Axial flow fans are often the largest contributor to the noise emitted from commercial refrigerator systems. This has driven the demand for quieter, yet higher performance axial flow fans. Turbulent fluid motion and aerodynamic forces interacting with the fan blades generate noise as the fan rotates. The aerodynamic noise emitted from a fan can be modelled with varying degrees of accuracy which are typically proportional to the computational cost. Application of a fan performance model to an optimisation algorithm can be used to develop new and higher performing fan blades. However, there is currently no rapid method to generate an optimal fan design of specified performance. The method developed in this project was experimentally validated and shown to be able to predict the flow rate accurately; however, the fan noise prediction was unsuccessful. Noise source identification showed that the dominant fan noise source was incident turbulence noise which was neglected in the aeroacoustic modelling and explains the poor noise prediction. Knowing the dominant noise source allows for future modelling to include all relevant physics which will allow for more accurate modelling that will increase the performance of axial flow fans and reduce refrigeration noise. This study also showed that fans with aerofoil blade shapes have superior performance to fans with cambered plate blades, which are commonly used in commercial refrigerators. iii