[American Institute of Aeronautics and Astronautics 1st AIAA Atmospheric and Space Environments...

American Institute of Aeronautics and Astronautics

1

Computational Investigation of a Bleed Air Ice Protection System

See-Ho Wong*, Michael Papadakis †, and Alonso Zamora‡ Wichita State University, Wichita, Kansas, 67260

A computational study was conducted to investigate the performance of a hot air anti-icing system under dry external flow conditions. The hot air system consisted of a piccolo tube and an inner-liner skin installed inside the leading edge of a 72-in span, 60-in chord business jet wing model. A conjugate heat transfer analysis was performed with a three-dimensional Navier-Stokes computer code to compute leading edge skin temperatures and hot air system internal flow properties. Simulations were conducted with a full-span model (FSM) to investigate the flow development inside the piccolo tube and with a partial-span model (PSM) to compute hot air system performance. The computational results obtained were compared with experimental data obtained from icing tunnel tests at the NASA Glenn Icing Research Tunnel. Computed leading edge skin temperatures, piccolo centerline total temperatures and pressures were found to be in good agreement with the experimental data.

Nomenclature 2D = Two-Dimensional 3D = Three-Dimensional Al = Aluminum AOA = Angle of Attack CFD = Computational Fluid Dynamics FSM = Full Span Model IPS = Ice Protection System IRT = Icing Research Tunnel LE = Leading Edge (Wing) NASA = National Aeronautics and Space Administration OD = Piccolo Outer Diameter PSM = Partial Span Model RTD = Resistance Temperature Detector WSU = Wichita State University Cp = Specific Heat Capacity at Constant Pressure k = Thermal Conductivity r = Radial Direction s = Surface Distance along LE x/c = Non-dimensional Chordwise Location y+ = Non-dimensional distance from the wall

I. Introduction

Aircraft icing is of concern because ice accumulation on aircraft surfaces can lead to considerable deterioration in aerodynamic performance and handling qualities and can compromise aircraft safety. Safe operation of aircraft and jet engines in icing conditions requires the use of ice protection systems which are typically installed on critical aerodynamic surfaces such as wings, tails, engine inlets, etc. * Research Scientist, Department of Aerospace Engineering, Campus Box 44, AIAA Member. † Professor, Department of Aerospace Engineering, AIAA Member. ‡ Graduate Research Assistant, Department of Aerospace Engineering, AIAA Student Member.

1st AIAA Atmospheric and Space Environments Conference22 - 25 June 2009, San Antonio, Texas

AIAA 2009-3966

Copyright © 2009 by Authors. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.


2

Ice protection systems (IPS) employ various means to remove or prevent ice accretion including freezing point depressant fluids, mechanical surface deformation, and thermal energy from electrical or hot air sources. From an operations point of view, ice protection systems can be classified as de-icing or anti-icing. In general, de-icing systems use mechanical or thermal means for periodic ice removal while anti-icing systems use thermal means for ice prevention.

Mechanical de-icing systems such as pneumatic inflatable boots, electro-impulse or electro-expulsive systems have low energy requirements and are suitable for small general aviation and commuter type aircraft where excess engine power and heat for ice protection are limited. De-icing systems are usually activated at regular intervals, for example every 0.5 to 3 minutes for periodic ice removal. Periodic ice-removal, however, results in higher aerodynamic penalties due to potential residual and inter-cycle ice formation.

Anti-icing systems use thermal energy from electrical or hot air sources to prevent ice formation. Most jet aircraft use hot air anti-icing systems where the thermal energy is provided by bleed air from the low or high-pressure compressor of the jet engine. Bleed air from the jet engine at elevated temperatures is directed to the protected area using a series of air jets from a piccolo tube installed inside the leading edge of the wing, tail or engine inlet.

Bleed air anti-icing systems are usually classified as evaporative or running wet. Evaporative systems use high heat fluxes to evaporate the impinging droplets so that both the protected areas and surfaces downstream remain dry and free of ice build-up. This calls for a surface temperature near or greater than 49°C. With the introduction of high bypass ratio jet engines it is becoming increasingly difficult to provide sufficient bleed air for evaporative anti-icing systems. In cases where engine bleed air is limited running wet systems are often used. These systems have lower thermal energy requirements and typically maintain surface temperatures between 4°C and 10°C. Running wet systems, however, can result in runback ice which is formed by water running back and freezing behind the protected region. To reduce the amounts of runback ice, some running wet systems have been designed to generate surface temperatures up to evaporative over part of their operating envelope.

The operation of a wing bleed air ice protection system relies on convective heat transfer via high temperature air supplied by a piccolo tube equipped with small diameter holes placed at regular spanwise intervals. Hot air jets emanate from the piccolo holes and impinge on the inside of the wing leading edge (LE) transferring heat to the metal skin. The design and optimization of these systems is a complex endeavor due to the large number of geometric and air flow (external and internal) factors that impact system performance and efficiency. 1,2,3 System and flow parameters that affect the performance of hot air systems include:

• Piccolo tube size and cross section shape, number of piccolo holes, piccolo hole diameter and circumferential placement, piccolo hole spanwise pattern, and piccolo vertical and horizontal placement within the leading edge cavity.

• Wing leading edge geometry and diffuser shape. • Wing skin thickness which affects heat distribution through conduction. • Bleed air flow properties such as temperature and mass flow. • Heat losses through the system components (e.g. inner-liner and wing skins in the chordwise direction). • External flow icing conditions including angle of attack, air speed, air temperature and pressure, liquid

water content and droplet size. This paper describes the application of a three-dimensional (3D) computational fluid dynamics (CFD)

commercial computer code to evaluate the effect of hot air system parameters and external flow conditions on system performance. The computational results obtained are compared with experimental data from tests conducted at the NASA Glenn Icing Research Tunnel (IRT). 1,2

II. Wing Model and Hot Air System Geometry

The two-dimensional (2D) wing model used in the computations and the icing tunnel tests had an airfoil section that was representative of business jet wings. The wing airfoil is shown in Fig. 1. The wing had 72-in span and 60-in chord. The leading edge skin was made of Al 6061 T6 aluminum and had a thickness of 0.1 inches. The leading edge was equipped with a bleed air system that consisted of a piccolo tube and an inner-liner skin as shown in Fig. 2. The inner-liner skin was designed to direct the hot air flow close to the leading edge interior surface. The piccolo tube had an outer diameter (OD) of 1.25 inches and a diamond hole-pattern as shown in Figs. 3 and 4. The diameter of each piccolo jet hole was 0.052 inches and the spacing (pitch) between adjacent holes was 2.44 in. The single


3

-0.1

0.0

0.1

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

holes in the diamond pattern were placed halfway between stations containing two holes. The heated region extended to 10% chord on both upper and lower surfaces as shown in Fig. 2.

Figure 1. Airfoil section.

Figure 2. Bleed air system details.

Figure 3. Details of P45-0-M45 piccolo tube. Figure 4. Details of P70-P10-M40 piccolo tube.

The piccolo tubes used in the icing tests included the P45-0-M45 configuration shown in Fig. 3 and the P70-P10-M40 configuration shown in Fig. 4. The only difference between these two piccolo tubes was the circumferential placement of the piccolo jet holes. For the P45-0-M45 tube the two holes were placed at +45º and -45º, and the single holes at 0º with respect to an axis through the center of the piccolo. For the P70-P10-M40 tube, the two holes were placed at +70º and -40º and the single holes at +10º.

