In the first attached file is it shown where we performed the various injections. From the analysis we can assert that:

• We couldn't pinpoint the scatterer to a specified location, but we could localize it to the east flange of the IB tower. In fact, when we did the tapping test on the beam dump and on the flange, both the accelerometer on the beam dump (BD ACC) and the one on the inner side of the flange (IB ACC X) see the same amount of noise. The 101 Hz peak came up, instead the 157 Hz peak was not excited (see plots 2 to 4, along x, y and z directions). Also while doing the injections with the shaker on the beam dump we excited both the beam dump and the flange, as shown on plot 5. To be noticed that this time also the structure at around 157 Hz has grown in Hrec.
• During the broadband injections we also observed an increase of coherence between PRCL and PMC TRA at 101 Hz, which already had been pointed out by the BRUCO even during quiet periods (look at plot 6). This could be explained through the PSTAB loop, which tries to compensate at those frequencies (in fact this increase in the coerence is also visible on the PSTAB error signal). However, we do not see any effect during the injections in the band 150-160 Hz.
• Plot 7 shows the injections of lines at 155, 156, 157 and 158 Hz that had been performed with the shaker installed on the beam dump. Only the 156 and 157 Hz are coupled in PRCL and Hrec.
• Plot 8 shows the tapping performed on the pipes linking SIB1 to SIB2. There is no coupling to Hrec. It is not surprising because there are the bellows to isolate the seismic noise. We should have done it on both sides of the bellows as well in order to verify if it is coming from the viewports.

Another test we could do is to increase the power going to the beam dump by turning the waveplates of the IPC1 (EIB) and 2 (SIB1). This would require a dedicated shift.

Looking at the tapping figures (2-4), especially the tapping in the z direction (figure 4), there is a broad bump between 120Hz and 140Hz that is increased in h(t) and in PRCL. If further studies are done with a shaker, then a noise injection in this band should also be done. That broadband bump, if present at lower level in normal operations, might be costing much more range than the clearly visible 101Hz and 157Hz peaks themselves.

Figure 1, shows that during the tapping (purple) a mechanical resonance at ~130Hz is excited as witnessed by the accelerometer. The excited motion is about a factor 10 above the normal conditions (blue). This correspond to a noise in h(t) that is a factor 3 above normal condition, so naively one could expect this noise to be  only a factor 3 below the sensitivity curve in normal conditions.

I tried to make the projection. Since there is no coherence between the accelerometer and Hrec I computed the coupling function and set a threshold of 2 on Hrec to distinguish between measured values and the upper limit. The measured values are the yellow points in the 1st plot, so the projection is meaningful only in the region 125-140 Hz

The 2nd plot shows the projection, where we can see that we are a factor 2 at worst (at 137 Hz) below the sensitivity.