ATST Thermal Design21 Oct 2002

ATST Thermal Design21 Oct 2002

Dr. Nathan DalrympleSpace Vehicles Directorate

Air Force Research Laboratory

Dr. Nathan DalrympleSpace Vehicles Directorate

Air Force Research Laboratory

2

Problem: Seeing

T. Rimmele & BBSO

Other solar telescopes:•Helium backfill•Evacuated optics

ATST must be open-air.Surfaces must be individuallytemperature-controlled.

3

Approach

• Define thermal requirements

– Flow down from SRD, error budgets, interfaces

– Connect image quality to surface temperature with modeling and empirical correlations

– Continue refining until early 2003

• Explore concepts

– Examine prior work

– Assemble short list of concepts

– Model/analyze concepts, list pros/cons

– Can we meet requirements?

– Examine interfaces/trades

– Select baseline concept (~CoDR, spring 2003)

Two concurrent tasks:

4

Big Picture

Thermal Problem Areas

5

Thermal Concerns

20280M6

Absorbed heat (W)Absorbed flux (W/m2)surface

150,000300Exterior shell

22694M5

25271M4

272034M3

3060M2

1866 + 10,271 = 12,137

303,000Heat stop(reflective)

1382110M1

6

Heat Stop Seeing

See Beckers and Melnick [1994] and Zago [1995 & 1997]

Not very restrictive on ∆T. Shoot for ∆T = 10 – 20 °C.

7

Heat Stop Concepts

(LEST)

(CLEAR)

(R. Coulter)

(ATST baseline)

8

Jet Cooling Scheme

Array of normal jets gives very large heat transfer coefficient:h = 30 kW/m2-K or larger not unreasonable.

Large h allows coolant to be near ambient temperature—no complex control systems needed.

9

Heat Stop Design Curves

Jet-cooled cone: d jets = 3 mm, N jets = 40, L jets = 13.5 cm, water coolant

0

50

100

150

200

250

300

0 5 10 15 20 25 30 35 40

∆T (K)

flow

rate

(gp

m)

or f

rict

ion

head

(ft

)

0

5

10

15

pum

p po

wer

(hp

)

Q (gpm)

head (ft)

power (hp)

Looks like we can achieve our goalof 10 – 20 K ∆T with ~50 gpmand less than 1 hp

10

Heat Stop Summary

• Seeing contribution expected to be low, although it is unclear how small-scale plume turbulence affects AO system performance

• Baseline concept: jet-cooled reflective cone

• 10 – 20 K ∆T with 50 gpm water coolant near ambient temperature

• May implement plume suction with larger ∆T

• Shape influences scattered light

11

Primary Mirror Seeing

Convection Regimes for D = 4m,T = 270K

0

1

2

3

4

5

0 1 2 3 4

V (m/s)

∆T(K

)

Mixed Convection

Nat

ural

Con

vect

ion

Fr =

10

Fr =

1.0

Fr =

0.1

Forced Convection

(a) Natural Convection

(b) Mixed Convection

(c) Forced Convection

Thin layers = good seeing.We want to be in forced convection:

low ∆T, high V.

GrRe

gLV

Fr22

=∆

≡ρ

ρ

12

M1 Thermal Requirements

Composite 4m mirror seeing estimateRacine [1991] used for natural convection; Zago [1995] used for mixed convection;

Gilbert et al. [1993] used for forced convection

0.00

0.05

0.10

0.15

0.20

0 1 2 3 4 5 6 7 8

V (m/s)

mirr

or s

eein

g (a

rcse

c)

0.2 K

0.5 K1.0 K

2.0 K5.0 K

GEMINI (0.2 K)

M1 error budget = 0.15 arcsec

13

M1 Thermal Models

h1

h2

q"abs (t).

q"rad.

ρ1,k1,cp1,T1(t)

ρ2,k2,cp2,T2(t)

ρ,k,c,T(r,t)

Tr(t)

M1

cold plate

•1D finite-difference on spreadsheet•3D FEM package (TMG)

Would like to use simplest, yet stillphysically realistic model.

14

1D Model Validation

Validation Case 7: Frontside solar loading, backside radiative cooling, convection both sides; h = 5 W/m^2-K, h r = 4.49 W/m2-K

-5

0

5

10

15

20

0 6 12 18 24

t (hours)

T (

K)

1D, back side1D, middle1D, front side3D, front side, center

1D and 3D models agree to within 0.3 °C.

