Leibniz MMS Days 2023 - Abstract
Janke, David
Estimating emissions and barn climate in animal housing systems is necessary and it requires the usage of multiple sensors to measure and monitor various parameters like velocity vectors, temperature- and gas concentration fields. An alternative approach to reduce the amount of acquired information is to combine sensors with computational fluid dynamics (CFD) to make real-time predictions. Digital twins and highly efficient solvers are required for this task. The lattice Boltzmann method (LBM) is an explicit solver for the incompressible Navier-Stokes equations, which has been demonstrated to be highly scalable and suitable for massive parallelization, and it is hence a promising candidate. The purpose of this study is to benchmark and validate a lattice Boltzmann solver based on the open-source OpenLB library for simulating flows around typical livestock housing systems. We conduct reference measurements in an atmospheric boundary layer wind tunnel to validate the simulations' accuracy. We also investigate the solver's efficiency and speed gain through GPU acceleration. Our results indicate that the OpenLB solver is a feasible, efficient tool for simulating airflows in livestock housing systems.