DESIGN, CONSTRUCTION AND EVALUATION OF A SUBSONIC WIND TUNNEL by MILAN VLATINAC S.B., Massachusetts Institute of Technology (1966) SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY dune, 1970 Signature of Author. Department of Keronautics and Astronautics, June 1970 Fnegis Supervisor Certified by Accepted by chairman, vohertmanefh Graduate Committee Archives MASS. Mai) >be fe DESIGN, CONSTRUCTION AND EVALUATION OF A SUBSONIC WIND TUNNEL by Milan Vlajinac Submitted to the Department of Aeronautics and Astronautics on June 4, 1970 in partial fulfillment of the requirements for the degree of Master of Science. ABSTRACT The design, construction and calibration of a subsonic wind tunnel is summarized. The design philosophy is dis- cussed and methods for wind tunnel calculations are outlined. Comparison of measured wind tunnel parameters are shown to be in excellent agreement with design calculations. Thesis Supervisor: Eugene E. Covert Title: Professor of Aeronautics and Astronauticsiii ACKNOWLEDGEMENTS The author wishes to thank Professor Morton Finston for making this thesis possible through the use of the facilities of the Aerophysics Laboratory. Special thanks are due to Professor Eugene E. Covert for the hours of patient discussion, his excellent sugges- tions, and encouragement offered during the course of this thesis. The author is grateful to Timothy Stephens, Jim Coffin and many other members of the laboratory personnel for the many hours of technical assistance. For the final preparation of this thesis, the author wishes to thank Helen Putnam, Cathy Callahan and Linda Wainionpaa. This work was carried out under the sponsorship of the Full-Scale Research Division, NASA-Langley Research Center, Hampton, Virginia, under Contract NAS1-4421. This contract was monitored by Mr. Harleth Wiley of the NASA-Langley Vehicle-bynamics Section.iv TABLE OF CONTENTS Chapter No. Page No. L INTRODUCTION 1 1.1 General Considerations and Requirements 1 1.2 Components of an Open Circuit Wind Tunnel 2 2 WIND TUNNEL DESIGN CONSIDERATIONS 4 2.1 Power Requirements 4 2.2 Calculation of Pressure Losses Due to Skin Friction 6 2.3 Additional Pressure Losses 7 2.4 Effects of Wind Tunnel Component Geometry on the Power Required 8 2.5 Turbulence 10 2.6 Turbulence Reduction 12 2.7 Test Section Boundary Layer Growth 13 3 WIND TUNNEL CONSTRUCTION 15 3.1 Construction of Wind Tunnel Components 15 3.2 Power Requirement Estimate 16 3.3 Wind Tunnel Energy Ratio Estimate 17 3.4 Fan Selection 17 3.5 Motor Selection 17TABLE OF CONTENTS (continued) Chapter No. Page No. 3 WIND TUNNEL CONSTRUCTION (continued) 3.6 Construction of Power Supply and Motor Speed Control 17 4 WIND TUNNEL MEASUREMENTS AND CALIBRATION 19 4,1 Initial Operation 19 4.2 Static Pressure Measurenents 19 4.3 votal Pressure Measurements 19 4.4 Dynamic Pressure Measurement 20 4.5 Dynamic Pressure Variation Across Test Section au 4.6 Axial Dynamic Pressure Variation 21 4.7 Test Section Speed Setting 21 4.8 Wind Tunnel Speed Resolution 22 4.9 Turbulence Measurements 23 5 EVALUATION OF TUNNEL PERFORMANCE 25 5.1 Comparison of Wind Tunnel Performance with Predicted Performance 25 5.2 Pressure Losses in the Effuser 25 5.3 Test Section Losses 26 5.4 Diffuser Performance 26 5.5 Fan Pressure Rise 26 5.6 Power Requirements 27 5.7 Energy Ratio 27 5.8 Fan Efficiency 27TABLE OF CONTENTS (continued) Page No. 6 CONCLUSIONS AND RECOMMENDATIONS 28 6.1 Summary of Wind Tunnel Design and Construction 28 6.2 Wind Tunnel Operation with a Magnetic Suspension System 29 30 A.1 Hot-Wire Construction 30 A.2 Instrumentation 30 A.3 Hot-Wire Calibration 31 A.4 Hot-Wire Response to Velocity Fluctuations 31 Figures 1 Open circuit tunnel configuration 33 2 Friction coefficient versus Reynolds number (from Ref 2) 34 3 Pressure loss coefficient for screens (from Ref 2) 35 4 Contraction section losses for various con- traction ratios 36 5 Diffuser losses for various exit diameters 37 6 Turbulence factor versus turbulence level (€xom Ref 2) 38 7 Wind tunnel and magnetic balance (photo) 39 8 Wind tunnel assembly 40 9 Required power versus mass flow rate 41TABLE OF CONTENTS (continued) Figures (continued) Page No. 10 Required fan pressure rise versus mass flow rate 42 1 Tunnel energy ratio versus mass flow rate 43 12 Fan operating characteristics and predicted tunnel requirements 44 13 Motor windings schematic 45 14 Power supply block diagram 46 15 Control circuit schematic 47 16 Motor speed versus tachometer voltage 48 17 Spanwise test section dynamic pressure variation 49 18 Dynamic pressure variation between inlet and exit versus centerline dynamic pressure 50 19 Speed setting manometer fluid density versus temperature 51 20 Test section dynamic pressure versus speed setting pressure 52 21 Test section pressure loss coefficient versus dynamic pressure 53 22 Effuser pressure loss coefficient versus test section dynamic pressure 