the role of physical experiments in advancing the … role of physical experiments in advancing the...

The Role of Experiments in Advancing

the Design of Gas Turbine Compressors

The Role of Physical Experiments in Advancing the Design

of Aircraft Gas Turbine Compressors

Pratt & Whitney Aircraft Fellows Seminar

October 9, 2003

Tony Strazisar

NASA-Glenn Research Center



Comments on Experiments vs Tests

The Changing Roles between Analysis and Experiment

Examples of Impact from Experiments:

• Rotor blade geometry variations

• Rotor blade surface finish

• Rotor hub leakage

• Stator hub loss

• Rotor tip flow and stability



Experiments vs Tests

Experiment

• Designed to explore a particular problem or hypotheses

• Often performed in simplified geometry (low speed, single

stage, isolated blade row, etc)

• Often conducted at university, government, industrial

research labs

• Difficult to do in a product development environment

Test

• Designed to explore/document performance, validate design

• Performed on complex components (multistage, high speed)

• Conducted in industrial environment

• Essential component in a product development environment



Impact of Experiments on Compressor Design

Path 1:Experimental data ! improved CFD/analysis ! improved design

methods

Path 2:Experimental data ! improved understanding ! improved design/

fab methods



Analysis vs Experiment Which is the Horse and which is the Cart?

The 1980’s

Emerging CFD

• 2D ! 3D

• Inviscid ! Viscous

• Steady ! Unsteady

Emerging measurement

methods

• Holography

• LDV

The ability to calculate within blade rows created a drive that fueled

advancement of methods that could measure within blade rows



The 1990’s

LDV data sets DLR, NASA

• Rotors (steady)

• Stages (unsteady)

Refined CFD methods

• Turbulence models

• Grids & numerics

• Leakage models

The existence of high-quality, non-intrusive data sets enabled a

drive for continued refinement of CFD methods and best practices




1984

NASA Fan

Rotor 67

1994

NASA Core Compressor

Rotor 37

Data Requests

Known Publications

Organized CFD

Validation Efforts

28

14

32 during ASME exercise

100+ by 2003

28 (1998)

ASME 12 participants

AGARD WG26 11

ERCOFTAC 5 (as of 1997)

APACETT 7

Use of NASA LDV data for CFD development & assessment

1994 1996 1997

Nothing works! There is hope! Is there hope?

Charles Hirsch, ERCOFTAC comments




Horse + Cart = Team !

Analysis &

ComputationsExperiments

Modeling of Physics

Validation of Analysis

Understanding of Fluid & Thermal Physics

Accurate Predictive Capability



Rotor Blade Geometry Variations

Myths:

• All blade passages in a rotor have the same geometry

• Strain gauges are minimally intrusive



Rotor Blade Geometry Variations

Rotation

pitch

Re

lati

ve

Ma

ch

Nu

mb

er

0.8

1.6

1.2

Part-span damper manufacturing tolerance leads

to significant passage-to-passage shock variations

LDV data, PWA 1800 ft/sec fan



Rotor Blade Geometry VariationsImpact of strain gauge installation,

NASA fan rotor 67

Peak Efficiency

Near Stall

Re

lati

ve

Ma

ch

nu

mb

er

pitch

1.5

1.3

1.1

0.9

pitch

1.5

1.3

1.1

0.9Strain gauge

Strain gauged

passage



Rotor Blade Surface Finishan unplanned experiment

A lesson in the importance of leading edges

A take-away message for fabrication techniques



Rough paint coating

Metal blade

0.001 in.

Rotor Blade Surface FinishNASA Rotor 37, Utip = 1492 ft/sec

SS

PS

LE LE

JT8D fan Coated R370.1 in.




Baseline rotorSmooth-coated rotorRough-coated rotor

Ad

iab

ati

c E

ffic

ien

cy

1.9 2.0 2.1 2.20.7

0.8

0.9

1.0

Approximate

throttle line

Pressure ratio




Baseline

Smooth coating

Rough coating

% Rotor Chord

.6

.8

1.0

1.2

1.4

1.6

0-50 50 100

Mrel

A

A

bow shock passage shock

Relative Mach number along line A-A




0-100% PS

0-100% SS

I

2-10% PS

2-10% SS

D

20-30% SS

E F

2-100% PS

55-100% SS

0-10% PS

0-10% SS

G

0-10% PS

0-50% SS

H

18 19 20 21

2.2

2.1

2.0

1.9

I

F

G

H

Pre

ssu

re r

ati

o

Massflow (kg/s)




E

0 50100 50 100

H I

G

F F

D

4

6

8

10

2

0

Pressure Surface,

% chord

Suction Surface,

% chord

Red

ucti

on

in

to

tal

pre

ssu

re r

ati

o, %

Leading edge

rough

Leading edge

smooth



Rotor Hub Leakagean experiment driven by CFD assessment

The devil is in the details…

A lesson in the power of leakage flows

A take-away message for rotor seals

An example of flow feature modeling



Rotor Hub LeakageNASA Rotor 37, Utip = 1492 ft/sec

massflow

Pre

ssu

re R

ati

o

Aero probe dataCFD

Meas. uncertainty

2% error in Pressure Ratio

1.9 2.0 2.1 2.2 2.3 2.40

20

40

60

80

100

% s

pan

Rotor total pressure ratio

Measurement plane

Operating point



1.9 2.0 2.1 2.2 2.30

20

40

60

80

100

% s

pan

Rotor total pressure ratio

Baldwin-LomaxCMOTT k-!

Data


Turbulence model cannot

account for measured

rotor hub pressure deficit



Nominal Gap

Rotor

diskForward

centerbody

Gap=0.030 in.

Blade leading edge

Reduced Gap

Balsa

wood

Rotor

disk

Rotor Hub LeakageNASA Rotors 35 & 37, Utip = 1492 ft/sec

But…

The test rig contained a blind

upstream cavity Rotor hub chord

1.9 inches

Flow



0.0

20.0

40.0

60.0

80.0

100.0

% S

pan

1.82 1.86 1.90 1.94 1.98 2.02

Total Pressure Ratio

Reduced gap

Nominal gap

0.25% leakage

0.33% leakage


Experimental data confirms

sensitivity to hub leakage



100

80

60

40

20

0

% S

pan

1.8 2.0 2.2

Pressure ratio


0.9

0.8

1.2

1.1

1.0

1.2

1.1

1.00.8

0.8

0.9

0.7

PS

SS

Leading edge

Hub Gap

Pit

ch

wis

e d

ista

nce

Static

Pressure

CFD, no leakage

CFD, net leakage=0

Data



Leakage

flow




Rotor Hub Leakage

“Discourager”

seal



The devil is in the details…

A lesson in the power of leakage flows

A take-away message for stator labyrinth seals

An example of flow feature modeling

Stator Hub Lossa parametric study



Stator shroud leakage flow

NASA-Glenn Low Speed Axial Compressor

“No leakage” 0.2%

Baseline leakage 0.5%

Increased leakage 0.9%

Maximum leakage 1.3%

Stator Hub LossSensitivity of loss to clearance is comparable to that for rotor tip clearance

0.00 0.05 0.10 0.15 0.20 0.25

60

40

20

0

% s

pa

n f

rom

hu

b

Stator total pressure loss coefficient

Peak pressure

operating condition



Rotor Tip Flow & Stabilityexperiments lead the way

Understanding a detrimental flow feature does not

necessarily mean it can be eliminated by design

Understanding a detrimental flow feature can lead

to an engineering solution



3D Navier-Stokes Analysis

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

Mrel

estimated clearance vortex path.

blade

blade

Predicted vortextrajectory

Laser AnemometerMeasurements, 95% span

Rotor Tip Clearance Flow & StabilityThe rotor tip clearance vortex causes large total pressure loss and blockage

in the tip region, limiting the stable operating range of the rotor.



Rotor Tip Clearance Flow & Stability

Solution Pro Con

Reduce rotor tip loading Increased operating Reduced pressure

range ratio

Casing treatment Increased range !!Reduced efficiency

throughout the mission

!!Difficult to “turn off”

!!Increased temperature

at the tip

Discrete tip injection !!Increased range !!Additional plumbing

!!Can be “turned off” !!Rotor excitation

!!No temperature !!"#$%&#$!#''!()#*!%+#$

!!increase at the tip




3-12 injectors around circumference




NASA Rotor 35, Utip = 1492 ft/sec

Early experiments

were focused on

controlled unsteady

injection




Later experiments

focused on steady

injection and reducing

the amount of injected

flow required




Rotor Stator

Injector

InjectorAxial velocity contours

Full-annulus MSU Turbo

Simulation

12 injectors

36 rotors

46 stators



Summary

Experiments often play an indirect role in advancing turbomachinery

technology

Experimental data ! CFD validation ! Improved design methods

Experiments play a direct role when designed and executed to explore

specific fluid mechanic processes within turbomachinery

Experimental data ! improved understanding ! improved design/

fab methods

The ultimate return on investment is realized when experimental data,

numerical analysis, and modeling are used synergistically



Summary

“Having a wind tunnel

doesn’t tell you what to

do with it.

Having a computer

doesn’t tell you what to

do with it.

The insight about what to

do with a facility is the

important thing.

And it is still a human

insight.”

Wygnanski

Analysis &

ComputationsExperiments

Modeling of Physics

Validation of Analysis

Understanding of Fluid & Thermal Physics

Accurate Predictive Capability

the role of physical experiments in advancing the … role of physical experiments in advancing the...

Documents