Jet-edge interaction: linear and non-linear frequency-selection mechanisms

This study investigates jet-edge interactions in a round turbulent jet grazing a $45^\circ$ plate, classifying tonal dynamics into distinct regimes based on linear and non-linear frequency-selection mechanisms and identifying a robust, non-hysteretic mode-switching phenomenon at Mach 0.84.

The Gist

Imagine you are holding a garden hose, spraying a powerful stream of water. Now, imagine you hold a flat board just next to the stream, but not touching it. If you get the angle and distance just right, the water doesn't just splash randomly; it starts to hum. It makes a loud, pure musical note, like a flute or a whistle.

This paper is a scientific investigation into exactly how and why that humming happens, and why sometimes the sound changes from a gentle hum to a screaming roar, or even switches notes entirely.

Here is the story of the research, broken down into simple concepts:

1. The Setup: The "Jet" and the "Edge"

The researchers used a high-speed jet of air (like a jet engine, but smaller and quieter) shooting out of a nozzle. They placed a metal plate nearby, angled at 45 degrees. The air stream "grazes" the sharp edge of this plate.

• The Analogy: Think of the air stream as a river flowing downstream. The plate edge is a rock sticking out of the river. When the river hits the rock, it creates ripples. But in this case, the ripples don't just go downstream; they bounce back upstream, hit the source of the river, and start a loop.

2. The Feedback Loop: The "Echo Chamber"

The core of the mystery is a feedback loop.

1. Downstream: Tiny, invisible waves travel down the jet stream (like ripples in the river).
2. The Bounce: When these waves hit the sharp edge of the plate, they get "scattered" and turn into waves that travel backwards (upstream).
3. The Return: These backward waves travel all the way back to the nozzle.
4. The Spark: When they hit the nozzle, they create new downstream waves.

If the timing is perfect, this loop reinforces itself, creating a loud, steady tone. This is called resonance.

3. The Three Types of "Humming"

The researchers found that depending on how fast the air is moving (Mach number) and how close the plate is, the sound behaves in three very different ways:

A. The "Static" (Broadband Noise)

• What it sounds like: Like a hiss or a roar (think of a hair dryer).
• Why: The feedback loop is broken or messy. The waves are chaotic and don't lock into a rhythm. It's just random noise.

B. The "Linear Choir" (Linear Frequency Selection - LFS)

• What it sounds like: A few distinct, pure musical notes happening at the same time, but they don't harmonize. It's like a choir singing different notes that don't quite fit together.
• Why: The feedback loop is working, but it's "linear." This means the waves are just bouncing back and forth without changing each other. The speed of the air and the distance to the plate determine exactly which notes are sung. The researchers could predict these notes perfectly using math.

C. The "Rock Star Soloist" (Non-Linear Frequency Selection - NLFS)

• What it sounds like: One incredibly loud, dominant note, followed by a series of harmonics (like a guitar string vibrating). It's much louder than the "Linear Choir."
• Why: Suddenly, one of the notes from the "Linear Choir" gets super-charged. It becomes so loud that it starts to bully the other notes, suppressing them. It also starts interacting with itself to create harmonics (multiples of the main note).
• The Metaphor: Imagine a room full of people talking (LFS). Suddenly, one person starts screaming so loudly that everyone else stops talking, and the room fills with the echo of that one scream.

4. The "Magic Switch" (Mode Switching)

The most exciting discovery was a switch that happens very suddenly.

• The Scenario: As the researchers slowly increased the speed of the air, the system was humming one way. Then, at a very specific speed (Mach 0.84), the sound snapped to a completely different note.
• The Cause: It's like a radio tuning. At that specific speed, a new type of backward-traveling wave "turns on" (like a light switch flipping on). This new wave creates a better feedback loop than the old one, so the system instantly switches to the new note.
• The Cool Part: This switch is incredibly stable. If you speed up or slow down the air, the switch happens at the exact same point. There is no "lag" or "hysteresis" (it doesn't get stuck). It's a reliable, repeatable phenomenon.