54 23 Test section pressure loss coefficient versus dynamic pressure 55 24 Diffuser pressure recovery 56 25 Fan characteristics and measured wind tunnel requirements 57viii TABLE OF CONTENTS (continued) Figures (continued) 26 Wind tunnel pressure rise versus mass flow 27 Measured power required versus mass flow 28 Measured energy ratio versus mass flow 29 Fan efficiency versus motor speed Al Hot wire calibration versus freestream velocity Page No. 58 59 60 61po U0 Sew Rw Re ix SYMBOLS area constants speed of sound perimeter Drag coefficient DRAG FORCE diameter hydraulic diameter (44) flow kinetic energy energy ratio hot wire voltage hot wire current screen pressure loss coefficient Pressure loss coefficient “E length Mach number power dynamic pressure 1/2 pv? hot wire resistance Reynolds number &¥ hot wire cold resistance resistance fluctuationSYMBOLS (continued) turbulence level turbulence factor velocity mean velocity mean square velocity fluctuation velocity components « diffuser divergence angle EB energy loss P pressure loss ae speed setting pressure » friction factor é power loss factor fan efficiency P density % specific heat ratio Subscripts ° test section conditions i integersCHAPTER 1 INTRODUCTION 1.1 General Considerations and Requirements The evolution of the second generation magnetic balance! at the M.I.T. Aerophysics Laboratory indicated the need for a small subsonic wind tunnel which would be simple to operate and could simulate the aerodynamic loads (similar dynamic pressure) on models tested in the supersonic (M = 4.25) facility. ‘The immediate advantage of such a wind tunnel would be to provide the initial operational experience and evaluation of the magnetic balance at a considerably reduced operating cost. The goal in the construction of this wind tunnel was to incorporate into its design as many desirable features as possible in order to make it a high efficiency, low turbulence wind tunnel for gener laboratory use. The principal geometric requirement for the tunnel was that the test section be compatible with the magnetic balance "test" region. This limited the maximum test sec- tion size to 7 1/2 inches in diameter. In addition, the area in which the wind tunnel would be located restricted the overall tunnel length to less than 35 feet. An open circuit tunnel configuration was chosen both for simplicity in construction as well as the reduction in overall size by omission of the flow return duct. A practical consideration in a closed circuit wind tunnel design is the addition of a method for cooling the air in the tunnel circuit”. since this tunnel was to be installed in a reasonably large test room, this aspect of the design was unnecessary.The dynamic pressure in the test section was originally chosen to be 1.5 p.s.i. as a basis for power requirement and performance calculations. This value corresponded to the maximum dynamic pressure of the Hypersonic Flow Apparatus at NASA Langley at which a magnetic balance installation was proposed. From these calculations and cost considerations the maximum tunnel dynamic pressure was extended above this value for the particular fan and power plant selection that matched the calculated wind tunnel requirements. 1.2 Components of an Open Circuit Wind Tunnel The principal parts of an open circuit wind tunnel are shown in figure 1. A functional description of the various components follows. i.) Effuser. The effuser is located ahead of the test section and should consist in part of a large constant area xegion known as the stilling section. This is followed by a contraction section, where the flow is accelerated to its maximum value at the test section. The stilling section, because of the low flow velocity in it, is a practical loca~ tion for placing screens and honeycomb to reduce the flow irregularities which eventually go through the test section. The contraction section should create a continuously increas- ing flow velocity along its length. ) Test Sectior The test section should be the region of maximum tunnel velocity and is where models are normally tested. It should be easily accessible as well as provide viewing access to the model being tested. It is desirable to obtain uniform steady flow velocity along the test section length and a minimum variation of the axial velocity across the test section. iii.) Diffuser. Downstream of the test section the flow should be decelerated as efficiently as possible in order to minimize the loss of flow kinetic energy. Particular effort should be made to avoid flow separation in the diffuser, whichcan markedly decrease its efficiency and affect the overall tunnel performance. iv.) Power Plant. For subsonic tunnels the power plant usually consists of a fan driven by an electric motor. It provides the necessary pressure rise to overcome the loss in pressure in the tunnel circuit. The proper choice of fan and motor is