Difference: OlivierStatus ( vs. 1)

Revision 12013-12-16 - OlivierMartineauHuynh

Line: 1 to 1
Added:
>
>
META TOPICPARENT name="TrendSimu"
In order to reach our initial objective of producing teh expected skymap for EAS in TREND-50,
I list the following tasks ahead of us:
1)
- insert/test core position randomization (code is ready)
- insert/test minimal number of antenna fired for the selection of core position (see Valentin's email on December 3rd).
- insert mechanism to select showers from given theta, phi.
All of this consists in (minor) modification in the (python) code of register-tr50-fields & submit-tr50-jobs.py, and should be a question of few hours of work.
2) Then we want to produce fields for a given set of (phi, theta). For example phi 20->360กใ for theta=60กใ and theta=0->85กใ for phi=0กใ.
This will allow us to check very quickly that simulation behaves as expected (strong geomagnetic effect)
3) Then we can launch mass prod on ALL showers available. We need to organise for this (define a common db, decides who does what, etc)
4) Then we can launch mass prod on other energies (either random choice of energy, or for specific energy values (1e17eV, 1e18eV, 5e18...), which may be sager in our case where we have a very limitted knowledge
of our threshold. Anyway, we still have time for this.

5) Parallel to that shower+field simulation, we also have to complete the chain and go down to shower detection (or not), which is the final step to produce the expecetd EAS skymap.
A code (based on nec2) has been written by Valentin to generate the voltage values at the antenna output from the field values.
As far as I know (Valentin?) there is no existing tool to automatize & control this step.
In the long run (ie neutrino simulation), that will certainly be necessary, given the large amount of data.
At present we may want to skip that if nobody is ready to write the corresponding code.
Indeed nec2 is fast enough to do that part in interactive mode. I wrote some bash script to do that some time ago (in /sps/trend/omartino/eva/Ulastai).

Updated script matching the prsent data structure is now available in /sps/trend/trend-50.
Works for the available job outputs, but some paramaters in the script are hardcoded.
It will be easy to adapt for all job output formats.
I also put in the same directory one matlab script (PlotEVaSimu.m) to plot these traces.

Talk to you tomorrow,
Cheers
Olivier
6) Then we have to perform the antenna->DAQ step.
That means producing S_daq=H*S_ant+N, where H is the typical DAQ transfer function, N the typical noise (both taken from actual data)
and S_ant and S_daq teh signals at antenna (output of previous stage) and DAQ level.

7) Last step is just trigger decision. As for the actual acquisition we should consider that one shower is detected if 5+ antenna have a signal above 8 sigmas.
Again I have some code ready, but that was not so satysfying (very strongly depending on the choice of the input noise). I hope the output of teh apmlitude calibration tests will bring some
positive input into that.

All in all, I think the roadmap is quiet clear, and that there should not be any major obstacle for this to do list (at least from 1 to 5).
The total amount of work to be provided is very reasonnable, anda first map could be produced within few weeks (depending on available CPU)
if we work seriously. I propose a meeting early next week (Monday or Tuesday) to discuss that and how to share the task.

Cheers
Olivier

-- OlivierMartineauHuynh - 2013-12-17

 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback