ligo-g000193-00-m ligo i science run barry barish pac meeting - lho december 13, 2000
Post on 19-Dec-2015
220 views
TRANSCRIPT
2LIGO-G000193-00-M
The LIGO I Science RunData & Computing Group Operations Plan
9th Meeting of the LIGO PACLIGO Livingston Observatory
Livingston, Louisiana13 December 2000
Albert Lazzarini
LIGO Laboratory Caltech
3LIGO-G000193-00-M
LIGO Plansschedule
1996 Construction Underway (mostly civil)
1997 Facility Construction (vacuum system)
1998 Interferometer Construction (complete facilities)
1999 Construction Complete (interferometers in vacuum)
2000 Detector Installation (commissioning subsystems)
2001 Commission Interferometers (first coincidences)
2002 Sensitivity studies (initiate LIGO I Science Run)
2003+ LIGO I data run (one year integrated data at h ~ 10-21)
2005+ Begin ‘advanced’ LIGO installation
5LIGO-G000193-00-M
Significant Events
Hanford 2km
interferometer
Single arm test complete installation complete interferometer locked
6/00 8/00 12/00
Livingston 4km
interferometer
Input Optics completed interferometer installed interferometer locked
7/00 10/00 2/01
Coincidence Engineering Run (Hanford 2km & Livingston 4km)
Initiate Complete
7/01 7/02
Hanford 4km
interferometer
All in-vacuum components installed interferometer installed interferometer locked
10/00 6/01 8/01
LIGO I Science Run (3 interferometers)
Initiate Complete (obtain 1 yr @ h ~ 10-21 )
7/02 1/05
6LIGO-G000193-00-M
•operating as a Michelson with Fabry-Perot arms
•reduced input laser power (about 100 mW)
•without recycling •noise level is a factor of 104-105 above the final specification
•sources of excess noise are under investigation
Strain SensitivityNov 2000
2-km Hanford Interferometer
7LIGO-G000193-00-M
LIGO Isteps to science run
commissioning interferometer » robust locking» three interferometers» sensitivity» duty cycle
interleave engineering runs» implement and test acquisition and analysis tools» characterization and diagnostics studies» reduced data sets» merging data streams» upper limits
8LIGO-G000193-00-M
LIGO/LSC Data Analysis Model
Now: » Initial engineering runs starting to set the stage for how
science, research is done
» Data being archived at Caltech in HPSS
» Access from archive according to LIGO Laboratory MOUs
» “Stress testing” of software and hardware systems - both LDAS and GDS/DAQS/CDS
» Initial data analyses focus on – sorting out commissioning issues, – understanding environment, – Calibrations, data conditioning, pre-processing
9LIGO-G000193-00-M
LIGO/LSC Data Analysis Model
Near-term (2Q2001): » LIGO science will focus on using engineering runs to extract
meaningful first upper limits» Organized around 4 upper limits papers using ~1 week of data
in 2Q2001» Opportunity to set current best upper limits on these classes of
sources» Provides a basis to "exercise" the LSC data analysis groups» Provides a basis for future organization of the LIGO I Science
Run search teams– groups will expand as interest grows in LIGO science.
Problems:» LDAS readiness to support the engineering run goal» Strategy is to limit scope primarily to LDAS supported goals
10LIGO-G000193-00-M
Astrophysical Signaturesdata analysis
Compact binary inspiral: “chirps”» NS-NS waveforms are well described» BH-BH need better waveforms » search technique: matched templates
Supernovae / GRBs: “bursts” » burst search algorithms – eg. excess power; time-frequency patterns» burst signals in coincidence with signals in electromagnetic radiation » prompt alarm (~ one hour) with neutrino detectors
Pulsars in our galaxy: “periodic”» search for observed neutron stars (frequency, doppler shift)» all sky search (computing challenge)» r-modes
Cosmological Signals “stochastic background”
11LIGO-G000193-00-M
LIGO/LSC Data Analysis Model
LIGO I Science Run (2Q2002):» Key astrophysical searches follow the LSC Data Analysis
White Paper plan:
» Organized around teams, as in near-term upper limit studies– Open to all who are willing to work
» LIGO Lab LDAS resources to be used for searches will be shared among the teams
» LSC member institutional resources used by individual researchers
» Longer term: establish 5 LIGO/LSC Tier 2 centers (“University Research Centers” or URCs) to provide additional computational, data distribution resources across collaboration
12LIGO-G000193-00-M
Inspiral Sources
LSC Upper Limit
Group
Inspiral Sources Co-chair Patrick Brady, Gabriela Gonzalez
-------------------------------------------------------------------
Bruce Allen [email protected]
Sukanta Bose [email protected]
Douglas Boyd [email protected]
Patrick Brady [email protected]
Duncan Brown [email protected]
Jordan Camp [email protected]
Nelson Christensen [email protected]
Jolien Creighton [email protected]
S.V. Dhurander [email protected]
Gabriela Gonzalez [email protected]
Andri Gretarsson [email protected]
Gregg Harry [email protected]*
Syd Meshkov [email protected]
Tom Prince [email protected]
David Reitze [email protected]
B.S. Sathyaprakash [email protected]
Peter Shawhan [email protected]
13LIGO-G000193-00-M
Interferometersastrophysical sources
Compact binary mergers
Sensitivity to coalescing binaries
Binary inspiral ‘chirp’ signal
2002
2007
future
14LIGO-G000193-00-M
Interferometer Data40 m
Real interferometer data is UGLY!!!(Gliches - known and unknown)
LOCKING
RINGING
NORMAL
ROCKING
15LIGO-G000193-00-M
The Problem
How much does real data degrade complicate the data analysis and degrade the sensitivity ??
