Intro:
The p50 laser scanner was used to scan the
sides of the towers to see if they are still vertical after
the platform collapse.
Timeline:
- 10aug20: Tower 4auxNorth pulls out of socket
- -6nov20: tower 4 MainCable 4 (southern main) breaks
- 01dec20: platform collapses.l
The tower configuration was/is:
- Prior to the collapse:
- T12,T4 had 4 step backs
- T8 had 6 step backs
- After the collapse
- T12,T4 had 3 step backs (one stepback broke off)
- T8 had 4 step backs (2 stepbacks broke off)
Labeling the faces of the towers:
When the p50 looks perpendicular to
a tower face, it will see 3 flat faces.
The image below labels these faces:
Each color shows the 3 faces that are visible when looking
perpendicular to a face.
- face 1 always points toward the center of the dish
- When looking at the central face, the other two faces are
labeled Left and Right
- The face number increases clockwise from the face pointing
inwards.
- As an example, take tower 12
- face 1 points south,
- face 2 points north.
Processing the point cloud.
The steps taken to process the point cloud
were:
- Scan a particular set of tower faces.
- For each face there will be N stepbacks
- project the points onto the xy, yz, and xz planes.
- remove points that don't belong to the faces
- do a robust linear fit (throwing out 3sigma
outliers) to the tower faces projected on the xy plane
- If there are N stepbacks in the tower, there will be N
projected lines and N linear fits.
- Use the points within 3sigma of the fit as points
belonging to a face/stepback
- rotate the N step back datasets to the x,z plane using the
rotation angle computed for the bottom most step back face.
- eg: after rotation T12 southern face:, x-east, y-north ,
z up.
- For each face/step back generate a histogram in z (with
either .1 or .2 meters bin size).
- Include a reverse_index option so we have ptrs to all
the points in a bin
- compute the average y value for each zbin to see how the
tilt changes as a function of z
- compare the values computed with p50 scans prior to
the platform collapse (if the data is available).
Table of tower scans:
tower
|
date
|
face
|
.las file
|
link to writeup
|
p5 location
|
Notes
|
plots
|
12
|
201209
|
1
|
.las
|
201209
|
visitor center balcony
|
After collapse
|
.ps
.pdf
|
|
200904
|
1
|
.las |
201209
|
eastSide rim roade
|
after aux4North
failure
|
.ps
.pdf |
|
201223
|
3
|
.las |
201223
|
road next to Visitor center
|
after collapse
|
.ps
.pdf |
|
201124
|
3
|
.las |
201223
|
road next to visitor center
|
after aux4N and T4M4
failure
|
.ps
.pdf |
4
|
201228
|
2
2L
2R
|
.las
|
201228
|
southeast side of helipad
|
after collapse
|
.ps
.pdf
.ps .pdf
.ps .pdf |
|
200915
|
2
2L
|
.las
|
201228
|
southeast side of helipad
|
pre collapse
|
.ps
.pdf
.ps .pdf
|
8
|
201111
|
1
1l
|
.las
|
201111
|
west side of helipad
|
pre collapse
|
.ps .pdf
.ps .pdf
|
09dec20: T12 (face 1) scan south
looking face from vc balcony
The scanner was located on the visitor
center balcony looking north to the southern face of T12.
- The scanner was about 10 meters from the base of tower 12
(bottom step back).
- At the step backs, there was some shadowing from the
scanner being so close to the tower.
- There were 163 million pnts before excluding outliers
We compared these results with data taken 04sep20
before the collapse
- The scanner was located on the east edge of the rim road
(about 250 meters away)
- there were a lot fewer points on the tower since the
longer distance was done at at lower resolution.
After rotation the coordinate system was:
- X - east
- Y - north
- Z - up
- + angle is CCW looking down the Z axis.
The first plots shows the
scanned data results for 9dec20 (.ps) (.pdf)
- The y value is averaged over z bins of .1 meter.
- The scanner was about 10meters from the base of the bottom
step back.
- Page 1: face rotation relative to bottom step back
- Each step back face was rotated to the xz axis. The
plots shows the angle difference relative to the bottom
step back face.
- Positive angle is CCW looking down from the top
- Page 2: y value vs z. 0 is defined by the mean
value of the lower face.
- top: Y value vs z . + y is northward.
- you can see the 3rd step is bowed a bit.
- it starts out south, bends northward (inward) and then
comes back out.
- Bottom: y value vs z, with face mean removed.
- This lets you see the deviation with finer detail (x
axis is now cm)
- Since the mean is removed from each step back face, the
transitions between stepbacks won't line up.
- the 3rd step back has a variation of 7 cm.
The 2nd plot shows the scan of
the T12 southern face on 04sep20 (pre collapse) (.ps) (.pdf)
- the y value is averaged over .2 meters.
- the scanner was 250 meters from the base of the bottom
step back.
- T4 auxNorth cable had already fallen, but the T4 M4 cable
was still attached.
- Page 1 face rotation relative to bottom face.
- The face angles are rotated CCW relative to the 9dec20
measurements.
- The 2 aux main cables from tower 12 had a large tension
difference.
- The west cable tension was much higher than the east
aux cable tension (about 400 kips vs 720 kips)
- This rotated the tower CCW. especially the upper
section
- Page 2:
- The data was binned to .2 meters in z (there were fewer
points).
- Top: Average y position vs z
- Bottom: average y value vs z with face mean value
removed.
09dec20 Summary:
- The faces are rotated by less than a degree relative to
the bottom face
- sep20: the rotation goes 0 to -.3 degrees and then jumps
up to .85 degrees (CCW)
- The tower was being rotated by the aux main cable
tension difference (from t4auxN failure)
- the largest rotation occurred at the upper step back
- I wonder if the rotation was part of the reason why
the 2nd to 3rd transition had cement failure during the
collapse
- dec20:
- It looks like the entire tower unwound after the
cables were released and the top section fell off.
- Y deviation vs z
- The 3rd step back has a y variation of 7 cm
- This was seen before and after the collapse. This may
have been present from the start.
processing:
x101/p50/towers/t12/f1south/getdata_20xxxx.pro,
plotit_20xxxx.pro
23dec20:T12 North looking face (face 3)
Data was taken from the road next to the
visitor center. The same location was used on 24nov20 prior to
the collapse (the 24nov20 spot was about 20 feet farther
away).
The plots show the results of the measurements:
The results are similar to the south looking face. There is
an unwinding of the tower after the collapse.
processing:x101/p50/towers/t12/f3north/getdata_20xxxx.pro,
plotit_20xxxx.pro
Tower 4
28dec20: T4 face 2 from SE side of
helipad
Data was taken from the se side of the
helipad looking toward tower 4 (face 2 facing NE).
- The 201228 position had a little better view of the
tower than the 200915 spot.
- 201228 T4 post collapse
- 200915 T4 pre collapse
- The face 2L had large rotation angle for upper steps backs
before and after collapse.