Reports 1-1 of 1 Clear search Modify search
VIS (SR3)
fabian.arellano - 16:53 Thursday 01 December 2022 (23093) Print this report
Intercalibration between the seismometer on the ground and the F0 LVDT.

I calculated the intercalibration factor between the seismometer on the ground and the F0 LVDT, following the method described in entry 22998.

Fig. 1 shows the TF from the seismometer to the F0 LVDT, the coherence and the ASD of the readout of both devices. An advantageous feature of the F0 LVDT in SR3, is its high sensitivity, namely around 2 nm/rtHz above 1 Hz. This allows us to acquire information of the three peaks expected in the TF.

Fig. 2 shows the fit of a transfer function mathematical model to the measured data (see entry 22998). The data used was between 0.1 and 1.5 Hz with a coherence higher than 0.95. In the script, I used the seismometer readout in units of velocity and without a sensor correction filter. 

The value of the intercalibration factor, that must multiply the seismometer readout, is a = 1.0838.

This is information on the Jupyter notebook I used:

  • File name: seismometer_F0_lvdt_intercalibration.ipynb
  • Directory: /kagra/Dropbox/Subsystems/VIS/vis_commissioning/sr3/IP_diag/SENS_corre/Vertical/notebooks/
  • Conda environment: vishack.
  • Diaggui file: sr3_seis2f0_tf_ol_221130.xml
Images attached to this report
Comments to this report:
fabian.arellano - 15:18 Monday 05 December 2022 (23124) Print this report

In the context of the intercalibration between the ground seismometer and the F0 LVDTs, a calculation suggests that the effect of missing data from the displacement transfer function, can be compensated using the force transfer function.

As reported in klog 23095, the lack of a complete set of data does not allow us to calculate the intercalibration factor between the ground seismometer and the F0 LVDT in SR2. A complete set of data includes data points of the three peaks expected in the transfer function, whereas in the SR2 measurement, only one peak stands out from the LVDT noise level.  A workaround, proposed by Ushiba-san, is to compensate the missing information using the force transfer function, of which we always have reliable measurements.

I tested this idea using SR3 because we already have a complete set of data (klog 23093), and we can compare the results given by the two approaches.

The procedure I followed is this:

  • I fitted a mathematical model to the force transfer function measurement T (Figs. 1 and 2).
  • The function T was normalized, and the amplitude a was written as a separate parameter: T →  a × T.
  • Per klog 22998, I constructed the mathematical model a × ( T - 1 ).
  • Then, using a as a free parameter, I fitted the model to the measured displacement transfer function, that goes from the seismometer on the ground to the F0 LVDT (see klog 23093).

Fig. 3 shows the measured data points, the fitted model previously shown in klog 23093, and the fitted model constructed with the force TF data. The values of the intercalibration factor calculated using both methods are:

  • With a model with all its parameters free (klog 23093): a = 1.0838.
  • With the model given by the force TF:  a = 1.0789.

Despite I used only one free parameter for the latter, above 80 mHz the coincidence in the amplitude is very good, and the difference between the two values of the parameter a is only 0.4%!

This suggests that it is possible to compensate the missing data in the displacement TF, with information from the force TF.

Images attached to this comment
Search Help
×

Warning

×