Today we performed a shutdown test at the NEB, with the aim of better understanding the ACOUSTIC and SEISMIC noise that could affect the frequency band of the Newtonian Noise.
Here is the timetable:
Until 07:58:00 UTC in normal working conditions
from 08:03:00 (+ 200 seconds) the airflow (fan) is off; all remaining HVAC devices (eg water pumping systems) are left on
from 08:08:00 (+ 200 seconds) all HVAC devices are switched off (water pumps, chillers, heaters ...)
from 08:13:00 The HVAC system is again set in standard working conditions.
Some comments:
all ACOUSTIC noise (from infrasounds below 1Hz up to 100Hz) is due to airflow; no appreciable reduction is visible when the remaining devices are also switched OFF (see plot 2).
On the contrary, all the SEISMIC noise is due to what remains ON except for a ~ 13.5Hz line which is known to be the fan rotation speed (clearly visible in the seismometer on the north axis ENV_NEB_SEIS_N) which disappear at the fan switch off
From this test it seems that we have to work on the air flow and air ducts for acoustic noise reduction, while we have to work on the ground vibration induced by the flow of water (and by other devices) to reduce the seismic noise.
Note that some 'bumps" are visible in the 0 to 10Hz seismic sepctrum, they are probably due to some truks; these explain the singularity that when switching OFF only the air flux some seimic noise increase.
(apologies for the non omogeneus colours used in plots)