⚡ electrical engineering

Abrasive media viscosity regulation in abrasive flow finishing of additively manufactured waveguides: balancing boss surface smoothing and edge profile retention

This study demonstrates that regulating abrasive media viscosity to the 2000–2500 Pa·s range in abrasive flow finishing of additively manufactured waveguides effectively balances boss surface smoothing and edge profile retention by reducing edge removal and improving machining uniformity while maintaining a surface roughness reduction rate above 70%.

Original authors: Xuang Zhao, Yuli Sun, Zhe Yan, Guoqing Jiang, Wenzhuang Lu, Yebin Sun, Dunwen Zuo

Published 2026-07-15

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Xuang Zhao, Yuli Sun, Zhe Yan, Guoqing Jiang, Wenzhuang Lu, Yebin Sun, Dunwen Zuo

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a brand-new, 3D-printed metal tunnel shaped like a complex maze. This isn't just any tunnel; it's a waveguide, a high-tech pipe designed to guide invisible radio waves for things like radar and satellite communication. But because it was printed layer-by-layer, the inside is rough, bumpy, and full of tiny "stair-steps" left over from the printing process. If you don't smooth it out, the radio waves will get messy and the device won't work right.

To fix this, the researchers used a method called Abrasive Flow Machining. Think of this as pushing a giant, thick blob of toothpaste mixed with tiny, sharp sand grains through the tunnel. As this gritty paste squishes through, it scrapes the walls, smoothing out the bumps.

The Big Problem: The "Bosses" and the "Edges"
Inside this tunnel, there are special raised platforms called bosses (the target areas that need to be super smooth) and sharp corners called edges (the boundaries that define the tunnel's shape).

The Goal: You want the paste to scrub the flat tops of the bosses until they are glassy smooth.
The Danger: You don't want the paste to eat away the sharp edges. If the edges get rounded off or shaved down too much, the tunnel changes shape, and the radio waves get confused.

The Experiment: How Thick Should the "Toothpaste" Be?
The researchers asked a simple question: How thick (viscous) should the gritty paste be to get the best result?

They tested five different "thicknesses" of their abrasive paste, ranging from a runny 1,000 Pa·s to a very thick, gloopy 3,000 Pa·s. They used a computer simulation first to guess what would happen, and then they actually ran the machines to see.

What They Discovered
The results showed a classic trade-off, like trying to choose between a fast car and a safe car:

The Runny Paste (Low Viscosity - 1,000 Pa·s):
- The Good: It was a super-efficient scrubber! It smoothed the boss surfaces incredibly well, reducing the roughness by 83.80%. The surface went from being very rough (over 10 μm) down to a smooth 1.838 μm.
- The Bad: It was too aggressive. Because it flowed so fast and easily, it also chewed up the sharp edges. The edges got shaved down by an average of 0.1537 mm. That's like shaving off a significant chunk of the tunnel's boundary, which could ruin the device's performance. Also, the paste didn't clean every part of the tunnel evenly; some bosses got polished more than others.
The Super-Thick Paste (High Viscosity - 3,000 Pa·s):
- The Good: It was very gentle. It barely touched the edges, shaving them down by only 0.0417 mm. It also cleaned every part of the tunnel very evenly, so all the bosses ended up with the same finish.
- The Bad: It was too lazy. It didn't scrub hard enough. The boss surfaces only got 67.76% smoother, leaving them with a roughness of 4.222 μm. That's still too rough for high-performance radio waves.

The "Sweet Spot" Found
The researchers realized that neither the super-runny nor the super-thick paste was perfect. They needed a "Goldilocks" zone.

By simulating the flow and testing the paste, they found that a viscosity between 2,000 and 2,500 Pa·s was the winner.

In this range, the paste was thick enough to protect the sharp edges (shaving them down only 0.0951 mm to 0.0641 mm) but still strong enough to scrub the bosses effectively.
The bosses got smoothed by more than 70%, bringing the roughness down to below 3.7 μm.
The cleaning was also much more uniform across the whole tunnel.

Why Does This Happen?
The researchers used math (called a power-law fluid model) to explain this. Imagine the paste flowing through the tunnel like a river.

Thin paste: The middle of the river flows much faster than the water hugging the banks. This speed difference causes the paste to swirl and hit the sharp corners hard, rounding them off.
Thick paste: The whole river moves more like a solid block. The speed difference between the middle and the edges is small. This makes the flow smoother and more predictable, so it scrapes the flat tops without violently attacking the corners.

The Bottom Line
The paper concludes that for these specific 3D-printed waveguides, you shouldn't just use the thickest or thinnest paste you can find. Instead, aiming for a viscosity of 2,000–2,500 Pa·s strikes the perfect balance. It smooths the important surfaces enough to work, keeps the sharp edges sharp enough to function, and ensures the whole tunnel gets cleaned evenly. It's a careful balancing act between scrubbing hard and being gentle.

Technical Summary: Abrasive Media Viscosity Regulation in Abrasive Flow Finishing of Additively Manufactured Waveguides

1. Problem Statement

Additively manufactured (AM) special-shaped waveguides with continuous blind cavities are critical for aerospace and defense communication systems due to their complex internal geometries. However, the layer-by-layer fabrication process often results in significant internal surface defects, including powder adhesion, spheroidization, and stair-stepping effects, leading to high initial surface roughness. This roughness compromises electromagnetic wave transmission stability and device performance.

While abrasive flow machining (AFM) is effective for finishing complex internal cavities, a specific challenge arises in waveguides containing coupled geometric features such as bosses, grooves, and sharp edges. The finishing requirements are conflicting:

Boss Surfaces: Require sufficient and uniform material removal to reduce roughness.
Edge Regions: Must retain their precise profile and sharpness, as excessive material removal (edge rounding) alters boundary conditions, affecting key electromagnetic properties like cutoff wavenumber and bandwidth.

Existing studies often adjust medium viscosity by changing abrasive content, making it difficult to isolate the specific effects of viscosity from particle parameters. Furthermore, the flow behavior of abrasive media in narrow, variable-section blind-cavity channels and its coupled impact on boss smoothing versus edge preservation remains unclear.

2. Methodology

This study employs a combined theoretical, numerical, and experimental approach to regulate abrasive medium viscosity to balance boss surface smoothing and edge profile retention.

Theoretical Analysis

Modeling: The complex internal cavity was simplified into an equivalent circular straight pipe to analyze flow characteristics.
Constitutive Model: The abrasive medium was treated as an incompressible, non-Newtonian shear-thinning fluid described by the power-law model ( $\tau = K\dot{\gamma}^n$ ).
Derivation: A theoretical relationship was established between medium viscosity and the velocity difference ( $v_d$ ) between the channel center and the near-wall region. The analysis posits that higher viscosity reduces the velocity gradient, leading to a flatter velocity profile.

Numerical Simulation

Software: ANSYS Fluent was used to simulate two-phase flow (Mixture model) within the actual waveguide geometry.
Parameters: Simulations were conducted for viscosities ranging from 750 to 3000 Pa·s.
Metrics: The study analyzed near-wall velocity, dynamic pressure, and strain rate distributions. A "PV value" (product of near-wall pressure and velocity) was defined to characterize material removal capacity. An evaluation coefficient ( $k = PV_{top} / PV_{edge}$ ) was introduced to assess the balance between boss finishing and edge removal risk. Standard deviation of PV values across different boss locations was used to evaluate machining uniformity.

Experimental Investigation

Setup: Single-factor experiments were conducted on AM special-shaped waveguides using a two-way abrasive flow machining system.
Variables: Medium viscosity was the sole variable, tested at 1000, 1500, 2000, 2500, and 3000 Pa·s. All other parameters (pressure: 2.5 MPa, cycles: 40, SiC grit size: 36#, concentration: 70%) were held constant.
Evaluation:
- Boss Surfaces: Measured via laser confocal microscope to determine surface roughness reduction rate ( $\Delta Ra\%$ ).
- Edges: Measured via coordinate measuring machine (CMM) to determine edge removal height ( $\Delta H$ ) and profile rounding.
- Uniformity: Assessed by comparing roughness reduction across multiple boss locations (Bosses 1, 4, 5, and 8).

3. Key Results

Theoretical and Simulation Findings

Viscosity-Velocity Relationship: Increasing viscosity reduces the velocity difference between the central flow and the near-wall region, flattening the velocity profile.
Material Removal Capacity: Higher viscosity leads to lower near-wall velocity, dynamic pressure, and strain rate in both boss top surfaces and edge regions. Consequently, the PV value (material removal potential) decreases as viscosity increases.
Uniformity: As viscosity increases, the standard deviation of PV values among different boss locations decreases, indicating improved machining uniformity across the cavity.
Edge vs. Boss Trade-off: While high viscosity reduces removal capacity everywhere, it disproportionately protects edge regions from the high-velocity gradients that cause rounding, thereby increasing the evaluation coefficient $k$ .

Experimental Findings

Surface Roughness: As viscosity increased from 1000 to 3000 Pa·s, the average surface roughness reduction rate of boss surfaces decreased from 83.80% to 67.76%. The final surface roughness ($Ra$) increased from 1.838 μm to 4.222 μm.
Edge Preservation: Conversely, the average edge removal height decreased significantly from 0.1537 mm (at 1000 Pa·s) to 0.0417 mm (at 3000 Pa·s). Low viscosity caused pronounced edge rounding, while high viscosity preserved the edge profile.
Uniformity: The standard deviation of post-machining roughness across key bosses decreased from 0.320 μm to 0.032 μm as viscosity increased, confirming that higher viscosity improves consistency across the cavity.

4. Key Contributions

Mechanism Clarification: The study elucidates the coupled effect of medium viscosity on boss surface material removal, edge removal, and machining uniformity in continuous blind-cavity waveguides.
Theoretical Framework: It establishes a power-law fluid-based relationship linking viscosity to near-wall velocity differences, explaining why higher viscosity flattens velocity profiles and reduces local shear stress.
Optimization Strategy: It identifies a specific viscosity range (2000–2500 Pa·s) that successfully balances the conflicting requirements of effective boss finishing and edge retention.
Experimental Validation: It provides empirical data demonstrating that while low viscosity maximizes efficiency, it risks geometric distortion; high viscosity ensures uniformity and edge safety but reduces efficiency.

5. Significance and Claims

The paper claims that regulating abrasive medium viscosity is a critical lever for finishing additively manufactured special-shaped waveguides. The authors conclude that a viscosity range of 2000–2500 Pa·s is the preferred operating window under their specific processing conditions. Within this range:

The surface roughness reduction rate remains above 70%.
The final surface roughness ($Ra$) is reduced to below 3.7 μm.
The average edge removal height is maintained between 0.0641 mm and 0.0951 mm, significantly lower than low-viscosity conditions.
Machining uniformity among internal bosses is maximized.

The study asserts that this approach allows for the effective finishing of complex internal cavities while preserving the critical structural boundaries necessary for electromagnetic performance, offering a practical solution for the post-processing of AM waveguides without the need for complex tooling or multiple process steps.