This page has info on hurricane georges as it
passed ao. It was cobbled together on 5may21. I had trouble finding
a lot of the original monitor data (it must be backed up on tape
- When did it occur:
- the eye passed over AO on 21sep1998 around 23:30
- What monitor data is available:
- Wind data was taken on the platform (until i think the
anemometer blew away:).
- I have not located it yet (.. the data, not the
- td tension data:
- there was tension in the tiedowns during the hurricane. I
remember seeing the total tension (sum of all tiedowns)
oscillate by about 150kips in 3 seconds.. I remember this
because i was in the control room talking to paul goldsmith
up in ithaca. He asked me how things were going.. I remember
glancing at the tiedown tension monitor and telling him i
think we might have a problem :).
- I have not been able to find the tiedown tension data yet.
It must be backed up on tape somewhere.. arun said he'd look
- Tilt sensor data:
- the pitch tilt sensor was running in the dome during the
storm. We lost data from 21sep/14:24 to 22sep/01:00 (we must
have lost power). So there is no recording when the eye
passed.. but there was plenty of oscillations after
- the data is at /share/obs4/tilt/tilt.georgesN
- see below for
Tilt sensor data
The tilt sensors were taking data during the
storm. They were located in the dome on the turret floor. When
the turret floor was at 75 degrees, the pitch tilt sensor (+/- 15
degrees) pointed along the azimuth arm and the roll tilt sensor was
pointed perpendicular to the azimuth arm.
Looking at the recorded data we must have used the pitch tilt
I don't have an online version of the turret
location. I'm assuming it was at 75 degrees so the pitch tilt sensor
was pointing along the azimuth arm.
The az, za was 257 az, 8.8 za (stow position with pins in). The
rotary floor is level at za=10deg, so the tilt sensor should have
been reading -1.2 degrees. The dc value was around -5.9 degrees, so
the angle amplitudes are not correct:
- The tilt sensor had been replaced at some point. I am using
the current (pre dome collapse) calibration routine. This data
may have been taken with the older sensor.
- When plotting the data I've removed the dc tilt value so it
will only show changes. The amplitudes may be off a bit. The
frequencies should be ok.
When was data taken:
- 20sep98 22:18 started
- 21sep98 21:36 .. stopped recording data (power problems?)
- 21sep98 23:30 eye passed
- 22sep98 01:00 started taking data again
- 22sep98 20:30 stopped recording
Plots showing the pitch tilt sensor value and
spectra (.ps) (.pdf)
- Top Frame: pitch value vs day of month
- the tilt sensor was sampled at 5 Hz.
- the vertical scale is uncalibrated degrees (see above).
- The x axis is day of month for sep1998 (21,22sep)
- there is no data taken between 21.6 and 22.04
- the pairs of vertical colored lines delimit data used to
- the purple line at 21.95 is when the eye of the hurricane
passed over AO.
- I computed the spectra of the tilt as: spc=abs(fft(pitch -
dcpitch)). (i did't square the result)
- The bandwidth is 2.5Hz.
- Any oscillations greater than 2.5 Hz will alias back into
- I used 4.8 hours of data. I then smoothed the spectra by 9
channels to give a spectral resolution of about .5 milliHz.
- The dashed vertical line is at .8 Hz. The dewars have a cold
head that pumps helium down to the cold head. the cycles at .8
Hz. There are about 10 of these running on the turret floor.
The signal looks strong since the cycling is pretty stable.
- Frame 2: (green) 21.0 to 21.2 (00:00 to 04:56) on 21sep
- This was before any strong winds arrived
- Frame 3: (blue) wind causing a few oscillations
- looks like a few bands of the hurricane were passing over
- frame bottom: (red) strong oscillations.
of strong oscillations (.ps) (.pdf)
- This shows the early morning of 22sep after the eye passed.
- I've flagged and labeled the strongest frequencies:
- .633 Hz this is the widest oscillation
- .683 Hz
- .723 Hz (pretty narrow)
- .8 Hz (this is the crossheads on the floor)
spectra of 22sep98 01:12 to 06:00 (.gif)
- 84 1024point spectra were computed from the tilt data 22sep98
01:12 to 06:00.
- the resulting spectra has a freq resolution of 2.5/512=.005
- You can see the time variation of the strong peaks.
- ignore the .8 hz from the crossheads.
- The pitch tilt sensor on the turret floor measured
oscillation frequencies of:
- .633 Hz (wide and strong)
- .683 Hz
- .723 Hz (pretty narrow).