The wing model for the icing tunnel tests was designed with a removable leading edge. The leading edge used for thermal measurements is shown in Fig. 5. Note that the thermal LE was insulated from the remaining wing structure using Teflon® inserts. This was done to reduce heat loss through conduction to the main body of the wing.

Pitch Pitch

Channel between Leading Edge Skin and Inner-Liner Skin

Diffuser BaySection

Piccolo Tube

Inner-Liner Skin

Leading Edge Skin Lower Surface

Upper Surface

Chord Line

10% chord

Diffuser Passage Exit


”


4

Figure 5. Wing cross-section showing the wing LE and bleed air system details – Thermal LE.

The wing model tested in the IRT is shown in Fig. 6, and was extensively instrumented for thermal measurements. Leading edge skin temperatures were measured at four spanwise stations as shown in Fig. 6. Station 1 was 27.725 inches above the IRT test section floor and contained 12 micro-foil(R) heat flux gages as shown in Fig. 7. Each heat flux sensor was equipped with a thermocouple to monitor surface temperatures and correct the heat flux measurements. The sensors were 0.003 in. (0.08 mm) thick, 0.276 in. (7 mm) wide and 0.709 in. (18 mm) long. Station 2, also referred to as Station B, was located 30.175 inches above the tunnel floor and was instrumented with 32 T-type thermocouples. The thermocouples were embedded inside the leading edge skin approximately halfway between the inner and outer skin surfaces as shown in Fig. 8. Station 3, also referred to as Station A, was located 42.425 inches above the tunnel floor and was instrumented with 32 T-type thermocouples as shown in Fig. 8. Station 4 was 52.225 inches above the tunnel floor and was instrumented with 15 platinum RTDs embedded inside the skin as shown in Fig. 9.

Figure 6. Spanwise locations of leading edge skin thermal instrumentation.

Figure 7. Chordwise distribution of heat flux sensors attached to the skin interior surface.

Teflon®Inserts

{

Two sets of inserts (I1, I2) were fabricated. One set was made }

I1

I2

I3 FrontSparTeflon®

Inserts

FrontSpar

Insert 1

Insert 2


Diffuser PassageExit

15 Heat Flux Gages on LE Skin Interior Surface

4 Heat Flux Sensors on back of Inner-Liner Skin

RTD’s

Rib

Rib

TC’s - A

TC’s - BHFG’s

52.225 in

42.425 in

30.175 in

27.725 in

Taps 36 in

Floor

Two sets of inserts werefabricated. One set was madeof aluminum and the other wasmade of Teflon®

LE Skin Interior Surface


5

Figure 8. T-type thermocouples embedded inside the leading edge skin (Stations A and B).

Figure 9. Chordwise distribution of RTD’s embedded inside leading edge skin.

Figure 10. T-type thermocouples installed in diffuser and diffuser passages.

Figure 11. T-type thermocouples installed on back of inner-liner.

Figure 10 shows the instrumentation installed in the diffuser and diffuser passages formed by the inner-liner and

the leading edge skin. Pressure taps installed on the inner-liner skin were used to monitor pressures and T-type thermocouple were used to measure hot air temperature. Another set of thermocouples (see Fig. 11) and four heat flux gages were installed on the back of the inner-liner skin to quantify heat loss through the skin. Inner-liner and diffuser pressure taps and thermocouples were installed at spanwise locations corresponding to Stations A and B. Heat flux gages were placed at Station B only.

In addition to the instrumentation described above, thermocouples and total pressure ports were installed inside the piccolo at spanwise locations 6.92, 30.18, 42.43, and 65.59 inches above the tunnel floor. A flow meter, two thermocouples and a pitot-static probe were used to monitor bleed air mass flow properties upstream of the piccolo inlet.

Groove in Skin for RTD Installation

RTD Embedded Inside Skin

T-type Thermocouple Embedded Inside Skin

Grooves in LE Skin for Thermocouple Installation

Thermocouple Placement in Diffuser Region

Thermocouple Placement on Back of Inner-Liner Skin

S = 0

S = 13.81-in

S = 6.87-in

S = 6.59-in

S = 9.61-in

S = 4.13-in Wing LE Surface Distance, S


6

III. Computational Methodology A. Computational Models

In general, the simulation of anti-icing systems requires modeling of the interior hot air region, the external (dry or wet) flow region, the runback water region (for wet external flow) and the solid (skin) region. An iterative process is needed to determine the flow properties that would satisfy all the equations used to model the physical phenomena in these four regions. The computational methodology involves the solution of the internal and external flows in a coupled fashion. For a wet external flow the methodology requires evaluation of the collection rate of droplets, the computation of the thin film that develops from the impingement of the water droplets, the conduction through the solid walls, and the use of a conjugate heat transfer procedure to couple the conduction-convection problem. The CFD studies presented here were performed for dry external flow conditions. A 72-in Full Span Model (FSM) and a 4.88-in Partial Span Model (PSM) were used in the computations performed with the commercial Navier-Stokes code FLUENT4. The geometric specifications for the two models are listed in Table 1. The FSM provided pressure and temperature conditions at the piccolo jet hole exits and at the piccolo tube exterior surface. These values were used in the PSM, allowing a considerably higher grid resolution (particularly in the jet regions), lower storage requirements, and reduced computational effort to obtain a converged solution.

Table 1. Geometric specification of computational models Span

length (in) AOA (deg)

Piccolo hole angular

placement

Total number of piccolo

holes modeled

Piccolo hole pattern

Piccolo interior/skin

modeling Full Span

Model 72 3 P45-0-M45 84 3-in-line

(28 × 3) Yes

Partial Span

Model 4.88 3 P45-0-M45,

P70-P10-M40 6 Diamond (1/2-2-1-2-1/2) No

The FSM, shown in Figs. 12 and 13, was modified to reduce the meshing effort and the grid cell density. It included a partial tunnel (6-ft high × 9-ft wide × 2.33-ft long) and a truncated wing. The whole computational domain was terminated at a downstream plane normal to the end of the diffuser passage formed between the wing skin and the inner-liner skin. The circular 0.052-in diameter piccolo holes in the IRT test model were converted to a set of 0.046-in × 0.046-in square holes. The exit area of each square hole was the same as the area of the circular holes. The square holes were used to reduce the meshing effort. The piccolo jet pattern was modified from a diamond type to three-in-line such that the +45°, 0º and -45° holes were at the same plane normal to the piccolo centerline axis. The spacing (pitch) in the three-in-line hole arrangement was kept at 2.44 inches, which is the same as for the IRT test model.

