Data-driven modeling and uncertainty quantification in infrastructure science and engineering

Few centuries ago, building palaces and castles were based on past experiences and observations, hence, resulted in very inefficient designs with wall sometimes as thick as 1~2m. Today, computer modeling provides a low-cost way of predicating what may happen when engineering structures are subjected to different loads and conditions. We know that some loads are such as weight of the structure, however, others like wind can have a probabilistic nature. Considering all possible combinations of probabilistic loads or parameters demand many computer simulations which was not feasible before cloud computing. Combined with today’s availability of low-cost sensor such as thermostats and accelerometers, this project is set to investigate next generation of simulation tools that can account for probabilistic loads and sensor readings. This type of simulations can not only better quantify safety of a structure during the design phase, but also help to monitor health of structures and plan their required maintenance.

Faculty Supervisor:

Reza Vaziri

Student:

Ehsan Haghighat

Partner:

Intelligent Simulation Technologies

Discipline:

Engineering - civil

Sector:

Construction and infrastructure

University:

University of British Columbia