基于PIV技术的压力场重构算法实现与研究

Implementation and research on the reconstruction algorithms of pressure fields based on PIV

  • 摘要: 介绍了有限容积法、直接积分法和Poisson方程法3种基于PIV瞬时速度场重构压力场的基本原理以及相应的计算方法,选取管流突扩流场和偏置方块绕流流场两个不可压缩流场的瞬时速度场数据,采用上述3种压力场重构算法,分别研究了图像噪声、速度场精度、插值算法以及边界条件的类型与精度对重构压力场的影响。最后针对管流突扩过程第20ms的流场,给出了3种重构算法下的压力场云图以及对应的CFD模拟结果。研究表明,有限容积法和直接积分法容易受到噪声的影响而产生剧烈震荡,但是可以在较大的速度场误差范围内保持较高的精度,通过采用双线性插值可以获得更高精度的重构压力场;Poisson方程法不易受到噪声的影响而产生震荡,同时在高精度PIV速度场下的优势较为突出,通过采用双三次差值可以获得更高精度的重构压力场;混合边界条件仅仅测定边界上有限个点的压力值,就获得了接近狄利克雷边界条件下重构压力场的精度,远高于诺依曼边界条件;边界条件的误差严重降低重构压力场的精度,其影响程度比速度场误差还要大。

     

    Abstract: The basic principles and the corresponding algorithms of the finite volume method, the direct integral method and the Poisson equation method are introduced in detail, which are used to reconstruct the pressure fields based on PIV velocity fields. The instantaneous velocity fields of two incompressible flows, including the pipe flow with a sudden expansion and the flow around a square, are selected to study the influence of picture noise, velocity error, interpolation methods, the type and the precision of boundary conditions on reconstructed pressure fields by using different reconstruction algorithms. Finally, the transient pressure distributions of the pipe flow with a sudden expansion at 20ms are obtained by using the three algorithms respectively as well as the CFD. It shows that the finite volume method and the direct integral method are easily affected by noise to produce rude shocks, but maintain high accuracy in a larger range of error in velocity fields while they can get higher precision of reconstructed pressure fields with bilinear interpolation; the Poisson equation method isn't easily affected by noise so it produces few shocks, and has great advantages with the accurate PIV velocity fields while it can get higher precision of reconstructed pressure fields with bicubic interpolation; by measuring only several pressure points on the boundaries, the mixed boundary condition gets the accurate reconstructed pressure fields which are close to those of the Dirichlet boundary condition and far better than those of the Neumann boundary condition; the error of boundary conditions reduces the precision of reconstructed pressure fields, which is more severe than the error of velocity fields.

     

/

返回文章
返回