Jet-Surface Interaction – High Aspect Ratio Nozzle Test ...

National Aeronautics and Space Administration!

www.nasa.gov

Jet-Surface Interaction – High Aspect Ratio Nozzle Test Nozzle Design and Preliminary Data

Cliff Brown * Vance Dippold **

NASA Glenn Research Center

October 20, 2015

1

* [email protected] ** [email protected]


www.nasa.gov

Jet-Surface Interaction (JSI) Noise Sources and Effects

•  Measured far-field noise includes: –  Jet-surface interaction noise sources –  Jet mixing noise (isolated) –  Shielding/Reflecting effect

•  Types of JSI noise sources –  Surface loading (“scrubbing”) noise –  Trailing edge (“scattering”) noise –  Surface vibration noise

2

StDj

1/12

Oct

ave

PSD

(dB

)

10-2 10-1 100 101

ShieldReflectIsolated

633273986859

5 dB

Shielding Effect

Jet / SurfaceInteraction Noise

Reflected Noise

JSI Source

Reflecting Effect

Shielding Effect

StDj

Jet Mixing Noise 1/

12 O

ctav

e PS

D (d

B)

Ground Observer


www.nasa.gov

Jet-Surface Interaction Noise Test Programs

3

JSI1044 (2015)

JSI-HAR (2015)

(TBD) (TBD)

Multi-Stream

Asp

ect R

atio

JSIT (2011-2013)

ERN(2013) /JSIT(2013)

JSI1044 (2015)

JSI1044 (2015)

* Covered by AATT and CST Projects


www.nasa.gov

Jet-Surface Interaction Noise Test Programs

4

JSI1044 (2015)

JSI-HAR (2015)

(TBD) (TBD)

Multi-Stream

Asp

ect R

atio

JSIT (2011-2013)

ERN(2013) /JSIT(2013)

JSI1044 (2015)

JSI1044 (2015)

* Covered by AATT and CST Projects


www.nasa.gov

Turbo-electric Distributed Propulsion Concept

•  32:1 aspect ratio slot •  Divided into 2:1 at exit •  Electric fan has low pressure ratio, low temperature ratio

–  Test conditions 1.2 ≤ NPR ≤ 1.86, TTR = 1

•  Aft deck extends (estimated) 1-4 slot heights downstream

5

Superconducting Turbogenerators

Asymmetric Flow Path

* Kim et. al., AIAA 2015-3805


www.nasa.gov

Nozzle Design for JSI-HAR Testing •  Problem specific to model scale testing

–  TeDP has other issues but each fan is round to approximately 2:1 aspect ratio •  Limited by flow rate and scale factor to 16:1 aspect ratio •  Must transition from round to rectangular

–  Low noise - minimize internal flow separations and exit shocks –  Uniform flow profile at exit –  Minimize nozzle length and weight

•  Allow parametric variations of septa/internal flow –  Rapid prototyped plastic inserts (not load bearing)

6


www.nasa.gov

First Efforts Using SUPIN •  SUPIN* is a parametric inlet design tool

–  Assume “backward” inlet is a nozzle •  Observations:

–  Lines not always smooth near inflow. –  Thick boundary layers and separation

along side walls as major axis spread. –  Normal shock along centerline

•  Greater control to parameterize nozzle designs required (SUPIN is not for nozzles!)

7 7

Weak normal shock at exit

Vortex pair at exit plane

Thick BL

24.26 in

* AIAA Paper 2012-0016


www.nasa.gov

CFD for Design Evaluation

•  Wind-US v4 used for all simulations presented here. –  General purpose, compressible Reynolds-Averaged Navier-Stokes solver –  SST turbulence model used –  Steady flow simulations, i.e. constant CFL number

•  Flow conditions: –  Quiescent Freestream: p∞=14.3 psi; M∞=0.01 –  NPR=1.861 → Mjet=0.98 (Ma=0.9) –  Unheated Jet: T0=529.64°R (TTR=1)

•  Grid: –  9 – 24.5 million cells

•  Simulations performed on NAS: –  5 Ivy Bridge nodes (20 cores/node) –  Converged solution < 60 hours

total wall time.

8

Extruded grid along wall

Vane lines


www.nasa.gov

Next Approach: Parameterized Nozzle Design •  Idea: Transition flow in segments rather than all

at once to gain greater control over design –  Created new code to generate flow lines

•  Example: Four segments 1.  Transition from circular to order 10 superellipse; grow

major axis to nozzle exit width via cubic polynomial; maximum divergence angle<33°; constant area

2.  Transition from order 10 superellipse to order 100 via exponential function; constant area

3.  Contract area to nozzle exit area using cubic polynomial for minor axis

4.  Constant area and shape to nozzle exit to accommodate septa inserts

•  Include capability to add turning vanes –  Minimize BL growth and flow separation by distributing

flow out to side walls as major axis grows. –  Modeled vanes with inviscid boundary condition

(infinitely thin, slip surface) for ease of gridding and improved run time during design evaluation stage

9

24.22 in

1 2 3 4


www.nasa.gov

Final Design: A16-10 •  Three segments:

1.  Transition from circular to order 10 superellipse; grow major axis to nozzle exit width via cubic polynomial; maximum divergence angle<33°; linear area contraction through segment 2 (80% of total)

