Urban wind simulations are very important for designing smart cities, in which urban heat island is mitigated by designing wind flows, and for improving nuclear security by the prediction of plume dispersion in nuclear accidents. Since a lot of tall buildings and complex structures make the air flow turbulent in urban cities, multi-scale CFD simulations are needed. We develop the CityLBM code based on a Lattice Boltzmann Method (LBM) with a block-based Adaptive Mesh Refinement (AMR) method. The CityLBM code was optimized on GPU-based supercomputers using a communication reduction algorithms, and real time wind simulations were enabled. In this study, we introduced a nudging based data assimilation method for the boundary conditions of temperature and wind, which are given from the Weather Research and Forecasting (WRF) model, and plume dispersion simulations were conducted at more realistic wind conditions. In order to validate the CityLBM code, we analyzed the Joint Urban 2003 (JU2003) field study in Oklahoma City, and details of plume distribution were compared against the experiment.
抄録全体を表示