In the autumn of 1996 ITI installed
a remote monitoring system to perform a four-month-long test
on the tied arch span of the Milwaukee Harbor I-794 (Hoan) Bridge
in downtown Milwaukee, WI. The sensors for this test consisted
of both strain and temperature sensors. The objective of the
test
was to determine if thermally driven stresses were sufficient
to drive fatigue cracks that were discovered in the structure.
The
cracking was discovered in the tie chords in the area that is
penetrated by the arch. Prior to the remote system installation,
strains and
acoustic emission were monitored in the vicinity of the cracks
during controlled load testing. The load tests showed that live
traffic testing was insufficient to account for the crack growth.
The remote system allowed us to conduct a test over an extended
time period during which the bridge experienced wide temperature
excursions while avoiding the need to send engineers out to the
bridge on a periodic basis. This test expanded the technology
that was used for the Michigan Street bridge. A total of 18 strain
gages
and 9 thermocouples were installed on a section of the main tie
chord of the bridge. The sensors were spread over a section of
the box girder that was approximately 200 feet in length. Three
Somat field computers that provided the signal conditioning and
data logging functions were networked and distributed over the
200-foot length using a hard wired network link. The Somat "Global
Data Link" (a hardened PC and a cellular telephone/modem)
was used to provide remote communication. The system operated
flawlessly for four months in the autumn and winter of 1996/1997.
The data
indicates that thermally induced stresses are probably the primary
driving force for the fatigue cracks.
On December 13, 2000, two
girders in a three-girder approach span on the Milwaukee Harbor
I-794 (Hoan) Bridge failed. This
failure occurred in a completely different location that
the previous test site and was un-related to the arch problem.
A major effort to determine the cause of the failure was
initiated
by Wisconsin DoT. The failure analysis effort involved Lichtenstein
Engineering Associates, Lehigh University, The Federal Highway
Administration, The University of Michigan and Northwestern
University's Infrastructure Technology Institute. The failure
analysis utilized the wide range of experience and expertise
of bridge experts from the above mentioned institutions and
involved both analytical and experimental techniques including
metallography, 3-D finite element modeling, materials testing,
and load tests. The conclusions of this massive effort indicate
that the primary cause of the failure was a combination of
very high tri-axial stress combined with lowered material
properties caused by low temperatures. The high tri-axial stress
results
from the geometry of the connection detail used for the lateral
bracing system for the Hoan Bridge. The resulting high stress
in the "web gap" of this detail produces a condition
where a ¼-inch crack becomes un-stable and can lead
to explosive failure. The retrofit design consists of removal
of the connection detail. |