Test with real data by setting an upper limit on galactic neutron star inspiral rate using 40 m data
16LIGO-G000193-00-M
“Clean up” data stream
Effect of removing sinusoidal artifacts using multi-taper methods
Non stationary noise Non gaussian tails
17LIGO-G000193-00-M
Inspiral ‘Chirp’ Signal
Template Waveforms
“matched filtering”687 filters
44.8 hrs of data39.9 hrs arms locked25.0 hrs good data
sensitivity to our galaxyh ~ 3.5 10-19 mHz-1/2
expected rate ~10-6/yr
18LIGO-G000193-00-M
Detection Efficiency
• Simulated inspiral events provide end to end test of analysis and simulation code for reconstruction efficiency
• Errors in distance measurements from presence of noise are consistent with SNR fluctuations
19LIGO-G000193-00-M
Setting a limit
Upper limit on event rate can be determined from SNR of ‘loudest’ event
Limit on rate:R < 0.5/hour with 90% CL = 0.33 = detection efficiency
An ideal detector would set a limit:R < 0.16/hour
20LIGO-G000193-00-M
Two Sites - Three Interferometers»Single Interferometer non-gaussian level~50/hr
»Hanford (Doubles) correlated rate (x1000) ~1/day
»Hanford + Livingston uncorrelated (x5000)<0.1/yr
Coincidencesbetween
LLO & LHO
21LIGO-G000193-00-M
Burst Souces LSC Upper Limit Group
Burst Sources Co-chair Sam Finn, Peter Saulson
-----------------------------------------------------------
Warren Anderson [email protected]
Barry Barish [email protected]
Biplab Bhawal [email protected]
Jim Brau [email protected]
Eric Black [email protected]
Kent Blackburn [email protected]
Ed Daw [email protected]
Ronald Drever [email protected]
Sam Finn [email protected]
Ray Frey [email protected]
Ken Ganezer [email protected]
Joe Giaime [email protected]
Gabriela Gonzalez [email protected]
Andri Gretarsson [email protected]
Bill Hamilton [email protected]
Warren Johnson [email protected]
Masahiro Ito
S. Klimenko [email protected]
Al Lazzarini [email protected]
Szabi Marka [email protected]
Genakh Mitselmakher [email protected]
Soumya Mohanty [email protected]
Benoit Mours [email protected]
Soma Mukherjee [email protected]
Fred Raab [email protected]
Ravha Rahkola
Peter Saulson [email protected]
22LIGO-G000193-00-M
pulsar proper motions
Velocities - young SNR(pulsars?) > 500 km/sec
Supernovaeasymmetric collapse?
23LIGO-G000193-00-M
LIGO Iscience run
Strategy» initiate science run when good coincidence data can be reliably taken and
straightforward sensitivity improvements have been implemented (~ 7/02)» Then, interleave periods of science running with periods of sensitivity
improvements
Goals» obtain 1 year of integrated data at h ~ 10-21
» searches in coincidence with astronomical observations (eg. supernovae, gamma ray bursts)
» searches for known sources (eg. neutron stars)» stand alone searches for compact binary coalescence, periodic sources,
burst sources, stochastic background and unknown sources at h ~ 10-21 sensitivities
Exploit science at h ~ 10-21 before initiating ‘advanced’ LIGO upgrades
24LIGO-G000193-00-M
LIGO/LSC Data Analysis Model
Throughout Engineering & Science Runs, the Laboratory’s Data & Computing Group fulfills the following roles:
» LIGO science, data analysis: scientific staff are actively engaged in the astrophysics searches
» Simulation & Modeling: detector support, data analysis» Continuous management and movement of large volumes of
data» Maintaining pipeline analyses running, archive running» Software maintenance/improvements/enhancements» LSC support, visitors» LIGO Laboratory-wide IT support
25LIGO-G000193-00-M
LDAS Staffing Plan
Personnel FTE LIGOScience
Run;AlgorithmPerf. Mon
LIGOScienceRun; HWPerf. Mon
LIGOScience
Run; DataArchival
LIGOScience
Run; DataAnalysis &Research
Operations: DB
Mgmnt
Operations:Sys. Admin
SW Dev& Maint.
