Upload
philip-sparks
View
223
Download
4
Tags:
Embed Size (px)
Citation preview
High-Power Density Target Design and Analyses for Accelerator Production of Isotopes
W. David PointerArgonne National LaboratoryNuclear Engineering Division
Workshop on Applications of High Intensity Proton AcceleratorsFermi National Accelerator LaboratoryOctober 20, 2009
Outline
Purpose Design Challenges Design Options Solid Target Concepts
– Lithium-cooled Tungsten Pebble Bed – Lithium-cooled Tungsten Plate
Liquid Target Concepts– Lead Bismuth Eutectic
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
2
Purpose
Develop a high-power accelerator target concept for use as the primary stage in a two stage isotope production target– Neutrons are generated in primary target stage
• Spallation of heavy metal target material by high energy charged particles – Protons, Deuterons, or Helium-3
– Neutrons are absorbed in second stage• Fissioning and decay processes produce desired isotopes in Uranium
Carbide target• High temperature (~2000 C) maintained in second stage to encourage
fission and decay products to diffuse out of second stage target
3
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Design Challenges
Primary stage target must meet a number of engineering constraints and challenges– Physical Constraints
• Must produce sufficient neutrons to power second stage isotope production target
• Must be small in size to maximize efficiency of neutron use– Ideally cylindrical, <5 cm in diameter and ~9cm long– Beam will be approximately 1 cm in diameter with uniform cross-section
– Structural Constraints• Must isolate coolant and target material from vacuum in beam line
– Must satisfy mechanical stress limits– Thermal Constraints
• Sufficient heat removal needed to maintain structure within acceptable limits
– Total beam power of ~400kW at 1 GeV» ~1/3 of power deposited in small target
• Must be thermally isolated from high temperature second stage
4
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Target material/coolant considerations
Solid Target Options– tungsten
• Good neutron yield• Good heat transfer and structural
characteristics in new target material• High melting point (3422 °C) • May be clad when in contact with
water or alkali metals– uranium alloys
• Better neutron yield than tungsten• Poorer heat transfer in new target
material• Lower melting point (1000-1400 °C)• Many alloys compatible with alkali
metals. Likely must be clad in water or air cooled systems
• Higher decay heat load and longer-lived decay heat load than tungsten target
Solid Target Coolant Options– Air
• Low heat capacity limits applicability– Water
• Low boiling point must account for two-phase flow
• Corrosion management– Lithium
• Excellent conductivity, but low heat capacity compared to other coolants
– Sodium• Better heat capacity than lithium
– Mercury• Power generation in coolant limits
applicability • Potential for two-phase and non-wetting
issues– Lead or Lead-Bismuth Eutectic
• Power generation in coolant limits applicability
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
5
Target material/coolant considerations
Liquid Target Options– mercury
• Good neutron yield• Good heat transfer characteristics• Low boiling point (357 °C) may
have two phase flows• Liquid at room temperature• Does not wet many materials well
careful surface preparation required– lead
• Better neutron yield per proton, but longer stopping distance than mercury
• Higher boiling point (1749 °C)• High melting point (327.46 °C)• Erodes structural materials in high
speed turbulent flows• Wets most structural materials
– Lead-Bismuth Eutectic• Similar neutron yield to pure lead• High boiling point (~1700 °C)• Low melting point (123.5 °C)• Erodes structural materials like lead• Wets most structural materials like
lead• Higher polonium production
higher decay heat load
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
6
Design Options
Liquid Lead-Bismuth Eutectic Target vs. Lithium-Cooled Solid Tungsten Target
Advantages Disadvantages
Lithium-Cooled Solid
Tungsten
Short stopping distance yields more neutrons in small target length
Most spallation products are contained within solid target material
High thermal conductivity and heat capacity in solid target
