Revolutionizing Turbine Cooling with

Micro-Architectures Enabled by Direct

Metal Laser Sintering
The Ohio State University
Aerospace Research Center
J.P. Bons, A. Ameri, J. Gregory, R. Prenter, A. Hossain
(4 Nov 2015 NETL UTSR Workshop)
1

Motivation
Turbine Cooling Where did we come from?

Motivation
Turbine Cooling Where did we come from?

Motivation
State-of-the-Art in Turbine Cooling Where are we now?

Motivation
State-of-the-Art in Turbine Cooling Where are we now?
Turbulated serpentine
internal cooling passages

(Gupta et al., 2012)

Double walled
impingement cooling.
Film Cooling.

Motivation
State-of-the-Art in Turbine Cooling Where are we now?

(from Bunker, IGTI2010)

Pin-fin arrays/cutback

Shaped film holes

(from Cunha and Chyu, 2006)

Motivation
Current Manufacturing Process Investment Casting

Lost wax

Ceramic cores
7

Motivation
Current Manufacturing Process Laser Drilling and plunge EDM

Critical Need
Topic #3 from the 2015 UTSR FOA: The key goal of this topic area is to support

advanced internal cooling strategies

the development of
including advanced impingement for airfoil cooling and advanced near wall
cooling techniques. .. The increased turbine inlet temperatures likely

required to achieve 65% combined cycle efficiency will further

increase turbine component heat loads, requiring even more advanced,

efficient, and effective cooling techniques. Therefore, research is needed in this

support manufacturers

topic area that can

as they design hot
gas path components with sufficient cooling capabilities.
Where will these advances come from

Direct Metal Laser Sintering

10

Direct Metal Laser Sintering

Can you fabricate a cooled turbine blade with DMLS?
DMLS Turbine Blade

Micro-Machining Process

Just how small can the features be?

11

Direct Metal Laser Sintering

QUOTES FROM UTSR WORKSHOP TUESDAY A.M.
To take advantage of additive manufacturing, you need to start with the
design. - Bill Brindley, Pratt & Whitney
DMLS enables novel designs. - Karl Wygant, Samsung Techwin America
Challenges become opportunities! David Teraji, Solar Turbines
Additive manufacturing moving from nicety to necessity. Boeing
Manufacturing as an enabler rather than as a burden. - Sanjay Sampath, Stony
Brook University
Ability to make macroscale parts with microscale features. Suman Das, Georgia
Tech
12

Objectives
Explore innovative cooling architectures enabled by additive
manufacturing techniques for improved cooling performance and
reduced coolant waste.
Leverage DMLS to better distribute coolant through microchannels,
as well as to integrate inherently unstable flow devices to enhance
internal and external heat transfer.
Demonstrate these technologies
1. at large scale and low speed.
2. at relevant Mach numbers in a high-speed cascade.
3. finally, at high speed and high temperature.
Complement experiments with CFD modeling to explore a broader
design space and extrapolate to more complex operating conditions.

13

Research Team
TEAM LEAD
Focus: Experimental
Fluid Mechanics and
Heat Transfer

Dr. Jeffrey Bons

Professor
Department of Mechanical and
Aerospace Engineering
Ohio State University
Columbus, OH

Co-PI
Focus: Computational
Fluid Dynamics and
Heat Transfer

Dr. Ali Ameri

Research Scientist
Department of Mechanical and
Aerospace Engineering
Ohio State University
Columbus, OH

Co-PI
Focus: Experimental
Fluid Mechanics,
Fluidic Oscillator
Development

Dr. Jim Gregory

Associate Professor
Department of Mechanical and
Aerospace Engineering
Ohio State University
Columbus, OH

Robin Prenter
PhD Candidate

Arif Hossain
PhD Candidate

14

Cooling Designs Enabled by DMLS

Blade is cooled from the
center YET only surface
needs cooling!

All film cooling holes fed from

the same reservoir YET not
all regions NEED the same
coolant flowrate!

Bunker (IGTI 2013) showed that skin cooling could yield

25% cooling flow reductions
50% thermal gradient (stress) reductions.
40% coolant savings.

15

Cooling Designs Enabled by DMLS

Microchannels provide
unparalleled coverage.
Bunker (IGTI 2013)

Lee and Vafai (IJHMT 1999) showed

microchannel cooling is superior to backside jet
impingement cooling
16

Cooling Designs Enabled by DMLS

Sweeping Fluidic Oscillators for flow control

and many other applications

17

Cooling Designs Enabled by DMLS

Sweeping Fluidic Oscillators

Thurman et al. (IGTI2015) experimentally studied application to film cooling.

