Table 5: Column tilts.
Column tilts (cm)
Average
July
September February August
Average
direction
()
Column
Direction
1987
1987*
1988
1988
resultant
A4
Along
32
24
33
33
38
306
Across
24
19
21
25
A3
Along
33
33
30
35
42
308
Across
27
22
24
29
A2
Along
27
22
17
23
36
322
Across
32
24
27
29
A1
Along
19
14
15
16
32
330
Across
30
24
26
29
N4
Along
37
29
20
30
33
298
Across
17
11
16
16
N3
Along
23
25
20
28
29
305
Across
22
14
14
17
N2
Along
15
15
17
19
26
321
Across
18
17
23
23
N1
Along
15
9
20
15
25
323
Across
25
13
17
24
Average
Along
25
21
22
25
32
313
Across
24
18
21
24
32
Avg. resultant (cm)
35
28
30
35
32
Avg. direction ()
314
310
314
314
313
Avg. tilt (radians)
0.019
0.015
0.017
0.019
(0.017)
*Middle section of column only.
displacements divided by the measurement
DISCUSSION
length appears in the data figures. A plumb bob
During the sway bolt surveys in 1987 and 1988,
hanging from a magnetically mounted reel was
concerted efforts were made to remove any pinch-
used in conjunction with a measuring tape
ing or spreading loads at each collar level. These
to find the vertical position of the data points as
opposing loads serve no structural purpose, so in
well as the two horizontal displacements. The
those years, most were eliminated. Sway bolts all
data for September 1987 were for the section be-
around the column collars were brought to bear
tween the building and enclosures only.
against the column to ensure no looseness in the
Because so few data are available for the col-
structure after pinching and spreading loads were
umn tilts, it is difficult to draw any hard conclu-
relieved.
sions. Several points are worth noting from the
data that are available, however. The first is that
On some collars, normal sway bolt measure-
the columns are tilting in the same general direc-
ments could not be taken. The reasons were either
tion and with the same magnitude as the pro-
bad sway bolts that could not be turned (heads
gression of the footing and column settlements
missing, weld metal on threads, distorted threads,
for all data sets. A second point is that the magni-
etc.) or sway bolts that had been turned all the
way into their nut blocks. When this occurred, a
tude of the lean seems to correlate with the base
load-deflection measurement had to be performed.
level stress factors (Table 5 and Fig. 12). The data
To conduct a measurement, a deflection is meas-
for September 1987 are not complete, which may
be why the data do not correlate as well as that of
ured for specific loads applied between the column
August 1988. A third point is that after an initial
and collar opposite the dial indicator. An example of
rapid increase in tilt following the shimming
a load-deflection measurement is shown in Figure
19. Although not highly accurate due to the influ-
operation in 1987, the rate of tilt leveled off to an
ence of such uncontrollable factors as friction loads
average of about 0.3 in. per 10 ft/year (0.25 cm/
and the effect of adjacently loaded sway bolts, this
myr) (Fig. 18). We do not have a large enough
measurement procedure gives approximate loads
database to confirm points 2 and 3, but the trends
through linear extrapolation sufficient for analysis.
are present and worth mentioning.
14
Previous Page