High thermal conductivity in coolant Potentially simplified target handling procedures
Small stopping distance leads to much higher power density
Coolant channels must be built into target, increasing target length
Higher activity in waste products, decay heat removal issues
Low heat capacity in coolant Oxygen content must be controlled to prevent
fire Interaction between liquid lithium and common
structural materials largely unknown Tungsten must be clad in stainless steel
Liquid Lead-Bismuth Eutectic
Higher neutron yield per incident proton Single coolant and target material No need for complex structure to accommodate
coolant channels Good thermal conductivity and heat capacity Very high boiling point Spallation products can be removed on-line using
cold trap technology No danger of fire with oxygen exposure
Longer stopping distance requires longer target Oxygen content must be carefully controlled to
limit corrosion Fluid velocities must be carefully controlled to
prevent erosion High density coolant requires more robust
structure Potentially more complex target handling
procedures7
Solid Tungsten Target Design
Tungsten Plate Target vs. Tungsten Pebble Bed Target
Advantages
Simple Construction Low Pressure Drop Separate Window
Cooling Channel
Disadvantages
Low Volume Fraction of Target Material
Plate Deformation Limits Target Life
Advantages
High Volume Fraction of Target Material
Annular Design Provides Thermal Isolation
Target Material Can Deform Freely
Disadvantages
More Complex Construction
Manufacture of Very Small Pebbles May Not Be Feasible
High Pressure Drop Integrated Window
Cooling8
Heat Transfer In Tungsten Pebble Bed Target Assumptions
– Use average heat generation rate within beam radius throughout target (130 kW)
• Total target power is over estimated• Peak target power is underestimated
– Neglect conduction between pebbles– Calculated pebble-averaged temperatures
• Does not account for hot spots at contact points– Coolant Inlet Temperature = 523 K
• ~50 K above melting point– Cladding Thickness = 0.1 mm
Consider three limiting cases to examine effects of pebble size on performance
• Constant centerline temperature• Constant pressure drop• Constant inlet fluid velocity
9
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
10
Pebble Bed – Constant Centerpoint Temperature
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 0.02 0.04 0.06 0.08 0.1
Pebble Radius (cm)
Inle
t V
elo
cit
y (
m/s
)
800 degrees C
600 degrees C
400 degrees C
0
25
50
75
100
125
150
0 0.02 0.04 0.06 0.08 0.1
Pebble Radius (cm)
Pre
ss
ure
Dro
p (
ps
i)
800 degrees C
600 degrees C
400 degrees C
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Pebble Bed – Constant Pressure Drop
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 0.02 0.04 0.06 0.08 0.1
Pebble Radius (cm)
Inle
t V
elo
cit
y (
m/s
)
400
450
500
550
600
650
700
750
800
0 0.02 0.04 0.06 0.08 0.1Pebble Radius (cm)
Av
era
ge
Te
mp
era
ture
(C
)
Pebble Centerpoint
Clad Inner Surface
Clad Outer Surface
Target Pressure Drop = 30 psi
11
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
12
Pebble Bed – Constant Inlet Velocity
0
100
200
300
400
500
600
700
800
900
1000
0 0.02 0.04 0.06 0.08 0.1
Pebble Radius (cm)
Pre
ss
ure
Dro
p (
ps
i)
0
100
200
300
400
500
600
700
800
900
1000
0 0.02 0.04 0.06 0.08 0.1
Pebble Radius (cm)
Ave
rag
e T
emp
erat
ure
(C
)
Pebble Centerpoint
Clad Inner Surface
Clad Outer Surface
V = 0.5 m/s
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Pebble Bed Target Concept Conclusions Pebble bed target could be developed to satisfy thermal
requirements Many manufacturing challenges
– Construction of steel clad tungsten pebbles• Diameter less than 1 mm
– Arrangement of small pebbles in target• Randomly distributed pebble beds introduce large uncertainties
– Neutron production– Thermal performance
Many development challenges– Analysis of peak temperatures
• Contact points between pebbles• Heat deposition distribution
– Analysis of mechanical behavior of pebble/clad Pursue plate-type target as primary option
13
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Tungsten Plate Target Concept
Easy to construct Plates are easily clad if
necessary Lower pressure drop than a
