A method for identification and correction of bad vectors in velocity field based on intrinsic orthogonal decomposition

Abstract

The invention discloses a velocity field bad vector identification and correction method based on intrinsic orthogonal decomposition, comprising: A: partitioning the original flow field to obtain a large sample velocity field; B, performing intrinsic normalization on the large sample velocity field Intersection decomposition (POD), when it is judged that the iterative convergence condition is not satisfied, the flow field reconstruction is carried out according to the calculated optimal reconstruction order and the mode and coefficient obtained by POD; C. Calculate the difference between the original flow field and the reconstructed flow field The modulus of the absolute error of the velocity vector, when judging that the modulus of the absolute error of the velocity vector is greater than the preset threshold, determine that the corresponding velocity vector is a bad vector, and remove it; D, before the next iteration, at the position where the bad vector has been removed Perform POD-based interpolation to correct the rejected vector; then return to B and continue to perform POD on the corrected large-sample velocity field until the iterative convergence condition is met, thereby realizing the identification and correction of bad vectors in a single instantaneous velocity field.

Description

technical field [0001] The invention relates to the technical field of velocity measurement of fluid mechanics, in particular to a method for identifying and correcting bad vectors of a velocity field based on intrinsic orthogonal decomposition. Background technique [0002] Particle Image Velocimetry (PIV for short) is a modern laser velocimetry technology, which is mainly used in flow field velocity measurement. The flow field velocity field is obtained by tracking the movement of tracer particles in the flow field. PIV technology can realize two-dimensional or three-dimensional velocity field measurement through sheet light source or volume light source. Usually, the velocity field of the flow field measured by the PIV experiment will have an abnormal velocity vector, which is often called a bad vector. There are many reasons for bad vectors in the velocity field of the flow field, mainly due to the poor imaging quality of particle images and the failure of correlation a...

AI Technical Summary

Problems solved by technology

However, since the normalized median detection method is based on the detection of local adjacent velocity vectors, there are two outstanding problems: first, the detection method is a purely mathematical processing method for the vector field, and does not take into account the vector field The flow mechanism behind it makes the velocity field bad vector removal and subsequent corrections introduce non-physical deviations; second, the local inspection method often fails when there is a large bad vector area in the velocity field, especially when the algorithm does not introduce iteration Especially problematic when identifying mechanisms

However, this method is very unstable, and there will be a phenomenon that the number of selected modes does not increase monotonously, especially in the late iteration, when the energy spectrum curve approaches the real flow field energy spectrum curve, it is difficult to find a clear intersection point, which is the reason for the mode The selection of the state order brings great uncertainty, which often leads to the failure of the algorithm.

Images