Page 214 - Forensic Structural Engineering Handbook

6.16 ENGINEERING RESPONSE TO FAILURES

Full-scale load testing of recreated subsections of the structure are sometimes used to
evaluate possible failure sequences and determine the actual collapse loads.
If geotechnical or foundation issues are relevant to a structural failure, both laboratory
and in-situ testing of soils may be necessary and the structural analysis may need to take
into account soil-structure interaction.
The American Society for Testing and Materials has a number of publications, including
Refs. 10 through 18, on the standard methods of testing structural components and assemblies.

CONCLUSIONS OF THE INVESTIGATION

Identifying Possible Causes of Failure
The basic approach of the failure investigation is to identify at what load individual struc-
tural elements reach failure and what sequence of individual element failures leads to col-
lapse. Typically, a number of different scenarios emerge that could lead to collapse. Each
needs to be further compared to the documents, field observations, laboratory testing, and
possible in-situ testing to determine all likely collapse scenarios. This process is iterative
and differing possible outcomes will emerge over time.

Determining What Caused the Failure
Initially, it is important to identify all possible failure scenarios. Subsequently, facts gathered
from the observations, documents, analysis, and testing should be used to eliminate unlikely
scenarios, and give some weight to the likelihood of all remaining scenarios. Knowledge
gleaned from this process will lead to further areas of analysis and testing that hopefully will
lead to reducing the list of possible failure scenarios. The goal of this iterative process is to
converge on a small set of factors that are the most likely technical causes of the collapse.
Ultimately, the facts unearthed in the investigation must support the likely technical causes of
failure identified by the engineer. The key to all structural failure investigations is to compare
actual loads to in-place structural capacities and behavior (Fig. 6.12).

FIGURE 6.12 Collapse of the Hartford Civic Center roof in January 1978
occurred under an actual snow load of about one-half the code-required snow load,
thereby pointing to causal factors other than extreme overloading of the structure.