# Insect-scale flapping flight: a first-order envelope

**Status: calculation, not discovery.** Momentum theory with explicit assumptions. No experiment.

Assumptions: air density 1.2 kg/m³, lift coefficient 1.2, aspect ratio 3, kinematic viscosity 1.5×10⁻⁵ m²/s,
single actuator disk, profile drag coefficient 0.02. Wing loading swept 5–50 N/m².

| Mass | Wing loading | Re | Stroke speed | Specific power | Feasible vs insect muscle (100 W/kg)? |
|---|---:|---:|---:|---:|---|
| 1 mg | 5 N/m² | 142 | 2.63 m/s | 14.6 W/kg | yes |
| 1 mg | 10 N/m² | 142 | 3.73 m/s | 20.6 W/kg | yes |
| 1 mg | 20 N/m² | 142 | 5.27 m/s | 29.2 W/kg | yes |
| 1 mg | 50 N/m² | 142 | 8.33 m/s | 46.1 W/kg | yes |
| 10 mg | 5 N/m² | 449 | 2.63 m/s | 14.6 W/kg | yes |
| 10 mg | 10 N/m² | 449 | 3.73 m/s | 20.6 W/kg | yes |
| 10 mg | 20 N/m² | 449 | 5.27 m/s | 29.2 W/kg | yes |
| 10 mg | 50 N/m² | 449 | 8.33 m/s | 46.1 W/kg | yes |
| 100 mg | 5 N/m² | 1,420 | 2.63 m/s | 14.6 W/kg | yes |
| 100 mg | 10 N/m² | 1,420 | 3.73 m/s | 20.6 W/kg | yes |
| 100 mg | 20 N/m² | 1,420 | 5.27 m/s | 29.2 W/kg | yes |
| 100 mg | 50 N/m² | 1,420 | 8.33 m/s | 46.1 W/kg | yes |
| 1 g | 5 N/m² | 4,492 | 2.63 m/s | 14.6 W/kg | yes |
| 1 g | 10 N/m² | 4,492 | 3.73 m/s | 20.6 W/kg | yes |
| 1 g | 20 N/m² | 4,492 | 5.27 m/s | 29.2 W/kg | yes |
| 1 g | 50 N/m² | 4,492 | 8.33 m/s | 46.1 W/kg | yes |
| 10 g | 5 N/m² | 14,207 | 2.63 m/s | 14.6 W/kg | yes |
| 10 g | 10 N/m² | 14,207 | 3.73 m/s | 20.6 W/kg | yes |
| 10 g | 20 N/m² | 14,207 | 5.27 m/s | 29.2 W/kg | yes |
| 10 g | 50 N/m² | 14,207 | 8.33 m/s | 46.1 W/kg | yes |

## What the envelope says

1. **Reynolds number falls with size**, from ~10⁴ at 10 g to ~10² at 1 mg. Below Re ≈ 10³ viscous forces dominate and steady-state lift coefficients no longer describe the flow — which is why insects use unsteady mechanisms (leading-edge vortices, clap-and-fling) rather than bigger wings.
2. **Specific power is nearly constant across mass at fixed wing loading.** Induced power scales with mass, so the ratio does not move; it scales with the square root of wing loading instead. The size penalty is therefore aerodynamic, not energetic — revised after the first run contradicted the draft text.
3. **The binding constraint is low-Re aerodynamics, not power.** These estimates sit at 15-46 W/kg, below insect muscle, so a motor is not the limiting factor. What is hard is generating useful lift at Re 100-1000, where steady aerodynamics fails and unsteady mechanisms are mandatory.

## Honest limits

- Momentum theory is a hover idealisation; it ignores unsteady lift entirely, which is the dominant mechanism at these Re.
- Real insects vary wing loading by an order of magnitude and use elastic storage that this model omits.
- The actuator figures are order-of-magnitude, not measured for a specific device.
- Nothing here is new physics. It is a checkable envelope that says where to look, not what will work.

## The reverse-engineering point

The interesting question is not 'can we build an insect' but 'which insect solutions are load-bearing'.
Three candidates worth copying, in order of how much they buy you:

1. **Compliant wing hinges** — insects get pitch reversal passively from thorax deformation. A printed flexure replaces an actuator and a control loop.
2. **Elastic energy storage** — resilin returns a large fraction of stroke energy; at 100+ Hz this is the difference between feasible and not.
3. **Compound-eye optics** — wide field of view with no distortion at the edge, at the cost of resolution. Useful for collision sensing, not for imaging.

Each is a design hypothesis with a measurable claim; none is tested here.