PARAMETRIC ANALYSIS OF 3D TURBOMACHINERY FLOWS VIA REDUCED ORDER MODELING

摘要

Parametric and sensitivity analyses are essential tools in evaluating aerodynamic performance of multistage turbomachines, under different conditions. A CFD-based approach is usually carried out to test optimal configurations during the design phase. In order to substantially lower the cost of 3D computations, a Reduced Order Modeling (ROM) approach is proposed. The technique employs Proper Orthogonal Decomposition and Kriging interpolation to obtain the flow solution in a fraction of the time required by CFD methods. The value of the approach is judged based on a trade-off between the accuracy of the ROM solution with respect to a reference CFD one, and their relative computational cost.