Title: An a posteriori data-driven method for phase-averaged optical measurements Show Chinese title

Title: 一种后验数据驱动的相位平均光学测量方法

Abstract: Phase-averaging is a fundamental approach for investigating periodic and non-stationary phenomena. In fluid dynamics, these can be generated by rotating blades such as propellers/turbines or by pulsed jets. Traditional phase-averaging approaches often rely on synchronized data acquisition systems, which might require high-speed cameras, light sources, and precise delay generators and encoders, making them expensive and sometimes unfeasible. This work proposes an a posteriori data-driven approach that reconstructs phase information from randomly acquired uncorrelated photographic frames (snapshots) using the ISOMAP algorithm. The technique enables accurate reordering of snapshots in the phase space and subsequent computation of the phase-averaged flow field without the need for synchronization. The framework was validated through numerical simulations and experimental fluid dynamics datasets from an optical setup featuring single- and multi-propeller configurations. The results demonstrate that the proposed method effectively captures the periodic flow characteristics while addressing the challenges related to synchronization and hardware limitations. Furthermore, the ability to apply this technique to archival datasets extends its applicability to a wide range of experimental fluid dynamics studies. This approach provides a scalable and cost-effective alternative to traditional methods for the analysis of periodic phenomena. Show Chinese abstract

Abstract: 相平均是一种研究周期性和非平稳现象的基本方法。 在流体力学中，这些现象可以由旋转叶片如螺旋桨/涡轮机或脉冲射流产生。 传统的相平均方法通常依赖于同步数据采集系统，这可能需要高速摄像机、光源和精确的延迟发生器和编码器，使得它们成本高昂且有时不可行。 本文提出了一种事后的数据驱动方法，通过ISOMAP算法从随机获取的不相关图像帧（快照）中重建相位信息。 该技术能够准确地在相空间中重新排序快照，并随后计算相平均流场，而无需同步。 该框架通过数值模拟和来自具有单螺旋桨和多螺旋桨配置的光学设置的实验流体力学数据集进行了验证。 结果表明，所提出的方法有效捕捉了周期性流动特征，同时解决了与同步和硬件限制相关的问题。 此外，将该技术应用于归档数据集的能力扩展了其在广泛实验流体力学研究中的适用性。 这种方法为周期性现象的分析提供了一种可扩展且经济的替代方法。