Sweeping film cooling yields higher midpitch film effectiveness. More uniform coverage.

18

Cooling Designs Enabled by DMLS

Pulsed impingement cooling jet (Camci & Herr, JHT 1999)
- 40-60% heat transfer enhancement compared to
steady jets for x/d<30
- No external input required to produce oscillation

Pulsing Fluidic Oscillators (Gregory)

19

Potential Concerns with DMLS

Stimpson et al. (IGTI2015)
Microchannel array additive manufacturing.
Elevated roughness levels
High pressure drop for same heat transfer
augmentation
Natural roughness obviates need for ribs.

20

Innovative Cooling Designs

Reverse Film Cooling

IR Images of Surface Temperature (Ameri and Shyam, unpublished 2015)

Calculation vs. Experiment (Li et al., Energy and Power 2013)

21

Innovative Cooling Designs

Combine all promising technologies on single NGV.

(Bunker, IGTI 2013)

(Notional DMLS NGV Model)

22

Turbine Heat Transfer Facilities

For innovative concepts to be viable, must be
vetted in facilities that simulate the real
operating environment
Graduated complexity
Low speed, large scale
High speed, smaller scale
High speed, high temperature, small scale

23

Low Speed Large Scale Tunnel

Flat plate film cooling studies

Vane leading edge studies

IR Camera provides film effectivenes and heat transfer

PIV for velocity field and vorticity (shown below)

24

Transonic Turbine Cascade

Cascade operates in blow-down mode. High pressure supplied
from large high pressure reservoirs, exhausts to ambient (without
exit ejector).
Size: compromise between adequate resolution for flow
investigations and capacity of air supply system.
Maximum optical access
Modular construction to allow new
blade designs
Ejector allows reducing exit pressure
(Reynolds number)

25

Transonic Turbine Cascade

Adjustable tailboards to insure periodicity

Replaceable
endwall plate

Choked bar array in exit duct insures Mach number

distribution in the cascade independent of Reynolds
number
Inlet and outlet
Flow conditioned by screens and honeycomb: 4.4:1
contraction: inlet flow uniformity 1.5%. Tu = 1%
Tu augmented with upstream passive grids
Inlet and exit wall pressure taps
Traverse slot for wake surveys

Traverse Adjustable
tailboards
slot

pressure taps

Stagnation
Pressure and
temperature

Turbulence Grid

Choke bars array

Screens
Honeycomb

26

Turbine Reaction Flow Rig (TuRFR)

Natural gas burning

combustor rig
Combustor exit flow
accelerated in cone nozzle
Transition from circular to
annular sector
Real vane hardware
(industry supplied) installed
in annular cascade sector
Tt4 up to 1120C (2050F)
Inlet Mach number ~ 0.1
300,000 < Recex< 1,000,000
Adjustable inlet
temperature profiles
Adjustable inlet turbulence
profiles (through dilution
jets)
Film cooling from vane
casing and hub (density
ratio 1.6-2.0)

Vane Holder
View Section
Viewports
Transition
Piece

Sealing
System

Equilibration
Tube

Cone

Spool Piece

Steel Base

27

27

TuRFR II
Exhaust Duct

Traverse Probes

Test Unique
Nozzle Box
Viewing Arms
E-duct
Particulate
Manifold

Shroud Cooling
Bypass Air Ducts.
Insulated around.

Shroud Cooling
Air Supply
Combustion
Chamber

Max Gas Path Temp = 2700F

Max Coolant = 1100F

Burner

Refractory and
Steel Casing

Combustor Stand &

Leveling Adjustment

28

TuRFR I & II

Gas Skid

Damper
Blower

29

TuRFR II

30

Optical Access

CAMERA IN LOWER POSITION

CAMERA IN UPPER POSITION

31

PHASE 1: Concept Exploration

Use available literature to identify most promising cooling designs:

o Pulsed fluidic oscillators for internal cooling of leading and trailing edges
o Sweeping fluidic oscillators for external film cooling
o Reverse flow film cooling from microchannel circuits for pressure surface
o Microcooling circuits replace showerhead cooling
Low-speed wind tunnel testing with scaled geometry
o Characterize cooling effectiveness and heat transfer
o Test variants of geometry to determine optimum
o Test sensitivity of each design to manufacturing tolerances
Develop computational models of each cooling design
o Generate flow solutions for each initial geometry
o Validate solutions with experimental data from initial geometry
o Explore design space and aid in optimization of geometry for each design
Determine most promising and feasible technologies for Phase 2 based on
experimental and computational results
32