Graduate Student 2.0 2
Post DoctoralResearcher
3.0 0.45 0.33 0.25 1.97
Sci. 3.8 1.38 0.2 0.45 1.24 0.53
Sw. Eng 3.8 1.42 2.38
Contract SW Eng. 1.0 1
DB Admin 1.0 0.5 0.5
HPSS Admin 1.0 0.5 0.5
LDAS-SysAdmin 1.0 1
CIT Totals -> 16.6 3.25 0.53 1.7 5.21 0.5 1.5 3.91
LHO Sci* 1.0 0.2 0.2 0.2 0.2 0.2
LLO Sci* 1.0 0.2 0.2 0.2 0.2 0.2
LIGO Lab Totals -> 18.6 3.65 0.93 2.1 5.61 0.5 1.9 3.91
Personnel Students 2
Classifications Post Docs 0.45 0.33 0.25 1.97
Scientists 1.38 0.2 0.45 1.24 0.53
SW Eng. 1.42 3.38
SysAdmin 1 0.5 1.5
* NOTE: LHO and LLO personnel are in the budgetary proposals from the respective observatories
Data & Computing Group Principal LDAS activities during operations
26LIGO-G000193-00-M
Data & Computing Group Principal Modeling & Simulation
activities during operations
Modeling & Simulation Staffing Plan
Personnel FTE Code environ.Development:Architecture
Code environ.Development:
EM Fields,Optics
Code environ.Development:
Mechanicalsystems
Code environ.Development:GUI Develop.
DetectorPerformance
Modeling
DataAnalysis
Simulation
Graduate Student 2.0 1 1
Post DoctoralResearcher
1.0 0.5 0.5
Sci. 2.0 0.5 0.5 0.25 0.5 0.25
Sw. Eng 1.0 0.75 0.25
Contract SW Eng. 1.0 1
CIT Totals -> 7.0 1.25 0.5 0.75 1.25 2 1.25
Personnel Students 1 1
Classifications Post Docs 0.5 0.5
Scientists 0.5 0.5 0.25 0.5 0.25
SW Eng. 0.75 0.25 1
27LIGO-G000193-00-M
Data & Computing Group Principal General Computingactivities during operations
General Computing Staffing Plan
Personnel FTE Server & DataSystems
DesktopWorkstations
LAN & WANAdministration
TelecommunicationsAdministration
Web & DBAdministration
Student Asst. 0.40 0.4
www, DB Admin 0.50 0.5
Sci. 0.10 0.1
Sr. Sys. Admin 1.00 0.2 0.15 0.3 0.35
Asst. Sys. Admin 1.00 0.3 0.6 0.1
Contract Sys.Admin.
1.00 0.25 0.75
CIT Totals -> 4.00 0.75 1.9 0.5 0.35 0.5
MIT Sys. Admin* 1.00 0.3 0.6 0.1
LHO Sys. Admin.* 1.25 0.3 0.6 0.1 0.13 0.12
LLO Sys. Admin.* 1.25 0.3 0.6 0.1 0.13 0.12
LIGO Lab Totals -> 7.50 1.35 3.1 0.7 0.48 0.62
Personnel Students 0.4
Classifications DB Admin. 0.5
Sci. 0.1
SysAdmin 0.75 1.5 0.4 0.35
* NOTE: MIT, LHO, and LLO personnel are in the budgetary proposals from the respective sites
29LIGO-G000193-00-M
* MIT, LHO, and LLOhave local General
Computing staff
* LHO, and LLOhave local LDAS staff
31LIGO-G000193-00-M
Hardware Class
ProcurementExpense
Maintenance @ 30%/yr
HPSS $1428.0K $428.4K
Servers $272.0K $81.6K
Switches $163.0K $48.9K $/yr $/qtrFY2001Q4-FY2004Q1
Maintenance
FY2001Q4-FY2004Q1
Maintenance
HW RAID, Small Robots
$735.0K $220.5K $779.4K $194.9K $194.9K $107.1K
Maintenance @ 15%/yr
FY2001Q4-FY2004Q1 Equipment
FY2001Q4-FY2004Q1 Equipment
Workststions, small switches, disks
$1387.0K $208.1K $208.1K $52.0K $52.0K $139.8K
Totals-> $3985.0K $987.5K $987.5K $246.9K $246.9K $246.9K
From Procurement Plan C90755-A
LDAS Operations BudgetHardware Support
33LIGO-G000193-00-M
Conclusionsscience run
Short term -- » implement LDAS
– 4 sites; computing; archiving
» engineering runs– data handling and access, reduced data sets
– diagnostics; characterize instrument and data
– algorithms; statistics
Longer Term» LIGO Lab support for Science Run
Support Required» LDAS procurement and implementation» incremental resources requested
– manpower
– maintenance and networking
– support of LSC