The paper addresses the current state of the development of a computational steering environment (CSE) for interactive indoor thermal comfort simulation by utilizing high-performance supercomputing facilities. The CSE consists of a parallel CFD kernel, a fast spacetree-based 3D mesh generator and an integrated virtual reality-based visualization engine. The numerical method is based on a hybrid thermal lattice Boltzmann (LB) method with extensions for large eddy simulations of turbulent convective flows. We use a multiple-relaxation-time LB scheme for solving the mass and momentum equations numerically and a finite difference scheme for the heat equation. The CSE allows for modifying both the geometric model and the boundary conditions during runtime with immediate visualization of changes in the results. The application is demonstrated by two industrial applications with complex geometries, turbulent natural convection in the separator room of a ferry boat and turbulent convection in a train's passenger carriage. We currently enhance our model using a radiosity method with a fast spacetree-based visibility check and integrate a new local thermal comfort model developed by our partners.     «

The paper addresses the current state of the development of a computational steering environment (CSE) for interactive indoor thermal comfort simulation by utilizing high-performance supercomputing facilities. The CSE consists of a parallel CFD kernel, a fast spacetree-based 3D mesh generator and an integrated virtual reality-based visualization engine. The numerical method is based on a hybrid thermal lattice Boltzmann (LB) method with extensions for large eddy simulations of turbulent convecti...     »