The PSM had a full-length tunnel (4.88-in high × 108-in wide × 3072-in long) and full-chorded wing model (see Fig. 14). The tunnel height and the wing span in the PSM were set to 4.88 inches (twice the piccolo jet hole pitch). Six circular 0.052-in diameter jet holes were modeled on the piccolo tube surface, arranged in a diamond pattern. Only half a hole was modeled at each spanwise end of the computational domain as required for symmetry boundary conditions (see Fig. 15). The hot air domain in the PSM was terminated at the end of the diffuser passage that was formed between the wing and inner-liner skins while the 0.1-in thick LE wing skin was extended to the trailing edge of the wing. Note that in all the experiments referenced in this paper, Teflon® inserts were used to reduce the heat transfer between the LE and the wing main body. As a result the wing skin in the PSM was divided into five parts to allow the specification of low thermal conductivity in the regions where the insulation inserts were located (See Fig. 14). These insert regions were positioned at between 13.5% and 14.5% chord of the wing on both upper and lower surfaces.


7

Figure 12. Full Span Model (FSM) geometry in 6-ft x 9-ft x 2.33-ft tunnel. Figure 13. Close-up of the wing geometry in FSM.

Figure 14. Partial Span Model (PSM) wing geometry. Figure 15. Close-up of the wing geometry in PSM.

B. Grid Generation Gambit5, a grid generation package, was used to create the computational grids presented in this paper. Triangular and quadrilateral elements (quads) are available in Gambit for surface meshing while hexahedral, tetrahedral, wedges and pyramidal elements are used for volume meshing. The three-dimensional grids for both the full and partial span models included the IRT walls and had the wing model centered inside the tunnel at AOA=3º. For the FSM, a hybrid volume mesh consisting of hexahedra and wedges was generated by sweeping the triangles and quads of a 2D surface mesh across the span. The mesh, shown in Figs. 16 and 17, contained triangles in the diffuser and external flow regions, and quads in the inner-liner passage, piccolo interior, wing skin, piccolo tube skin, and boundary layer regions. The total number of grid cells was significantly reduced by using hexahedra and wedges to allow large aspect ratios and stretching along the span. A total number of 763 points were distributed across the 72-in piccolo length with grid point clustering toward the jet planes. Five points were equally distributed across the width of each square piccolo jet hole. The total number of cells, nodes and faces of the full span grid are reported in Table 2.

Table 2. Grid details of the Full Span Model (FSM) Number of

Cells Number of

Nodes Number of

Faces y+average

Full Span Model (FSM) 15,156,180 14,360,738 44,225,796 1.23

Floo

r

Cei

ling

Tunnel Flow Out

Tunnel Flow In

Tunnel Flow Out

Wing skin

Piccolo tube *Tunnel geometry not shown

Insulation insert

Wing skin

Piccolo tube

Inner liner

-45º/0/+45º Diamond

Wing skin

Piccolo tube

-45º/0/+45º 3-in-line x

y

z

6-ft

9-ft

Insulation insert 0-deg half jet

0-deg half jet


8

For the PSM, two hybrid volume grids consisting of hexahedral, wedges and tetrahedral elements were generated for the two different piccolo configurations: P45-0-M45 (Figs. 18-19) and P70-P10-M40. Tetrahedral elements were used mainly in the diffuser section and portion of the external flow region outside of the diffuser, while wedges and hexahedral elements were used in regions including the inner-liner passages and wing skin. Tetrahedral cells sized at 0.003-in were used to mesh the custom designed conical volumes containing the jets and jet impingement regions inside the diffuser. This provided a node distribution of 48 nodes along the circumference of each circular piccolo jet hole (or 18 nodes across the diameter). The size of these tetrahedrons increased gradually to a maximum size of 0.05-in in connecting the high density jet regions to the rest of the diffuser volume. Grid density near walls was high also to ensure proper resolution of the flow and heat transfer properties. Grid resolution selected for the PSM was based on the findings of a previous study on the effects of grid resolution and spatial discretization on recovery factor in an impinging jet (not reported in this paper). The total number of cells, nodes and faces for the two partial span grids are reported in Table 3.

Table 3. Grid Details of the Partial Span Model (PSM) Number of

Cells Number of

Nodes Number of

Faces y+average

Partial Span Model - PSM (P45-0-M45) 13,039,161 5,055,499 30,251,163 0.06

Partial Span Model - PSM (P70-P10-M40) 13,047,346 4,983,428 30,152,870 0.06

Figure 16. Isometric view of the FSM mesh. Figure 17. Section view of the FSM mesh.

Figure 18. Isometric view of the PSM mesh. Figure 19. Section view of the PSM mesh.

Tetrahedrons in conical volume containing the jet and jet impingement region


9

C. Boundary Conditions Typically, the boundary conditions required to compute the internal and external viscous flows include no-slip

conditions at wall boundaries and inflow or outflow conditions at fluid boundaries. The various boundary conditions employed in the FSM and PSM are described below. 1. Full Span Model (FSM)

Boundary conditions included adiabatic walls on four sides of the 6-ft × 9-ft × 2.33-ft tunnel, pressure inlet and pressure outlet conditions at tunnel inlet and outlet respectively. A 2D total pressure profile (in the direction of the y-axis shown in Fig. 12) and total temperature was specified at the 6-ft × 9-ft inlet of the tunnel, while a 2D static pressure profile was specified at the tunnel outlet. At the inlet of the piccolo tube, a pressure inlet boundary condition was used in which the total pressure and total temperature of the hot air entering the tube were specified. At the closed end of the piccolo tube, an isothermal (constant temperature) wall boundary condition was employed. At the exits of upper and lower diffuser passages, shown in Fig. 2, pressure outlet boundary conditions were used where the static pressure of the hot air leaving the system was specified. The inner-liner skin was modeled as a zero thickness wall surface with a 2D chordwise heat flux profile applied to the rear (downstream end) surface of the wall. This profile was computed from the experimental data obtained from the heat flux gages attached to the rear inner-liner surface as shown in Fig. 7. A heat flux profile was used since the experimental inner-liner skin temperatures (measured at two spanwise stations) were not sufficient for a full-span simulation. All wing surfaces at the two spanwise ends were defined as adiabatic wall boundaries. Note that flow properties at the piccolo tube exterior and interior wall surfaces and at the piccolo jet holes were computed during the solution and did not have to be specified.

2. Partial Span Model (PSM) Symmetry conditions were utilized at both spanwise ends of the computational domain which consisted of a 4.88-in span wing model inside a 4.88-in × 108-in × 3072-in tunnel. The two side walls of the tunnel, facing the upper and lower surfaces of the wing model (model was installed vertically in the IRT), were defined as adiabatic walls. Total pressure and total temperature were specified at the tunnel inlet (pressure inlet condition) while static pressure was specified at the tunnel outlet (pressure outlet condition). An isothermal wall boundary condition was used at the piccolo surface and the total pressure and total temperature (pressure inlet condition) were specified at the piccolo jet holes. At the exits of upper and lower diffuser passages, shown in Fig. 2, the static pressure of the hot air leaving the system was specified (pressure outlet condition). A wall boundary condition was specified at the zero thickness inner-liner skin. The wall temperature at the rear (downstream) surface of the inner-liner skin was specified from the thermocouple measurements performed at Station B of the test model (see Fig. 11). The wall portion of the wing LE skin, that extended from the inner-liner exit to the insulation insert (see Fig. 5), was modeled as a wall boundary with a temperature profile specified from the experimental data obtained. D. Turbulence Model

