Numerical Methods for Internal Heat Transfer in Airplanes

Abstract

Understanding the limits of an airplane is crucial for ensuring the reliability of the airplane and its systems. One limiting factor is the thermal loads from various internal heat sources in the airplane. It is necessary to understand how heat is transported in the internal bays of the airplane to ensure that the airframe and avionics are rated for the thermal environment that is developed. Since the process for designing an airplane is very time-consuming, simulations of the airflow using computational fluid dynamics (CFD) have become the norm in the industry. However, highly resolved turbulent simulations take months or even years to run through. Simplified simulations are often more feasible in industrial applications due to a great reduction in time and computational cost. This thesis explores different numerical methodologies for simplified CFD simulations, evaluating their accuracy and time consumption compared to a highly resolved turbulent simulation. This comparison shows not only which flow structures and flow properties that are captured, but also to what degree. The largest simplification for the simulations in this thesis is the use of the steady state Reynolds-Averaged Navier-Stokes (RANS) solver for all the explored numerical methods. The numerical methods investigated are the turbulence modelling schemes Lag Elliptic Blending (EB) k − ε and Shear Stress Transport (SST) k − ω, the effects of a polyhedral versus a trimmed cell mesh, the use and number of prism layers for the near-wall regions in the mesh, the effects of two transition models for the SST k −ω turbulence model, and the effects of the inviscid flux discretization schemes Roe’s and Advection Upstream Splitting Method + (AUSM+) Flux-Vector Splitting (FVS).