Data-Driven Multiscale and Multiphysics Cloud Framework for the Prediction of Long-Term Behavior of Mass Timber

Each year, the global demands of construction cause the consumption of three tons of concrete per person. In the United States, the building sector alone accounts for nearly 40% of the nationÆs human-produced greenhouse gas emissions (GHG). Efforts to mitigate climate change must necessarily include a shift in focus within the engineering field to minimize this environmental impact. Timber has recently regained interest as a primary construction material, as production requires less energy than either steel or concrete, and wood contains sequestered CO2, lowering a buildingÆs cradle-to-grave environmental impact. Mass timber structures are also faster to construct in comparison to concrete, and thus a well-designed project adopting timber products can have both economic and environmental benefits. However, wood takes time and space to grow, and thus lumber regulations have placed limits on how many trees may be harvested each year. This in turn limits the timber supply such that it cannot meet the demand for structural materials each year. Furthermore, if available supply is used only in 100% timber structures (as opposed to hybrid structures), and concrete structures supplied the remaining demand, then the inefficiency of timber in certain structural capacities would lead to more concrete be used for the remaining buildings than if hybrid structures had been used. It is therefore clear that hybrid structural systems are necessary for modern sustainable structures. Hybrid systems though pose unique design considerations. Due to the disparate behaviors of wood and concrete or steel under sustained stresses, such systems are highly susceptible to differential shortening.ĀTo address the problems discussed, the main goal of the proposed Phase II work is to develop a data-driven multiscale and multiphysics cloud framework for the prediction of long-term behavior of mass timber. Towards this goal, the specific technical objectives of the proposed research are (1) To formulate and validate a multi-scale model for moisture transport, heat transfer, and creep that takes into account the morphology of wood internal structure based on the existing Connectors & Beam Lattice (CBL) model; (2) To obtain experimental creep data of wood and timber products at multiple length scales; and (3) To implement the validated framework in the open-source software Project Chrono and to deploy the framework in the Azure Cloud for integration with machine learning and digital twin applications.Ā