required in order to obtain efficient tunnel operation.CHAPTER 2 WIND TUNNEL DI IGN CONSIDERATIONS 2.1 Power Requirements The power required to maintain steady flow through the wind tunnel is equal to the total losses occurring in the flow through the tunnel. These losses are due to kinetic energy being dissipated by vorticity and turbulence. The loss in kinetic energy, which appears as a decrease in total Pressure, must be compensated by a pressure rise, usually provided by a fan. Thus, if the power input to the fan is P (i.e. motor shaft output) and the fan has an efficiency n , the equation balancing the energy input to the stream to the energy losses in the tunnel is AP = Xeirevit Losses (2.1) As has been pointed out in reference 2, the tunnel can be divided into sections with the energy loss of each section written as a drop in pressure 4p or a pressure drop coefficient k, =AP,/q, where q, is test section dynamic pressure (1/2 p,V,-). The flow energy through the test section is oe oI E, = % Pe Ao Vo (2.2) The energy loss in each tunnel section is AE.= > (EPAM) (2.3)Substituting equation (2.1) 2 = KiCePoVe ) AEL= Se CE PAVE & PVE ) (2.4) AEL= Ki (EpeVe DAVE (2.5) The continuity equation is PiVi Ad = PoVoAo (2.6) which gives - Fe 1 3 AEi= K(B \& poVerAo) (2.7) 6 For subsonic flow with M< + 4, > i = 1 (within 18) and equation 2.1 becomes nP=tpoeAs > K& (2.8) The required power for a given test section size and flow conditions depends on the sum of the pressure drop coefficients (K,) in the various tunnel sections. A reduc- tion in these coefficients improves the tunnel efficiency. A quantity relating to the tunnel performance is the energy ratio defined in reference 3 as $foVe As | QP EK and is a measure of the tunnel efficiency. ER.= (2.9)2.2 Caloulation of Pressure Losses Due to Skin Friction The methods for predicting pressure losses in ducts due to skin friction have been well documented in the literature’. Experimental values oi skin friction coefficients are avail- able for various wall roughnesses and thereby enable accurate estimates of pressure losses to be made. The pressure loss for a circular cylinder can be determined by the following AP= Sacepve) ° formula: (2.10) The friction coefficient \ is generally tabulated as a function of the Reynolds number. This is shown in figure 2. In the case of a non-circular cross section one can define a hydraulic diameter Dy = a where A is the cross sectional area and C is the wetted perimeter. It has been shown experimentally’ that using the hydraulic diameter for non- circular cross sections gives good results with laws for circular pipes in the turbulent regime (Re, > 2000). Pope” has shown that both wall friction and expansion losses occur in divergent sections. The combined losses for a constant divergence angle a are d» « ( of) o: saa—i«i «COO TAN S)\[—=t | Me (2.11 BTAN &/e. =\\ bale ) where Dy = small diameter, Dy = large diameter, Do = test section diameter. Differentiation of equation (2.11) leads to an optimum expansion when x TAN 3 = ae (2.12)2.3 Additional Pressure Losses Aside from skin friction losses in the tunnel circuit, the following losses occur. Damping Screens. Wire screens, which are used for the reduction of turbulence, produce losses in pressure which are dependent on the screen mesh size and diameter of wire used. Pressure drop coefficients for commercially available wire screens are measured as the ratio of the pressure loss through the screen to the local dynamic pressure. Pressure loss coefficient data for various screens is shown in figure 3. Honeycomb. In aédition to screens, honeycomb is sometimes used for reduction of turbulence, particularly transverse velocity fluctuations. Roberts” has measured honeycomb losses to be approximately 0,20 times the local dynamic pressure. This value is considerably lower than the recommended pressure loss coefficients for screens. Diffuser Exit. For an open circuit tunnel an additional pressure loss results from the discharge of the flow kinetic energy at the tunnel exit. The loss is just equal to the dynamic pressure at the diffuser exit. Corners. Losses that occur at corners or where the flow turns can be computed by methods described in references 2 and 3 which give values of pressure drop coefficients for various types of turning vanes. In open circuit tunnels, corners are not advisable unless the available space is limited. Model Drag. The effect of a model placed in the test section is usually considered as an additional power require- ment rather than a pressure loss. ‘The power required to overcome the drag of a model in the tunnel is? EpV SCE (2.13)where § is the model area on which the drag coefficient, C), is based. The factor € allows for the additional power requirements due to the interference of the model on the flow in and downstream of the working section. This factor is normally on the order of 1 but can be as high as 10 in some cases. 