Wing-Beat Aerodynamics of the Arrakis Ornithopter Under Desert Thermal and Particulate Loading: A Model Against Two Notional Configurations, Checked Against Carthag and Arrakeen Service Logs, 10180–10191 AG

1. Introduction

Throughout the Imperium the ornithopter is the standard light aircraft. House Atreides brought its craft to Arrakis in 10191 AG, and the Harkonnen administration had flown the same class from Carthag for decades. On Arrakis, however, the type operates close to its limits. Afternoon thermals over the open erg produce abrupt vertical gusts, the lower boundary layer carries a persistent load of fine mineral grains, and the Coriolis storms that cross the planet destroy craft caught in the open. Sorties supporting spice harvesters add a further constraint, because carryall extraction over worm country runs on time budgets measured in seconds.

This paper asks two linked questions. First, which features of the articulated-wing, jet-assisted design account for its tolerance of the Arrakeen flight environment? Second, where does that tolerance end? To isolate the contribution of the wing-beat, we compare the ornithopter against two notional configurations built for the purpose, a jet-driven fixed wing of similar mass and a turboshaft rotor craft. Neither is flown on Arrakis; both are controlled baselines inside the model.

Writing from about 10240 AG, we examine three outcomes: recovery time after a major thermal gust, the rate at which grains adhere to airfoil surfaces, and flight time before the jet-assist cooling intake loses half its flow. Only the last can be checked against field records.

2. System Description

The reference craft is a two-seat ornithopter with articulated wings of 5.2 m semi-span. In powered beat flight each wing strokes at 3.2 Hz, a period of about 0.31 s. The downstroke, which generates most of the lift and part of the thrust, occupies 47% of the cycle. On the upstroke the wing feathers and partly folds, so its leading edge travels inside the wake of the preceding downstroke. The pilot or the stability augmentation linkage resets wing pitch at the top of each stroke, giving a correction opportunity every 0.31 s.

Jet assist supplies the balance of thrust in cruise and the whole of it when the wings are swept back for a high-speed dash; we take the division of lift and thrust between wing and jet from the coupled dynamics of Marn (10235 AG). The jets draw air through a forward intake, an inertial grain separator and a finned cooling screen that rejects heat from the jet case and the wing actuators. Our measurements put the separator's capture efficiency at about 0.995 for grains near the modal diameter (Marn, 10231 AG). The grains that escape it settle on the cooling screen, whose face area in the reference craft is about 2,130 cm². As the screen loads, flow through it falls; when flow has dropped by half, the intake temperature warning trips and the craft must land for cleaning. Dust grounds the aircraft through this intake path.

The two comparison configurations share the reference craft's mass, cruise speed (28 m/s) and cooling screen. The notional fixed wing (4.8 m span, 1.2 m chord) draws all thrust from the jets and uses conventional surfaces with an actuator bandwidth of 0.8 Hz. The notional rotor (3.2 m diameter, 240 rpm) controls through disk tilt, which lags cyclic input by about two and a half revolutions, for an effective bandwidth near 1.6 Hz.

3. Analysis and Model

Thirty-eight in-flight updraft profiles over the Arrakeen interior gave a mean major-gust magnitude of 2.4 m/s (SD 1.2; 95% CI 2.0–2.8), a mean interval between major gusts of 12.3 s (SD 4.1) and a mean gust duration of 4.2 s (SD 1.8) (Threll, 10236 AG). Twenty-nine filter samples from the lower boundary layer gave a suspended-grain concentration of 1,240 grains/cm³ (SD 340; 95% CI 1,120–1,360) with a modal diameter of 9.4 µm (SD 1.1) (Threll, 10237 AG). At a grain density of 2.65 g/cm³ this is about 1.4 g/m³.

Each configuration flew 200 independent stochastic sorties of level cruise at 150 m, with gust and grain parameters drawn from the survey distributions. The replicate count keeps Monte Carlo error on each mean small against the differences of interest (Vessary, 10229 AG). Because the analyst sets the number of runs, simulation p-values carry little meaning; we report effect sizes with Monte Carlo uncertainty and use one-way ANOVA with Tukey HSD only to summarise contrasts.

Surface adhesion was modelled as the product of grain concentration, the airspeed of the surface relative to the grain stream, an impaction efficiency for the 9 µm mode, a retention fraction, and the exposure duty cycle, defined as the fraction of time a given airfoil surface faces the grain stream. The duty cycle is 1.0 for both the fixed wing and the rotor blade and about 0.47 for the flapping wing. The rotor blade meets grains at about 41 m/s at three-quarter radius against 28 m/s for the fixed wing, which raises grain flux by a factor of 1.46 and impaction efficiency by about 12%. Centrifugal loading, however, sheds loosely held grains from the spinning blade, and we set its retention fraction at 0.81 of the fixed-wing value. This retention value is assumed, not measured; together the three factors place rotor adhesion about 1.32 times the fixed wing's.

Intake loading was modelled separately. The rate at which grain mass settles on the cooling screen equals the mass concentration times the intake volumetric flow times the separator pass fraction, divided by screen area. Time to blockage is the screen loading at 50% flow loss, taken as 42 mg/cm² (SD 6 across screen-geometry variants), divided by that rate. Mean intake flow in cruise was 0.62 m³/s for the ornithopter, whose wing-beat carries most of the lift, and 1.24 m³/s for the fixed wing. The rotor's turboshaft draws the same 1.24 m³/s, but downwash recirculation raises the effective concentration at its intake by a factor of about 1.4. The resulting settling rates are about 7.5, 15.0 and 20.9 mg/cm² per hour, giving central blockage times of 5.6, 2.8 and 2.0 h. Every run continued until blockage, so no times are censored.