All turbulent regions of the flowfield were modeled with the two-equation k-ω Shear Stress Transport (SST) turbulence model in the FLUENT code. For the simulations presented in this paper, the k-ω SST model was operated as a low Reynolds number model by using the enhanced wall treatment option available in FLUENT. The enhanced wall treatment required a high resolution mesh near wall surfaces with a +y of 1. The k-ω SST model has been successfully validated for heat transfer applications including heated subsonic and supersonic air jets,6,7,8,9 and transonic and separated airfoil flows.10,11 Note that analysis performed with a wing section, showed that for the angles of attack used in the experimental investigation, transition from laminar to turbulent flow occurred at approximately 3% chord on the upper surface and 63% chord on the lower surface. Since the FSM had chord of approximately 6 inches (10% of wing chord) only laminar external flow was modeled in FLUENT. However, the interior flow was assumed turbulent and was computed with the k-ω SST two-equation turbulence model. E. Solver Settings The three-dimensional FLUENT CFD simulations were performed using the pressure based (segregated) solver with the Green-Gauss node based gradient option. Solution controls for the segregated solver included the default SIMPLE pressure-velocity coupling, the standard scheme for pressure discretization, and the first-order upwind scheme for the remaining flow and turbulence scalars. Properties of the fluid material (air) included ideal-gas and Sutherland’s law for density and viscosity specification respectively. Temperature based piecewise-linear functions were used to define the specific heat capacity (Cp) and thermal conductivity (k) of the fluid and solid materials. The


10

wing skin was set as solid material with thermal properties similar to those of aluminum alloy Al 6061-T6 while the Teflon® inserts modeled in the PSM were set as solid material with a low thermal conductivity of 0.5 Wm-1k-1. The iterative solution process was set to end either when the maximum number of iterations (set typically to 20 thousand) was reached or when the magnitudes of the residual of the governing equation were less than 10-6.

IV. Results and Discussion Simulations were performed for the experimental cases listed in Table 4. In all cases the wing model was set at AOA=3º and the Teflon® inserts were used to reduce heat loss from the leading edge skin to the wing main body. Results from a warm hold case are presented first, followed by parametric studies that demonstrate the effects of external air temperature, piccolo temperature, piccolo mass flow, and piccolo hole pattern on LE skin temperatures.

Table 4. Experimental test cases used in the computational studies

Test Run

Freestream Speed

kts (m/s)

Freestream Temperature

ºF (ºC)

Piccolo Inlet Mass Flow lbm/min (kg/min)

Piccolo Inlet Total

TemperatureºF (ºC)

Piccolo Configuration

026 115 (59.2) 20 (-6.7) 4.27 (1.94) 373.8 (189.9) P45-00-M45 042 115 (59.2) 20 (-6.7) 5.31 (2.41) 383.7 (195.4) P45-00-M45 058 110 (56.6) -22 (-30.0) 4.68 (2.12) 367.7 (186.5) P45-00-M45 062 115 (59.2) 20 (-6.7) 4.12 (1.87) 263.5 (128.6) P45-00-M45 066 115 (59.2) 20 (-6.7) 1.83 (0.83) 362.9 (183.8) P45-00-M45 076 115 (59.2) 20 (-6.7) 10.70 (4.86) 257.5 (125.3) P45-00-M45 107 115 (59.2) 20 (-6.7) 4.26 (1.93) 377.9 (192.2) P70-P10-M40

A. Warm Hold Case (Test Run 026) The 72-in Full Span Model (FSM) and the 4.88-in Partial Span Model (PSM) were used to perform simulations for the warm hold case corresponding to experimental Run 026. Results from the FSM are presented to demonstrate the flow development inside the piccolo. The computational results include piccolo centerline total temperature and pressure distributions, and piccolo exterior surface temperature distribution. For the CFD analysis performed with the PSM, the boundary conditions at the exit of the piccolo jet holes, and at the piccolo surface were specified based on the results obtained with the FSM. The tunnel operational pressure in the FSM and PSM simulations was set to 14 psi. Boundary conditions for the FSM computations are as follows:

• Pressure inlet at the opened end of the piccolo tube with a total pressure of 14.09 psig and a total temperature of 373.8ºF. This provided a mass flow rate similar to the experimental value.

• Wall condition at the closed end of the piccolo tube with a fixed temperature of 316ºF, a value extrapolated from piccolo interior temperatures from the experiment.

• Pressure outlet at the upper and lower exits of the diffuser passage formed between the wing skin and the inner-liner skin with a gauge static pressure of 0.47 psig. This value was obtained from the experimental measurements.

• Wall condition along the inner-liner surface with a chordwise heat flux profile based on the heat flux measurements from the experiment.

• Pressure inlet at the upstream end of the truncated tunnel (see Fig. 12) with a fixed temperature of 22.95ºF and a total pressure profile obtained from CFD analysis performed with the wing in the full-size IRT test section (6-ft × 9-ft × 20-ft).

• Static pressure at the upper and lower tunnel outlet surfaces (note the outlet of the truncated tunnel shown in Fig. 12 is divided by the wing into two parts) with a static pressure profile determined from CFD analysis performed with the wing in the full-size IRT test section. The tunnel airspeed computed from the pressures specified at the tunnel inlet and outlet boundaries was approximately 115 knot.

• No-slip adiabatic wall conditions at all four walls of the truncated tunnel.


11

The FSM simulation was conducted with the default pressure-based segregated solver, first order discretization scheme, and the k-ω SST turbulence model. Note that the wing external cold flow was assumed to be laminar. Selected FSM results from the warm hold simulation are shown in Figs. 20 – 29. Figures 20 and 21 show the spanwise total temperature contours of the piccolo outer surface and the piccolo interior flow respectively. The results presented in these figures show that both the piccolo skin and interior flow temperatures decreased with spanwise distance from the piccolo inlet. The temperature drop was approximately 110ºF between the inlet and the closed end of the piccolo tube. The surface temperature of the piccolo tube was approximately 50ºF lower than that of the fluid inside the tube at all spanwise stations. The spanwise distribution of the total temperature along the centerline of the piccolo tube is plotted against experimental data in Fig. 22. In this plot, the x-distance is measured from the piccolo inlet which was close to the tunnel floor. Figure 22 shows that the drop in the hot air temperature along the piccolo tube centerline was nonlinear and was characterized by a steep decline near the closed-end of the piccolo tube. The hot air total pressure along the piccolo centerline remained nearly constant as shown in Fig. 23. In general, the computed piccolo centerline total temperature and pressure distributions were in good agreement with the experimental data. Figures 24 – 27 present the radial distributions of various flow properties along the jet axis passing through the 0-deg piccolo hole located at z=30.52-in of the FSM (a spanwise distance close to Station B in the icing tunnel model). The start of the x-axis in Figs. 24 – 27 is at the piccolo centerline (r=0) and the end is at the exit of the piccolo jet hole located on the surface of the piccolo tube (r =0.5*OD). The dashed lines in the figures indicate the start of the piccolo hole. Figure 24 demonstrates that the hot air total pressure dropped by approximately 1.4 psig across the piccolo hole due to orifice discharge effects. The hot air total temperature decreased by approximately 10ºF from the center of the piccolo tube to the exit of the piccolo jet hole as shown in Fig. 25. The radial Mach number distribution from the piccolo axis to the plane of the jet exit is shown in Fig. 26. Figure 27 shows the variation in the static pressure in the piccolo radial direction.

