SAC Phase 1 Analytical Studies of Building Performance
Project Title:
Elastic and Inelastic Analysis for Weld Failure Prediction of Two Adjacent Steel Buildings
Sub-contractors:
Gary C. Hart, UCLA/Hart Consultant Group
Sampson C. Huang, Hart Consultant Group
Roy F. Lobo, Hart Consultant Group
Joe Stewart, KPFF Consulting Engineers
Project Summary:
Two adjacent building near the epicenter of the January 17, 1996 Northridge earthquake (one
with weld fractures and one without weld fractures) are considered in this study. Both buildings
are six stories in height, and have their primary lateral force resisting system comprised of
special moment resisting frames above grade.
The buildings considered are referenced in this
report as the East Building and the West Building. An elastic three-dimensional analysis using
SAP90, and an inelastic two-dimensional analysis using DRAIN-2DX is performed. Three different
models are created for the elastic analysis: two for the East Building, designated as the
Benchmark Model and the Probable Model, and one for the West Building, designated as the West
Building Model. The East Building was unoccupied at the time of the earthquake, hence two models
are created for this building. The Benchmark Model is representative of the design state of the
building, while the Probable Model is representative of the state of the building at the time of
the earthquake. The West Building, which was fully occupied at the time of the earthquake, was
modeled similar to the Benchmark Model. Elastic and inelastic studies are performed on these
models to determine if a correlation exists between analysis and observed behavior.
The findings from the analysis showed no damage to the West Building with possible damage to
the East Building. This result is corroborated by the damage survey. The analysis results,
however, were unable to predict the locations where damage to the connections may be expected in
the East Building. From the analysis performed, the maximum demand capacity ratios for the N-S
frames showed potential for weld failures at all floors above ground which is consistent with the
findings of the damage survey. No damage is predicted at the ground floor due to the low demand
capacity ratios from both the elastic and inelastic analysis. Unfortunately, large column cracks
were observed at this level. Another inconsistency is the number and locations of damage
occurring in the E-W frames of the building when no damage potential is shown from the analytical
results. The damage to the East Building is random, and this is confirmed by unequal scatter of
damage locations on identical frames though the building is symmetric. The results are
unsuccessful in predicting damage to the welded connections and further research is necessary.
