Research Area: Urban Systems Science

Exascale to enable new planning capabilities for evolving urban landscapes

This image from the NEK5000 simulation shows how turbulence and heat convection in the urban canyon affects building energy use. Courtesy: Rob Jacob and Alex Obabko, Argonne National Laboratory

More than 50 percent of the global population lives in urban areas and that number is rising. As city populations grow, so do demands for energy, water, transportation, healthcare, infrastructure, and physical and cyber security. At the same time, new technologies, such as sensors and mobile applications, are making it possible to provide real-time data on urban systems.

Many urban systems—from buildings to transportation to energy grids—have been modeled effectively, with mature computational modeling codes available. But these systems have interdependencies that are either not considered or are approximated as fixed boundary conditions.  By interconnecting these models, enabling them to exchange data representing their interdependencies as the models unfold, it will be possible to develop much more realistic forecasts of the impact of changes in areas such as infrastructure, technology, weather patterns, or particular patterns of growth.

By developing methods for coupling these models with operational data, new capabilities will be available to cities and utility providers, for instance, to reduce energy costs and stabilize supply by monitoring renewables on the grid and forecasting building energy consumption. But researchers first need to develop a framework by which these various urban systems models can exchange relevant data.

With such a framework it will be possible to execute multiple urban systems models in concert, resulting in much more holistic forecasts.  This coupled modeling approach will require exascale computing both because multiple models must run together and because integrating their interdependencies will increase the complexity of each of the models. To tackle the large-scale data challenges of developing models that will work for cities, researchers need big compute power and the next-generation computers that are designed with coupled models and data analytics in mind.