SUPERNOVA NEUTRINO SIMULATION IN THE AMANDA DETECTOR
INTRODUCTION and MOTIVATION
The main goal of this project is to provide a suitable code able to
simulate supernova neutrino events in the AMANDA detector.
AMANDA
is a high-energy neutrino telescope located at the geographical South Pole.
The main goal of Amanda is to search for high-energy neutrinos as a probe
for the highest energy phenomena in cosmos and to open new observation windows to the Universe.
Neutrinos are interacting only weakly, and undeflected by interstellar magnetic fields
they can travel astronomical distances providing hidden information complementary to that
obtained from electromagnetic observations.
Although Amanda is a high-energy neutrino telescope,
it can be used to search for low energy neutrino bursts from Galactic supernova explosions.
However, only a time profile of the neutrino burst during the collapse phase
can be measured. Additional information on the energy and direction of the
supernova neutrinos is unlikely to be reconstructed in such a
detector.
To simulate supernova events in the Amanda detector, two different issues have
to be considered:
- Simulation of dark noise background.
- Simulation of the supernova signal on top of that background.
We can approximate the distribution of the background events
as a Gaussian distribution.
On top of the background distribution we can combine the supernova signal
also normally distributed. Many input parameters have to be taken into
account, like: the given model for the profile of the neutrino
luminosity, the distance of the supernova,
the expected number of photo-electron in excess per PMT due
to a SN, and the time interval of the noise rate.
To select supernova candidates we set up a Maximum Likelihood
approach for each event in order to reject background which might fake
the supernova in our detector..
By minimizing the Likelihood function and solving for the Likelihood
parameter, i.e. the rate excess due to SN event, we can infer the time
and the distance of the obtained SN.
CODE and CLASSES
- First main program
The main program GenerateSupernova.cc
takes care of the generation and simulation of the
supernova events.
Set of the expected classes used by the main program GenerateSupernova.cc
- GaussGen.h this class generates normally randomly distributed
data given the parameters of the mean and the variance.
- NoiseGen.h this class takes the randomly distributed data and
does all calculation in order to produce background noise which
can simulate the noise seen in the AMANDA data
- SNmodel1.h This class does all calculation for the SN signal
which is also normally distributed but takes into account the strength
of the signal, the distance of the SN and the time of the SN burst.
- SNmodel2.h This second model does additional calculation
which takes into account the distribution of the stars in the Galactic
plane, and the model of the supernova SN1987A.
- GenData.h This class collects all generated data and produce
an output file which can be used for the analysis, which is the second
part of this project.
- Second main program
The second main program AnalyzeSupernova.cc is designed to analyze the data with the
statistical approach of the Likelihood maximum method.
Set of the expected classes used by the main program AnalyzeSupernova.cc
- MovingAverage.h This class will take as input the generated
data by the previous simulation and will calculate a moving average of
the combined noise rate background of all PMTs.
- Likelihood.h This class does the Maximum Likelihood
calculation over all individual PMTs. Crucial input parameters are the
mean and standard deviation of the noise rate distribution of the
PMTs. The Likelihood is so defined to extract the most probable value of
the collective excess of noise rate due to all PMTs in a given time window.
- FindSupernova.h This class takes the output of the calculated
moving average and Likelihood parameter, the collective excess of
noise rate of all PMTs and extract from this the final distribution
separating the SN signal by the background.