pressure . Static pressure during different flow regimes . . Computation of recovery an

9、d efficiency AIR DIFFUSER WITHOUT VANES General Description Experimental procedure . Variation of turbulence Boundary-layer trip . Separation Behavior Types of separation encountered . Separations at 7 and lo0 Separation at 12O Separati0nat15 Separation at 20 Separation at 22.5 . Separation at 30 Se

10、paration at 45 . . Comparison with water-table results Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 4080 iii Page . Results and Discussion of Recovery and Efficiency 56 Tabulated recovery and efficiency data . 56 Crossplotting a.e, 56 . Co

11、mparison with Reids data 56 Position of maximums with respect to first sepaxation . angle 56 Interpretation of recovery and efficiency graphs 57 AIR DIFFUSER WITH VANES AND RODS 58 Vanes . 58 . Vanes used 58 Diffuser angles investigated 59 Improvement of separation conditions 59 Improvement of perfo

12、rmance parameters 59 ROS . 60 . Two cases of rod insertion 60 Air-diffuser rod indications 61 APPENDIX A-DEFINITION AND DERIVATION OF ENERGY PARAMETERS AND THEIR COMPARISON WITH PRESSURE PARAMETERS . 65 . ENERGY RECOVEBY FACTOR 65 ENERGY EICIENCY 69 COARISON OF EGY TH PRESSURE PARAMETERS 70 APPENDIX

13、 B-EFFECT OF EXIT-VELOCITY- PROFILE SQUARENESS ON PRESSURE RECOVERY AND PRESSURE EFFECTIVENESS 72 APPENDIX C-ACCURACY . 77 BRIEF DISCUSSION OF PROCEDURE 77 ACCURACY OF PRESENT DATA 79 TABLES . 84 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NATION

14、AT, ADVISORY COMMITTEE FOR AERONAUTICS TECHNICAL NOTE 4080 SOME EFFECTS OF VANES AND OF TURBULENCE IN TWO-DIMENSIONAL WIDE -ANGLE SUBSONIC DlXFUSERS By Carl A. Moore, Jr., and Stephen J. Kline SUMMARY Tests on two aspects of the behavior of wide-angle, plane-walled, two-dimensional diffusers with es

15、sentially incompressible flow have been conducted in the Mechanical Engineering Laboratory at Stanford University. First, a thorough study of the flow mechanism has been made using dye injection in a water table. The four regimes of flow found are delineated on graphs in terms of the three important

16、 parameters. Test data from a large water-table unit and a small water-table unit are given. Second, data are presented which demonstrate means for producing efficient dif- fusers for total included angles up to at least 45O by use of simple, short, flat vanes. In the absence of vanes, or other mean

17、s of boundary-layer control, all of the following parameters are important in determining the behavior of the flow: (a) Divergence angle, (b) ratio of throat width to wall length, and (c) free-stream turbulence; divergence angle alone is defi- nitely insufficient. Variations in Reynolds number and a

18、spect ratio seem to have little effect on the flow regime for the range of aspect ratios normally encountered and for all Reynolds numbers in excess of a few thousand. Inlet-boundary-layer shape and thickness probably also have an effect on performance but have not been investigated in the present t

19、ests. Starting from very low divergence angles and maintaining other con- ditions constant, the following four entirely different regimes of flow are found as the divergence angle is increased from zero: (a) Unstalled flow, (b) transient, three-dimensional stalls, (c) steady, two-dimensional stalls,

20、 and (d) jet flow separated from both walls. With the turbulence level held constant, increasing the ratio of wall length to throat width from 4 to 20 decreases the angles at which both three-dimensional tran- sient and two-dimensional steady stall occur by a factor of the order of 2 or 3 to 1. Incr

21、easing turbulence level, with the ratio of wall length to throat width held constant, increases the angle at which transition occurs from three-dimensional transient separation to two-dimensional Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-2 NACA

22、 TN 4080 1 steady separation by roughly 1-to 1 but has little effect on the angle 2 at which three-dimensional separation begins. Any type of turbulence- promoting device inserted in the flow has about the same effect as an increase in the free-stream turbulence. Overall pressure recovery and effici

23、ency are high in the unstalled regime and drop only a small amount through the three-dimensional-stall zone, but they drop to very low values as soon as two-dimensional steady separation begins. The use of multiple, short, straight vanes placed just downstream from the throat can eliminate all sepas

24、ation up to angles as high as 45O as well as provide a means for controlling the exit velocity profile and for smoothing the flow. Water-table results on both vanes and flow regimes are confirmed by preliminary data from the air apparatus. INTRODUCTION The present report is a condensed and revised v

25、ersion of the Ph. D. dissertation of Dr. C. A. Moore (ref. 1). Reference 1 includes additional details concerning apparatus and procedures. w In internal flows of fluids in ducts, among which are included those of centrifugal compressors and turbines, jet engines, air-conditioning equipment, pumps,

26、and many other machines, it frequently becomes necessary to decrease the fluid velocity and, simultaneously, to increase the static pressure. pr iff us ion“ denotes the process of simultaneously decreasing the velocity and kinetic energy and increasing the pressure and flow work of the fluid. The pr

27、ocess of diffusion is usually an inefficient one because of the effect of the positive pressure gradient on the boundary layer of the flow. In particular, attempts to decrease the length and weight of diffusers by increasing the divergence angle results in very large losses and poor performance unle

28、ss some means of boundary-layer control is employed. In fact, the earliest systematic investigation by Gibson (refs. 2 and 3) showed that the losses can exceed those of a sudden enlargement for certain particularly bad, wide-angle geometries. Many other workers have investigated the performance of s

29、ubsonic diffusers including Jones and Binder (ref. 4), Tults (ref. 5), Patterson (ref. 6), and Reid (ref. 7). The combined papers of Tults, Patterson, and Reid provide an excellent summary of the literature, and it is con- sequently unnecessary to repeat that information here. Some of their Provided

30、 by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 40W 3 data, together with those of Vedernikoff (ref. 8), Nikuradse (ref. 9), Demontis (ref. lo), and Polzin (ref. 11) , are given in table I for con- venient reference. A great deal of attention has bee

31、n given by previous investigators to the problem of the divergence angle yielding optimum performance. The most comprehensive summasy and the most careful data on optimum per- formance are given by Reid (ref. 7). Reid noted that the angle for opti- mum efficiency was decidedly different from that fo

32、r optimum recovery. In addition, the data of different observers are in very poor agreement concerning the optimum angle. The present study had two main objectives: First, to study more systematically and in more detail the flow models in subsonic diffusers, and, second, to investigate the use of va

33、nes in the flow as a means for producing efficient wide-angle diffusers. The first objective has been largely accomplished. Using dye injection in a water table to visualize the flow, as described below, four regimes of flow were found. The bound- aries of these regimes are determined by three param

34、eters which are dis- cussed in detail. Quantitative graphs delineating these regimes of flow are given. These results have been verified by use of two different water- table units and an air apparatus. While the water table provides an excellent and simple means for studying flow mechanism and while

35、 it also provides quantitative data concerning the effects of the controlling parameters on the flow mecha- nism, it does not provide any data on overall recovery or efficiency. It is to provide these data that the air apparatus was constructed. Since the taking of data in the air apparatus is a lon

36、g and tedious process, the water-table results are also being used to guide the air-apparatus tests. Unlike the water-table tests, the air-apparatus tests are still in progress, and final results have not been obtained. However, included in this report are enough preliminary data from the air unit t

37、o: (1) Verify the effect of the governing parameters on the flow models as found in the water table; (2) demonstrate that the water-table flow- model studies do provide the proper indications concerning regions of high and low performance; and (3) indicate that the vane configurations developed in t

38、he water table show very considerable promise. Since the air-apparatus tests are only preliminary, and since the water-table studies provide very useful information by themselves, the present report deals primarily with the results of the water-table units and includes only a brief section on the ai

39、r-apparatus results. Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-4 NACA TN 40 This investigation was conducted at Stanford University under the sponsorship and with the financial assistance of the National Advisory Committee for Aeronautics. SYMB

40、OLS A cross-sectional mea of diffuser a distance along x axis from origin to center of vortex a- a line indicating points of transition to first stall (see figs . ) b-b line indicating points of transition from three-dimensional to two-dimensional separation (see figs. ) width of core, core being re

41、gion of flow having uniform velocity c-c line indicating points of transition from two dimensionally separated flow to jet flow (see figs.) C energy recovery factor C pressure recovery factor d-d line indicating points of transition from jet flow to two dimens ionally separated flow ( see figs . ) f

42、* arbitrary function of velocity u /2 depth of water above horizontal glass bottom in water table ( 12) c1 depth of water measured above glass plate on center line of diffuser at throat, in. G distance between parallel Plexiglas walls of air diffuser H height of lip of outlet gate or weir above glas

43、s plate, in. z-direction width of fluid flowing; replaced G and /2 in derivation of appendix B Id integral in denominator of equation (9) Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 4080 5 integral representing rate of flux of kinetic ene

44、rgy into diffuser integral in numerator of equation (9) integral used for mathematical manipulations integral representing flow work integral defined by equation (5) height of bluing jets above glass bottom of water table all jets 3 inches above glass plate five jets ahead of throat at 3-inch height

45、, downstream jets 114 inch above glass plate five jets ahead of throat 1/16 inch above glass plate, downstream jets at 114-inch height coefficient, a function of entering profiles and diffuser geometry diffuser length, distance from throat to diffuser outlet measured along center line of diffuser va

46、ne length length of flat-plate portion of diverging walls of diffuser number of vanes used distance from throat to diffuser exit measured along diverging wall, N = M + rsO normal to streamline static pressure total or “reservoir“ pressure volume flow rate or water flow rate, lb/sec dynamic pressure,

47、 p12/2 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 4080 area ratio, A/A rod-diameter Reynolds number Reynolds number based on diffuser length L and bulk average throat velocity V1 Reynolds number based on and coordinate r at point of stal

48、l Reynolds number based on throat width W1 and bulk average throat velocity V1 radius in general radial coordinate in polar coordinates radius of entrance-sect ion semicylinders uncertainty in a variable uncertainty in a result coordinate along a streamline temperature of flowing water, OF velocity

49、of core, constant at a cross section variable x-direction (or r-direction) velocity components of turbulence one-dimens ional mass flow velocity or bulk average velocity variable y-direction (or 8-direction) velocity velocity of a potential flow at wall of throat velocity at center line of throat diffuser