Figure 20. Spanwise temperature distribution on piccolo tube outer surface, Run 026.

Figure 21. Spanwise temperature distribution inside piccolo tube, Run 026.

Inlet

Closed-end

Inlet

Closed-end


12

0 10 20 30 40 50 60 70Spanwise Distance (inch)

240

260

280

300

320

340

360

380

400

Tem

pera

ture

in P

icco

lo (°

F)

Full Span ModelR026

ExperimentalCFD

0 10 20 30 40 50 60 70Spanwise Distance (inch)

20.0

22.0

24.0

26.0

28.0

30.0

32.0

34.0

36.0

38.0

40.0

Pres

sure

in P

icco

lo (p

sia) Full Span Model

R026

ExperimentalCFD

Figure 22. Total temperature spanwise distribution at piccolo centerline, Run 026.

Figure 23. Total pressure spanwise distribution at piccolo centerline, Run 026.

0.5 0.4 0.3 0.2 0.1 0r / OD

12.6

12.8

13.0

13.2

13.4

13.6

13.8

14.0

14.2

Tota

l Pre

ssur

e (p

sig)

0.5 0.4 0.3 0.2 0.1 0r / OD

358

360

362

364

366

368

370

372

374

Tota

l Tem

pera

ture

(ºF)

Figure 24. Total pressure distribution along 0-deg jet axis, Station B, z =30.52-in, Run 026.

Figure 25. Total temperature distribution along 0-deg jet axis, Station B, z =30.52-in, Run 026.

0.5 0.4 0.3 0.2 0.1 0r / OD

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Mac

h

0.5 0.4 0.3 0.2 0.1 0r / OD

2

4

6

8

10

12

14

16

Stat

ic P

ress

ure

(psi

g)

Figure 26. Mach number distribution along 0-deg jet axis, Station B, z =30.52-in, Run 026.

Figure 27. Static pressure distribution along 0-deg jet axis, Station B, z =30.52-in, Run 026.

r r

r r

Piccolo interior Piccolo exterior

Piccolo interior

Piccolo exterior

Piccolo interior

Piccolo exterior

Piccolo interior

Piccolo exterior


13

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5rexit / width

0.60

0.65

0.70

0.75

0.80

0.85

0.90

0.95

1.00

Loca

l / M

ax R

atio

Normalized Boundary Conditions DistributionPiccolo Jet Exit

Tt/Ttmax, Ttmax = 358.181°F Ttavg = 349.41°F

Pt/Ptmax, Ptmax = 12.61 psigPtavg = 10.15 psig

Ps/Psmax, Psmax = 2.33 psigPsavg = 2.1482 psig

Flow properties across the diameter of the piccolo hole (0-deg jet at z=30.52-in) are examined in Figs. 28 – 29. Static pressures, total pressures, and total temperatures at the exit plane of the piccolo hole were sampled at five equally spaced points (see Fig. 28 for the locations). The sampled properties were subsequently normalized with their respective jet centerline values and are plotted in Fig. 29. Figure 29 shows that the pressure and temperature profiles were symmetric about the jet axis as expected. Total pressure and temperature were highest at the center and decreased towards the edge of the hole. Furthermore, the total temperature distribution exhibited a large variation across the diameter of the piccolo hole (a 40% change). Both the jet centerline and averaged flow properties at the piccolo hole exit are listed in the legend of Fig. 29.

Figure 28. Cross section total pressure at 0-deg piccolo hole; FSM, z = 30.52-in (Station B).

Figure 29. Normalized flow properties at the piccolo hole exit, Run 026.

The FSM simulation was followed by two PSM simulations which were used to predict the wing skin temperatures measured at Station B. Two sets of boundary conditions were applied at the piccolo jet holes in the PSM simulations, one for the jet centerline and the other the averaged flow properties. The conditions at the piccolo holes as well as other boundary surfaces of the PSM included:

• Pressure inlet at the piccolo holes with two different sets of boundary values as follows: 1. Jet centerline values (Ptotal = 12.60 psig, Ttotal = 358.0ºF and Pstatic = 2.14 psig) 2. Averaged values (Ptotal = 10.15 psig, Ttotal = 349.4ºF, and Pstatic = 2.15 psig)

• A wall boundary condition at the piccolo tube surface with a fixed temperature of 303ºF. This value was obtained by averaging the piccolo exterior surface temperatures computed with the FSM at z=30.52-in.

• Pressure outlet with a gauge static pressure of 0.47 psig at the upper and lower exits of the diffuser passages.

• Wall condition along the inner-liner surface with a chordwise temperature profile based on the temperature measurements from the experiment.

• Wall conditions along portion of the wing LE skin that extended from the inner-liner exit to the insulation insert with a chordwise temperature profile specified from the experimental data.

• Pressure inlet at the inlet of the tunnel with a total pressure of 0.375 psig and total temperature of 22.95ºF as recorded in the experiment.

• Pressure outlet at the tunnel outlet with a zero gauge static pressure. The tunnel speed resulted from the pressure difference between the tunnel inlet and outlet was approximately 115 knot.

• Symmetry conditions at both span ends of the 4.88-in segment. • No-slip adiabatic walls at the interior surface of the wing skin downstream of the Teflon® insert. • No-slip adiabatic walls at the two side walls of the tunnel.

Jet Width

1

3

4

5

2

Jet Inlet


14

The PSM simulations were conducted with the pressure-based segregated solver and the default first order discretization. The k-ω SST turbulence model was applied to compute the turbulent internal hot-air flow. A laminar region was defined on the airfoil surface extending from the upper surface to the lower surface transition point. Laminar external flow analysis was performed over the laminar region. For the remaining part of the airfoil surface and for all other wall surfaces the k-ω SST turbulence model was applied to compute turbulent flow properties. The transition from laminar to turbulent flow was set to 2.97% and 63.37%, expressed in percent chord, on the upper and lower surfaces of the wing respectively. It was found from earlier analysis, not reported in this paper, that a fully turbulent external flow over the airfoil surface resulted in lower wing skin temperatures near the leading edge region compared with the experimental data due to higher external flow convection associated with turbulent flow. The Teflon® insert used to reduce the heat transfer between the LE and the main wing body in the experiments was modeled in the simulations through the specification of low thermal conductivity in regions of the wing skin where the Teflon® inserts were attached.

Figure 30 demonstrates the effects of using averaged (across the jet hole) and centerline (centerline of the jet hole) piccolo jet hole boundary values on the LE skin temperature. The experimental data shown in Fig. 30 correspond to the LE skin temperatures measured at Station B. The computed temperatures were taken from the PSM at z=1.83-in which corresponded to a spanwise distance halfway between the single- and double-jet stations. Note that the computed skin temperatures were extracted halfway across the leading edge skin thickness emulating the readings from the embedded thermocouples in the experimental setup. The surface distance (s) in Fig. 30 is the wrap distance measured chordwise along the wing LE surface from the furthermost downstream thermocouple on the lower surface (see Fig. 8 for locations of all the wing LE thermocouples). Figure 30 shows that both averaged and centerline piccolo jet hole boundary values predicted the correct experimental trend. Surface temperatures in both cases were highest near the leading edge and decreased gradually away from it. Lower temperatures were observed along the upper surface of the wing compared to the temperatures measured along the wing lower surface. However, using the averaged boundary values resulted in better agreement with the experimental data than when the centerline values were used. For the averaged boundary values case, the temperatures at the leading edge and along the upper surface were computed accurately with little discrepancy between the computational and the experimental data. But when the centerline values were used, the leading edge temperature was over-predicted by as much as 10ºF (5%). Figure 30 also shows that the wing skin temperatures computed using averaged and centerline properties were practically the same at the downstream end of the upper and lower surfaces. This was the result of imposing the same temperature profiles at the aft portion of the LE interior skin (-2.09-in < s < 0 and 13.81-in < s < 15.76-in).

