Proton Driver Options Cost Comparison

1
Proton Driver OptionsCost Comparison

• Rich Stanek
• March 15 , 2005

2
Outline

• Goal of the Comparison
• History of the Numbers
• Methodology Cost Estimating Approach
• Results Comparing the Estimates
• Cost Drivers
• Observations Contingency Range of Values
• Conclusions

3
Goal of the Comparison

• Provide a fair and consistent cost comparison
  between the Linac and Synchrotron PD options
• Independent of technical or physics related
  issues
• Consistent level of estimation and detail for
  each option
• Use the same labor rates, /sf for buildings and
  data format/cost roll up
• Assemble a reasonable level of back up
  documentation (Basis of Estimate) for both cost
  estimates
• Understand the Range of Values for the cost of
  a PD
• Provide the starting point/basis for the final PD
  cost estimate
• NOTE Costs are shown in FY04 dollars and
  without GA charges unless specifically noted
• GA and Contingency are added at the highest
  level

4
History of the Numbers (Synchrotron)

• PD I Study done for a 16 GeV Proton Driver (2000)
• PD II Study reduced the energy to 8 GeV and added
  a SC Linac option (2002)
• Much of the PD II Sync cost estimate was scaled
  from the original PD I numbers but then adjusted
  to include a replacement of the existing Front
  End Linac (Directors Charge Letter)
• Familiarity of FNAL personnel with building
  synchrotron machines and how things scale, helped
  this estimate maintain its accuracy
• The Synchrotron cost estimate has not had the
  same level of recent cost peer review nor working
  group participation as has the Linac cost
  estimate
• (Synchrotron technical design has been reviewed)

5
History of the Numbers (Linac)

• Linac options initial cost estimate relied
  heavily on scaling from TESLA and SNS numbers
  with appropriate modification for technical
  differences
• As a core project team was formed, the Linac
  estimate was further refined and developed by
  forming Working Groups and performing cost
  exercises (ex. Civil Working Group gt meets
  weekly and cryomodule cost estimate exercise gt
  resulted in RFQ)
• Experience based on FNAL/TESLA/SNS actuals was
  used where applicable

6
Cost Estimating Approach

• Take a conservative approach to estimating the
  costs of the PD (defendable numbers)
• Cost estimators started out with the knowledge
  that an overall contingency was going to be added
  to the entire project cost (30 across the
  board)gt some subsystems added scope contingency
  as part of their input estimates
• Cost estimates cover components up to and
  including injection into the Main Injector.
• Cost of the Main Injector Upgrades (ex. RF
  Upgrades) was not included in either cost
  estimate
• Assumption is that as the PD project moves
  forward and detail designs are further developed
  it will go through a Value Engineering stage
  where technical/cost trade-offs will be evaluated
  and a true contingency and risk matrix will be
  adopted

7
Maintaining the Accuracy

• Each option had a primary contact
• Synchrotron gt Weiren Chou
• Linac gt Bill Foster
• This allowed for a reasonable level of checks
  and balances in the comparison effort
• Tried to use the same people to make estimates
  for both systems (where applicable)

8
Methodology

• Evaluate documentation that was compiled for the
  PD II Study gt the starting point
• Put both estimates into a common form (Excel)
  using common costs where possible (labor rates
  based on TD actual averages and BTeV values)
• Improve the accuracy of the estimates
• Look for holes/missing items
• Look for obvious errors in the estimates
• Look for inconsistencies between the two options
• gt Led to new estimates for several key
  components
• Compare the two options (escalate old est. to
  2004)
• Perform reality checks internal review

9
Cost Comparison - High Level
Most of the initial installation work will be
Davis Bacon and shows up as MS costs GA rate
will be less on large purchasing actions
10
Comparing the Estimates

• Civil Construction
• Both use the same format of estimation (FESS
  std.)
• Linac is an all new estimate gt based on new
  location
• Includes more detail in terms of exact
  routing/requirements
• Synchrotron is a mixture of new/old estimates
• Old estimates (2002) were escalated to 2004
  (using DOE Guidelines)
• Use the same /sf for new/similar buildings
• Both use construction techniques very familiar to
  FNAL
• Site choices made to allow ease of accelerator
  ops and future growth
• Linac option has had external consultant look at
  feeder req.
• Sync placement is in close proximity to existing
  infrastructure
• Has both positive and negative implications
• Should make it easier to tap into existing
  utilities
• May have more of an impact on current operations

