This master thesis treats the further development of an artificial cloud, which includes a 3D numerical model and field research at the actual prototype. The process aims to produce snow of dendritic structure. Inside the cloud chamber two-component nozzles perform the atomization of water into small droplets. Some of the tiny water particles serve as ice nuclei and the rest of the water droplets convert into water vapour. The water vapour adheres to the ice nuclei so they can grow to snow crystals. For numerical modelling, the ANSYS software package Fluent was used which is suitable for parallel processing and therefore allowed the performance of calculations at the computational clusters of TU Wien and the European Organization for Nuclear Research. Within the master thesis, empirical measurements of the jets, using the medium air, were executed and compared to analytical equations at transonic conditions. ^Results of the measurements showed a high correlation to outcomes of the analytical calculations. Findings were used to define inlet boundary conditions of the numerical model. Furthermore, the work treats the generation and comparison of different meshes, aiming to represent the components and geometry used within the process. An airflow model was setup to compare two numerical solvers, the influence of different near-wall models and turbulence modelling using several Reynoldsaveraged Navier-Stokes-equations (RANS). Results of the comparison show that high near-wall resolutions are not favourable for grid qualities and therefore downgrade convergence behaviour of the model. Furthermore the influence of near-wall treatment to jet spreading needs to be treated carefully. Calculation results could be achieved in about 10 % of iteration steps with the pressure-based, coupled solver, compared to the density-based, explicit solver. ^Also, the k- turbulence models showed faster convergence behaviour, compared to the other RANS models, while results developed the same way after a certain calculation time with all tested turbulence models. Numerical results of the flow field could not be validated on-site because of missing measurement methods and components. Despite to the neglecting of water droplets, ice crystals and water vapour in the numerical model of the cloud chamber, outcomes of this thesis already represent the foundation of a numerical model for an artificial cloud. Further knowledge on the prototype performance relating environmental conditions could be achieved by the performance of field measurements.