5. The "Tug-of-War"

In a specific range of speeds, the researchers saw two different feedback loops fighting for control.

• The Analogy: Imagine two teams playing tug-of-war. One team uses a "Thick Rope" (one type of wave), and the other uses a "Thin Rope" (a different type of wave).
• The Result: As the wind speed changes, the "Thick Rope" team slowly gets stronger until, suddenly, they win the tug-of-war, and the "Thin Rope" team goes slack. The sound changes from one note to another.

Why Does This Matter?

This isn't just about making cool sounds. This is about noise control.

• The Problem: When jets fly near airports, or when engines are installed on aircraft wings, these "edge tones" create massive amounts of noise.
• The Goal: By understanding exactly how these tones are created and how they switch, engineers can design planes that avoid these "sweet spots." They might be able to tweak the shape of the engine or the wing to break the feedback loop, turning that screaming roar back into a quiet hiss.

Summary

The paper is a map of the "sound landscape" created by a jet hitting a plate. They found that the sound can be a hiss, a choir, or a soloist. They discovered that the system can suddenly snap from one state to another due to a hidden "switch" in the physics of the air waves. It's a story of how chaos can suddenly organize into a perfect, loud rhythm, and how that rhythm can change in the blink of an eye.

Technical Summary

1. Problem Statement

The study investigates the resonance behavior of a subsonic turbulent jet grazing the sharp edge of a rectangular metal plate inclined at $45^\circ$. This configuration, relevant to installed jet noise in aerospace applications, generates both broadband noise and distinct tonal components. While previous research (e.g., Jordan et al., 2018) established that tonal dynamics are often driven by a linear feedback loop involving downstream-traveling Kelvin-Helmholtz (K-H) wavepackets and upstream-traveling guided jet modes, the non-linear aspects of these interactions—specifically how multiple tones interact, amplify, and transition between regimes—remained unexplored. The authors aim to map the spectral signatures across a parameter space defined by jet Mach number ($M_j$) and the radial position of the plate edge ($R/D$), identifying the mechanisms governing regime transitions and frequency selection.

2. Methodology

Experimental Setup:

• Facility: Experiments were conducted at the Bruit & Vent facility of Institut Pprime.
• Configuration: A round jet ($D = 50$ mm) with a fully turbulent boundary layer grazes a rigid plate ($S=15D$, $C=5D$, thickness $8$ mm). The streamwise edge position is fixed at $L/D = 2$, and the plate angle is $\alpha = 45^\circ$.
• Parameters: The study varies the jet Mach number ($M_j$) from 0.4 to 1.0 and the radial edge position ($R/D$) from 0.5 to 0.8.
• Measurements: Far-field sound pressure was recorded using a B&K 4944-B microphone at $90^\circ$ to the plate edge. Data was acquired at a 200 kHz sampling rate for 30 seconds per condition.

Analysis Techniques:

• Power Spectral Density (PSD): Computed using the Welch method to identify spectral peaks and broadband characteristics.
• Bicoherence Analysis: Used to detect non-linear phase coupling (triadic interactions) between frequencies $f_1$, $f_2$, and $f_3 = f_1 + f_2$. A threshold of $b > 0.35$ was established to distinguish true triadic interactions from broadband noise artifacts.
• Linear Stability Modeling: A cylindrical vortex sheet model (Jordan et al., 2018) was employed to predict tonal frequencies. The model assumes a feedback loop where downstream K-H waves ($k^+$) scatter at the plate edge into upstream guided modes ($k^-$), which travel back to the nozzle to regenerate the K-H waves. The resonance condition is determined by the phase matching of these waves.