PHASE 2: Integrated SLA Vane

Implement most promising technologies into preliminary nozzle guide vane

design
Develop computational model of preliminary vane design in high-speed cascade
Generate flow solutions at various operating conditions
Modify preliminary vane design per computational results
Fabricate properly scaled plastic vanes with stereolithography (SLA) using
modified design
Test fabricated vanes in high-speed cascade
o Characterize flow and heat transfer at various operating conditions
o Determine compressibility effects
Validate flow solution using experimental data
Iterate back to low speed testing as necessary
Generate flow solutions for final Phase 3 design at higher inlet Mach numbers
and Reynolds numbers

33

PHASE 3: Fully Simulated NGV

Fabricate high-temperature alloy vane using DMLS (e.g. EWI)

Coat vane in thermal barrier coating (TBC)
Characterize surface roughness and tolerances due to manufacturing method
Test full material system in the TuRFR turbine test facility
o Characterize cooling performance and pressure drop at various coolant
mass flow rates
o Characterize cooling performance at various main flow conditions
Compare new vane design performance to conventional vane at same coolant
and main flow operating conditions to determine improvement
Develop computational model of coated NGV
o Generate and validate flow solution in context of TuRFR testing
o Generate simulations at higher temperatures and pressures not possible in
the facility

34

Motivation for CFD

CFD can be used to elucidate and complement
experimental results and to inform the flow
physics.
Allows for extrapolation of flow outside the
pressure and temperature limits of experiments.
Understand unsteady and rotating frame of
reference effects
Allows exploration of the broader design space to
find promising combinations of feasible variables
for the application.
CFD will be used at every stage of our research.
35

CFD Methods Utilized

CFD is a research tool not a goal.
The CFD, as much as possible, will be validated by
the experiments to ensure accuracy.
Our team has demonstrated capability to use
various CFD methods for solving fluid flow and
heat transfer problems relevant to gas turbine
flows.
Any of RANS, URANS, DES or LES will be used, as
needed, with structured or unstructured or
meshless methods.
Examples follow:
36

Capabilities
Tip Gap Modeling
Computed flow
traces and heat
transfer in a
turbine rotor
tip clearance
gap

Blade and Tip Heat Transfer

(Shyam and Ameri, 1998)

37

Capabilities
Film Cooled Heat Transfer Results

Blade film cooling

Three Dimensional filmcooled blade analysis

(Garg, 1999)

38

Capabilities
Internal Coolant Passage Modeling

computed
heat transfer
in internal passages

Grid

Internal Heat Transfer

(Rigby and Bunker, 2002)

39

Capabilities
Conjugate Heat Transfer
wall temperatures
Tw/To

Grid
Solution

Conjugate

Conjugate Heat Transfer

(Heidmann, Rigby and Ameri, 2003)

40

Capabilities
Hot Streak Clocking Study

One streak per vane doublet

Contours of total temperature in deg K

(Casaday, Ameri, and Bons, AIAA 2012)

41

Capabilities
Unsteady Particle Tracking in Rotating Frame

2 Vanes per 3 Blades

Contours of Absolute
and Relative Total
Pressure

(Zagnoli, Prenter, Ameri, and Bons, IGTI 2015)

42

Capabilities
Representative Mesh Topologies

(Prenter and Ameri, 2013 - 2015)

43

Efficient Cooling
Will seek to improve cooling by using methods that
are more coolant efficient.
Using fluidic devices can reduce blow off and the
sweeping action can improve spanwise uniformity .
Fluidic devices can be made to film cool by
sweeping or impingement cool by pulsing.
Reverse blowing may be an effective way of film
cooling at high blowing-ratios.
Internal micro-channels are shown to be capable of
being more effective than impingement cooling.
44

CFD - Conclusions
Validated CFD will be used, side by side, with
bench top and more physically realistic
configurations to extend the design space and
explore more realistic physical conditions.
We have the availability and have developed
the expertise and gained the experience to
perform such analyses using various steady
and unsteady CFD methods to fulfil this task.

