Abstract A combination of the Linear Stochastic Estimation (LSE) and a modified version of the Extended Proper Orthogonal Decomposition (EPOD) is proposed to link multi-plane particle image velocimetry (PIV) data describing transitional separated boundary layers. Measurements have been obtained in two orthogonal planes, which share a common line providing the spatial consistency between the data sets. A wall-normal plane, capturing the Kelvin-Helmholtz (K-H) rolls shed downstream of the bubble maximum displacement and a wall-parallel plane, capturing the main cross-flow fluctuations occurring during the break-up process, were considered. Data on the wall-normal plane were acquired at high sampling rate to track the formation and propagation of the K-H rolls. On the other hand, the structures observed in the wall-parallel plane are characterized by longer wavelength in the spanwise direction, implying that the PIV frame has to be enlarged reducing the sampling rate due to the actual PIV system constrains. Combining LSE and EPOD allows the estimation of the time-resolved dynamical features of the separated boundary layer in both planes, increasing the temporal resolution of the wall-parallel plane data. Results will clearly show that the reconstructed pseudo 3D time-resolved field preserves the statistical properties of the original (and independent) acquired datasets in the two planes.

中文翻译：

---

基于混合 LSE 和 EPOD 的技术，用于在分离流条件下进行多平面 PIV 测量同步

摘要 提出了线性随机估计 (LSE) 和扩展正交正交分解 (EPOD) 的修改版本的组合，以链接描述过渡分离边界层的多平面粒子图像测速 (PIV) 数据。测量是在两个正交平面上获得的，它们共享一条公共线，提供数据集之间的空间一致性。考虑了一个壁面法向平面，捕获了气泡最大位移下游的开尔文-亥姆霍兹 (KH) 辊脱落，以及一个壁面平行平面，捕获了破裂过程中发生的主要横向流动波动。壁法线平面上的数据以高采样率采集，以跟踪 KH 辊的形成和传播。另一方面，在壁平行平面中观察到的结构的特征是展向方向上的波长较长，这意味着由于实际 PIV 系统的限制，必须扩大 PIV 帧以降低采样率。结合 LSE 和 EPOD 可以估计两个平面中分离边界层的时间分辨动力学特征，提高壁平行平面数据的时间分辨率。结果将清楚地表明，重建的伪 3D 时间分辨场保留了两个平面中原始（和独立）采集数据集的统计特性。结合 LSE 和 EPOD 可以估计两个平面中分离边界层的时间分辨动力学特征，提高壁平行平面数据的时间分辨率。结果将清楚地表明，重建的伪 3D 时间分辨场保留了两个平面中原始（和独立）采集数据集的统计特性。结合 LSE 和 EPOD 可以估计两个平面中分离边界层的时间分辨动力学特征，提高壁平行平面数据的时间分辨率。结果将清楚地表明，重建的伪 3D 时间分辨场保留了两个平面中原始（和独立）采集数据集的统计特性。