CFD Simulations of Mixing using a Magnetically Driven Impeller- Method Development and Validation Study
Ladda ner
Typ
Examensarbete för masterexamen
Master Thesis
Master Thesis
Program
Applied mechanics (MPAME), MSc
Publicerad
2019
Författare
Olofsson, Caroline
Sundell, Emma
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
Mixing with impellers is common in many industries. By using CFD methods, more knowledge about the mixing process can be gained since it is relatively difficult to examine experimentally. Two common turbulence models used for active mixing are the RANS and LES models. RANS methods have the advantage of being relatively cheap in terms of computational effort, whereas the more demanding LES method usually results in higher accuracy. FS Dynamics has performed CFD simulations of active mixing using a RANS method. In this study, the purpose was to improve and validate this existing method. The method has been extended by implementing LES as turbulence model. The LES was implemented together with wall functions to lower the computational demand. A tracer, inserted using two different methods, has been used for tracking the mixing processes. The first method replaces some water inside the tank with a fixed volume of tracer. The second method has an inlet of tracer to mimic pouring from physical tests, which were also conducted during this project. Both methods were used for the RANS simulations whereas only the second method was used for LES. Using a fixed volume of tracer placed beneath the surface resulted in a mixing process starting with a higher concentration at the top, spreading downwards. Using an inlet of tracer resulted in a higher tracer concentration in the bottom of the tank, which was then mixed upwards. Regarding the mixing time, the LES gave results that correspond to the physical tests and the RANS model overpredicted the mixing time by 50%. The robustness of the method using a fixed volume of tracer was evaluated by changing both the volume of tracer and its position and shape. Changing the volume from approximately 3 to 25 litres did not affect the mixing time substantially. Using different shapes and positions resulted in varying mixing time results. Changing the Schmidt number between 1 8000 or the dynamic viscosity between approximately 0:9 10 3 1:7 10 3 kg/ms did not affect the mixing times.
Beskrivning
Ämne/nyckelord
Strömningsmekanik och akustik , Hållbar utveckling , Livsvetenskaper , Fluid Mechanics and Acoustics , Sustainable Development , Life Science