1

Improving Durability of Turbine Components Through Trenched Film Cooling

and Contoured Endwalls

Principal Investigator:

Prof. David G. Bogard University of Texas at Austin

Co-Principal Investigator:

Prof. Karen A. Thole Pennsylvania State University

Graduate Research Assistant: Amy Mensch

Background image: [Hamed, A., Tabakoff, W., and Wenglarz, R., 2006]

DOE Award Number DE-FE0005540 UTSR Project Number 07-01-SR127

UTSR Workshop, October 20-22, 2014

PSU Completed Milestones: DOE Award DE-FE0005540, UTSR Project 07-01-SR127

Measure Endwall Overall Effectiveness Completed Q8

Measure Endwall Overall Effectiveness with Deposits Completed Q10

Measure Endwall Overall Effectiveness with TBC Completed Q11

Computational Predictions of Conjugate Heat Transfer, with and without TBC

Completed Q12

Measure Overall Effectiveness with Optimized Endwall Design (Contoured)

Completed Q13

Measure Contoured Endwall Overall Effectiveness with TBC

Completed Q14

Measure Velocity Fields with and without Film Cooling Completed Q14

Computational Predictions of Contoured Endwall Conjugate Heat transfer, with and without TBC

Completed Q16

Supplementary Tasks (not in original proposal)

Fla

t E

nd

wa

ll

Co

nto

ure

d

2

3

Better understanding of endwall cooling and its interaction with endwall contouring is needed to predict performance

air

ax

k

hCNu

Heat Transfer

Coefficients

Metal

Temp.

Design

Codes

exitc,

aw

TT

TTη

Adiabatic Wall

Temperature

Praisner et al. [2007] – Pack-B contour

Lynch et al. [2011]

Conjugate heat transfer measurements and predictions of flat and contoured endwalls will be presented

Contoured Endwall Effectiveness and Flow Measurements

Endwall Effectiveness with TBC

Overall Effectiveness with Deposition

Flat Endwall Overall Effectiveness

4

φTBC

φ

5

Matched Parameters

Typical Engine Model

Re∞,in (Cax) 1.25 x 105 1.25 x 105

h∞/hi 1 0.5 - 2.3

M = (ρcUc/ρ∞U∞) 1 – 2 0.6, 1, 2

Bi∞ = h∞t/kw 0.27 0.30 - 0.77

Overall effectiveness (metal temperature)

Matching the geometry, Biot number and h∞/hi to engine conditions allows direct measurement of metal temperature

geometryM,Re,fχη

Blade

tw

Tc,in

T∞ φ, Tw

Impingement plate

Corian Endwall, kw

hi

h∞

χη, Tfilm χηhhBi1

χη-1

TT

TTφ

iinc,

w

6

Mainstream flow is heated, and coolant flow is chilled to maximize driving ΔT = T∞-Tc,internal

Wind Tunnel Side View

Plenums Tc,in~14°C Heat Exchanger

Auxiliary

Chiller

Des

icca

nt

Dri

er

T∞ ~ 54°C

Chilled air section of wind tunnel

Coolant Blower

LFE

Test Section

Heated Mainstream

Flow

Coolant Flow

6

5

4

3

2

1

Test Section Top View

Corian®

Film and impingement cooling

Only

impingement

cooling

Only film cooling

7

endwall thermocouples

PS

x

y

SS D = 4.4 mm

Thermocouples were installed on the endwall surface under the TBC to measure φTBC along two streamlines

Film cooling holes, 30°

P/D = 4.7

S/D = 4.7

Top view of endwall within the passage

Incoming Flow Direction

Exit Flow

Direction

8

in-hole convection

exiting coolant

uniform high effectiveness

The measurements of overall effectiveness demonstrated the key features of film cooling and internal impingement

Mavg = 2.0 inc,

w

TT

TTφ

Combined Film and Impingement, φ

Film Cooling Only, φf

Impingement Cooling Only, φo

9

Increasing blowing ratio improved average φ for impingement more than for film cooling

