Font Size: a A A

Reduced-order modeling of incompressible jet flow using proper orthogonal decomposition and Galerkin projection

Posted on:2011-09-14Degree:M.SType:Thesis
University:Northeastern UniversityCandidate:Hou, YimanFull Text:PDF
GTID:2440390002469817Subject:Engineering
Abstract/Summary:
Jet flows are complex and therefore challenging to be described accurately. This work aims at reduced-order modelling (ROM) of a three dimensional incompressible turbulent jet. The Galerkin approximation is utilized to project the governing partial differential equations (PDEs) onto a smaller size of ordinary differential equations (ODEs). A set of mesh grids are created in a cylindrical coordinate system with a nonlinear centerline in the streamwise direction to restore axial symmetry. Azimuthal decomposition is performed first as a pre-purification procedure, followed by the proper orthogonal decomposition (POD) with respect to kinetic energy norm and pressure norm. The capabilities of the POD modes to capture energy of the flow are presented in terms of eigenvalues. Different growing and rotating structures are found in the first few POD modes. The POD modes behaviors are further discussed in the spatial domain and the temporal domain.;The simulation database used in this work requires 50 Mb of binary data to represent a single time snapshot of one flow quantity. Therefore computationally speaking, a significant challenge is the efficient processing of high volume data in a reasonable time with standard PC-capabilities. The method of snapshots is used in the POD procedure to reduce the computational cost.
Keywords/Search Tags:POD, Flow, Decomposition
Related items