2.  Transition from order 10 superellipse to order 100 via exponential function; continues linear area contraction from segment 1 (80% or total); constant major axis length

3.  Linear area contraction (20% of total) with constant major axis length and constant superellipse order; longer segment length (5.5 inches) to accommodate septa inserts

•  No turning vanes –  CFD showed turning vanes did not do much once

outer flow lines were refined –  CFD showed significant wakes from turning vanes –  Center vane retained for structural support

10

24 inches


www.nasa.gov

A16-10: Design Evaluation

•  Significant vorticity near corners •  Attached flow along outboard edge of major

axis (BL thickness still significant) •  No normal shocks at nozzle exit •  Continuous area contraction helps •  Significant wake from center vane (added

for structural support)

11

Axial Velocity

Vorticity


www.nasa.gov

Septa Inserts

•  Inserts are rapid prototyped using solid ABS plastic

•  Inserts sit in small recess •  Septa are airfoil shaped

–  NACA0003, chord = 6” –  R=1 mm leading edge fillet,

R=0.5 mm trailing edge fillet –  2 mm fillet at root and stem

•  Half airfoil at center vane around sheet metal

12

2:1 / 7 Septa

1:1 / 15 Septa


www.nasa.gov

6 8 10 12 1412

14

16

18

20

22

24

26

28

Flow Profile at Nozzle Exit (1)

•  2:1 / 7 septa insert installed for JSI-HAR but not in WIND-US

•  Total pressure measured 0.25” downstream of nozzle exit

•  No indication of vortex in JSI-HAR data

–  1 Hz averaged pressure data would not likely pick this up even if present

•  Flat profile between septa •  Losses slightly higher in JSI-HAR

data

13

P 0 (l

bf/in

2 )

Position (in)

Septa wake (no septa in CFD)

WIND-US JSI-HAR

Ma=0.9, Unheated

2:1 / 7 Septa


www.nasa.gov

Flow Profile at Nozzle Exit (2)

•  1:1 / 15 septa insert in both •  Total pressure measured 0.25”

downstream of nozzle exit •  More losses at nozzle edge in

JSI-HAR than predicted •  Deeper wake deficits in

SolidWorks result –  JSI-HAR probe may not be directly

behind septa

•  Reasonable comparison for approximately 2 hours invested in SolidWorks simulation

14

Z

p0[lb

f/in2

]

0 5 10 1512

14

16

18

20

22

24

26

28SWORKSAAPL

P 0 (l

bf/in

2 )

Position (in)

SolidWorks JSI-HAR

Ma=0.9, Unheated

* Thanks to Dennis Eck, AAPL Facility Engineer, for SolidWorks result

1:1 / 15 Septa


www.nasa.gov

102 103 104 10555

60

65

70

75

80

85

90

Far-Field Noise – 1:1 / 15 Septa

•  Spectra: 1-ft lossless at Θ=90º •  No surface •  Increased broadband noise on

minor axis •  High frequency tonal content

–  Strouhal shedding from septa

15

Ma=0.7, Unheated

* Bridges, AIAA 2015-3119

1/12

Oct

ave

PSD

(dB)

Frequency (Hz)


www.nasa.gov

102 103 104 10545

50

55

60

65

70

75102 103 104 10555

60

65

70

75

80

85

90

Far-Field Noise – 1:1 / 15 Septa

•  Spectra: 1-ft lossless at Θ=90º •  8:1 smaller scale but similar

septa width* •  Trends follow from 8:1 to 16:1

–  Increased broadband noise on minor axis

–  Stourhal shedding from septa gives high frequency tonal content

•  Planning additional septa for 16:1 to separate aspect ratio from septa effects

16

Ma=0.7, Unheated

* Bridges, AIAA 2015-3119

1/12

Oct

ave

PSD

(dB)

Frequency (Hz)

16:1

8:1


www.nasa.gov

102 103 104 10550

60

70

80

90

17

Ma=0.7, Unheated

Far-Field Noise – 1:1 / 15 Septa with Surface

•  Spectra: 1-ft lossless at Θ=90º •  JSI noise source increasing with

longer surfaces –  Follows previously observed trends

•  Shielding only at the higher frequencies –  Approximate scale factor 25:1

based on slot height (h) –  JSI noise very low frequency at

full-scale (acoustic loading) –  Shielding could benefit EPNL

when taken to full-scale 1/12

Oct

ave

PSD

(dB)

Frequency (Hz)

xE/h=0.8 xE/h=2 xE/h=4 Isolated


www.nasa.gov

Summary

•  A round-to-rectangular convergent nozzle with aspect ratio 16:1 was designed for acoustic measurements –  Minimized potential noise sources from: (1) internal flow separation and

(2) shock cells •  16:1 aspect ratio nozzle fabricated for testing

–  Inserts to simulate TeDP concept details (septa) rapid prototyped •  Pressure traverse at nozzle exit shows expected flow profile •  Preliminary analysis of noise data consistent with previous experiments

–  JSI noise source prominent at low frequencies –  Shielding at only the highest frequencies

•  Test on-going through October –  Baseline (no septa), 2:1 / 7 Septa inserts planned

18

* [email protected] ** [email protected]

Jet-Surface Interaction – High Aspect Ratio Nozzle Test ...

Documents