Area

Averaged x

y

0.0

6

0.1

7

0.2

2

0.3

9

0.6

1

x/Cax =

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

-0.2 0.0 0.2 0.4 0.6 0.8 1.0

Lat. Avg. Film Cooling, f

Lat. Avg. Impingement, o

Lat. Avg. Film and Impingement, Lat. Avg. Film and Impingement, , repeat

x/Cax

Mavg = 1.0

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.5 1.0 1.5 2.0

Measured Area Avg. Film Cooling, f

Measured Area Avg. Impingement, o

Measured Area Avg. Film and Impingement, Predicted Area Avg. Film and Impingement,

Mavg

10

Conjugate RANS simulations used the SST k-ω model and an unstructured computational grid with wall prism layers

Grid

Surfaces Velocity inlet

Periodic boundaries

Blade

Plenum

Outflow y

x z

Conducting endwall

Adiabatic Adiabatic

TBC grid, 0.3 M cells

Endwall grid, 1.5 M cells

Impingement plate

Endwall

Blade

y z

FLUENT – SIMPLE RANS, SST k-ω, & energy – 2nd order Flow grid – 9.8 M cells y+ < 1 Flow and solid domains thermally coupled

11

There is good overall agreement between the measured and predicted φ, except for the attachment of the jets

inc,

wTBC

TT

TTφφ,

Measured Tw Predicted Tw

Mavg = 1.0 Mavg = 2.0

Measured Tw Predicted Tw

12

CFD temperature results show the three-dimensional conduction and steep gradients within the endwall

Velocity Magnitude inc,ΤΤ

ΤΤ

plenum

x/Cax = 0

1st row of impingement holes

x

y

viewing direction

y

z mainstream

plenum imp. plate

endwall y z

x/Cax = 0.09

2nd row of impingement holes

viewing direction

x y

Mavg = 1.0

Mavg = 2.0

13

We simulated deposition with wax, matching the Stokes number, Thermal Scaling Parameter and conductivity ratio

6.54Lμ 18

Udρ

scale time fluid

time relaxation particleStk

c

p2pp

0.3UL

ttTSP

i,

21

1220

1240

1260

1280

1300

1320

0.0001 0.001 0.01 0.1 1 10

10m Fly Ash

Fly

As

h P

art

icle

Te

mp

era

ture

(C

)

Tinitial

=

Tsolid

= 1260oC

Tgas

=

Particle Phase

t1 t2

i,UL

t

TSP < 1 Solid

TSP > 1 Molten

9.1 Cax upstream

0.069k

k

wall

deposit 0.11

k

k

Corian

wax

14

The cooling systems mitigated some deposition, but effectiveness was reduced everywhere in the passage

orinc,

w

TT

TTφ

inc,

wax

TT

TTω

Post deposition, w

No deposition, φ Post deposition, ω

Passage inlet

Mavg = 1.0

φ, Tw ω, Twax

tdep h∞

15

We observed clear areas due to the film cooling jets, and deposition on the blade from the passage vortex

intc,

fw,f

TT

TTφ

intc,

fwax,f

TT

TTω

Evidence of

passage vortex

Post deposition, wf

No deposition, f

Mavg = 1.0

16

Roughness from the deposition degrades the cooling performance, resulting in higher endwall temperatures

0.0

0.1

0.2

0.3

0.4

0.5

0.6

i, Mavg = 0.6

i,dep, Mavg = 0.6 w/ deposition

i, Mavg = 1.0

i,dep, Mavg = 1.0 w/ deposition

i, i,dep

Film Only Film & Impingement

Mavg = 0.6 Mavg = 1.0 Mavg = 0.6 Mavg = 1.0

i i,dep

h∞ increased by 30-100%

with deposition

17

0.0

0.1

0.2

0.3

0.4

0.5

0.0 2.0 4.0 6.0 8.0 10.0 12.0

meas, Mavg = 1

pred, Mavg = 1

meas, Mavg = 2

pred, Mavg = 2

NuD,i,pred, Mavg = 1

NuD,i,pred, Mavg = 2

0

3

6

9

12

15

H/D

NuD,i

Although the Nu peaks at H/D = 2.9, the area averaged φ, with film and impingement, is relatively insensitive to H/D

