Proper Orthogonal Decomposition (POD) analysis of CFD data for flow in an axisymmetric sudden expansion

CFD simulations of swirling flow in a 2:1 axisymmetric sudden pipe expansion are performed and used to understand the structure of the flow. The k-ε and SAS-zonal LES models available in ANSYS CFX are used to model turbulence. Comparison of these results shows that the scale-resolving simulation captures much more detail of the flow and provides much more physical results in terms of the flow structures that are resolved. As well as conventional analysis using typical CFD post-processing techniques and Fast Fourier Transforms (FFTs), Proper Orthogonal Decomposition (POD) is used to extract reduced order models of the wall pressure data. This technique is used to extract dominant structures in the spatial domain and to investigate transient behaviour. Data from a finite set of monitor points located at the pipe wall are also examined as a potential means of detecting flow behaviour for use in flow control models that utilise differential wall pressure input data. Such models are required to control, for example, inlet swirl to minimise wall impaction in spray dryers.