The ornithopter recovered level flight after a major gust in a mean of 3.2 s, against 5.8 s for the rotor and 8.1 s for the fixed wing (F(2, 597) = 395, η² = 0.57). By Tukey HSD the ornithopter was faster than the fixed wing by 4.9 s (95% CI 4.5–5.3) and faster than the rotor by 2.6 s (2.2–3.0), and the ordering follows the control-update rates of 3.2, 1.6 and 0.8 Hz. Airfoil adhesion followed the duty cycle: 12.4 grains/cm² per hour for the flapping wing, 23.7 for the fixed wing and 31.2 for the rotor (F(2, 597) = 648, η² = 0.68), reductions of 47.7% and 60.3% for the ornithopter. The fixed wing does not rotate, yet it gathers dust almost twice as fast as the flapping wing, so intermittent exposure is the operative mechanism. Median times to intake blockage were 5.6, 2.8 and 2.0 h and means 5.7, 2.9 and 2.0 h, so the ornithopter flew about 2.0 times as long as the fixed wing and 2.8 times as long as the rotor.

4. Validation Against Field Data

Only the intake prediction can be tested against records, and only for the ornithopter. Two register series survive. Harkonnen service registers from Carthag, covering roughly 10180–10190 AG, record cleaning entries for nine airframes (Harkonnen Air Wing, Carthag, 10180–10190 AG). The Arrakeen field logs of the Atreides household guard cover only the few months of the Atreides tenure in 10191 AG and five airframes (House Atreides Household Guard, 10191 AG); they survive as duplicate sheets forwarded to Caladan by household courier during the tenure, the Arrakeen originals having been lost in the Harkonnen attack of that year. Both series mix scheduled cleanings with cleanings forced by the intake temperature warning. We retained only warning-triggered entries that give flight hours since the previous cleaning.

That filter left 31 Harkonnen entries (mean 5.2 h, SD 1.3) and 14 Atreides entries (mean 4.9 h, SD 1.2). The Atreides mean is nominally lower, but the difference of 0.3 h (standard error about 0.4 h) cannot be distinguished at these sample sizes. Pooled, the 45 entries give a mean of 5.1 h (SD 1.3; 95% CI 4.7–5.5). The model mean of 5.7 h exceeds the observed mean by about 0.6 h (95% CI 0.2–1.0), an over-prediction of about 12% of the observed value. That interval combines the uncertainty of the observed mean with Monte Carlo error in the model mean; it omits any systematic error in the 42 mg/cm² threshold. Two features of the records plausibly account for much of the gap. Logged hours were rounded to the quarter hour and often summed across sorties that included landings on open sand, whereas modelled sorties were continuous cruise; and the warning may have been set below 50% flow loss. The registers support the model's scale for the ornithopter, but with a single configuration at one intake flow and coarse logging they cannot test its dependence on intake flow or any finer feature.

No register records gust recovery or adhesion, so those results rest on the model alone.

5. Failure Modes

Storm-scale gusts are the principal limit. We extended the model by sweeping mean gust magnitude from the survey baseline of 2.4 m/s to 15 m/s in seven steps, with 100 runs per step, and scored a run as failed if level flight was not recovered within 30 s. Below about 7 m/s the per-stroke pitch reset kept ornithopter failures under 1%. Near 9 m/s, 4% of runs failed as the required pitch change began to exceed one stroke's authority, and at 15 m/s failures reached 61%. The notional fixed wing was already failing 18% of runs near 9 m/s. Coriolis storms, with winds of several hundred kilometres per hour in Imperial and Fremen accounts, lie far beyond this range and destroy craft they overtake (Kessandre, 10233 AG). Storm frequency varies seasonally and between desert basins, as recent basin-scale storm records show, so the operational defence is forecasting and avoidance. The wing-beat advantage offers no margin at that scale.

Intake blockage, the second failure mode, is worst near the ground, where the jets run at high throttle while the craft's downwash lifts sand. A sortie profile with four landings per hour shortened the ornithopter's median time to blockage from 5.6 h to 4.1 h in the model, which matters for harvester-support crews.

Wear at the wing articulation is the third. The same folding motion that shelters the leading edge on the upstroke opens the hinge line to the grain stream at the top of each stroke. The model omits abrasion, but lengthening the feathered phase to reduce adhesion would raise hinge exposure in proportion.

6. Conclusion

The articulated-wing, jet-assisted ornithopter owes its tolerance of Arrakis to two properties of the wing-beat. A pitch reset every stroke lets it recover from thermal gusts about 2.5 times faster than a notional fixed wing, and intermittent exposure of each airfoil surface roughly halves grain adhesion. A third property belongs to the propulsion balance: because the beating wings carry most of the lift, the jets draw about half the intake flow of a jet-driven fixed wing, which doubles the flight time before the cooling screen clogs. Carthag and Arrakeen service registers put the observed interval about 0.6 h, or 12%, below that prediction, which confirms its scale for the ornithopter alone. The advantage is bounded. Gusts above roughly 9 m/s begin to overwhelm per-stroke control, and a Coriolis storm lies well beyond any regime in which flight technique is the relevant defence.