The computational time and resources required to calculate an accurate solution is the key concern in the field of CFD. Especially in the CFD analysis of turbomachines many simulations are required to validate the CFD code and to predict the performance of the turbomachines. In this thesis, the typical computational domain was remodelled and the best computational settings were identified to compute the flows. By modifying the numerical domain, improved grid distribution with less number of nodes was achieved and the results predicted were within the limits specified by NASA for the validation of CFD codes. The modified model with the best computational settings required 28.3% less computational time and 20.5% less computer memory than the typical model and baseline methods.