|M.Sc Student||Tatyana Bloch|
|Subject||Towards Building Information Modeling of Damaged Buildings|
to Guide Search and Rescue Operations
|Department||Department of Civil and Environmental Engineering||Supervisors||Full Professor Sacks Rafael|
|Full Professor Rabinovitch Oded|
|Full Thesis text - in Hebrew|
After an earthquake in an urban area, the number and location of collapsed buildings are unknown, as is the number and medical state of the people trapped inside these buildings. Trapped occupants are most likely to be found in void spaces that are naturally formed in the ruins. Information about the structural elements, their sizes, composition and material, and their new locations in the damaged building, would enable SAR teams to search selectively and thus minimize the risk and accelerate the search process.
This preliminary study suggests a method for gathering new information about the collapsed structure and examines and demonstrates its capability by means of an initial feasibility assessment. The suggested method uses the BIM model of the original building, in a suitable "collapse engine" to create different possible collapse patterns of the building. A series of parametrically differing ground movements is applied to the model to trigger a series of potential collapse mechanisms. The result is a large set of candidate solutions, which contain the exterior and interior geometry of the deformed state of the building.
A procedure developed by Zeibak-Shini et al. (2012) provides real post-damage information about the exterior components of a damaged building using terrestrial laser-scanning. In the research reported here, each candidate solution in the solution set is evaluated using an appropriate mathematical algorithm to provide a quantitative measure of the degree of similarity of each candidate solution to the model generated from the scan. The solution with the highest degree of similarity should be the one that provides the best information on the damage pattern of the exterior and interior geometry of the building.
In each of three simulations, the as-built model was imported to the “collapse engine” and the space of candidate solutions was generated. A single candidate solution was randomly chosen to represent the real damage pattern of the building. Because it was created artificially, this model contains both the exterior and interior geometry. The rest of the candidate solutions can be evaluated in two ways; once relying on information about exterior elements only, and once relying on information about the whole building. In all three simulations, the same candidate solution was identified at the highest degree of similarity when calculated based on external elements only, as well as when calculated based on the whole building. This demonstrates that the proposed method is feasible, and encourages further development with the goal of guiding SAR teams.