11
Comparing the Estimates (contd)

• Linac/Front End Improvements
• Both only have high level estimates
• Based on TESLA, SNS, FNAL Linac Upgrade or vendor
  communication
• Magnets (Main Synchrotron and all transport
  lines)
• Both Based on FNAL/TD estimates (using std.
  methodology which has been accurate in the past)
  gt use current steel and conductor prices
• Sync gt RD of rapid cycling, large aperture
  magnets not included
• Both assume existing MTF (conventional) is fully
  operational
• Utilities
• Linac gt new estimate based on conceptual design
  FNAL experience
• Sync gt scaled from MI actuals/modified using
  FNAL experience
• Power Supplies RF Equipment
• Both based on FNAL experience, purchase orders
  and vendor quotes
• Linac/Klystrons gt uses costs from recent TESLA
  purchase orders

12
Comparing the Estimates (contd)

• Cryomodules (Linac)
• Based on vendor RFQ (55 of MS est.) and
  TESLA/SNS experience
• Assembly estimates use TESLA LHC Project
  actuals
• Assumes CM assembly infrastructure in place
  (synergy with SMTF/ILC)
• Modulators Pulse Transformers (Linac)
• Based on actuals for equipment supplied by FNAL
  to TESLA (TTF)
• Modified to take into account quantity/price
  reduction
• Instrumentation Controls
• Sync estimated down to individual components
• Linac contains estimates at a high level
  (beginning to get more detail)
• Infrastructure Integration
• Often neglected when estimating individual
  subsystems
• Both used equal estimates (high level) where
  applicable
• Most of the initial installation will be
  Davis-Bacon work

13
Comparing the Estimates (contd)

• RD Costs
• Linac cost estimate assumes that the
  infrastructure to test cavities and build
  cryomodules will be in place either at FNAL or at
  other collaborating Labs (ex. LANL/ANL for spoke
  cavities) at the start of the Project gt synergy
  with SMTF and ILC RD project
• Therefore the cost to set up these facilities is
  not included
• Linac effort will continue to take advantage of
  ongoing work at SNS, TESLA TTF and other SCRF
  facilities
• Synchrotron assumes the RD to develop rapid
  cycling, wide aperture magnets and power supplies
  will be completed before the start of the Project
• gt Therefore the cost to perform this work is not
  included

14
PD Option Cost Comparison
15
Cost Drivers

• Synchrotron K
• Civil Construction 64,710 (26.5)
• Magnet System 40,841 (16.7)
• Linac Improvements 34,564 (14.2)
• Power Supply Network 30,921 ( 12.7)
• Linac
• Cryomodules 101,608 (31.9)
• Civil Construction 81,168 (25.5)
• Cryogenic System 28,589 ( 9.0)
• RF Klystrons Distribution 25,171 ( 8.2)

16
Synchrotron Cost Drivers
17
Linac Cost Drivers
18
Ratio of SWF/MS
Includes contracted EDIA for Civil Engineering
Design
19
Reality Checks

• Main Injector Project vs. Linac PD Estimate
• 230M gt escalated to 2004 273M PD 319M
• Project Management gt 5.8 PD 7.8
• Civil/Total Costs 44.8 PD 25.4
• Technical Sys./Total Costs 49.4 PD 67
• SWF/MS Costs 22.6 PD 38.8
• SNS (estimate)
• SNS Linac cost 550-600M (1.5 MW power)

20
Observations

• There is a difference in the level of recent
  detail attention given to the two PD options
  and the approach taken when estimating. However,
  both options were initially given an across the
  board 30 contingency (for consistency reasons)
• This way of estimating contingency for the
  overall project is somewhat arbitrary and does
  not reflect either the technical risk of the
  individual components nor the conservatism of the
  estimates
• The actual contingency analysis will do just that
• Look at another methodology which focuses on a
  Range of Values for the cost of each option
  gt reflects the confidence/uncertainty of
  the estimate

21
Synchrotron Cost Estimate Analysis
22
Linac Cost Estimate Analysis
23
Range of Values
Linac
Sync
24
Conclusions

• A fair and consistent cost comparison was
  performed between the two Proton Driver Options
  (Synchrotron SC Linac)
• A Basis of Estimate was established for both cost
  estimates that draws on relevant experience with
  similar equipment both at FNAL and other Labs
• This cost estimate will be a solid starting point
  for additional cost and schedule work that will
  be performed as the PD project moves forward
• Range of Values for the cost estimate of a Proton
  Driver to be built at FNAL is approximately 400M
  to 500M (including contingency and GA)