Update on Systems Modeling and Analyses

1
Update on Systems Modeling and Analyses
UCRL-PRES-212978
Wayne Meier LLNL

• HAPL Program Meeting
• LLNL
• June 20-21, 2005

Work performed under the auspices of the U. S.
Department of Energy by University of California
Lawrence Livermore National Laboratory under
Contract W-7405-Eng-48
2
Outline

• Driver cost and efficiency trades in support of
  DPSSL studies
• COE scaling with net power and sensitivity to
  assumptions for KrF
• Info from other studies on common subsystems

3
Target gain curves are a key part of the analyses
Direct-drive gain curves for different laser
wavelengths
Constant yield curves
Ref. John Perkins
4
Laser power consumption likely to be 20 of
gross power
Laser (DPSSL) efficiencies 9.5 at 2w 8.4 at
3w Power conversion eff. 45 System energy
mult. 1.08
5
Total capital cost (TCC) vs laser energy on
target based on Sombrero study scaling
Net power 1000 MWe Assumes laser total capital
cost 500/J (Orth study had 1871M/3.68MJ
504/J) TCC 1.94 ? Direct Capital Cost
6
Normalized COE and rep-rate vs laser energy

• TCC_laser 500/J
• Net power 1000 MWe
• Normalized to 3 MJ result
• using 2w gain curve

COE curve shows broad minimum 2.0 3.9 MJ
within 5 (25 5.2 Hz)
7
COE dependence on laser total capital cost
Net power 1000 MWe Laser eff. 9.5 2w gain
curve
8
COE dependence on laser efficiency
Net power 1000 MWe Laser TCC 500/J
9
Reducing laser cost and improving efficiency are
both important
A 2x decrease in laser efficiency is equivalent
to 82 increase in laser cost B 2x increase
in laser efficiency is equivalent to 27 decrease
in laser cost
Normalized to result at E 3 MJ, 2w gain
curve Pnet 1000 MWe TCC_laser 500/J Laser
eff. 9.5

• Sensitivity to cost is linear
• Sensitivity to efficiency decreases with
• increasing efficiency small beyond 10

10
Outline

• Driver cost and efficiency trades in support of
  DPSSL studies
• COE scaling with net power and sensitivity to
  assumptions for KrF
• Info from other studies on common subsystems

11
How does COE vary with net power and other
assumptions?

• These analyses also used old Sombrero chamber and
  BOP models
• Evaluated for different KrF gain curves
• KrF laser efficiency was fixed at 7.5 and
  independent of rep-rate (need this dependence for
  future models)
• Evaluated various scenarios for scaling with net
  power
• held rep-rate fixed (5 Hz)
• held target yield fixed (350 MJ)
• varied both RR and Y to get minimum COE at a
  given net power

12
Estimates of target gain for KrF vary widely
13
For all cases (fixed RR, fixed Y and optimum),
COE decreases rapidly with increasing net power
Schmitt-derated gain curve
Normalized to current base case for chamber
studies E 2.5 MJ, G 140 (Schmitt-derated
gain) Y 350 MJ RR 5 Hz Pnet 650
MWe (assumes 45 thermal conversion efficiency)

• Results at Pnet 650 MWe
• E, Y, RR, Norm COE - 1
• 2.2 MJ, 268 MJ, 6.78 Hz, -2.1
• 2.5 MJ, 350 MJ, 5.00 Hz, 0
• 2.5 MJ, 350 MJ, 5.00 Hz, 0

14
COE vs Net Power
Schmitt-high gain curve
Normalized to current base case for chamber
studies E 2.5 MJ, G 140 (Schmitt-derated
gain) Y 350 MJ RR 5 Hz Pnet 650
MWe (assumes 45 thermal conversion efficiency)

• Results at Pnet 650 MWe
• E, Y, RR, Norm COE - 1
• 1.94 MJ, 263 MJ, 6.69 Hz, -6.4
• 2.17 MJ, 341 MJ, 5.00 Hz, -4.8
• 2.19 MJ, 350 MJ, 4.85 Hz, -4.5

15
COE vs Net Power
Perkins KrF gain curve
Normalized to current base case for chamber
studies E 2.5 MJ, G 140 (Schmitt-derated
gain) Y 350 MJ RR 5 Hz Pnet 650
MWe (assumes 45 thermal conversion efficiency)

• Results at Pnet 650 MWe
• E, Y, RR, Norm COE - 1
• 2.33 MJ, 262 MJ, 7.12 Hz, 0.6
• 2.70 MJ, 356 MJ, 5.00 Hz, 2.9
• 2.68 MJ, 350 MJ, 5.10 Hz, 2.7

16
Rep-rates for fixed yield and optimal COE vs net
power
Schmitt-derated gain curve

• 5 Hz is below optimal over entire range
• This conclusion may change when driver cost and
  efficiency dependence on rep-rate are accounted
  for. Both will tend to favor lower rep-rates.