H/D = 0.6

H/D = 2.9 (nominal)

0.0

0.1

0.2

0.3

0.4

0.5

0.0 2.0 4.0 6.0 8.0 10.0 12.0

meas, Mavg = 1

pred, Mavg = 1

meas, Mavg = 2

pred, Mavg = 2

H/D

18

Contouring reduces effectiveness for impingement only, since hi decreases and h∞ increases from the flat endwall

0.0

0.1

0.2

0.3

0.4

0.5

-0.2 0.0 0.2 0.4 0.6 0.8 1.0

Flat, Mavg = 0.6Flat, Mavg = 1.0

Flat, Mavg = 2.0

Contoured, Mavg = 0.6Contoured, Mavg = 1.0

Contoured, Mavg = 2.0

o

x/Cax

Impingement Cooling Only

Flat Contoured

M = 2.0

Lynch et al. [2011]

flat

flatcontour

h

hh

19

0.0

0.1

0.2

0.3

0.4

0.5

-0.2 0.0 0.2 0.4 0.6 0.8 1.0

Flat, Mavg = 0.6

Flat, Mavg = 1.0

Flat, Mavg = 2.0

Contoured, Mavg = 0.6

Contoured, Mavg = 1.0

Contoured, Mavg = 2.0

x/Cax

Film and Impingement Cooling

Overall effectiveness does not change much between the flat and contoured endwall with film cooling

0

0.1

0.2

0.3

0.4

-0.2 0.0 0.2 0.4 0.6 0.8 1.0

f

x/Cax

Film Cooling Only Flat Contoured

Film Cooling

Only, M = 2.0

20

Although the average overall effectiveness is the same for the flat and contoured endwall, there are local differences

inc,

w

TT

TTφ

Flat Measured φ Flat Measured φ Contoured Measured φ

Contoured Measured φ

Mavg = 1.0 Mavg = 2.0

21

Other than film attachment, the contoured endwall simulations predict the same trends as the measurements

inc,

w

TT

TTφ

Measured Tw Measured Tw Predicted Tw Predicted Tw

Mavg = 1.0 Mavg = 2.0

22

The trailing edge flowfield was measured using a time resolved particle image velocimetry (PIV) system

Flowfield was sampled for 3000 or 6000 images at Δt = δ/U∞

U(t), V(t) t

t +Δt

Cross correlation

LaVision

Double pulsed laser

Blade

Light sheet Blade

flow

High speed camera

Image 1

Image 2

interrogation window = .035D ≈ 4δ

23

Plane C

Pla

ne

C

3D

5D U∞,in

y/C

ax

x/Cax

The trailing edge flowfield was measured for three vertical planes to capture the passage vortex development

1.08Cax

30°

Pla

ne

C

Mavg = 2.0

Plane A Plane B

Plane C

Endwall Endwall

Trai

ling

Edge

Trai

ling

Edge

Endwall

Trai

ling

Edge

Endwall Endwall Endwall

z = 0

z = 0

U/U∞,in

z = 0.32S

z = 0.32S

The passage vortex, indicated by the low velocity region, moves farther away from the wall with film cooling

No Film Cooling

24

Pla

ne

C

Mavg = 2.0

Plane A Plane B

Plane C

Endwall Endwall

Trai

ling

Edge

Trai

ling

Edge

Endwall

Trai

ling

Edge

Endwall Endwall Endwall

z = 0.32S

z = 0.32S

z = 0

z = 0

Contours of turbulent kinetic energy show two bands of peak tke, indicating the presence of two vorticies

