Investigation of non-linear projection for POD based reduced order models for Aerodynamics

Citations Over TimeTop 10% of 2001 papers

Abstract

this paper a methodfor inviscid air analysis and design optimization using r - duced or der models has beenr eviewed, modifications to the methodology ar e pr oposed, and new r sults of flow computations and design optimization a r pr esented. The goal of theser educed or der models is to obtain solutions to the Euler equations with su#cient accur yfor design optimization, at a computational cost which isfar lower than that fr om tr aditional Computational Fluid Dynamics (CFD) methods. Design wor k with these models begins with a ser ies of flow solutions (snapshots) in which the design var les of inter est have been per tur bed: this is the up-fr ont cost of the designpr ocedur e.Pr oper Or thogonal Decomposition (POD) is used to compute the optimal linear basisfor the snapshots. This linear basis can then be used to compute new, appr oximate flow solutions at significantlyr educed computational cost for ar bitr ar y geometr ies.Our pr evious wor k with these models had shown that while a significantr eduction in computational cost was achieved the tendency of all flow solutions computed with the linear basis to be ver similar to theor al data significantly impair d their use in the design fr wor

Related Papers

• → Numerical analysis of aerodynamic characteristics of multi-pod hyperloop system(2022)8 cited
• → Geometric Filtration Using POD for Aerodynamic Design Optimization(2008)25 cited
• → Proper Orthogonal Decomposition for steady aerodynamic applications(2010)14 cited
• Aerodynamic Analysis of an Airborne Equipment Pod(2012)
• → ON A PROJECTION THEOREM FOR MULTIFUNCTIONS(1983)1 cited