3. Key Contributions

The paper provides a comprehensive classification of jet-edge interaction dynamics and elucidates the transition mechanisms between linear and non-linear regimes:

1. Spectral Classification: Identification and categorization of four distinct dynamical regimes:
   • Broadband: No distinct tones; purely stochastic noise.
   • Linear Frequency Selection (LFS): Multiple non-harmonic tones with zero bicoherence, governed strictly by linear feedback loops.
   • LFS with Weak Non-linear Interactions: Dominant LFS tones accompanied by weak, non-harmonic triadic interactions or single harmonics.
   • Non-Linear Frequency Selection (NLFS): A single fundamental tone (originating from an LFS mechanism) undergoes strong amplification, suppressing other LFS modes and generating a harmonic series via self-interaction.
2. Mechanism of Regime Switching: Demonstration that the transition from LFS to NLFS is driven by the massive amplification of a specific linear mode, leading to a "strong-oscillator" dynamic.
3. Mode Competition and Switching: Discovery of a robust, hysteresis-free mode-switching mechanism in the range $0.82 \leq M_j \leq 0.86$. This involves a competition between two distinct feedback loops closed by different upstream-travelling waves ($k^-_d$ and $k^-_p$).
4. Robustness of Transitions: The transitions between regimes are shown to be repeatable and independent of the direction of the Mach number sweep (no hysteresis), occurring over very small increments ($\Delta M_j = 0.01$).

4. Key Results

Spectral Regimes:

• Broadband: Observed at high $M_j$ ($>0.99$) and large $R/D$.
• LFS: Dominates at lower $M_j$ and larger $R/D$. Multiple tones appear, predicted accurately by linear stability theory using upstream waves $k^-_{th}$ (for $M_j < 0.82$) or $k^-_d/k^-_p$ (for $M_j \geq 0.82$).
• NLFS: Occurs in a specific region ($0.56 \leq M_j \leq 0.87$ and $R/D \leq 0.57$). Here, one LFS tone amplifies by $>15$ dB, becoming the fundamental of a non-linear oscillator. This generates harmonics and weak non-harmonic triads.
• Transitions: The transition from LFS to NLFS (e.g., at $M_j \approx 0.87$) is abrupt, with a 2-order-of-magnitude increase in amplitude. The reverse transition (NLFS to LFS) is less abrupt, often passing through an intermediate "weak non-linear" regime.

Mode Switching ($M_j \approx 0.84$):

• In the range $0.82 < M_j < 0.86$, two feedback loops compete: one closed by the $k^-_d$ wave (higher frequency) and one by the $k^-_p$ wave (lower frequency).
• As $M_j$ increases, the system switches dominance from the $k^-_d$ loop to the $k^-_p$ loop.
• Cause: The switch is attributed to the "cut-on" of the $k^-_p$ wave at $M_j \approx 0.82$. The $k^-_p$ wave possesses greater radial support and higher reflection coefficients at the nozzle lip compared to $k^-_d$, making it the preferred feedback mechanism at higher Mach numbers.
• This switching is robust, repeatable, and free of hysteresis.

Amplitude Characteristics:

• Sound Pressure Levels (SPL) are highest in the NLFS region, reaching up to 150 dB.
• Amplitude variations across Mach number transitions can reach three orders of magnitude.

5. Significance

This work significantly advances the understanding of installed jet noise by bridging the gap between linear stability theory and observed non-linear dynamics.

• Predictive Capability: It confirms that even strong non-linear oscillators originate from linear frequency selection mechanisms, validating the use of linear models for initial tone prediction.
• Control Implications: The identification of robust, hysteresis-free transitions suggests that small perturbations in operating conditions (Mach number or geometry) could be used to trigger or suppress specific tonal regimes, offering potential pathways for noise control strategies.
• Physical Insight: The study clarifies the role of specific upstream-guided modes ($k^-_p$ vs. $k^-_d$) in feedback loop selection, highlighting the importance of wave radial structure and reflection coefficients in determining resonance dominance.
• Methodological Rigor: The combination of high-resolution spectral analysis, bicoherence for non-linearity detection, and rigorous linear modeling provides a robust framework for analyzing complex fluid-structure acoustic interactions.

In conclusion, the paper establishes that jet-edge interaction dynamics are governed by a complex interplay of linear feedback loops and non-linear amplification, with regime changes driven by the availability and amplification of specific upstream-travelling wave modes.