2.4 Effects of Wind Tunnel Component Geometry on the Power Required The required power for a specified test section size and dynamic pressure can be varied by altering the geometry of the remaining wind tunnel components. Since the test section area and length for the proposed tunnel were fixed by geomet- rical constraints, the effect of various effuser and diffuser geometries on the power required were analyzed. The test section conditions were the following: Area = 32.6 sq. in. Length = 36 in. Dynamic Pressure = 1.5 pounds per square inch a.) Contraction Section Losses The losses in the contraction section are due to skin friction and can be computed from formula (2.10). Assuming a constant taper for the contraction and integrat- ing gives AP _ A _Le C- a) foe 4 D,-D UF (2.14) where L, is the length, D, the test section diameter and D, the inlet cone diameter. Note that for non-circular sec- tions one can replace the diameters with the hydraulic diameter defined in section 2.2. The results obtained for various contraction ratios are shown in figure 4 for both constant length and L, = 2D,. The contraction conelength is normally chosen to be twice the inlet diameter. This choice gives a minimum value of logs coefficient for a 30:1 contraction ratio. b.) Diffuser Losse The diffuser losses are due to both skin friction and expansion. In order to avoid separation in the diffuser, the maximum divergence angle should be approximately 5° total angle“. This value agrees with the optimum expansion angle (equation (2.12)) for a friction factor ) = 0.011. Calculations of pressure loss coefficients were made for a 5° divergence angle diffuser using equation (2.11). The addition of exit losses results in a total diffuser loss coefficient equal to ap 2) DY tout = Ca14a +0.0262\( - 3) + (@) (2.15) where D, is the diffuser exit diameter. Pressure loss coefficients for various exit diameters are shown in figure 5. It can be seen that the pressure loss coefficient decreases at a reduced rate with increasing exit diameter above an exit diameter of approximately 20 inches. c.) Test Section Losses The test section loss coefficient for a constant area section using equation (2.10) and a mean value of the friction factor reduces to AP Lyk (2.16) ZPNo Ds For the test section design chosen at a dynamic pressure of 1.5 1b/sq in, the pressure loss coefficient is AP. =0.0584 2 ZoNS (2.17)10 d.) Stilling Section Losses The losses in the stilling are primarily due to screens used for turbulence reduction. The recommended screen mesh should give a pressure loss equal to twice the local dynamic pressure (see section 2.5). Here, again, the contraction ratio affects the total wind tunnel pressure loss due to the screens. For a given choice of screen pressure loss coefficient k, the tunnel pressure loss coefficient is 4 ap = K( 2) (2.18) range Tyove B Since the losses are inversely proportional to the square of 2 Ds the contraction ratio, (;~) , it is desirable to have a large o contraction ratio for power economy. 2.5 Turbulence One of the most important quantities which affects measurements in a wind tunnel is the turbulence or unsteady motion superimposed on the mean flow. In order that simula- tion of free flight might be properly made, it is desirable to reduce the turbulence in the test section to a minimum, The effect of increasing the turbulence in the test section is to increase the effective Reynolds number. A higher Reynolds number tends to produce natural turbulence. The boundary layer transition point on a model will move forward with increasing tunnel turbulence and produce a turbulent boundary layer over a larger portion of the model with pro- nounced changes in the force and moment coefficients. The separation characteristics of some models will be markedly changed with increased tunnel turbulence, particularly at low Reynolds numbers where laminar separation is expected.a The level of turbulence in a wind tunnel can be expressed as the ratio of the mean velocity fluctuations to freestream velocity. For isotropic turbulence the mean velocity fluc- tuations in the three coordinate directions are equal: Poe Ve (2.19) The turbulence level can then be defined as Vo Vo (2.20) where V, is the mean tunnel velocity. The turbulence may be found by measuring the critical Reynolds number of a sphere at which the drag coefficient decreases sharply. This critical Reynolds number of a sphere at which the drag coefficient decreases sharply. This crit- ical Reynolds number depends strongly on the turbulence level in the tunnel. A high intensity of turbulence leads to boundary layer transition at low Reynolds numbers, caus- ing the separation point to move downstream and thus decreasing the wake, which in turn reduces the drag. The critical Reynolds number for a sphere in free flight has been measured to be 385,000. The turbulence factor is defined” as 385,000 T.F, = 385,000 _ (2.21) Re. where Re, is the critical Reynolds number at which the sphere drag coefficient equals 0.30. The effective Reynolds number, Re,, in the test section is defined as Re.