This wiki page is created to note all the motivations for performing specific runs, links to the result pages and codes. The main motivation behind this study is to assess the reliability of IMRPhenomHM waveform to recover physical parameters from injected NR waveforms in zero-noise. We inject NR waveforms in zero-noise so as to suppress all errors from noise contributions to the data.

• The total response of a detector network (2 detector) is taken to be : sqrt( F²_{+, H} + F²_{x, H} + F²_{+, L} + F²_{x, L}). If a third detector is added, then we include F_+/x for that detector to compute the total response.

• The response function depends on the ra, dec values of the sky-position along with the polarisation of the waveform and gps trigger time. So, we fix the polarisation to 1.75 radians (~ 100 deg) and the gps time to 1187128000. Constraining the values of ra, dec between (0, 2 pi) and (-pi/2, pi/2) respectively, the plot of response function for HL and HLV. The blue point is the chosen sky-position for setting up the PE runs. The blue dot has the values (ra, dec) = (0.1843687374, -0.68820842) and we will be using these values for creating NR frame files.

• HL detector network response

• HLV detector network response

• First, create a h5 file from the psi4 data for the different simulations. The script to create the h5 file is at : /home/c1549390/work/PhenomHM_pe/src/nrutils-make-nrh5-BAM.py for the BAM waveforms, and /home/c1549390/work/PhenomHM_pe/src/nrutils-make-nrh5-GaTech.py for the Georgia tech waveforms. All modes upto and including l=4 are stored in the h5 file.
• As of 18/09/2017; I have created h5 files for q1, q2, q4, q8 and q18 non-spinning NR waveforms. q1 and q2 are Gerogia Tech waveforms while q4, q8 and q18 are BAM.
• I wanted to keep the f_low (as calculated from the NR waveform) of the waveform for all the systems to be the same, so I adjust the mass of the injection to adjust the f_low ~ 20Hz for all runs.
• Created HL and HVL (inclination series) frame files for the q4 and q8 non-spinning systems. To ensure f_low is same, q4 is injected at 80 Msun and q8 at 115 Msun.
• The two detector network injection (H-L) frames can be found at : /home/c1549390/work/PhenomHM_pe/frames/HL_Frames
• The three detector network injection (H-L-V) frames can be found at: /home/c1549390/work/PhenomHM_pe/frames/HLV_Frames

• In some previous runs, we had noticed that the improvement of PhenomHM vs PhenomD to constrain the distance was increased by ~ 50 % (i.e., the posteriors of PhenomHM were narrower by 50% compared to PhenomD ) for the two detector network and for the HLV configuration, the same value was ~ 40%.
• One caveat between those two runs was that for the HL run, the source was placed at a sky-position where the response was the maximum, but for the HLV run the source was at a sky-position where the response was about half of that.
• I wanted to check the effect of changing the inclination of the source on the recovery and to check for any biases on the posteriors and to check whether including a third detector would improve the results.
• With that in mind, I have started a set of runs for the BAM q4 & q8 non-spinning waveforms for two and three detector configuration, at three different inclinations for each. So it's a total of 24 runs.
• I've used the aLIGOzeroDetHighPower and aVIRGO PSD's for the runs.
• The base directory for all the run directories is at: /home/c1549390/work/PhenomHM_pe/runs/18_09_2017/