pebble bed
Less surface area available for heat removal
Lower target material volume fraction than a pebble bed
Must give special attention to beam window cooling
14
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Tungsten Plate Target Concept
Easy to construct Plates are easily clad if
necessary Lower pressure drop than a
pebble bed
Less surface area available for heat removal
Lower target material volume fraction than a pebble bed
Must give special attention to beam window cooling
15
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Heat Transfer In Tungsten Plate Target Assumptions
– Use average heat generation rate within beam radius throughout target
• Total target power is over estimated• Peak target power is underestimated
– Calculate plate-averaged temperatures• Does not account for radial temperature distribution
– Coolant Inlet Temperature = 523 K• ~50 K above melting point
– Cladding Thickness = 0.1 mm Consider effects of four parameters on performance
• Plate thickness• Inlet velocity• Coolant gap width• Clad thickness
16
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Plate Target – inlet velocity and plate thickness
0
200
400
600
800
1000
1200
1400
0 0.002 0.004 0.006 0.008 0.01
Plate Thickness (m)
Ou
ter
Cla
d S
urf
ace
Tem
per
atu
re (
K)
0.1 m/s0.5 m/s1.0 m/s1.5 m/s2.0 m/sLimit
0
500
1000
1500
2000
2500
3000
3500
4000
0 0.002 0.004 0.006 0.008 0.01
Plate Thickness (m)
Pla
te C
en
terl
ine
Te
mp
era
ture
(K
)
0.1 m/s0.5 m/s1.0 m/s1.5 m/s2.0 m/sLimit
Clad thickness = 0.1 mmGap width = 1.0 mm
17
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
18
Plate Target – gap width and plate thickness
0
500
1000
1500
2000
2500
0 0.002 0.004 0.006 0.008 0.01
Plate Thickness (m)
Cla
d O
ute
r S
urf
ace
Tem
per
atu
re (
K)
1 mm
1.5 mm
2 mm
2.5 mm
3.0 mm
limit
0
500
1000
1500
2000
2500
3000
3500
4000
0 0.002 0.004 0.006 0.008 0.01
Plate Thickness (m)
Pla
te C
en
terl
ine
Te
mp
era
ture
(K
)
1 mm1.5 mm2 mm2.5 mm3.0 mmlimit
Clad thickness = 0.1 mmInlet velocity = 1.0 m/s
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
19
Plate Target – clad thickness and plate thickness
0
200
400
600
800
1000
1200
1400
0 0.002 0.004 0.006 0.008 0.01
Plate Thickness (m)
Cla
d O
ute
r S
urf
ac
e T
em
pe
ratu
re (
K)
0.1 mm
0.5 mm
1.0 mm
1.5 mm
2.0 mm
surface
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0 0.002 0.004 0.006 0.008 0.01
Plate Thickness (m)
Cla
d O
ute
r S
urf
ac
e T
em
pe
ratu
re (
K) 0.1 mm
0.5 mm
1.0 mm
1.5 mm
2.0 mm
limit
Gap width = 1.0 mmInlet velocity = 1.0 m/s
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Channel width optimization for fixed coolant channel velocity
9.9880E-04
9.9900E-04
9.9920E-04
9.9940E-04
9.9960E-04
9.9980E-04
1.0000E-03
1.0002E-03
0 5 10 15 20 25
Gap Number
Ga
p W
idth
(m
)
Plate thickness = 3 mm21 plates
Channel Velocity = 1.0 m/sPressure Drop = ~800 Pa
20
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Computational Fluid Dynamic Studies
Use the commercial CFD code Star-CD Apply 3-dimensional CFD simulations for
– Confirmation of conclusions drawn from scaling studies– Evaluation of effects of radial conduction of heat in solid components and convection of
heat in the coolant away from the region heated by the particle beam– Approximation of localized peak temperatures
Modeling Strategy– Solid Target
• Consider tungsten plate cooling and beam window cooling separately• Consider one symmetric half of the target geometry
– Liquid Target• Consider a 10° wedge of the target geometry
Modeling assumptions– Uniform volumetric heat source in target materials
• Limited to region actually heated by the particle beam– Constant velocity condition at model inlet– Zero gradient condition at model outlet
21
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Solid Target Plate Cooling
Model– Approximately 300,000 computational
volume elements– Vin= 1 m/s
– Tin= 250 °C
Results– Channel velocities
• 0.1 < VC< 1.1 m/s
– Peak surface temperature• 485 °C
– Peak solid temperature• 527 °C
22
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Solid Target Plate Cooling
Model– Approximately 300,000 computational
volume elements– Vin= 1 m/s
– Tin= 250 °C
Results– Channel velocities
• 0.1 < VC< 1.1 m/s
– Peak surface temperature• 485 °C