= (+-F.) Re ate)12 The effective Reynolds number thus increases with increasing turbulence level. The relation between the turbulence factor and turbulence level is shown in figure 6. The tur- bulence level can be measured using hot wire techniques. These are described in a later section. 2.6 ‘Turbulence Reduction The conventional methods for reducing the turbulence level in the test section involve the use of screens and honeycomb. The effect of a screen on the turbulence is dependent on the screen pressure loss coefficient K defined as AP ZPve (2.23) where AP is the pressure drop across the screen and 1/2 pV is the dynamic pressure of the flow through the screen. Collar® has shown that the fluctuating axial component of velocity,V,, is reduced as Ve = 20K Vx Zr et (2.24) where V', is the fluctuating velocity dovnstream of the screen and V, the fluctuating velocity ahead of the screen. The fluctuating component disappears downstream of the screen for K = 2.0. Batchelor’ has shown that the mean square velocity fluctuations, whose mean is zero, decrease as ye 2 fe = (I~ Wee Ve 142K (2.25) where ve and Vy are the mean square axial turbulent fluc- tuations upstream and downstream of the screens, respectively. From this V2 vanishes for K = 2.5 which is comparable to that required for steady flow. A similar expression for the transverse components is mpopents_f Ys Ve K = == alle (2.26) NS VE 10(1+ 2s K)13 which indicates that the transverse components are less reduced than the axial component. Dryden and Shubauer® have shown experimentally that the intensity of turbulence is reduced by TO. 1 TO YVeK (2.27) They have also shown that the turbulence is further reduced if several screens are used with the same total pressure drop as for one screen. If n is the number of screens, each with a pressure drop K, the turbulence is reduced by ‘ (2.28) i C+ K)% as compared to DT.e.ii_, (2.29) T Cir nk)® for a single screen with the same total loss coefficient. It is therefore advisable to use several screens for a given total pressure drop coefficient. The use of honeycomb in reducing the wind tunnel turbu- lence appears to be primarily effective in reducing the transverse components of velocity fluctuation, more so than the axial components. ‘This tends to complement the lesser effectiveness of the screens to reduce the transverse components of velocity fluctuations (equation (2.26)). The combined use of screens and honeycomb will provide a reduction of turbulence in the tunnel. Their location should be in the region of least velocity, ahead of the test section both to reduce the tunnel pressure losses and the unit Reynolds number in that region. 2.7 Test Section Boundary Layer Growth Due to the growth of the boundary layer on the test sec- tion walls, the flow velocity will increase along the length of the test section for a constant area test section. If14 uniform velocity distribution is desirable along the test section length, the cross-section area nust be gradually increased along its length to allow for the boundary growth on the walls. The divergence of the walls should be equal to the boundary layer displacement thickness (4*). Experimental data for obtaining values of boundary layer displacement thickness versus length for various Reynolds numbers are tabulated in reference 3. It should be noted that a choice of wall divergence will be proper for only one test Reynolds number. A constant taper is usually used for the entire test section based on a mean value of Reynolds number.15 CHAPTER 3 WIND TUNNEL CONSTRUCTION 3.1 Construction of Wind Tunnel Components An open circuit wind tunnel was designed and constructed incorporating features discussed in Chapter 2. The assembled wind tunnel attached to the magnetic balance is shown in figure 7. A sketch of the assembled tunnel in figure 8 shows the overall dimensions and components. A brief description of the tunnel components, as well as the Aerophysics Laboratory drawing numbers, is given below. a.) Test Section (D-8-868. D-8-871, D-8,273, D-8-883) The test section was designed to be compatible with the magnetic balance. Transparent plexiglass was used in its construction to provide viewing of the model. An octagonal cross section was chosen to optimize the available space inside the balance while at the same time providing plane surfaces on the walls to eliminate optical distortion of the model. The test section has a constant taper on its walls to allow for boundary layer growth. The test section is supported inside the magnetic balance with flanges attached at either end of the balance. A frame was con- structed to hold these flanges when the tunnel