Journal of Performance of Constructed Facilities

The writers' firm, Thornton‐Tomasetti Engineers (T‐T), was retained by the city of Bridgeport within hours of the collapse of L'Ambiance Plaza to investigate its cause. As the city's consultant, T‐T had unrestricted access to project documentation, to witness statements, to the collapse site, and to the collapse debris. T‐T personnel were present around the clock at the collapse site during the rescue operation documenting and preserving perishable evidence. This paper is a summary of a detailed report. The paper includes background information, field observations, the results of laboratory tests and computational analyses, and findings concerning the most probable cause of the collapse. Recommendations for reducing the risk of a similar lift‐slab failure in the future are also provided. New physical evidence is presented that may affect the findings of previous investigations. The most probable cause of the collapse was found to be instability of the steel wedges at a particular column. This caused a loss of support to the slabs at that column, which triggered the progressive collapse. This mechanism has not been identified by previously published investigations.

Nonlinear analysis of the construction collapse of the L'Ambiance Plaza lift‐slab structures is presented. During construction, the temporary connection between the columns and slabs is distinctly nonlinear. Due to that nonlinear response, overall structural behavior changes rapidly with horizontal displacement, leading to reduced lateral stiffness and stability of the structure as a whole and large changes in load path with relatively small lateral movements. It is hypothesized that jacking operations intended to plumb the towers provided lateral displacement that altered the stability and load path sufficiently to trigger collapse. It is concluded that failure was the result of one or several specific deficiencies in the design and construction identified by various investigators, and not any inherent shortcoming of the lift‐slab method. Implications of the collapse on issues of professional responsibility, safety assurance during construction, and continued use of lift‐slab construction are discussed.

When a construction‐fall investigation is undertaken, the success of arriving at a conclusion depends on the available information and the expert's judgment to establish the possible and probable causes of the fall. In the first of a two‐part study, fault‐tree models are developed to explain the etiology of construction falls. These models represent the frame of knowledge concerning causal relationships of reasonable and possible causes of falls. Causes are classified as the enabling, triggering, and support‐related causes. Each cause is expanded to reach the basic and conditional causes that contribute to the fall. The analysis is limited to fault‐tree qualitative modeling for construction falls from elevated floor openings. The study resulted in 17 basic causes, four conditional causes, and 28 sets of combined basic and conditional causes that have the potential to contribute to a construction fall. Our study concludes the usefulness of fault‐tree models to systematically and logically represent an expert's knowledge.

In this study, a fault‐tree expert system for construction fall (FTES‐FALL) was developed to simulate and investigate unintentional construction falls from floor openings. Fault‐tree models constructed in an earlier study were used as the knowledge representation for FTES‐FALL. The knowledge base accommodates expertise obtained from experts and literature. A step‐by‐step process in developing the knowledge base was presented. Prompts were designed to lead the user to finding information/evidence needed to reach a conclusion. Parameters were identified and classified, leading to the establishment of frame‐subframe relations. Each frame contains production rules that represent the expert knowledge. In addition, abundant help statements were provided to guide the user through the consultation process. FTES‐FALL was tested by the knowledge engineer for its accuracy, completeness, efficiency, and user‐friendliness. Although FTES‐FALL is still in its development stage, it has demonstrated its capability to assist an investigator in finding the cause(s) of construction falls.