– Peak solid temperature• 527 °C
23
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Solid Target Plate Cooling
Model– Approximately 300,000 computational
volume elements– Vin= 1 m/s
– Tin= 250 °C
Results– Channel velocities
• 0.1 < VC< 1.1 m/s
– Peak surface temperature• 485 °C
– Peak solid temperature• 527 °C
24
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Solid Target Window Cooling
Model– Approximately
125,000 computational volume elements
– One symmetric half of geometry
– Parametrically evaluate effect of stainless steel beam window thickness
Result– Limit thickness to no
more than 2.0 mm
25
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Solid Target Window Cooling
Model– Approximately
125,000 computational volume elements
– One symmetric half of geometry
– Parametrically evaluate effect of stainless steel beam window thickness
Result– Limit thickness to no
more than 2.0 mm
26
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Solid Target Window Cooling
Model– Approximately
125,000 computational volume elements
– One symmetric half of geometry
– Parametrically evaluate effect of stainless steel beam window thickness
Result– Limit thickness to no
more than 2.0 mm
(a) Adiabatic Surface
(b) Cross Section
CoolantFlow
27
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Beam Window Performance
0
500
1000
1500
2000
2500
0.1 0.3 0.5 0.7 0.9
Steel Power Density to Tungsten Power Density Ratio
Pe
ak
Win
do
w T
em
pe
ratu
re (
K)
Thickness = 3.5 mm
Thickness = 3.0 mm
Thickness = 2.0 mm
Thickness = 1.0 mm
Inlet Temperature
Temperature Limit
0
500
1000
1500
2000
2500
0 2 4 6 8 10
Window Cooling Channel Inlet Velocity (m/s)
Te
mp
era
ture
(K
)
Wetted Surface Temperature (K)
Peak Window Temperature (K)
Inlet Temperature
Temperature Limit
Tungsten plate target concept conclusions
Tungsten plate target concept can potentially satisfy thermal requirements– Plate thickness ≈ 3.0 mm– Clad thickness < 0.5 mm– Gap width ≈ 1.0 mm– Channel velocity ≈ 1.0 m/s
No real need to optimize gap width for target of this size
Need to consider realistic power distribution– Optimize plate thickness
29
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Liquid Lead-Bismuth Eutectic Target Concept High neutron yield per proton Coolant is target material
– No stress issues in target material Simple design
Need careful oxidation control Lower density → less target material per unit volume
– Nearly equivalent to lithium cooled tungsten plate concept
30
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
LBE target scaling studies
Assume uniform volumetric heat source in LBE Thermal analyses
– Target coolant temperature rise = 40 °C• Average inlet velocity = 2 m/s• Mass flow rate = 3.6 kg/s• Can be reduced to 20 °C
– Mass flow rate = 8.0 kg/s– Increase total target diameter from 4.0 to 4.5 cm
Stability analyses– Annular turning flows are inherently unstable
• Leads to flow induced vibration issues when using heavy liquid metal • Follow stability guidelines from Idelchik’s Handbook of Hydraulic
Resistance– Develop turning vane concept for leading edge of central flow baffle.
31
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
LBE target turning vane
32
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
LBE Target Cooling
Model– Approximately 35,000 computational volume elements– Vin= 2.0 m/s
– Tin = 250 °C
Results– Peak velocity occurs at inlet and outlet
• Implies good fairing design– Peak surface temperature
• 493 °C – Peak temperature
• 608 °C
Detailed physics analyses needed to provide enthalpy source distribution for further optimization
33
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Conclusions
The high power density neutron converter will drive the two-stage ISOL target Many design options were considered in the development of a conceptual design A lithium-cooled tungsten concept has been developed for application as the
neutron converter stage of a two-stage high-Z ISOL target A liquid LBE target is being developed as one alternate to the lithium cooled neutron
converter concept
34
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009
Questions?
35
Workshop on Applications of High Intensity Proton Accelerators, Fermilab, October 20, 2009