is operated without the magnetic balance. b.) Contraction Section and Stilling Section (Dwgs E= 7 D=8=869, 0 The contraction section and stilling section were made of wood (pine), and both have octagonal cross sections. contraction ratio section was designed using contours16 similar to those recommended by Tsien? at the small end of the contraction section. The stilling section contains three screens (20-mesh, 0.016 in. wire) and a 2-inch long honeycomb section with 1/4-inch cells. The stilling « tion has a faired inlet to avoid an abrupt boundary to the incoming flow. Both sections are mounted on a stand with roller tracks to permit fore and aft movement with r sspect to the test section. The contraction section is joined to the test section by a mating flange and three snap-clamps which compress an O-ring, thereby sealing the joint. For- ward movement of the inlet section provides access to the test section. c.) Diffuser (Dwg E-8-881) The diffuser was made of wood (mahogany) and is octagonal in cross section. A constant divergence angle of 5° between opposing walls was used. The diffuser attaches to the downstream test section flange. The diffuser exit matches the fan inlet with a 9.5 in. inscribed radius. Coupling between the diffuser exit and fan inlet is accom- plished by the use of two inflated rubber tubes which are compressed between flanges on the fan and diffuser. This provides an air-tight seal between the two as well as vibration isolation. Three access hatches are located along the diffuser, and a coarse screen (2-mesh) is located at the diffuser end to protect the fan. a.) Power Plant (Dwg D-8-885) A centrifugal blower driven by a 20 h.p. direct- current motor provides the power to the tunnel. The biower is mounted on an elevated platform with the fan discharging through the platform and into a baffle. 3.2 Power Requirement Estimate The power requirements for the tunnel were calculated using the methods described in Chapter 2. The required power for various mass flow rates through the test section is shown in figure 9. The tunnel maximum dynamic pressure17 could be increased from 1.5 1b/aq. in. to 2.5 lb/sq. in. with a 10% cost increase for the driving unit. The additional power versus mass flow rates for the extended tunnel opera- tion is also shown in figure 9. The required fan pressure rise versus mass flow is shown in figure 10. 3.3 Wind Tunnel Energy Ratio Estimate The tunnel energy ratio defined in equation (2.9 ) is shown in figure 11 for various mass flow rates. 3.4 Fan Selection The fan chosen for this tunnel is a centrifugal fan, type No. 20 SISW, made by the Chicago Blower Corporation. The operating characteristics of the fan are shown in figure 12, as well as the estimated tunnel operating require ments. 3.5 Motor Selection A 20 h.p. direct-current motor was selected to drive the centrifugal blower. The motor is a type CD-85 made by the General Electric Company. The maximum speed for this motor is 3600 r.p.m. 3.6 Construction of Power Supply and Motor Speed Control In order to have motor speed control from zero to 3600 revolutions per minute, two separate power supplies were designed. The supplies provide a variable voltage to the motor armature and field windings, respectively. The motor windings for the type used in this wind tunnel have a “stabilized shunt" arrangement shown in figure 13. The motor speed can be increased by increasing the armature current or decreasing the field current. The power supply block diagram is shown in figure 14 and was designed to convert the available 3-phase 440 volt power line into two variable voltage direct-current power supplies for the motor. The power supplies are described below.18 a.) Motor Armature Power Supply The power supply for the armature windings can pro- vide a maximum of 230 volts at 115 amperes direct current to the windings. The output voltage can be varied from 0 ~ 710 volts by manually rotating a wheel attached to a 3-phase variable transformer. b.) Motor Field Power Supply The power supply for the motor field windings can provide a maximum of 230 volts at 0.835 amperes direct current to the windings. In order to increase the motor speed beyond the base speed of 1750 r.p.m., a series vari- able rheostat is used to decrease the field current, thereby increasing the motor speed to 3600 r.p.m. c.) Control Circuit For Power Supply A control network was designed for the motor power supply to provide the safety features needed to prevent overloading of any part of the circuit. The control network circuit diagram is shown in figure 15. 