This project aims to realize an interactive scenario-based simulation framework with integrated microclimate simulation, data analysis, and visualization for decision support in urban planning. We want to enable workflows for planning mitigation measures interactively during a live session. This requires significant advancements in the execution of and interaction with simulation scenarios and insightful visualization of the resulting data, which we will research in this project.

One goal of this project is designing a highly efficient urban microclimate model system and its integration into a dynamic data flow system enabling interactive simulation steering. We seek to adapt PALM-4U, the state-of-the-art climate model system PALM 6.0 for applications in urban areas, to multi-GPU systems in order to combine its high reliability and functionality with the high computational performance needed for interactive workflows. For even higher performance in tasks that do not require the highest accuracy, such as parameter space exploration, we aim at a surrogate model based on a neural network to provide near-instant feedback.