Page 271 - Forensic Structural Engineering Handbook

DESIGN ERRORS, CONSTRUCTION DEFECTS, AND PROJECT MISCOMMUNICATION 8.11

all glass, concentrating the building’s lateral load-resisting systems in the central core.
Torsional response of the building enhances motion-induced accelerations that the occu-
pants experience, and produces more pronounced visual cues of building motion which in
turn are more perceptible by the occupants. Although the threshold of motion perception
varies from person to person, the tangential accelerations produced by a building’s tor-
sional response are additive to the linear acceleration components, generating higher levels
of uncomfortable accelerations. Annoying motion-induced accelerations occur at service-
level loads at which inherent structural damping, which is amplitude dependent, is also
smaller. Several options for mitigating such a problem are generally available, such as
increasing building’s stiffness, mass, or damping. Each option has to be considered and the
effect of the proposed retrofit on the building’s response to other load types has to be con-
sidered as well. For this building various damping schemes were explored to accomplish a
reduction of motion. Three kinds of supplemental damping systems—viscous, viscoelastic,
and friction—were analyzed. The final design employed 40 viscous dampers, which did not
induce a spring type force and had a good track record under high cyclic loading, and pro-
vided damping for a wide range of wind loadings.
Our current engineering practice recognizes the complexity of wind dynamics, and our
standards recommend the use of boundary-layer wind tunnel testing for high-rise buildings
of irregular shape, for structures whose flexibility makes them subject to across-wind or
torsional loading, vortex shedding, instability due to flutter or galloping, and for buildings
subject to either buffeting by the wake of near-by upwind buildings or to accelerated flow
caused by channeling or local topographic features. 3
Did our scientific knowledge finally conquer this formidable foe? Not quite, as our
recent experience indicates.
A month after the tragic events of September 11, 2001, Weidlinger Associates was
retained to investigate the question of whether the collapse of the south World Trade Center
Tower (2 WTC) caused or contributed to the collapse of the north WTC Tower (1 WTC), as
part of the ensuing insurance litigation that became known as the “one or two occurrences”
litigation. One of the arguments raised by the insurance companies’ engineering experts in
defense of the “one occurrence” assertion was that, in the hypothetical scenario of a single
airplane attack and a single tower collapse, the remaining tower would not be viable
because of the loss of the wind shielding effect of the adjoining tower–an effect for which
it had not been designed. This argument led to the most extensive wind tunnel tests probably
ever undertaken for any structure, involving the three best known wind tunnel laboratories
in North America: Cermak Peterka Petersen (CPP); Rowan Williams Davies and Irwin,
Inc. (RWDI); and the Alan Davenport Boundary Layer Wind Tunnel Laboratory (BLWTL)
at the University of Western Ontario. CPP’s report indicated design overturning moments
between 40 and 70 percent larger than those obtained by RWDI! Litigation issues aside,
how is a structural engineer to design a building, if such discrepancies can exist in the eval-
uation of design wind loads? We retained BLWTL to undertake a third independent study
and to explain the differences in results in the prior studies. BLWTL noted that the mathe-
matical techniques used in the combination of wind tunnel and climatic data involve com-
plex statistical techniques that are still evolving. These mathematical techniques essentially
translate the building model data measured in the wind tunnel to the wind-induced response
of the prototype by considering the full-scale characteristics of wind speed and direction at
the building site. Since the basic wind tunnel data measured at all three laboratories com-
pared well with each other, the difference in the statistical treatment used in each laboratory
was determined to be the source of the differences in the design loads. RWDI’s “upcrossing
method” compared well with BLWTL’s Monte Carlo-based “storm passage” technique, but
differed substantially from CPP’s directionally modified design wind speed or “sector”
4
approach. All these reports and analyses were turned over to the National Institute of
Standards and Technology (NIST) for consideration in its WTC investigation. NIST’s