45

Accomplishments to Date
Literature Search
CFD Study
Fluidic Oscillator development/preliminary
study
Reverse film cooling preliminary study

46

Preliminary CFD Studies

47

777 Hole A DES Simulation

Tool development and validation: shaped hole film cooling via DES

Choice of model and method is critical to

accurate predictions.
48

Fluidic Devices - Simulation

3D grid used for fluidic-oscillator
driven film hole computation

SNAPSHOTS OF MACH NUMBER IN

MID-PLANE OF FLUIDIC HOLE FROM
UNSTEADY 3D CFD AT BR=2.0
(BLUE=0, RED=0.45)

Thurman et al. IGTI2015

49

Preliminary Testing Oct. 2015

Reverse Film Cooling

50

Preliminary Testing Reverse Film Cooling

Test Conditions and Scaling Parameters:
, = 1.625m
= 1.75cm

= 4.5m/s
= 22C

Boundary layer:

= 1.4

=
= 1.38

5800

,
4.5 5
=

, =

,
4.5 5
(45% of total chord length)
=
6
,,
1

Suggests turbulent boundary layer

51

Preliminary Testing Reverse Film Cooling

Experimental Facility

Freestream conditions:

= 4.5m/s (can increase to 20m/s)

= 22C (can be heated up to 60C)

Coolant conditions:
=

= 0.5, 0.75, 1, 1.4

1.065

Infrared camera: Cedip SILVER 480M

320x256 pixel Indium Antimonide

(InSb) sensor
Accuracy: 1C
Sensitivity: 0.02C
Max frame rate: 400Hz
52

Preliminary Testing Reverse Film Cooling

Analysis (steady state test)
T T f

T Taw
=

T Tc
T Tc
=0
=1
min

max

T (Freestream temperature) - Measured with thermocouple

T (Coolant temperature) Measured at exit of film cooling holes with thermocouple.

c

Taw (Surface temperature) Measured using infrared camera.

53

Preliminary Testing Reverse Film Cooling

Analysis h (transient test)
2T

1 T
=
x 2 t

1D heat equation

T
T
= 2

x
(4t )

and

init

"

x =0

s ,cond

T ( ) T ( )

[ t t

(t ) =

"

rad

i =1 1
2

lim
T ( x, t ) = Tinit
x

T ( x, t ) T
x
= erf

T T
2 t

(Ts Tinit )

Duhamels Superposition

1/ 2

Boundary conditions (semi-infinite solid)

T ( x,0) = Tinit

dT
dx

qs",cond (t ) =

i 1

"
conv

Boundary condition (step

change in T_surface)

i 1

init

"
cond

=q +q
"

"

rad

conv

"
cond

= h(T T )

"

conv

T (0, t ) = T

h =
n

1 2
T ( ) T ( )

T T
t + t
n

sm

s m 1

i 1

m =1

sn

i 1

54

Preliminary Testing Reverse Film Cooling

Analysis Heat Load
q = h (T T )

Without film cooling

q = h(T T )

With film cooling

"

"

h (T T
q
=

q
h (T T

"

"
with . film . cooling

"

"

without . film . cooling

(
T Tc )
q" h
= 1
(T TS )
q0 h0

=
"

"

q
h
=
q h
0

q
< 1 Net cooling benefit
q
q
> 1 Net cooling loss
q

Overall cooling
effectiveness

0.6

"

"

"

55

Preliminary Testing
Results Conventional Blowing (P/D = 3, = 30)
M = 0.5

M = 0.75

M=1

M = 1.4

Main
flow

56

Preliminary Testing
Results Reversed Blowing (P/D = 3, = 30)
M = 0.5

M = 0.75

M=1

M = 1.4

Main
flow

57

Preliminary Testing
Results Comparison M = 0.5
1

0.4
Forward M=0.5
Conven.
Backward M=0.5
Reversed
Backward M=1
Reversed

0.9

0.8

Forward M=0.5
Conven.
Backward M=0.5
Reversed
Backward M=1
Reversed

0.35

0.3
0.7
0.25

0.6

0.5

0.2

0.4

0.15

0.3
0.1
0.2
0.05

0.1

0
1

x/D

Centerline Traces

-3

-2

-1

z/D

Lateral Traces at x/D = 4

Promising initial comparison, however:

h/h0 is likely augmented
Compare to more representative
geometries (e.g. 777 Hole)
Effects of higher FST levels
Effects of pressure gradient
58

