Robust statistical damage assessment of infrastructures

Structural health monitoring is regarded as the main tool in assessing the functionality of existing structures. The importance of these techniques emerges by considering that failure of an infrastructure results in catastrophic loss. With such techniques, damages in a structure can be detected, before reaching dangerous levels. Those are purely based on measurement data and no structural models are required. In recent years, a statistical framework for damage detection has been developed with significant success on real structures. In order to infer the location of a damage after its detection, a structural model is required. To achieve an efficient damage localization method, interaction between signal processing specialists at Inria and civil engineering experts at UBC is required. Models of investigated structures will be coupled with the previously developed statistical approach in this research. The objective is to develop the necessary tools to bring damage localization into practice, aiming at a robust method for real applications such as bridges and tall buildings.

Faculty Supervisor:

Carlos Ventura

Student:

Saeid Allahdadian

Partner:

Discipline:

Engineering - civil

Sector:

University:

University of British Columbia

Program:

Globalink Research Award