17
Optimum yield vs. net power
Schmitt-derated gain curve

• 350 MJ, our base case, is higher than optimal for
  Pnet lt1150 MWe
• This conclusion may change when driver cost and
  efficiency dependence on rep-rate are accounted
  for. Both will tend to favor lower rep-rates and
  higher yields.

18
Summary of fixed rep-rate results for different
gain curves

• RR 5 Hz fixed
• For fixed 650 MWe, COE varies from 0.952 to 1.029
• For fixed COE (1.00), net power varies from 590
  to 690 MWe

19
Summary of fixed yield results for different gain
curves

• Y 350 fixed
• For fixed 650 MWe, COE varies from 0.955 to 1.027
• For fixed COE (1.00), net power varies from 600
  to 680 MWe

20
Outline

• Driver cost and efficiency trades in support of
  DPSSL studies
• COE scaling with net power and sensitivity to
  assumptions for KrF
• Info from other studies on common subsystems

21
Cost info is being gathered from other studies in
order to improve laser IFE plant models
Refs Meier and C. W. von Rosenberg, Jr.,
"Economic Modeling and Parametric Studies for
SOMBRERO - A Laser-Driven IFE Power Plant,"
Fusion Technol., 121, 1552 (1992). F. Najmabadi
et al., Overview of the ARIES-RS reversed-shear
tokamak power plant study, Nucl Eng. Design,
38, 3 (1997) J. Delene, An Assessment of the
Economics of Future Electric Power Generation
Options and the Implications for Fusion,
ORNL-TM-1999-243 (1999)
22
Updated and improved models are needed for HAPL
systems modeling

• Drivers working on updates for different
  options
• Capital cost models including scaling as function
  of energy, rep-rate and key design parameters
  (number of beams, J/cm2, etc.)
• Driver efficiency as function of design choices
  (gain media, aperture size) and operating
  parameters (energy, rep-rate, etc.)
• OM costs (e.g. optics replacement) and
  dependencies
• Chamber typically estimated based on /kg of
  structures, breeding blanket, etc.
• Cost scaling as a function of radius (depends on
  yield, rep-rate) and design choices (breeder,
  coolant, structure)
• OM (periodic first wall replacement)
• Target Factory - based on proposed high
  production rate methods and required equipment
  and building sizes.
• Need to incorporate results from GA study for
  capital and OM
• BOP - everything else (buildings, heat transfer
  and power conversion)
• Based on fission plant cost estimates
• Need Models for combined liquid metal (e.g., Li)
  primary with Brayton (gas) cycle power conversion
  system

23
Back-ups
24
IFE power balance
Fusion Chamber
E driver energy
Driver h efficiency

RR Rep-rate
G Target gain M Multiplication factor
Pt Thermal power
Power Conversion e conversion efficiency Pg
gross power Pa auxiliary power
Pd Driver input power
Recirculating power fraction
Pd / Pg 1/(hGMe)
Pn Net electrical power
25
Some basic relationships
Pt ERRGM thermal power, MW RR pulse
repetition rate, Hz M overall energy
multiplication factor (due to neutron reactions),
1.08 Pg Pte gross electrical power, MWe e
thermal conversion efficiency, 0.45 Pn Pg -
Pa - Pd net electrical power, MWe Pa faPg
plant auxiliary power, MWe fa auxiliary power
fraction, 0.04 Pd ERR / h driver power,
MWe h driver efficiency
Pd / Pg 1 / hGMe Driver recirculating power
fraction Example h 10, G 100, M 1.08, e
45 Pd / Pg 21
26
Cost of electricity (COE)
COE Cost of electricity, /kWeh FCR Fixed
charge rate, 0.0966/yr TCC Total capital cost,
OM annual operations maintenance costs,
(function of plant power) F annual fuel cost,
106 D decommissioning charge, 0.05
/kWeh) 0.0876 (8760 h/yr) ? (0.001 kW/MW) ?
(0.01 /) Pn Net electric power, 1000 MWe CF
annual capacity factor, 0.75
Fusion plant COE is a useful figure of merit for
self-consistent design trades and optimization.
It is far less useful as a predictor of future
reality due to large uncertainties in
technologies and costing.