Transcript ppt - IN2P3

Searching for
correlations in global
environmental noise
Karl Wette, Susan Scott and Antony Searle
ACIGA Data Analysis
Centre for Gravitational Physics
The Australian National University
[email protected]
Outline
Global correlations impact global
networks
 Southern hemisphere monitoring station
established
 Data acquisition from LIGO, Virgo
 Software to handle datasets
approaching Tb

K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
2
Global gravitational astronomy



Multiple observatories required
to triangulate source and veto
false alarms
Southern hemisphere detector
could complete a global network
Confidence in results requires
knowledge of correlated
disturbances to detectors
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
3
Local environmental monitoring

In The ANU Gravitational Wave
Research Facility

Optical cleanroom on damped slab
(VRM, HPTF etc.)
 Not at Gingin, but still far from existing
observatories

Three seismometers (XYZ)

Supplied by Dept. Earth Sciences, from
the field
 One down for repairs

Tri-axial magnetometer (XYZ)


Off-the-shelf instrument
Electrical supply voltage monitor


Locally built
Australia has 60 Hz power grid
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
4
Local data acquisition


Outside clean-room
Power supplies



Anti-aliasing filters
GPS receiver


1 PPS signal
Signal generator


Several supply different components,
minimising cross-talk
2048 Hz locked to GPS
NI ADC card

Sampling triggered by generator
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
5
Local data management


NI LabView software
extended to write
acquired channels
directly to standard
Frame files
Script uploads
completed frames to
ANUSF multi-terabyte
Mass Data Storage
System (MDSS)
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
6
Global data acquisition

Basic environmental channels are large


1 Gb, per day, per site
Important to download only needed channels, not
files

Use LIGOtools (getFrames) + LDAS to produce and
download an environmental monitoring RDS



LIGO does not archive these channels outside science
runs, so this must be done in near real time
Use Benoit’s goodwill
Data grid technologies

Local work on extensions to permit distribution at
channel, not file level
 Future feasibility will depend on relative costs of internet
and physical (couriered tape or disk) transfer
 Part of major Australian data grid effort, with
astronomers and particle physicists
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
7
ACIGA Data Analysis Cluster
16 P4 processors
 1 Tb RAID1 storage
 14 Gb ECC RAM
 LDAS
 Data grid
 Environmental data
analysis

K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
8
Pipeline





Accesses all data
Highly parallel
Runs on ACIGA Data
Analysis Cluster
Produce spectrograms
Automated detection of
unusual features
 Ensemble
of statistical
tests
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
9
Future

Upgrade
 AU$10k
this year for better acquisition system
 Postdoc 2005-7 includes this project

Analysis
 Long
timescale coherences
 Coherence bursts in spectrograms
 Impact on global astronomy
 Veto recommendations
K Wette, S Scott, A Searle
Searching for correlations in global environmental noise
ACIGA Data Analysis
10