Could Rough Be the New Smooth?

Making surfaces smoother is often a reliable way to reduce resistance in engineered systems. A smoother pipe, wing, vehicle body, or machined surface usually sounds like the obvious path to lower drag or friction. In aerodynamics, especially, smoother surfaces have long been associated with cleaner flow and reduced drag. But a recent study from Tohoku University adds an interesting wrinkle to that rule of thumb: under certain test conditions, an almost imperceptibly rough surface reduced aerodynamic drag on a streamlined model.

The keyword here is certain. This is not evidence that roughening airplane surfaces would directly cut fuel use. The 43.6% figure discussed below refers to aerodynamic drag measured on a specific wind-tunnel model under controlled transitional-flow conditions, not to fuel consumption, operating cost, or total drag on a real aircraft. What the Tohoku team demonstrated is narrower and arguably more interesting: a specific microscopic, irregular surface texture changed airflow around a streamlined test model in a way that reduced measured drag in the transitional-flow regime.

The surface treatment is called distributed micro-roughness, or DMR. Unlike ordered grooves or riblets, DMR consists of randomly distributed micron-scale roughness features. The experiments used a glass-bead DMR in phase I and two sandblasted concave DMR surfaces, called DMR1 and DMR2, in phase II. The glass-bead version used beads with diameters of 38-53 micrometers; the sandblasted surfaces had different measured roughness parameters and depression patterns.

That matters because this was not ordinary roughness. The DMR height was approximately 1% of the local boundary-layer thickness, small enough that the paper describes it as much smaller than typical roughness elements. Tohoku’s Institute of Fluid Science also characterizes the surface as nearly “smooth” from a fluid-engineering standpoint. In other words, the notable finding is that very small, deliberately distributed roughness affected the boundary layer in a measurable, drag-reducing way under these conditions.

To understand why that is intriguing, it is helpful to zoom in on the boundary layer: the thin region of air that clings to a surface as flow moves over it. In simplified terms, air in the boundary layer may remain orderly and layered (laminar flow), or it may become more chaotic (turbulent flow). Laminar flow generally produces less skin-friction drag, while turbulent flow usually increases it. The messy middle, where the flow changes from one state to the other, is called the transition region.

The Tohoku study focused on that transition behavior. The researchers tested a streamlined model using the university’s 1-meter magnetic suspension and balance system, or MSBS. Instead of holding the model in the wind tunnel with rods or wires, the system suspends it electromagnetically.

The test article itself was a streamlined body about 1.069 meters long, with a central cylindrical section about 0.40 meters long and 0.10 meters in diameter. The paper reports that two rows of artificial tripping tape were applied to the leading edge to induce boundary-layer transition in the low-turbulence wind tunnel. The headline result comes from a controlled model experiment, not a full-scale vehicle operating in everyday conditions.

In phase II of the experiment, using tripping-tape configuration 2, the smooth model’s critical Reynolds number for drag rise was about 1.9 × 10⁶. With the DMR1 and DMR2 surfaces, that shifted to about 2.2 × 10⁶. The largest reported reduction was 43.6%, observed for DMR2 compared with the plain, smooth surface at approximately Re = 2.25 × 10⁶ in the transition region. The paper also reports that DMR1 and DMR2 maintained lower drag coefficients than the smooth surface up to the highest measured Reynolds number, about 3.6 × 10⁶.

The 43.6% figure should not be read as a 43.6% reduction in fuel consumption, operating cost, or total vehicle drag for any real aircraft or vehicle. It also does not mean that roughness is always good. In fact, the same paper notes that initial experiments without artificial disturbances found that a glass-DMR coating generally promoted earlier transition, consistent with the conventional idea that roughness can trip a boundary layer into turbulence.

Another useful comparison is a dimpled sphere, like a golf ball. Dimples can reduce pressure drag by changing the separation behavior around a blunt body. The Tohoku result appears to be different. The researchers used wall-resolved large eddy simulations and oil-flow visualization to examine whether the DMR benefit came from suppressing separation. Their analysis estimated the pressure-drag budget at about Cp ≈ 0.00021 at Re = 3.6 × 10⁶, while the observed drag reduction was around ΔCD ≈ 0.001. Even eliminating that pressure-drag contribution entirely could explain only about 20% of the observed reduction, according to the study.

That points back to the boundary layer. The paper attributes the main benefit to changes in skin-friction drag through modification of the boundary-layer state, rather than to separation control. Oil-flow visualization supported that interpretation. At low Reynolds number (Re = 1.2 × 10⁶), both smooth and glass-DMR cases showed localized oil accumulation near the tail, yet their measured drag coefficients were identical. At a higher Reynolds number (Re = 3.4 × 10⁶), the oil moved downstream smoothly for both surfaces, supporting the conclusion that separation was not the dominant explanation.

For engineers, that may be the most satisfying part of this story. Surfaces interact with flow in ways that depend on scale, geometry, velocity, disturbance environment, and boundary-layer state. A texture that is harmful in one regime may be helpful in another. A feature that looks insignificant to the eye may be large enough for the air to notice. And a design rule that is perfectly useful most of the time may still have exceptions worth exploring.