NGAO: NGAO: Cost Comparison with First Cost Comparison with First Light TMT AO Light TMT AO Peter Wizinowich, Richard Dekany, Don Gavel Peter Wizinowich, Richard Dekany, Don Gavel with input & review by Brent Ellerbroek with input & review by Brent Ellerbroek SSC Meeting SSC Meeting November 3, 2008 November 3, 2008
22
Embed
NGAO: Cost Comparison with First Light TMT AO Peter Wizinowich, Richard Dekany, Don Gavel with input & review by Brent Ellerbroek SSC Meeting November.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
NGAO:NGAO:Cost Comparison with First Light TMT AOCost Comparison with First Light TMT AO
Peter Wizinowich, Richard Dekany, Don GavelPeter Wizinowich, Richard Dekany, Don Gavelwith input & review by Brent Ellerbroekwith input & review by Brent Ellerbroek
• SDR Reviewer comments leading to this cost comparison• NGAO SDR Cost Estimate & Methodology • Cost Comparison Goal, Products & Approach• High-level Summary of Results• Cost Comparison• Conclusion
Conclusion Preview:- We & TMT agree that NGAO is traceably less expensive than the 1st light TMT AO system & we understand why
- Larger NGAO contingency needed in some areas- Quote needed for laser & will be obtained soon
3
SDR Reviewer Comments
• “Based on the cost and schedule of past and planned projects of lower or similar complexity, the review panel believes that the NGAO project cost and schedule are not reliable and may not be realistic.
Contingencies are also too tight. In particular, the time of 18 months allocated for manufacturing and assembly and 6 months for integration and test, is probably optimistic by a large amount.”
• Relevant to this point they also said:– “The review panel believes that Keck Observatory has assembled an
NGAO team with the necessary past experience … needed to develop the Next Generation Adaptive Optics facility for Keck.”
– “The proposed schedule and budget estimate have been carried out with sound methodology”
• Clarification: Reviewers thought our lab and telescope I&T durations were smaller by 2x than our plan (they are 6 & 12 months, respectively).
4
Response to SDR Reviewer Comments
• “We will develop another level of cost estimation during preliminary design.” – During system design, we considered project costs composed of hundreds
of tasks and specific procurements
– Reviewers noted “the budget book showed some pages not complete or no estimates for the words in the task”. Our choice to produce cost sheets for each WBS in each of 4 project phases resulted in blank sheets
• e.g., no subsystem work during the delivery & commissioning phase
– During PD, we will revise the basis of estimate to move more items from engineering judgment and analogy to direct vendor quote basis
– Reviewers felt that the laser procurement needs more time & should start much earlier. We agree.
• Requested next level of cost information for TMT NFIRAOS project to better understand the reasons for the cost estimate differences– This brings us to this presentation …
5
Cost Estimation Methodology (KAON 546)
• Cost estimation spreadsheets– Based on TMT Cost Book approach, simplified for SD phase– Prepared for each WBS element (~75 in all)– Prepared for each of 4 phases
• Preliminary design, detailed design, full scale development, delivery/commissioning
– Prepared by technical experts responsible for deliverables– Process captures
• WBS dictionary• Major deliverables• Estimates of labor hours• Estimates of non-labor dollars (incl. tax & shipping) & travel dollars• Basis of estimate (e.g. vendor quote, CER, engineering judgment)• Contingency risk factors & estimates• Descope options
– Standard labor classes, labor rates & travel costs used
6
Cost Comparison Goal, Products & Approach• Goal of Comparison:
– Increase confidence in the NGAO cost estimate
• Products– Updated cost estimation data base to take advantage of comparison information
– Identify similarities & differences in effort & procurement estimates, & resolve/justify these differences
– As appropriate update the NGAO cost estimate
– Identify other support for NGAO estimates
– Respond to reviewer sense that “methodology good, but not fully executed”
• “NFIRAOS” used as shorthand for the full TMT 1st light AO configuration– Updates to TMT cost estimate since 2006 (globally < 5%) not incorporated
7
Cost Comparisons(in FY08 $ as reported at the SDR)
NGAO Cost Estimate Comparisons
0
10
20
30
40
50
60
70
80
90
100
NGW
FC
PALM3000
K2 LG
SGPI
GEMS
NGAO
NFIRAOS
FY
08 $
M
Upgrades1st Gen 8-10m AO
2nd Gen 8-10m AO
1st Gen 30m AO
8
