1. on behalf of the Virgo Collaboration
Python and the Commissioning of the Advanced Virgo Gravitational Waves Detector
Franco Carbognani
on behalf of the Virgo Collaboration
PyCon 8
Florence, 09 April 2017
11/01/2019
Franco Carbognani

2. Outline Detecting GWs with Interferometers A Bit of History
The Global Detectors Network
Advanced Virgo (AdV)
AdV Commissioning
The Control of AdV and Python Role
Conclusions
Franco Carbogna

3. Detecting GWs with Interferometers
Effect of GWs:
Squeeze and stretch the space in perpendicular directions: strain h = DL/L
We’ll need:
A set of free test masses, far apart,
A means to measure their relative motion, and
Isolation of the masses from other causes of motion.
Here’s the challenge:
Even for the most tremendous events in Universe, h~10^-21
If test masses are separated by 3 km, that means a length change less than 10-18m! 3

4. …causing the interference pattern to change at the photodiode
Detecting GWs with Interferometers
Laser used to measure relative lengths of two orthogonal arms
Arms are few km
Measure difference in length to one part in 1021 or meters
…causing the interference pattern to change at the photodiode
As a wave passes, the arm lengths change in different ways….

5. 1st generation GW detectors
LIGO and Virgo achieved the nominal sensitivity set more than a decade ago with good duty cycle during science runs
No detections were made and a next generation of machines, an order of magnitude more sensitive, was designed and constructed 5

6. Advanced Detectors
10 x sensitivity improvement over 1st generation detectors
1000 x increase of observation volume
1 day of AdV data >> 1 yrs of Virgo data

7. A Bit of History

8. The GW astronomy era has begun (and is awaiting for us…)
LVT151012
GW151226
Franco Carbognani

9. Sky location GW laser interferometers are not pointing telescopes
Sky location can be reconstructed through the time of arrival of GW radiation at the different detector sites, as well as the relative amplitude and phase of the GWs in different detectors.
GW sky location is with large uncertainty600 deg2

10. In the future, we’ll be more precise
2019+
2022+
As sensitivity progresses, so does the localization In the design LIGO-Virgo network, GW could have been localized to less than 20 deg2

11. Present/Future GW detectors network

12. Start of multi-messenger astronomy
“LOOC-UP” STRATEGY
trigger, information GW
data
telescopes, satellites
or other detectors
“ExtTrig” STRATEGY
trigger, information
Really profits from 3rd (Virgo!) detector….

13. Advanced Virgo
Franco Carbognani

14. AdV Detector Design MAIN CHANGES wrt Virgo larger beam heavier mirrors
higher quality optics
thermal control of aberrations
thorough stray light control
better vacuum
200W fiber laser
signal recycling
(postponed)

15. AdV Commissioning progress
Last mirror (WE) suspended in July 2016
November 2016, full ITF available for commissioning
First lock at half fringe (December 30th)
February 2017, ITF locked at dark fringe for 15 minutes 15

16. Commissioning path to join Ligo on O2
Improve ITF stability at dark fringe
Improve alignment loops
Lock OMC and switch to DC readout
Suspend/evacuate all detection benches
Low noise actuation
Build noise budget and get sensitivity curve
Define TCS needs
Week end engineering runs
More commissioning and noise hunting 16

17. Python application areas
AdV Detector Fast and Slow Controls
real-time control
Acl
Acl
supervisory control
Python application areas

18. Link to Fast Control
Dealing with an hybrid system using both and our legacy communication middleware Cm
Python
Virgo
Tools
Tango Server
Tango Software Bus
Pure Cm Client
Cm Server
Cm Software Bus
Pure Tango Client
PyFd/PyCm
PyTango
Problem managed by exploiting the existing PyTango and generating a Cm (and Fd) C library bindings using ctypegen (PyCm and PyFd)

19. Guardian
In Ligo Guardian provide a complete top level framework for the detector automation based on hierarchical finite state machines
Fully developed in Python. The automation logic is also written in Python 19

20. From Guardian to Metatron
Guardian (to be referred as Metatron: the name of a Virgo zodiac Guardian angel) integration in Virgo performed.
Fully reusing the pre-existing python scripts applications layer already in use and built on top the PythonVirgoTools library
Using the ConfigParser library and the .ini files format for managing configuration parameters (another good example of the Python “batteries included” philosophy)
The development of automation state machines took immediately momentum 20

21. DAQ integration
Automation nodes as DAQ nodes getting data directly from shared memory and synchronized with the one second period frame availability. 21

22. DAQ integration
Each Automation node is a state machine. To each state are associated a main (executed once) and a run executed cyclically and synchronized with the frames generation.
A generic mechanism to read and write DAQ channels has been introduced
wait loop for frame
get_frame()
…….
Inside run
ezca.get_channel(channel)
……..
ezca.set_channels(prefix, channels) 22

23. Metatron architecture23

24. Metatron workflow 24

25. Automation Nodes Hierarchy25

26. Nodes Documentation
All nodes are documented inline by the use of docstrings allowing the auto-generation of the documentation (via make html)
Documentation available as part of the AdV Python documentation: gw.eu/PythonDoc/html/doc/Automation/userapps/doc/index.html
On this page the states/transitions diagrams and the list of states (and the associated indices) for each node are also available 26

27. Current Developments Very closely supporting the ITF locking progress
Calibration automation
Generation of ITF Mode and DetChar flags
Automation of Noise Budgets generation
….. 27

28. ITF Mode and DetChar Flags generation28

29. INJ Noise Budget Automation29

30. Automated Version Control
An automatic checkin in the SVN archive of the .py and .ini file (and all other Python files imported from the userapps area) has been introduced.
It is triggered when:
the corresponding automation server start ('DAEMON INIT‘)
A reload of the automation code is performed (‘RELOAD‘)
A dedicated user interface has been made available for querying the SVN archive about files differences. It allows to show:
difference between the latest 2 committed revisions
difference between two dates 30

31. GUIs Development

32. Conclusions
The scientific community is eagerly awaiting AdV entering Science operations and join the international detectors network.
AdV is striving to finalize the current commissioning phase and significant step forwards have been made on the last months.
Joining Ligo for the last part of the O2 Science Run is becoming a realistic goal.
Python demonstrated to be very effective in supporting the detector supervisory control automation (in particular for quickly progressing on the ITF Locking sequence) and keep extending its fields of use in our Control Room.

33. THANKS FOR YOUR ATTENTION!
For more information and for staying up to date on the VIRGO project: