Volumetric density measurement in buoyant plumes using Tomographic Background Oriented Schlieren (TBOS)

This paper presents a novel method for measuring three-dimensional density fields in buoyant plumes using an eight-camera Tomographic Background Oriented Schlieren (TBOS) system, successfully validating the technique against theoretical models and demonstrating its capability to capture complex phenomena such as puffing in lazy plumes.

The Gist

The Big Picture: Seeing the Invisible Smoke

Imagine you are watching a candle flame or smoke rising from a chimney. You can see the smoke swirling and dancing, but you can't actually see the air itself changing density. The air gets lighter as it heats up or mixes with helium, but it remains invisible to the naked eye.

Scientists have long wanted to take a "3D X-ray" of these invisible plumes to understand exactly how they move, mix, and spread pollutants. The problem? Traditional cameras only see 2D slices, and the air is transparent.

The Solution: This paper introduces a new way to take a 3D "CT scan" of rising air using a technique called Tomographic Background-Oriented Schlieren (TBOS). Think of it as turning the invisible air into a visible, measurable 3D map.

---

The Setup: The "Eight-Eyed" Camera Rig

To do this, the researchers built a special rig that looks like a high-tech observatory.

• The Subject: They created a "lazy plume" (a gentle, rising column of air) by mixing helium (which is lighter than air) with regular air and shooting it out of a nozzle. It's like blowing a giant, invisible bubble of helium into a room.
• The Eyes: Instead of one camera, they used eight cameras arranged in a circle around the rising plume, like the lenses on a security camera dome.
• The Background: Behind the plume, they hung a wall covered in random black dots (like a starry night sky or a polka-dot curtain).

How It Works: The "Distorted Mirror" Analogy

Here is the magic trick:

1. The Baseline: First, the cameras take a picture of the dotted background with no plume. They see the dots clearly.
2. The Distortion: When the helium plume rises, it acts like a giant, invisible lens. Because the helium-air mixture has a different density than the surrounding air, it bends the light passing through it (just like a glass lens bends light).
3. The Shift: When the cameras take a picture with the plume, the dots on the background look slightly shifted or "wobbly." The denser the air, the more the dots move.
4. The Math: The computer measures exactly how much every single dot moved. By combining the views from all eight cameras, it can calculate the density of the air at every single point in the 3D space.

Analogy: Imagine looking at a fish tank. If the water is still, you see the rocks at the bottom clearly. If you start swirling the water, the rocks look like they are wiggling. If you had eight people standing around the tank, all taking photos of the wiggling rocks, a super-computer could figure out exactly how the water is swirling inside the tank, even though you can't see the water itself.

The "Lazy" Plume and the "Puffing" Phenomenon

The researchers studied three different types of rising plumes. Two of them were "lazy" (slow and unstable), and one was stable.

• The Stable One: This plume rose smoothly, like a steady stream of smoke from a cigarette. It just got wider and wider as it mixed with the air.
• The Lazy Ones (The Puffers): These plumes were chaotic. They didn't just rise; they puffed.
  • What is "Puffing"? Imagine a balloon inflating and then suddenly pinching off a piece of itself. The researchers found that these lazy plumes form giant, donut-shaped rings of air (vortices) near the bottom. These rings grow, pinch off a pocket of low-density helium, and shoot it upward like a bubble.
  • The Discovery: Using their 3D density map, they could actually see these "Low-Density Pockets" (LDPs) detaching and floating away. It's like seeing the individual bubbles in a soda can, but for air.

Why Does This Matter?

You might ask, "So we can see invisible air. So what?"

1. Pollution Control: If we understand exactly how smoke or toxic gas mixes with the air, we can design better chimneys and ventilation systems to keep cities clean.
2. Disaster Prediction: This helps us model volcanic ash clouds or wildfire smoke. If we know how the "puffing" works, we can predict where the ash will go and how fast it will spread.
3. Fixing the Math: Scientists have been using math models to predict these flows for decades, but they often guess wrong about how much air gets sucked into the plume (called "entrainment"). This new 3D data gives them the real numbers to fix those models.

The Takeaway

This paper is a breakthrough because it's the first time anyone has successfully taken a full, 3D "density photo" of a rising air plume. They turned a transparent, invisible phenomenon into a colorful, 3D data set that reveals the hidden "donut rings" and "puffing" behavior of rising air.

It's like finally putting on 3D glasses to see the invisible world of air currents, giving us a much clearer picture of how our atmosphere works.

Technical Summary

1. Problem Statement

Buoyant plumes are ubiquitous in natural phenomena (e.g., volcanic ash, wildfires, mantle plumes) and engineered systems (e.g., industrial emissions, HVAC). While theoretical models (such as the Morton-Taylor-Turner model) exist to predict plume dynamics, they rely on assumptions like "top-hat" profiles and a universal entrainment coefficient.

• The Gap: There is a critical lack of three-dimensional (3D) volumetric density data for buoyant plumes in the literature. Existing measurements are predominantly point-based or 2D, often limited to symmetric flows.
• The Consequence: Without accurate 3D density fields, it is difficult to validate theoretical models, calculate entrainment coefficients directly, or understand complex unsteady phenomena like "puffing" (periodic vortex formation) in "lazy" plumes.

2. Methodology

The authors developed an experimental rig and a processing pipeline to measure the 3D density field of buoyant plumes using Tomographic Background-Oriented Schlieren (TBOS).

A. Experimental Setup

• Plume Generation: A He-Air mixture was used to generate buoyant plumes. Two Mass Flow Controllers (MFCs) mixed Helium and Air in a settling chamber with a honeycomb structure to ensure laminar flow before discharging through a 35 mm nozzle.
• Imaging System:
  • Cameras: Eight synchronized IMPERX cameras (2352 × 1768 px) with 85 mm lenses were arranged circumferentially around the plume (radius 600 mm, 22.5° spacing).
  • Background: Random-dot patterns were placed diametrically opposite the cameras.
  • Illumination: A single 1000 W halogen lamp illuminated the background.
  • Synchronization: Cameras were triggered at 10 fps (Release I & II) or 21 fps (Release III) with a 100 µs exposure time.

B. Data Processing Workflow (The "Three-Step" Approach)

The authors employed a modified classical three-step workflow:

1. Ray Deflection Estimation:
   • Cross-correlation (using the PIVlab toolbox) was performed between reference images (no plume) and object images (plume present) to calculate 2D ray deflection (dot displacement).
   • This yielded the line-integrated density gradient ($\int \nabla \rho \, dz$).
2. Poisson Integration:
   • A finite-difference-based Poisson solver was used to integrate the density gradients to obtain the line-integrated density ($\int \rho \, dz$).
   • Neumann boundary conditions were applied, with a Dirichlet condition later used to correct the integration constant.
3. Tomographic Reconstruction:
   • Sinogram Interpolation: To mitigate streak artifacts from sparse angles (8 views), cubic spline interpolation was used to synthesize data, effectively increasing the view count from 8 to 15.
   • SART Algorithm: The Simultaneous Algebraic Reconstruction Technique (SART) was used to reconstruct the 3D density field plane-by-plane. This iterative approach was chosen over Filtered Back Projection (FBP) for better accuracy with limited projections and noisy data.
   • Post-Processing: Included smoothing (5×5×5 box filter) and enforcing physical constraints (e.g., setting density outside the plume to ambient air density).

3. Key Contributions

• First 3D Density Measurements: This study presents the first reported 3D volumetric density measurements of buoyant plumes using TBOS.
• Lazy Plume Characterization: The technique successfully captured the dynamics of "lazy" plumes (where the source momentum is low compared to buoyancy), specifically visualizing the puffing phenomenon (periodic formation and detachment of toroidal vortices).
• Methodological Refinement: The authors refined the TBOS pipeline by implementing cubic spline sinogram interpolation and a plane-by-plane SART reconstruction strategy, which reduces memory requirements compared to full 3D voxel reconstruction.
• Validation Framework: The measurements were rigorously validated against similarity solutions and existing Large Eddy Simulation (LES) data from literature.

4. Results

The study examined three release cases with varying source strengths (defined by the plume function $\Gamma_0$):

• Release I & III (Lazy/Unstable): Exhibited strong puffing behavior. The 3D density field clearly visualized the formation of Low-Density Pockets (LDPs) that pinch off from the source and convect downstream.
• Release II (Forced/Stable): Exhibited laminar, non-puffing behavior with no detached LDPs.

Key Findings:

• Profile Validation: Time-averaged transverse density profiles followed a Gaussian distribution (consistent with theory), while the near-field showed a flat-topped "potential core" characteristic.
• Similarity Solutions: The centerline reduced gravity ($g'$) and plume half-width ($b_g$) showed good agreement with similarity solutions for lazy plumes, particularly in the far-field.
• Puffing Dynamics: The 3D visualization revealed that puffing intensity correlates with the Richardson number ($Ri$). Release III (lower $Ri$, higher instability) showed more pronounced necking and larger detached LDPs compared to Release I.
• Entrainment: The density fields confirmed the entrainment of heavier ambient air, causing the plume density to increase and the profile to narrow (necking) in the near-field before transitioning to self-similarity.

5. Significance and Future Directions

• Scientific Impact: This work provides the "ground truth" data necessary to refine theoretical models of buoyant plumes, particularly regarding the entrainment coefficient ($\alpha$), which has historically been a source of discrepancy in literature.
• Quantitative Potential: By combining 3D density with 3D velocity (e.g., via Tomographic PIV), researchers can directly compute volume, momentum, and buoyancy fluxes without assuming geometric profiles (top-hat or Gaussian), allowing for the direct calculation of entrainment coefficients.
• Limitations & Future Work: The current study lacks simultaneous 3D velocity measurements. Future work aims to integrate Tomographic PIV and Particle Tracking Velocimetry (PTV) to achieve full 3D flow field characterization (density + velocity) for a complete analysis of buoyancy-driven flows.

In summary, this paper establishes a robust experimental and computational framework for measuring 3D density in complex, unsteady buoyant plumes, bridging a significant gap between theoretical modeling and experimental observation in fluid dynamics.