Abstract:

To reduce the 0.2 Hz motion of GAS filter, which caused TM pitch motions, I tried to implement modal damping into ETMY.
Currently, modal damping successfully reuced the pitch RMS around 0.2 Hz but induced noise at high frequency (0.3-2 Hz), though the total RMS is reduced by a factor of 2.
So, the modal damp might make the pitch RMS worse if it will be implemented on a quiet day
To improve the performance, actuator diagonalization is necessary.

Detail:

First, I performed sensor diagonalization by exciting each resonant mode and measured TF of M2/M1, M3/M1, and so on.
Based on the measurement results, I update the sensor matrix for modal damping.
Figure 1 shows the comparison of the TF from modal actuator to modal sensor.
Blue and red show the TF before and after sensor diagonalization, respectively.

Then, I implemented modal damping for the lowest resonant frequency (M1)
Figure 2 shows the spectrum of each GAs filter and modal sensor.
As you can see, 0.23 Hz peak was reduced a lot but each GAS filter spectrum have some excess at high frequency.
Figure 3 shows the spectrum of ETMY pitch OpLev (blue: without modal damping, red: with modal damping).
Though peak at 0.23 Hz was reduced, some excess also can be seen on the OpLev signals.

According to the fig4, which shows the TF from M1 actuator to all modal sensors.
Since I didn't perform actuator diagonalization, there are clear peaks in M2-M4 sensors even exciting M1 actuators.
So, to reduce thecoupling at high frequency, it is necessary to decouple the modal actuators as well as sensors.

Note:

Current modal damping control induced the noise at high frequency but reduced the RMS around 0.23 Hz.
So, it works well when the microseismic noise is large.
However, it would just inject the noise at high frequency if the modal damping is engaged on the quiet day.
So, I didn't implement the modal damping into the guardian yet.