4.) Motor Tachometer A tachometer is coupled to the motor shaft to pro- vide the motor (and fan) speed. The motor speed can be read on the power supply console by meas of a voltmeter. A calibration of motor speed versus tachometer voltage is shown in figure 16.19 CHAPTER 4 WIND TUNNEL MEASUREMENTS AND CALIBRATION 4.1 Initial Operation Following the assembly of the various tunnel components, a series of tests were performed to obtain a tunnel calibra- tion and evaluation of its performance. Preliminary measure- ments indicated that a static pressure of approximately 2.5 1b/in.? below ambient pressure could be obtained at maximum fan speed. This value was useful in determining the neces- sary range of pressure-measuring equipment needed for the tunnel calibration. 4.2 Static Pressure Measurements The test section static pressure was measured along its length by means of nine static pressure orifices in the test section wall. The static pressure orifices were .020 in. dia~ meter holes drilled through the test section walls and equispaced 4 in. apart along the test section length. A 0.125 in. diameter spotface 0.188 in. deep on the test section outside surface was made with copper tubing (0.50 in. long) epoxied into each spotface. The static pressure at each station was read on a 50-inch manometer rack. 4.3 Total Pressure Measurements The total, or stagnation, pressure in the test section was measured with a pitot tube rake which spanned the test section. The rake consisted of seven pitot tubes spaced 1 in. apart and made from 0.35 diameter hypodermic tubing. The rake could be attached at various locations along the test section wall. The pitot rake was attached to a manometer20 pank with which the variation of total pressure across the test section could be measured. 4.4 Dynamic Pressure Measurement the test section dynamic pressure can be obtained from knowledge of the total and static pressure in the test section. Using the isentropic relations for steady flow, one obtains the following equation!® for the flow Mach number (4.1) where P, is the total pressure, P, the static pressure and @ the specific heat ratio (for air y = 1-4). Substituting the following relations eP. ve BP satay (4.2) Pp MW one obtains * Po- Ps 2 st yen Bate Sen (CE M +1) ut (4.3) where 1/2 eV’ than 27-1) equation (4.3) can pe expanded to give iis the dynamic pressure. For values of M less R-2 Mt (2-8) MA -%3-Z5)ME ies (+ Me 4 ES EEO 4 Jia ay Tove Zev 192. using y = 1-4 for air one obtains Po- Ps, 2. 4 6 wos a me MO + Mo 4.5) (+ B+ ge *igoo* ae apy As the Mach number approaches zero, the above equation reduces ‘to the incompressible Bernoulli equation pe teva (4.6) P.-Ps= 7 eV21 An interesting result from equation (4.4) is that, for Mach numbers up to 0.6, the right hand side of the equation varies linearly with the ratio PP, to within 0.5%. The equation for the dynamic pressure reduces to P,- Ps 1 via Zz = 5 < 2FP O34 ia 4 653 5 MCR-R3)32 Measurement of the root-mean-square voltage fluctuation of the wire will give the root-mean square velocity fluctuation once the wire has been calibrated as described in section A.3.33 4OLOW uotzernbTzu0o TeuuNz qqnozqo edo “1 sanbta Lina ONIAIHG yasnsid NOILO3S 1S3L y3asn4s3 NOILO3S NOILOVYLNOD NOILO3S ONITIS 7 130i SN3349S ‘SINODAZNOH:34 3 REYNOLDS NUMBER xI076 007 -009 Ol O13 SKIN FRICTION COEFFICIENT, \ Figure 2. Friction coefficient versus Reynolds number (from Ref 2)PRESSURE-DROP COEFFICIENT, K Figure 3. 6O-mesh screen, .007-in. wire 20-mesh screen, .O17 -in. wire 18-mesh screen, .OI!-in. wire 24-mesh screen, .0075- in, wire 35 4 8 12 16 20 VELOCITY (FT/SEC) Pressure loss coefficient for screens (from Ref. 2)36 100 CONSTANT LENGTH = 2 INLET DIAMETERS 40 50 60 70 80 CONTRACTION RATIO 30 20 a ° 2 © nu Ss 2 3 8 3 ° 3 a 6 a a 8 4N319145309 $SO1 34NSS3ud Figure 4. Contraction section losses for various contraction ratios37 AP qo x 8 PRESSURE LOSS COEFFICIENT, 3 io 12 14 ‘6 18 20 22 24 DIFFUSSER EXIT DIAMETER (IN.) Figure 8. Diffuser losses for various exit diameters38 3 T PERCENT TURBULENCE a @ Figure 6. 12 14 16 Rerrective 385,000 | CRITICAL, R 20 TURBULENCE FACTOR FTA AR 22 24 Turbulence factor versus turbulence level (from Ref 2)soueteq oTJsuBbeu pue ToutM; PUTM “7 sanbTE40 Inlet section and fairing Inlet section stand Honeycomb grid Screens Inlet section slide track Contraction section Test section Magnetic balance housing Figure 8. 9%. 10, 1. 12. 13. 14, 15. 16. Diffuser Diffuser access hatches Diffuser-fan coupling and vibration isolation Centrifugal fan Motor (D.C.) Motor tachometer Fan discharge baffle screens Support frame and discharge baffle @ Wind tunnel assemblyHORSEPOWER REQUIRED 41 DYNAMIC PRESSURE 1.5 psi ° ' Z 3 4 5 6 7 8 MASS FLOW (C.F.M. x 1073) Figure 9. Required power versus mass flow rateREQUIRED FAN PRESSURE RISE (INCHES WATER) 2 © o ~ o o s ow Figure 10. 