The magnitude 7.0 January 12, 2010, Haitian earthquake severely damaged several modern engineered buildings in the capital city of Port-au-Prince. This research project will gather perishable data on the observed performance of these structures for the purposes of comparison with, and calibration of, analytically-based response predictions used in seismic assessment procedures. The research is a partnership between academia and industry with university researchers working alongside professionals from the earthquake engineering consultant firm, Rutherford and Chekene. The project relies on close collaboration with a faculty member at the Université de Sherbrooke, Canada, who has established contacts and first-hand knowledge of the conditions on the ground in Haiti. The objectives of the research are to: (1) catalog and classify damage for a class of structures in the Haiti earthquake; (2) correlate damage to structural system, layout, design and detailing; and, (3) evaluate the efficacy of analytical techniques and procedures used in current assessment. The expected research outcome is a report documenting: (1) structure/site description; (2) design information; (3) documentation of post-earthquake state; (4) description of analytical techniques and model; (5) analytical results and comparisons; and (6) evaluation of the methods. The research will involve the documentation of the design, site conditions, and observed damage on a small set of important engineered structures (schools, institutes, government buildings), focusing on reinforced concrete construction. Candidate structures have been identified, all within the capital city of Port-au-Prince, on the basis of damage state, available design information, and accessibility. Seismic performance of the candidate structures will be obtained primarily from on-site visual inspection in Haiti by the research team and collaborators. Included in the inspection will be failure modes, damage intensity and patterns, and residual state. The available information will be used to construct analytical models using state-of-the-art techniques for nonlinear dynamic analysis. Seismic hazard levels will be estimated by indexing to the level of demand required to bring about the observed damage. In this process, the analytical models will be subjected to incremental dynamic analysis in order to map the observed damage to reference amplitudes available from USGS shake maps. Then, predictions of the observed response can be conducted using accepted seismic assessment techniques common in the U.S. (linear, nonlinear, static, and dynamic) and compared to the observed actual performance. In particular, the effectiveness of the structural models to produce similar behavior modes, overall patterns of damage, and appropriate assessment decisions will be evaluated. Conclusions on the efficacy of these methods will be drawn through comparisons of the predicted response to actual outcomes. Preliminary evaluation of the sensitivity of the response to various parameters and model calibration will be performed. Where possible, the sensitivity of the predictions to different variables and estimations of uncertainty will be made. This work gathers perishable information on building performance in actual seismic events and has the potential of providing information on a key step in seismic assessment, namely calibration of models. This research additionally extends knowledge on seismic performance of reinforced concrete structures in earthquakes. The potential impact of this research is increased public safety and more informed decision making pertaining to infrastructure retrofit and upgrading through improved analytical models and assessment techniques, as well as improved retrofit evaluations. The information provided by this project can be used widely by earthquake engineering researchers. Project results will be made available to the profession through the industry partners' participation in professional groups and committees. Project investigators will also present results of the study at a workshop for Haitian earthquake RAPID awards in summer or fall of 2010.
" involved an earthquake reconnaissance trip to Port-Au-Prince Haiti in May 2010 to document the damage state of engineered building structures that were heavily damaged yet did not collapse. The team was composed of academic researchers and practicing seismic design structural consultants. The term "engineered" refers to buildings that went through some form of a design process, as opposed to many structures in Haiti that are simply built based on local practices. Among the handful of structures that met these criteria, four candidate structures were selected for survey based on the availability of structural blueprints and ability to gain access to the site. Comprehensive annotated photographic damage surveys were produced from detailed floor-by-floor observation of each structural element in the building. The primary objective of the trip was to document and archive perishable data on the performance of engineered buildings built according to practices in underdeveloped countries such as Haiti. These photographic surveys are available to researchers and the public at http://civil.arizona.edu/robert-b-fleischman. After the trip, computer models of these buildings were created based on blueprint information and measurements taken, and using test data and accepted estimates on the typical strength of Haitian building materials. The models were subjected to computer simulations intended to represent the January 2010 earthquake event. These computer simulations followed accepted methods employed by structural design consultants when assessing a building’s risk to a future earthquake event, and were performed as if the actual damage that occurred in the Haiti earthquake was unknown. Then, the computer simulation results were compared to the actual damage observed on a member-by-member basis, in terms of intensity, type and location. A classification system was used to relate the computer simulation numerical results to the visual data from the building photographic survey. The primary objective of this research phase was to evaluate the state-of-the-art software packages used to assess existing buildings by researchers and practitioners. A key aspect of the research was the estimation of the January 2010 Haiti ground motion. Unlike other earthquake events where seismographic records are available, no seismographs were deployed in Haiti. Thus, this project evaluated ground motion estimation in data-scarce environments, and examined the sensitivity of assessment methods to knowledge of the earthquake. The Haiti earthquake was estimated by matching similar historic earthquakes based on projections by the U.S. Geological Survey, oriented relative to the buildings based on the Haitian fault, and used a simple shade canopy structure near one of the buildings with clearly defined damage to calibrate the motion. The key observations of the reconnaissance trip include: Low to mid-rise commercial buildings (offices, ministries, hospitals, hotels, banks, shopping centers) and low rise residential buildings (one to two story homes) suffered significant and widespread collapse. The buildings were typically reinforced concrete frames with or without masonry infill walls, or masonry bearing wall buildings. Given 150 years since the last devastating earthquake in Haiti, many buildings seemed designed for gravity loads only. These buildings were typically dominated by heavy floors and lighter columns, often with insufficient steel ties to keep the column intact; a potentially lethal combination for moderate to strong ground motions. The material used to construct these buildings was in general poor, including weak concrete and unreinforced masonry using fine aggregate containing calcium carbonate; smooth reinforcing bars, and uneven concrete fill. Most buildings employed heavy unreinforced masonry partition walls (as opposed to drywall) that attracted significant seismic forces, but are quite weak. In cases, these in-fill walls were partial height, inducing a shear failure of the adjacent column above the wall. A significant obstacle to rebuilding is the ubiquitous presence of large amounts of rubble. A concern is that without training, the same poor materials will be used to rebuild structures in the same manner as previously. The major findings of the post-trip research include: The state-of-the-art computer assessment tools for earthquake simulation, which model the building three-dimensional geometry and possible failure modes, were able to reproduce most of the damage observed in the structures in terms of intensity, type and location. The research methodology developed, including ground motion calibration for an unknown event, and quantitative metrics for observed damage, seemed effective for these evaluations. The modeling of partitions, not typically considered structural elements, was a key part of the Haitian building simulations. These stiff, weak, brittle elements are typically viewed unfavorably. However, for poorly constructed frames, these elements may prevent collapse. As most loss of life in this earthquake was associated with collapsing floors in frame structures, more study is needed on the role of masonry partitions in these collapse-critical structures. The PI participated in an international workshop on issues facing rebuilding Haiti and is working with student groups on a Haiti service project.