0 2 4 6 8 10 12 14Surface Distance, s (in)

60

80

100

120

140

160

180

200

220

Skin

Tem

pera

ture

(ºF)

Boundary Condition StudyRun 026

ExperimentalAverage Inflow ConditionsCenterline Inflow Conditions

Figure 30. Leading edge skin temperature, PSM, Station B, Run 026.


15

Other selected results from the PSM simulation are shown in Figs. 31-37. The streamlines shown in Fig. 31 indicate that the external flow accelerated over a longer distance on the upper surface downstream of the stagnation point. The accelerated flow subsequently lowered the static temperature and increased the cooling effect on the upper surface. The figure also shows that the hot air from the piccolo holes impinged on the wing skin and circulated inside the diffuser region before leaving through the upper and lower diffuser passages formed between the inner-liner and the wing skin. The temperature contours shown in Fig. 32 correspond to a surface created halfway between the exterior and interior surfaces of the wing LE skin. This interior surface corresponds to the thermocouple plane inside the LE skin. The LE skin temperature contours in Fig. 32 show the locations of the hot air jet impingement regions. Figure 33 shows the surface heat flux on the interior of the wing skin at five spanwise stations corresponding to the double-jet station (z=1.22-in), the single-jet station (z=2.44-in) and three additional stations equally spaced in-between (z=1.525, 1.83 and 2.135 inches). Experimental data recorded at 27.725 inches above the tunnel floor (a spanwise station nearly halfway between the single- and the adjacent double-jet stations) are also provided in Fig. 33 for comparison. At each spanwise jet station, the computed heat flux was highest at the location of jet impingement and decreased rapidly away from the jet. Note that the heat flux distribution computed at each spanwise station was not symmetrical along the upper and lower surface. The -45º jet contributed a higher local heat flux than the +45° jet as seen in the results taken at z=1.525-in. This is attributed to higher wing skin temperatures along the lower surface. In general, the experimental data were best correlated with the computed surface heat flux taken at z=1.83-in, the station exactly halfway between a single- and double-jet stations. The surface heat flux values computed at spanwise station z=1.83-in was similar to the recorded experimental data (i.e., between 5 and 15 W/in2). Note that the accuracy of the surface heat flux measurement in the experiment depended on both the thickness of the sensor and the thickness of the glue layer between the sensor to the wing interior surface. Figures 34 and 35 show sectional contours of total temperature and gauge total pressure, respectively, at five spanwise stations containing the hot air jets in the PSM. Figure 36 shows that the 0-deg jet impinging on the concave surface of the wing skin was non-axisymmetric. More flow was directed towards the lower surface due to the inclination of the wing surface with respect to the jet axis. Figure 37 shows that the total pressure remained high throughout the jet core and over the impingement region. The jet shear layers that bounded the jet core are also visible in the contour maps presented in Figs. 36 and 37.

Figure 31. Streamlines colored by total temperature in PSM, Run 026.


16

Figure 32. Wing LE skin temperature, Run 026. Figure 33. Interior skin heat flux at different spanwise stations, Run 026.

Figure 34. Total temperature contours at jet stations in PSM, Run 026.

Figure 35. Total pressure contours at jet stations in PSM, Run 026.

Figure 36. Mach number contours of 0-deg jet in PSM, Run 026.

Figure 37. Total pressure contours of 0-deg jet in PSM, Run 026.

4 5 6 7 8 9 10Surface Distance (in)

0

5

10

15

20

25

30

35

40

45

50

Hea

t Flu

x (W

/in2 )

Surface Heat FluxR026

z = 1.22 in (2 jet station)z = 1.525 inz = 1.83 in z = 2.135 inz = 2.44 in (1 jet station)z = 1.83 in (exp)

Piccolo jet hole surface

Piccolo jet hole surface


17

B. Anti-Icing System Parametric Studies A total of seven cases were analyzed with the FLUENT code to investigate the effects of external flow

conditions, piccolo circumferential jet placement, and hot air flow conditions on LE skin temperatures. The experimental flow conditions used in the FSM simulations are listed in Table 4. The boundary conditions obtained from the FSM results along with the mass flow (per hole) reported in the PSM simulations are presented in Table 5. The mass flows reported in the PSM simulations were found to be proportional to the piccolo inlet mass flow used in the experiments. The effects of each system parameter on the wing skin temperatures are presented below. The computed skin temperatures shown in Figs. 38-43 were extracted at the spanwise station z=1.83-in (halfway between the single- and double-jet spanwise stations) and halfway across the leading edge skin thickness. The CFD results are compared with experimental data recorded at Station B of the wing model. Note that the surface distance, s, in Figs. 38-43 is the wrap distance measured chordwise along the wing LE surface from the furthermost downstream thermocouple on the lower surface.

Table 5. PSM Hot air boundary conditions

Test Runs

Piccolo Holes Piccolo Tube Total

Temperature (ºF)

Total Pressure

(psig)

Static Pressure

(psig)

Computed Mass Flow per hole

(lbm/min)

Surface Temperature

(ºF) 026 349.41 10.15 2.15 0.056 303.45 042 365.86 14.85 3.36 0.067 330.00 058 346.50 10.02 2.01 0.056 308.00 062 252.88 8.65 1.77 0.055 232.33 066 306.49 2.38 0.77 0.026 227.00 076 253.86 38.03 11.87 0.131 246.80 107 356.22 12.36 2.60 0.061 314.50

1. Effect of External Flow Conditions (Test Runs 026, 058)

Computed and experimental wing skin temperatures were compared for a cold hold case (Run 058) and a warm hold case (Run 026). The cold hold case was characterized by external airspeed of 110 knot, and -22ºF static external flow temperature. The wing skin temperatures computed from the two PSM simulations performed are compared with corresponding experimental data in Fig. 38. The results in Fig. 38 show that for a similar piccolo hot air temperature and mass flow rate, the wing skin temperatures decreased as the external cold air temperature was decreased (increased external cooling). The drop in skin temperature due to the drop in the external flow static temperature was approximately 39ºF and 36ºF at the downstream ends of the upper and lower wing surfaces respectively, and 24.5ºF at the leading edge. The computed and experimentally recorded temperatures at three regions of the LE skin namely the forward region, upper region and lower region were averaged and are reported in Table 6. The forward LE region was located between a surface distance (s) of 6.59-in and 6.87-in. This region contained the two thermocouples installed near the most forward portion of the wing LE as shown in Fig. 8. The upper region was defined as 9.61-in ≤ s ≤ 13.81-in. This region corresponded to the portion of wing upper surface positioned above the inner-liner. The region of the wing lower surface over the lower inner-liner passage was defined as 0-in ≤ s ≤ 4.13-in. The upper and lower regions, as defined above, contained six thermocouples each. These thermocouples were installed inside the wing LE skin. The experimental values reported in Table 6 were obtained from the thermocouple data. The CFD values were computed from the PSM simulations. The percent errors of the computed temperature with respect to experimental data are also provided in Table 6. The following equation was used to compute the percent error:

100ExpAvg,

ExpAvg,CFDAvg,×

−

T

TT (1)

The percent errors in Table 6 for all three wing regions indicate that in general, the CFD results for the warm hold case (Run 026) were in better agreement with the experimental data compared to corresponding results for the cold hold case (Run 058).