Preliminary Testing
Oscillating Impingement Jets

59

Preliminary Testing Oscillating Jets

Initial Design:
Geometry from
literature (Raman et al.,
1999)
= = 4.5mm
AR = w/h = 0.82

1. CAD design

CNC Fabrication

Fabricated test article

from acrylic using CNC

60

Preliminary Testing Oscillating Jets

Fluidic Oscillator Characterization:
Determine frequency-mass flow
rate relationship
Place microphone outside of the
hydrodynamic field

Microphone

Use FFT to determine frequency

of oscillations at different mass
flow rates

Fluidic Oscillator

61

Preliminary Testing Oscillating Jets

Fluidic Oscillator Characterization:
Determine frequency-mass flow
rate relationship
Place microphone outside of the
hydrodynamic field
Use FFT to determine frequency
of oscillations at different mass
flow rates

62

Preliminary Testing Oscillating Jets

Fluidic Oscillator Characterization:

63

Preliminary Testing Oscillating Jets

Fluidic Oscillator Characterization:

Schlieren of oscillator operating at 430Hz

64

Preliminary Testing Oscillating Jets

Impingement Heat Transfer
Initial testing includes:

IR
Camera

Fluidic oscillator vs. straight jet

A single mass flow rate
x/D = 1.5 and 2.0

There are several parameters to

explore:

Jet aspect ratio

x/D
Frequency
Mass flow rate
Pressure drop

Quartz plate
x

Fluidic
oscillator
design

Black paint

65

Preliminary Testing Oscillating Jets

Mass flow rate: 2.96 g/s
y/D

f = 480 Hz

=

z/D

z/D

Stationary
Jet

y/D

For oscillating jets,

augmentation of h may
be the real advantage

Fluidic
Oscillator

z/D

y/D

Initial Results:

x/D = 2

y/D

x/D = 1.5

z/D

66

Preliminary Testing Oscillating Jets

Initial Results:

0.9

0.9

0.8

0.8

0.7

0.7

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3
Straight Jet x/D = 1.5

Straight Jet x/D = 1.5

0.2

0.2

Straight Jet x/D = 2

Straight Jet x/D = 2

Fluidic Osc. x/D = 1.5

Fluidic Osc. x/D = 1.5

0.1

0.1

Fluidic Osc. x/D = 2

Fluidic Osc. x/D = 2

0
-6

-4

-2

y/D

-5

-4

-3

-2

-1

z/D

67

Preliminary Testing Backward Blowing

Analysis Heat Load
q = h (T T )

Without film cooling

q = h(T T )

With film cooling

"

"

h (T T
q
=

q
h (T T

"

"
with . film . cooling

"

"

without . film . cooling

(
T Tc )
q" h
= 1
(T TS )
q0 h0

=
"

"

q
h
=
q h
0

q
< 1 Net cooling benefit
q
q
> 1 Net cooling loss
q

Overall cooling
effectiveness

0.6

"

"

"

68

Next Steps

69

Eventual Integration of Promising Designs

Reverse Cooling on PS:
- Fed by upstream microchannel
- Better surface coverage with lower
massflow?

Fluidic Oscillator Impingement

Cooling on LE:
- Eliminate showerhead
- Lower massflow required?
- Microchannel exhaust

Sweeping Fluidic Oscillator in TE:

- Improved coverage with lower
massflow required?
- Lower pressure drop than
pedastels?
Sweeping Fluidic Oscillator Film
Cooling:
- Improved coverage with lower
massflow required?

vs.
vs.
70

Manufacturing Challenges with DMLS

Surface finish could
Enable technology
naturally roughened microchannel

Or disable technology

Stimpson et al. (IGTI2015)

DMLS is still on a steep

learning curve.
Design for tomorrows
level of improved
geometric tolerance.
71

Gantt Chart
Year 1

Year 2

Year 3

Phase 1

- Literature Review

- Low speed model

testing
- CFD models
- Downselect for Phase
2 model

Phase 2

-- Incorporate designs
into NGV
- Model NGV in CFD
- Fabricate SLA model
- Test in transonic
cascade
- Iterate on design

Phase 3

- Fabricate DMLS NGV

with TBC
-Test DMLS NGV in
TuRFR
- Develop/validate CFD
model

72