15

M1 surface-air temperature excess for 100 mm thick ULE

-1.5

-1.0

-0.5

0.0

0.5

1.0

0 3 6 9 12

t (hours)

∆T (

K)

V = 0V = 5 m/sV = 10 m/sV = 19.2 m/s

M1 Thermal Model Results

Input Profiles

-30

-25

-20

-15

-10

- 5

0

5

10

15

20

0 6 12 18 24

t (hours)

T -

Ti (

K)

0

50

100

150

abso

rbed

so

lar

load

(W

/m2)

u1 (K)u2 (K)uR (K)qabs (W/m^2)

100 mm thick ULE

Results: •< 1 K ∆T over most of the day.•Mirror flushing assists temperature control.

16

M1 Thermal Model Results II

80 mm thick ULEInput Profiles

-20

-10

0

10

20

0 6 12 18 24

t (hours)

T -

Ti (

K)

0

50

100

150ab

sorb

ed s

olar

load

(W

/m2)

u1 (K)u2 (K)qabs (W/m^2)

Result: < 1 K ∆T over most of the dayEasier cooling than 100 mm case

M1 surface temperature excess, 80 mm thick ULE

-1.5

-1.0

-0.5

0.0

0.5

1.0

0 3 6 9 12

t (hours)

∆T (

K)

V = 0V = 5 m/sV = 10 m/sV = 19.2 m/s

17

M1 surface temperature excess, 200 mm thick ULE

-1.5

-1.0

-0.5

0.0

0.5

1.0

0 3 6 9 12

t (hours)

∆T (

K)

V = 0V = 5 m/sV = 10 m/sV = 19.2 m/s

M1 Thermal Model Results III

200 mm thick ULEInput Profiles

-50

-40

-30

-20

-10

0

10

20

0 6 12 18 24

t (hours)

T -

Ti (

K)

0

50

100

150

abso

rbed

sol

ar lo

ad (

W/m

2)

u1 (K)u2 (K)qabs (W/m^2)

Result: < 1 K ∆T over most of the day.Cooling more difficult.

18

M1 Summary

• Surface temperature requirement is a strong function of wind speed:

– ∆T < 0.5 K for V < 0.5 m/s

– ∆T < 1 K for 0.5 < V < 2 m/s

– ∆T < 2 K for V > 2 m/s

• ∆T < 1 K is achievable with 80 – 200 mm thick ULE mirrors

• Cooling difficulty increases with M1 thickness

• Wind flushing assists M1 temperature control

19

Enclosure Seeing

wind

ground layer uplift effect

natural convectionfrom heated shell

turbulentboundary

layer

hot air plume

turbulent shear layer

natural convectionfrom dome floor

apertureedge

vorticity

Ventilated Dome

breeze

internal b.l.'s

Variety of sources:•Shell•Ground layer•Internal nat. conv.•Shutter plume•Shear layer•Aperture edges

Passive ventilationreduces internal seeing sources

20

Other Concepts

wind

natural convectionfrom heated shell

turbulentboundary

layer

ground layer uplift effect

wind

ground layer uplift effect

apertureedge

vorticity

natural convectionfrom heated shell

hot air plume

turbulent shear layer

internalrecirculation

natural convectionfrom dome floor

turbulentboundary

layer

Retractable:•Eliminates many seeing sources•Still have shell seeing (reduced)

CLEAR concept:•Reduction in heated shell area•Snout may extend past shell layer•Internal ventilation more difficult

21

Shell Seeing

Convection Regimes for D = 20m, T = 270K

0

1

2

3

4

5

0 1 2 3 4

V (m/s)

∆T

(K)

Nat

ural

Con

vect

ion

Mixed Convection

Fr = 10

Fr =

1.0

Fr =

0.1

Forced Convection

Thin layers = good seeing.We want to be in forced convection:

low ∆T, high V. Dome shells veryoften in mixed convection.

GrRe

gLV

Fr22

=∆

≡ρ

ρ

22

Shell Seeing Requirement

0.00

0.10

0.20

0.30

0.40

0.50

0 1 2 3 4 5 6 7 8

V (m/s)

shel

l see

ing

(arc

sec)

0.2 K0.5 K

1.0 K

2.0 K

1.0 K (Ford)

2.0 K (Ford)0.5 K (Racine)

1.0 K (Racine)

2.0 K (Racine)

shell seeing budget 0.15 arcsec

23

Enclosure Summary

• Major trade in process

• All concepts suffer from shell seeing

• Wide spread in shell temperature requirement:

1 – 8 K ∆T for 0.15 arcsec, depending on source

• Guess: ~2 K sun-facing ∆T will be adequate (must refine)

• Must provide for internal flushing

• Performance: future work, must model/measure

24

Future Work (to CoDR)

• For every surface:

– Refine requirements

– Define thermal concepts

– Model/measure to predict performance

– Define interfaces/trades

• Complete major trades, e.g.

– Enclosure style

– M1, M2, etc thickness/material/thermal control

25

Discussion