Major Differences between NFIRAOS & NGAO• 30 m versus 10 m
– ~ 3x physical size of AO system– 3x smaller image more sensitive to image motion & vibrations – 3x more perspective elongation more laser power for same performance
&/or better detectors– 9x LGS focus change with zenith angle (1.4 m from zenith to 65)– Physical paths for laser beam transport much longer
• New versus existing telescope– Interfaces and observatory software not as well known for TMT– TMT, NFIRAOS & Instrument I&T could be overlapping– 2nd (or 3rd) Keck AO system versus 1st TMT AO system– Need to be conservative for TMT first light system
• Architecture & Technologies– Different lasers (new laser suppliers have emerged since TMT 2006 CoDR)
– MCAO versus MOAO (2 large high order DMs vs 1 lower order DM + MEMS; & different I&T issues)
– NFIRAOS cooled to -30C versus -15C for NGAO
9
Major Differences between NFIRAOS & NGAO
9 x 4 x 3 m(blue) box
NGAO (below) roughlyto NFIRAOS (left) scale
3.8 m
10
High-level Results of Cost Comparison
• NGAO $42.2M SDR estimate is 48% of NFIRAOS $88.2M CoDR estimate– NGAO $34.5M without contingency is 52% of NFIRAOS $66.7M – All in FY08 $
• Cost differences attributed to the following major factors:
Category $MContingency 13.8
Laser Procurement 9.4Labor Rates 6.5
Component Development 4.7DM & Tip/Tilt Procurement 4.5
RTC System 3.9LGS WFS 2.7
+/- other itemsTotal = 45.5
Largest items in $46M differenceModest contingency increase likely needed
No significant adjustments required - differences largely understood
Conclusions:
Need quote for new laser
11
Cost Category Comparison
• All costs in FY08 dollars– NFIRAOS costs have been inflated by 4%/year for 2 years
• NFIRAOS cost estimate is 2.1x NGAO– $32.1M difference without contingency ($46.0M with contingency)
• Major difference is in non-labor (e.g., procurements)– Factor of 2.7 or $28.3M, but less TMT work is performed “in-house”.
• Other significant differences:– Factor of 1.3 or $4.5M in labor even though NGAO has 14% more labor– NFIRAOS uses 32% contingency versus 22% for NGAO
• NFIRAOS average rate = 1.46x NGAO– NFIRAOS average rate = $101/hr– NGAO average rate = $69/hr– Benefit rates are the same– 24% burden rate versus 19% at CIT & 0% at WMKO & UCO– NGAO uses actual CIT, UCO & WMKO rates
• 1.46x represents $6.5M of the $20.5M of NFIRAOS labor
• Similar complexity & parallel processing approaches • NFIRAOS assumes $5.4M RTC contract based on quote
– Includes ~ 6.9 work-years at industrial rates
• NGAO assumes mostly in-house development– Includes 6 extra work-years of labor vs NFIRAOS – NGAO hardware is based on quotes
• Need to add PSF RT boards ($100k should meet NGAO needs)
– Vendor quotes to be obtained prior to PDR to better anchor our estimate
• Less complex systems for comparison:– NGWFC $2M, including $0.72M Microgate (3 RTC systems) subcontract– Gemini MCAO RTC subcontract was $0.88M– GPI $0.65M; PALM-3000 $0.60M
• Reduced by ~ $1-2M at PDR by assuming fixed gravity vector
– NGAO assumes 2x 50W SOR-type lasers• NGAO estimate based on our previous experience trying to set up a company to build
SOR-type lasers for Gemini & Keck– FASORtronics recently set up to commercialize this laser
• We are participating in ESO’s call for laser preliminary designs– ROMs will be available within a month– A fixed price quote will be available at PDR (Aug/09)
• The SDR reviewers questioned whether 100W would be adequate for the highest order correction
– After re-assessment we believe that 100W has sufficient margin by a factor of at least 1.5
• No major procurement differences ( > $1M)– TMT system physically larger
• Biggest difference is the launch facility labor– NGAO labor based on K1 LGS launch facility– NGAO asterism generator based on LGS WFS pick-off mechanisms
• NFIRAOS estimate may be somewhat conservative (CoDR level)• NGAO may be slightly optimistic
– Not obviously missing anything after comparison to NFIRAOS– At most could imagine adding 2 work-years to estimate
• NGAO has more labor in all but one category than NFIRAOS– NFIRAOS laser lab I&T includes a full off-telescope system test– NFIRAOS tel. I&T only covers to beginning of science commissioning
• ~ 25% of NGAO AO telescope I&T is for science commissioning
• NGAO estimate takes into account past experience with a detailed breakdown• Possibility that we are not comparing apples to apples
– NFIRAOS may cover some under system eng. & I&T contingency– NGAO assumes that each subsystem is complete & has met its requirements prior to lab I&T (presumably