Dynamic Mode Decomposition of Flow and Mixing Characteristics in a Rectangular T- Junction Under Wall Jet
32 Pages Posted: 6 Sep 2022
Abstract
Dynamic mode decomposition (DMD) is performed to achieve time evolution information and spatial mode information of the mixing velocity and temperature fields between hot and cold fluids in a rectangular T-junction. Time-resolved sequences of velocity and temperature fields are extracted from the large eddy simulation (LES) results and regarded as the original data for DMD analysis. Time growth rates, characteristic frequencies and spatial structures of velocity and temperature fields are analyzed. The results show that the main characteristic frequency of the velocity field in x-z cross section is in the range of St = 0.2~2 and most frequencies are in the range of St = 0~5. The characteristic frequencies of velocity field are slightly larger than those of temperature field, which implies the scales of the velocity spatial structures are smaller than those of temperature field. As for the spatial modes, the main spatial structures of temperature field grow from x/D m = 0 to 0.5, which come out ahead of those of velocity field that generate from x/D m = 0.5 to 1. Additionally, a new method named Growth Rate Filter method (GRF) is proposed to better sort the spatial modes and improve the accuracy of flow reconstruction.
Keywords: Dynamic mode decomposition (DMD), T-junction, Wall jet, Growth Rate Filter method (GRF)
Suggested Citation: Suggested Citation