This master thesis is a study of the flow field inside a custom-made hydrocyclone using Computational Fluid Dynamics (CFD) while operating with different flow rates leading to non-optimal conditions. The RNG k-epsilon model is used to model the turbulence and the mixture model is used to model the multiphase and an experiment was carried out to validate the CFD model. The flow field study is carried out with a set Pressure Drop Ratio of 1.37, a fixed droplet size of 400µm, and flow rates from 0.4 to 0.8 L/s and the increasing flow rate showed an increase in separation efficiency. The flow field inside the hydrocyclone turned out to be unlike that of basic hydrocyclone theory; the overflow vortex was shown to origin relatively close to the hydrocyclone inlet and it was traveling along the hydrocyclone wall instead of the center. Increasing the flow rate will develop a longer overflow vortex and will therefore collect more oil from the oil core of the hydrocyclone, which is why the efficiency increases with the flow rate.