TL;DR
This paper introduces a new method, PODHW, to analyze fluid wakes using single-point hotwire data, revealing distinct flow states in the transition regime that traditional PIV-based POD (PODPIV) missed. This advancement offers a more comprehensive understanding of complex fluid dynamics, particularly for intermittent phenomena.
What Problem Does This Paper Address?
Understanding the fluid dynamics around bluff bodies, like square cylinders, is crucial for designing structures such as bridges and buildings. When two square cylinders are placed in tandem (one behind the other), the flow becomes significantly more complex than around a single cylinder. This complexity depends on the spacing between the cylinders (L/w) and the flow speed (Reynolds number). Previous research identified distinct flow regimes (extended body, reattachment, transition, and co-shedding) based on L/w. However, capturing the nuances of the "transition" regime, where the flow can switch between reattachment and co-shedding, has been challenging. Specifically, the intermittent nature of these switches, occurring over seconds to minutes, is difficult to resolve with high-speed PIV data, which typically captures snapshots over much shorter durations. This limits our ability to fully characterize and understand the dynamics within this critical transition phase.
Methods
The core of this study is the application and development of Proper Orthogonal Decomposition (POD), a powerful statistical technique for identifying dominant coherent structures in turbulent flows.
- PODPIV (Conventional): This method uses Particle Image Velocimetry (PIV) data, which provides a full-field view of the flow at a given instant. POD is applied to these snapshots to extract spatial modes representing the dominant flow structures and their energy contributions.
- PODHW (Novel): The paper introduces a new approach, PODHW, designed to analyze single-point hotwire anemometer data. Since hotwire data provides high-frequency temporal information at a single location but lacks spatial detail, a "trajectory matrix" is constructed using a method of delays (Broomhead & King). This allows for the reconstruction of state spaces and the application of POD to capture temporal dynamics and dominant modes from this localized, high-resolution temporal data.
The researchers used both PIV and hotwire data to compare their findings and highlight the advantages of the new PODHW method.
Datasets
The study utilized experimental data collected from a closed-circuit wind tunnel.
- PIV Data: High-speed PIV measurements were taken to capture instantaneous velocity fields. A total of 1000 images were acquired for each L/w condition.
- Hotwire Data: Single-point hotwire anemometer data was collected at a high sampling frequency (2 kHz) for extended durations (over an hour) for each L/w condition.
These datasets were analyzed using both PODPIV and the newly developed PODHW method.
Key Insights and Ideas
- PODHW's Advantage for Intermittent Flows: The most significant insight is that PODHW, by leveraging long-duration, high-frequency hotwire data, can successfully identify distinct flow states (reattachment and co-shedding) within the transition regime (L/w = 2.8). This is a crucial finding because traditional PODPIV, limited by shorter acquisition times, failed to capture this intermittent behavior.
- Confirmation of Transition States: The study provides the first experimental confirmation of the reattachment and co-shedding states within the transition regime at L/w = 2.8, as previously proposed by Zhou et al. (2024).
- Energy Distribution Across Regimes: The first two POD modes consistently account for the majority of the turbulent kinetic energy across all flow regimes. However, their dominance decreases in the transition regime, suggesting a more complex distribution of energy among higher-order modes.
- Scale Differences in Flow Structures: At higher L/w ratios (e.g., L/w = 4.2), the vortex structures associated with the dominant POD modes become smaller and less intense compared to lower L/w ratios. Conversely, in the transition regime, modes 3 and 4 show more complex patterns, indicating a shift in dominant flow dynamics.
Key Results
- Identification of Transition States: PODHW successfully identified two distinct states in the transition regime (L/w = 2.8): reattachment (St ≈ 0.13) and co-shedding (St ≈ 0.10). These states were characterized by different amplitude fluctuations in the hotwire signal.
- Energy Contribution of Dominant Modes:
- In the extended-body (L/w = 1.4) and reattachment (L/w = 2.4) regimes, the first two PODPIV modes contributed approximately 58% and 54.5% of the total fluctuating energy, respectively.
- In the co-shedding regime (L/w = 4.2), this contribution dropped to about 42.5%.
- The transition regime showed the lowest contribution from the first two PODHW modes (around 31.8%), but higher-order modes (1-4) collectively contributed significantly due to the presence of both co-shedding and reattachment states.
- Strouhal Numbers: The primary vortex shedding frequencies (Strouhal numbers) for the dominant modes were identified. For modes 1 and 2, St typically ranged from 0.12 to 0.14 across different L/w values. Higher modes sometimes exhibited peaks at twice these frequencies (e.g., St ≈ 0.26).
Limitations and Future Work
- PIV Data Duration: The primary limitation acknowledged is the insufficient duration of PIV data acquisition, which prevented the capture of intermittent flow phenomena in the transition regime.
- Hotwire Measurement Point: While PODHW is powerful, it relies on single-point measurements. The authors note that the results from hotwire data are expected to be dependent on the measurement point, as coherent structures weaken with distance from the cylinders. However, they attribute the qualitative similarity in trends to the hotwire data's effectiveness.
- Future Work: The authors suggest that the PODHW method could be extended to analyze other types of single-point data, such as 3D vorticity measurements, which are notoriously difficult to obtain simultaneously at multiple points. This opens avenues for studying complex turbulent flows that were previously inaccessible. The ability to capture intermittent phenomena also suggests potential applications in understanding and controlling unsteady flow behaviors.
Summary updated