42 DYNAMIC PRESSURE 1.5 psi 2 3 4 5 6 7 MASS FLOW (C.F.M.x 10-3) Required fan pressure rise versus mass flow rateFigure 11. w z oO w « a i] z 2 > a 43 MASS FLOW (C.F.Mx!0-3) 0 t ” “ OILVa ADY3N3 T3NNNL ONIM Tunnel energy ratio versus mass flow rate3 a STATIC PRESSURE RISE (INCHES WATER) o 4a o 1 2@ 3 4 5 6 7 B8 9 WO Nt MASS FLOW (CFM x 10-3) Figure 12. Fan operating characteristics and predicted tunnel requirements BRAKE HORSEPOWER 345 7002 225 WATTS 220V.0.C. SHUNT FIELD 0-220 V.0.C ARMATURE GOMUTATING FIELD Figure 13, Motor windings schenaticwez6etp yootq Atddns remoa "pT eanbta ———Jsoinoo aoawe oy insite watuiiow [__Samwosswra ” —eor1 nig 3Aum= TINS el o f soon vanwoaswve | —fusnaoaswva] w31gi.03y NoiLw7081 wawoa | LS | woman avele ey 3temring NOzz—AOve 3SuHd-¢ Sone | L @ ‘M3L3WOHOWL YOM LIN ABNNOSSNVEL 25 HOLON “ouinoo Aoll—Ove47 DyaeMEYOS YTNOTTO TorWUCD “st eaNbta WAWYOISNVUL 110A 022 /Ob> 21 UBWNOISNVYL LIOAOII/Obb |L fy — Aadns 3SvHd-e w3MOd ~ Honore aunivwuyTACHOMETER VOLTAGE Figure 16. 1000 2000 MOTOR SPEED (R.P.M.) Motor speed versus tachometer voltage 300049 uorzeTzea aanssezd oTweusp uotzoes 3803 estmueds *~T eanbta (SBHONI) 3NINY3LN39 WOUS JONVLSIC é l ° t- Z- e- =X ! J! | | eu ‘NI 91 NOILVINYA 3YNSS3Yd ODIWVNAC - (%)(%) PRESSURE VARIATION DYNAMIC 50 0 ' 2 DYNAMIC PRESSURE AT X=O (psi) Figure 18. Dynamic pressure variation between inlet and exit versus centerline dynamic pressureaangezeduez snszea Aqtsuep pynTy xozsWoUEW SuT339s pesds “61 oanbty (De) 3YNLVYSdW3AL 82 22 92 se ve <2 zz T T T T T 09620" 51 02620 HONI ISd 086z0°52 aansserd sz Buyaaes peads snszea ernssezd oTwevdp wa) Sag oz 6 1 uoT}0es 3seL so ‘oz eznbta Mao zezes US uo) seer 2 on oz unSS3ué OIWUNAG NOILDIS 1S3L (se)53 VELOCITY (FT/SEC) 200 250 300 (%) TURBULENCE LEVEL 10 12 14 16 18 20 xto® FREESTREAM UNIT REYNOLDS NUMBER (FT) Figure 21. Test section turbulence versus Reynolds number and velocity54 ainsserd oqweukp uoyq09s 4593 snsxeA WUSTOTI}s00 ssoT eInsseid rosng3g (isd) °D *FuNSsadd OIWWNAG NOILI3S 1S3L 1 z ol s o v z *¢z eanSta TT AN3WIN3dx3 O asaLvinows 7 0 3 “1N3I9143309 dowd 3YyNSS3ud °b dV £055 si aansseid opweudp snszea zuUeTOT;3000 ssoT eanssezd uoT}OeS 3S9L (sd) Sb * 3uNSSBad “DIWWNAG NOILIIS LS3L 3 vl 2 ol 8 v I I NOILD3uuOD HLMOYS YSAVT AYVONNOS YVNINYT HLIM G3YunSv3W (22 NOILO3S) a3ivinoiws ao *e@ eanbta v0 so 30° 20 @ Q %b av ‘4N3191dd309 SSO1 3uNSS3ud a aa 8 3 w 2 = w = 3 a a iA c a 56 60 80 100 120 140 160 DISTANCE ALONG DIFFUSER (INCHES) Figure 24. Diffuser pressure recovery 18057 squowerqnbex [ouung puym pernseow pue soyyzsTreqoereyo ued “sz eaNBTA (WH) MOTd SSvW 008 0009 000 0002 ° T T ° ar N3d0 1 — \—13r NI ag0ud-} 000! LOLid HLIM z (a3LvM S3HONI) 2 asi 3uNSS3ud Nud 9 3 2B 3 = — 0002 coosfe} S WIND TUNNEL PRESSURE LOSSES (INCHES WATER) Figure 26. ——— CALCULATED (0 BLOCKAGE) 58 MEASURED (0 BLOCKAGE) MEASURED WITH PITOT TUBE RAKE 2 3 4 5 6 7 MASS FLOW (C.F.M. x 1073) Wind tunnel pressure rise versus mass flow59 16 I I I I —— CALCULATED 14 -- © MEASURED | nN 3 HORSEPOWER REQUIRED o @ ° 2 3 4 5 6 7 8 MASS FLOW (CFM. x 107) Figure 27. Measured power required versus mass flow60 Moly ssew snszen oTez ABTOUe pernseeN “gz ern6Ta (6.01% W499) MO14 SSVW 2 9 s v < % oO aaunsvaw ° aa.vindiws Ollve ASYSNS TANNNL GNIM(%) ¢ FAN EFFICIENCY, 9 100, 2 ° a ° 2 ° np ° 61 © TUNNEL OPERATING POINT (O BLOCKAGE) = [ CEM. " 6084 CEM. 500 1000 1500 2000 2500 3000 FAN SPEED (R.P.M.) Figure 29. Fan efficiency versus motor speed62 «10° (AMPERE)? =Rg 6 8 10 2 14 16 18 Tv crtsecy'”? Figure A-1. Hot wire calibration versus freestream velocity10. 11. 63 BIBLIOGRAPHY Stephens, T., “Design, Construction and Evaluation of a Magnetic Suspension System for Wind Tunnels," M.I.T, Aerophysics Laboratory TR 136 November 1969. Pope, A., Wind Tunnel Testing, John Wiley and Sons, Inc., November 1958, Pankhurst, R. C. and Hobler, D.W., Wind Tunnel Technique, Pitman Publishing Corp., New York,1952. Schlichting, H., Boundary Layer Theory, McGraw-Hill Book Co., 1955, Roberts, H. E., "Considerations in the Design of a Low Cost Wind Tunnel," Paper presented at the 14th Annual Meeting of the I.A.S., January 1946. Collar, A, R., "The Effect of Gauze on the Velocity Distribution in a Uniform Duct," R and M No. 1867, 1939. Batchelor, G. K., "On the Concept and Properties of Idealized Hydrodynamic Resistance," A.C.A. 13, 1945. Dryden, H. L. and Schubauer, G. H., “The Use of Damping Screens for the Reduction of Turbulence," J.A.S. Vol. 14, p. 221, 1947. Tsien, H. S., "On the Design of the Contraction Cone for a Wind Tunnel," J.A.S., Vol. 10, p. 68, 1943. Ames Research Staff, “Equations, Tables and Charts for Compressible Flow," NACA Report 1135, 1953. Rosenhead, L., Editor, Laminar Boundary Layers, Oxford University Press, 1963,12. 13. 14. 64 BIBLIOGRAPHY (Continued) Viajinac, M. and Gilliam, G. D., "Aerodynamic Testing on Conical Configurations Using a Magnetic Suspension System," M.I.1T. Aerophysics Laboratory TR 162 November 1969. Judd, M., Vlajinac, M., Covert, E.E., "Sting-Free Drag Measurements on Ellipsoidal Cylirders at Low Reynolds Numbers" (to be published) Kovasznay, L.S.G., Section F, Physical Measurements n_Gas Dynamics and Combustion, Volume IX, High Speed Aerodynamics and Jet Propulsion, Princeton University Press, 1958.