Gas turbines consume a large volume of ambient air during in-service condition. The quality of air incoming the system is essential to the performance and longevity of a gas turbine. A filtration mechanism is employed to regulate air quality by removing contaminants. For high-powered turbines, the filter system can be composed by hundreds of single filters. Therefore, to achieve the best performance, a carefully design is needed. However, a correct sizing of the filtering houses is often complex. Their dimensions and manufacturing costs make design and experimentation difficult. Numerical modeling has been playing a crucial role in the design and optimization of complex products/systems. In literature, various methods have been proposed to reduce computational resources and speed-up the simulation process. The hypothesis of bi-dimensional flow represents one of the most used strategy in different industrial cases. Much research in recent years were focused on establishing the results accuracy of 2D simulation models compared to 3D ones. The aim of this study is to propose a method to design and optimize filter houses, which integrates, in a unique framework, 3D and 2D simulations models. The latter are exploited to speed-up the design process, reducing computational time and resources. 2D simulation models have been defined according to physically validated 3D simulations. The error in estimating the temperature, pressure and velocity fields using 2D simulation models, considering 3D ones as reference, is less than 5%. Keywords: Filter House, Computational Fluid Dynamics, CFD, Virtual Prototyping, 2D Modelling, Power Plant Gas Turbine