18

Table 6. Computed skin temperatures, experimental data and percent errors

Test Run

Forward Region 6.59 ≤ s ≤ 6.87-in

Upper Region 9.61 ≤ s ≤ 13.81-in

Lower Region 0 ≤ s ≤ 4.13-in

TAvg, CFD ºF

TAvg, Exp ºF % error TAvg, CFD

ºF TAvg, Exp

ºF % error TAvg, CFD ºF

TAvg, Exp ºF % error

026 192.63 192.90 0.14 108.08 106.69 1.30 131.82 127.28 3.57 058 168.07 171.26 1.86 81.35 71.53 13.72 101.46 96.61 5.02

2. Effect of Hot Air Temperature (Test Runs 026, 062)

The effect of piccolo hot air temperature on wing skin temperatures is shown in Fig. 39. The piccolo inlet total temperatures in Run 026 and Run 062 were set to 373.8ºF and 263.5ºF, respectively. In the simulations performed with the PSM, the total temperature specified at the piccolo jet holes in Run 026 was 96.5ºF higher than that used in Run 062. For similar external flow conditions and piccolo inlet mass flow rates, increasing piccolo hot air temperature raised the skin temperature at the forward region of the LE by approximately 52ºF. The gain in wing skin temperatures was most prominent at the forward region, followed by the lower region and lastly by the upper region. The percent difference between the computed temperatures obtained from the PSM simulations and the experimental data at the forward, upper and lower regions for test runs 26 and 62 are listed in Table 7. In general, the observed differences were small.

Table 7. Computed Skin skin temperatures, experimental data and percent errors

Test Run




TAvg, CFD ºF


ºF TAvg, Exp



026 192.63 192.90 0.14 108.08 106.69 1.30 131.82 127.28 3.57 062 141.38 140.62 0.54 86.57 81.54 6.17 98.30 95.04 3.43


0

20

40

60

80

100

120

140

160

180

200

Skin

Tem

pera

ture

(ºF)

Effect of External Flow ConditionsLeading Edge Skin Temperature

Warm Hold (R026 - CFD)Warm Hold (R026 - Experimental)Cold Hold (R058 - CFD)Cold Hold (R058 - Experimental)

Figure 38. Effect of external temperature on wing LE skin temperature.


19

3. Effect of Mass Flow at High Temperature (Test Runs 026, 042, 066)

For similar external flow conditions and piccolo inlet temperature (on the order of 370ºF), wing skin temperatures obtained for three piccolo mass flow rates are compared in Fig. 40. The mass flow rates investigated included 1.83 lbm/min (Run 066), 4.27 lbm/min (Run 026) and 5.31 lbm/min (Run 042). Figure 40 shows that skin temperatures over the protected area increased as the piccolo mass flow rate was increased. Figure 41 shows that the gain in averaged temperatures at the forward, upper and lower regions of the wing LE skin were nearly linear with respect to the piccolo mass flow rate. Good correlations between the numerical and experimental data were observed at the upper region with a maximum percent error of 2.53% corresponding to Run 066 as shown in Table 8. The highest discrepancy was observed at the forward region of the wing LE for the low mass flow case (Run 066) where the computed skin temperature was 7.6ºF (or 5.88%) lower than the experimental data.


20

40

60

80

100

120

140

160

180

200

220

Skin

Tem

pera

ture

(ºF)

Effect of Hot Air Mass FlowLeading Edge Skin Temperature

4.27 lbm/min (R026 - CFD)4.27 lbm/min (R026 - Experimental)5.31 lbm/min (R042 - CFD)5.31 lbm/min (R042 - Experimental)1.83 lbm/min (R066 - CFD)1.83 lbm/min (R066 - Experimental)

1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5Piccolo Inlet Mass Flow (lbm/min)

60

80

100

120

140

160

180

200

220

Skin

Tem

pera

ture

(ºF)

Test Runs 026, 042, 066Forward TAVG (CFD)Forward TAVG (Exp)Upper TAVG (CFD)Upper TAVG (Exp)Lower TAVG (CFD)Lower TAVG (Exp)

Figure 40. Effect of piccolo mass flow at high temperature on wing LE skin temperature.

Figure 41. Skin temperatures at different regions of the wing as function of mass flows.


40

60

80

100

120

140

160

180

200

Skin

Tem

pera

ture

(ºF)

Effect of Hot Air TemperatureLeading Edge Skin Temperature

373.8°F (R026 - CFD)373.8°F (R026 - Experimental)263.5°F (R062 - CFD)263.5°F (R062 - Experimental)

Figure 39. Effect of hot air temperature on wing LE skin temperature.


20


Test Run




TAvg, CFD ºF


ºF TAvg, Exp



026 192.63 192.90 0.14 108.08 106.69 1.30 131.82 127.28 3.57 042 214.17 214.50 0.15 121.96 120.77 0.99 147.72 143.71 2.79 066 122.35 129.98 5.88 77.25 79.26 2.53 79.10 80.44 1.67

4. Effect of Mass Flow at Low Temperature (Test Runs 062, 076)

The effect of hot air mass flow was also evaluated at a lower piccolo inlet temperature (approximately 260ºF). The mass flow rates investigated included 4.12 lbm/min (Run 062) and 10.7 lbm/min (Run 076). Figure 42 shows that skin temperatures increased with increasing piccolo mass flow rates, as was the case for the high piccolo inlet temperature. Gains in LE skin temperature were observed over all three regions of the wing LE skin. The temperature gains at the forward region and downstream end of the lower surface were both approximately 39ºF higher for the 10.7 lbm/min case compared to the 4.12 lbm/min case. The percent differences between the experimental and analytical skin temperatures for all three wing regions are presented in Table 9 and were in the range of 0.54% to 6.17%.


40

60

80

100

120

140

160

180

Skin

Tem

pera

ture

(ºF)

Effect of Hot Air Mass FlowLeading Edge Skin Temperature

4.27 lbm/min (R062 - CFD)4.27 lbm/min (R062 - Experimental)10.7 lbm/min (R076 - CFD)10.7 lbm/min (R076 - Experimental)

Figure 42. Effect of piccolo mass flow on wing LE skin temperature.