25

TKE/U∞,in2

No Film Cooling

Matched Parameters

Typical Engine Model

Re∞,in (Cax) 1.25 x 105 1.25 x 105

h∞/hi 1 0.5 - 2.3

M = (ρcUc/ρ∞U∞) 1 – 2 0.6, 1, 2

Bi∞ = h∞t/kw 0.27 0.30 - 0.77

0.6–9.3 2.5

χη

1RRBi

hhBi1

χη-1

TT

TTφ

wTBC

iinc,

wTBC

Overall effectiveness with TBC

To accurately quantify the thermal effect of TBC, the thermal resistance was scaled to match the engine

TBCw

wTBC

w

TBC

kt

kt

R

R

Blade

tw

Tc,in

T∞ φTBC, Tw

Impingement plate

τ, TTBC TBC

Corian Endwall, kw

hi

h∞

inc,

TBC

TT

TTτ

geometryM,,Refχη

TBC effectiveness

cork

26

27

The conjugate simulations predict significant and uniform cooling with TBC for the flat endwall

Mavg = 1.0

inc,

wTBC

TT

TTφφ,

With TBC - Predicted Tw

Mavg = 2.0

No TBC -Predicted Tw

0.19φflat 0.32φflat

0.22φflat 0.39φflat

28

The conjugate simulations predict similar improvements with TBC for the contoured endwall with small differences

Mavg = 1.0

inc,

wTBC

TT

TTφφ,

With TBC - Predicted Tw

Mavg = 2.0

No TBC -Predicted Tw

0.19φflat 0.33φflat

0.38φflat 0.22φflat

29

-0.5 0.0 0.5 1.0 1.5

Flat TBC Flat Contoured TBC Contoured

s/Cax

SS, Mavg = 1.0

-0.5 0.0 0.5 1.0 1.5

s/Cax

SS, Mavg = 2.0

0.0

0.2

0.4

0.6

0.8

1.0

-0.5 0.0 0.5 1.0 1.5

,

TBC

s/Cax

SS, Mavg = 0.6

Endwall contouring measurements along the streamlines are similar to the flat endwall, with and without TBC

PS

SS

30

Adding TBC improves φ more than an increase in blowing ratio because TBC reduces heat transfer

0.05ΔφM

0.15φφΔφ TBCTBC w

TBCw,wr

q

qqΔq

Net Heat Flux Reduction – outer endwall surface

also increases with M

x

y PS

SS

TBCΔφ

0

0.05

0.1

0.15

0.2

0.25

0.3

Measured flat improvement w/ TBC

Predicted flat improvement w/ TBC

Predicted flat qr w/ TBC

Measured cont improvement w/ TBC

Predicted cont improvement w/ TBC

Measured contoured qr w/ TBC

0

0.1

0.2

0.3

0.4

0.5

0.5 1.0 1.5 2.0

TBC

Mavg

qr

0

0.05

0.1

0.15

0.2

0.25

0.3

Measured flat improvement w/ TBC

Predicted flat improvement w/ TBC

Predicted flat qr w/ TBC

0

0.1

0.2

0.3

0.4

0.5

0.5 1.0 1.5 2.0

TBC

Mavg

qr

31

Flat Measured, Mavg = 2.0

TBC temperature is less affected by the internal cooling and more affected by the film cooling performance

inc,

TBC

TT

TTτ

Flat Measured, Mavg = 1.0

TBC

Tc,in

T∞

τ , TTBC

endwall

Contoured Measured, Mavg = 2.0

Contoured Measured, Mavg = 1.0

32

This study demonstrates conjugate heat transfer trends for gas turbine endwalls and the secondary flow effects

Measured φ

Good agreement between conjugate measurements and CFD predictions

Predicted φ

Demonstrate trends for:

Cork - TBC

Contouring

Deposition

Unique flowfield measurements demonstrate interactions between passage vortex and film cooling

Temperature gradients Turbulent

Kinetic

Energy

with TBC with film

cooling