Personal tools
You are here: Home Virgo Data Analysis Software Engineering Runs SER1
Document Actions

Software Engineering Run 1

by Nicolas Leroy last modified 2011-12-01 20:44

Virgo proposition for SER1

First software engineering run will take place between 18 January 2012 and 15 February 2012


wiki page from DAC

first minutes from DASWG meeting

Goals for Virgo :

- Produce online Virgo data

- Produce online Data quality flags - test VDB

- Transfer data/DQ from and to Caltech

- run MBTA with similar results as what was done for VSR3 : send alerts to graceDB, produce web pages, ...

People involved:

- h(t) production and transfer : Benoit

- DQ production : Didier

- VDB: Gary

What frames are going to be simulated (Dec 1st proposal):

- Frames of 4 seconds long (to match the LIGO length) containing at least:

  • h(t) signal at 16384Hz generated as gaussian color noise using an Advanced Virgo sensitivity like the current baseline one. A 4096Hz signal will also be produced. Injection signals provided by the search group will be added to this h(t) signal, similarly to the hardware injection procedure.
  • The software injection signal (sum of all signals).
  • A state vector at 1Hz telling if the h(t) signal could be used or not because of unlock or any other reasons. The length of the "locked" segment will have a random flat distribution from 30min to 1 day, with an average duty cycle of 80%
  • A single data quality flag at 1Hz with some segment of bad quality data with a random flat distribution from 1s to 30s and an average duty cycle of 98% (see DQ description below)

Virgo frames will be generated at Cascina and transfer on line (using Fd) to Caltech to be made available as the LIGO frames. LIGO frames will be transfered from Caltech to Cascina (also using Fd), probably with some down sampling at 4096Hz, to be made available at Cascina for MBTA.

Frames files of 4000 seconds long (TBC), will be produced. The full set of ER1 Virgo frames will be stored at Cascina in the procdata buffers (bufferv33,34,35? TBC). It could then be accessible by remote sites using LDR. No commitment is made on storing the LIGO frames at Cascina.

What DQ are going to be simulated:

One DQ flag which will be produced at 1 Hz with the following characteristics:
  • The values will be 0 (good quality) or +1 (bad quality) . The +1 values will cover 2% of "locked" time (deadtime=2%).
  • The +1 segments will have a flat distribution from 1s to 30s and will be randomly distributed in time along all ER1
  • The values at 1 Hz will be stored in one FrAdcData of the h(t) frames. This "ADC" will be named "V1:DQ_ER1"
  • The frames will be passed (using Fd) to the SegOnline process which will produce the xml files and will fill VDB with the V1:DQ_ER1 segments.