Test Run




TAvg, CFD ºF


ºF TAvg, Exp



062 141.38 140.62 0.54 86.57 81.54 6.17 98.30 95.04 3.43 076 178.54 176.97 0.88 113.91 112.48 1.27 139.19 134.16 3.74

5. Effect of Piccolo Configuration (Test Runs 026, 107)

Leading edge skin temperatures were computed for two piccolo tube configurations. These configurations included the baseline P45-0-M45 (Run 026) and the modified P70-P10-M40 (Run 107) piccolo tubes tested at the NASA IRT. Figure 43 shows that for similar internal hot air and external cold air conditions, the P70-P10-M40 configuration produced gains in upper surface temperatures without compromising peak temperatures by directing


21

the air jets towards the upper surface. According to the computed temperatures listed in Table 10, the gain in skin temperatures over the wing upper region was 14.5ºF. The P70-P10-M40 piccolo tube resulted in a small reduction of approximately 5.6ºF in wing skin temperatures over the lower surface of the wing compared to the P45-0-M45 design. The percent difference between the experimental and computed temperature distributions ranged from 0.14% to 3.57% with the best correlation found near the forward region of the wing. The computed gain in the skin temperatures over the wing forward region due to the use of the P70-P10-M40 piccolo was 2.3ºF as compared with 7.9ºF recorded in the experiments. The reason for the higher peak temperature recorded in the experiment was due to a spanwise misalignment of the P70-P10-M40 piccolo jets with respect to the thermocouples embedded in the skin. Note that the computed temperatures correspond to a spanwise station halfway between the single- and double-jet stations. The experimental results presented correspond to a spanwise station close to the thermocouples at Station B.


Test Run




TAvg, CFD ºF


ºF TAvg, Exp



026 192.63 192.90 0.14 108.08 106.69 1.30 131.82 127.28 3.57 107 194.95 200.79 2.91 122.55 121.05 1.24 126.25 124.80 1.16

V. Summary and Conclusions

A parametric computational study was performed with a business jet wing equipped with a hot air ice protection system. The study examined the effects of external flow conditions, hot air temperature, hot air mass flow, and piccolo configuration on the wing leading edge skin temperatures. Three-dimensional CFD analysis data obtained with FLUENT, a commercial Navier-Stokes computer code, were compared with experimental dry air data obtained at the NASA IRT facility. Two computational models were used in the simulations. These included a full-span (72 inches) partial-chord (forward 10% of wing chord) wing model and a partial-span (4.88-in) full-chord (60-inch) wing model. Hot air system data computed with the full-span model and test data obtained from icing tunnel tests with a highly instrumented wing model were used to set up boundary conditions for the partial-span wing model. The partial-span wing model was used to investigate the effect of hot air system parameters on wing leading edge


80

100

120

140

160

180

200

220Sk

in T

empe

ratu

re (º

F)

Effect of Piccolo ConfigurationLeading Edge Skin Temperature

P45-00-M45 (R026 - CFD)P45-00-M45 (R026 - Experimental)P70-P10-M40 (R107 - CFD)P70-P10-M40 (R107 - Experimental)

Figure 43. Effect of piccolo configuration on wing LE skin temperature.


22

skin temperatures and bleed air system interior flow properties. Key test variables in the parametric study performed included external flow temperature, hot air temperature, hot air mass flow and piccolo hole circumferential placement. Key findings and conclusions from the simulation studies performed are as follows:

• Modeling simplifications of the icing tunnel wing model were incorporated in the computational models to reduce meshing and computational effort. The full-span model consisted of square piccolo jet holes and three-in-line piccolo jet hole pattern. The partial-span model had 4.88-in span and for this model the flow inside the piccolo tube was not computed.

• Grid resolution requirement was considerably lower for the full-span wing model than for the partial-span model. Five nodes were distributed across the width of each piccolo jet hole in the full-span model compared to eighteen in the partial-model.

• All simulations were conducted with the pressure-based segregated solver in the FLUENT code and the default first order discretization. The k-ω SST turbulence model was applied to compute the turbulent internal hot-air flow.

• Numerical results obtained with a full-span wing model showed good correlation with experimental data in terms of total piccolo hot air pressure and temperature along the piccolo centerline. For the low bleed mass flows investigated there was a significant reduction in total temperature along the piccolo span. The total air pressure, however, remained constant over the entire piccolo length.

• Piccolo flow properties were found to vary along the radial direction from the piccolo tube axis to the center of each piccolo jet hole. In addition, hot air flow properties varied across the jet hole plane. Leading edge temperatures obtained from computations performed with the partial-span model were in better agreement with the experimental data when averaged (across the jet hole) properties were specified at the piccolo jet orifices instead of centerline values. These properties were computed from the full-span model.

• Numerical results obtained with the partial-span wing model showed good correlation with experimental data in terms of wing leading edge skin temperatures. For the simulations performed, the percent difference between the experimental and computed temperature distributions over the forward, upper and lower surface regions of the LE ranged from 0.14% to 13.7%.

• Other parameters that had a measurable effect on the computational results obtained with the partial-span model included turbulence model, near-wall grid resolution, location of the external flow transition points, and variation in aluminum conductivity with temperature.

In summary, computed leading edge skin temperatures, piccolo centerline total temperatures and pressures were in good agreement with the experimental data.

References 1Papadakis, M., Wong, S.-H., Yeong, H.-W., Wong, S.-C., and Vu, G.T., “Experimental Investigation of a Bleed

Air Ice Protection System”, SAE paper 2007-01-3313, Sept. 2007. 2Papadakis, M., Wong, S.-H., Yeong, H.-W., Wong, S.-C., and Vu, G.T., “Icing Tunnel Experiments with a Hot

Air Anti-Icing System”, AIAA paper 2008-444, Jan. 2008 3Papadakis, M. and Wong, S-H.J., “Parametric Investigation of a Bleed Air Ice Protection System”, AIAA paper

2006-1013, Jan. 2006. 4FLUENT 6.3 User’s Guide, 2007, Fluent Incorporated, Lebanon, NH. 5GAMBIT 2.4 User’s Guide, 2002, Fluent Incorporated, Lebanon, NH. 6Georgiadis, N., Yoder, D., and Engblom, W., “Evaluation of Modified Two-Equation Turbulence Models for

Jet Flow Predictions”, AIAA paper 2006-490, Jan. 2006. 7Alvi, F. S., Ladd, J. A., and Bower, W. W., “Experimental and Computational Investigation of Supersonic

Impinging Jets”, AIAA Journal, Vol. 40, No. 4, April 2002. 8Franko, K. and Georgiadis, N., “Computational Investigation of Heated High-Speed Coaxial Jets”, AIAA paper

2004-2980, 2004. 9Zuckerman, N. and Lior, N., “Impingement Heat Transfer: Correlations and Numerical Modeling”, Journal of

Heat Transfer, Vol. 127, 2005, pp. 544-552. 10Peng, S-H., Eliasson, P., and Davidson, L., “Examination of the Shear Stress Transport Assumption with a

Low-Reynolds Number κ-ω Model for Aerodynamic Flows”, AIAA paper 2007-3864, Jun. 2007. 11Godin, P., Zingg, D.W., and Nelson, T.E., “High-lift Aerodynamic Computations with One- and Two-Equation

Turbulence Models”, AIAA paper 96-0567, Jan. 1996.

Date post:	14-Dec-2016
Category:	Documents
Upload:	alonso
View:	212 times
Download:	0 times

[American Institute of Aeronautics and Astronautics 1st AIAA Atmospheric and Space Environments...

Documents