Preprint
Article

This version is not peer-reviewed.

A Turbineless Hybrid Electric-Hydrogen Propulsion System with On-Demand Aluminum-Water Hydrogen Generation: Design and Static Thrust Characterization

Submitted:

22 October 2026

Posted:

23 October 2026

You are already at the latest version

Abstract
In this paper a turbineless hybrid electric-hydrogen propulsion system was built and tested under static conditions. Two electric ducted fans (EDFs) supplied the baseline thrust and combustion airflow. An alkaline aluminum-water reactor generated hydrogen on demand for combustion downstream of the fans. Rod immersion depth controlled hydrogen production, which was measured at each condition with a hydrogen-calibrated mass-flow meter. Aluminum consumption was also measured to determine the effective hydrogen yield. Tests covered ten immersion depths, with 76 independent paired EDF-only and hybrid runs per condition. At the highest hydrogen flow, thrust rose from 8.49 ± 0.15 to 9.94 ± 0.21 N, an increase of 17.1%. Hydrogen production increased approximately linearly with exposed aluminum area (R2 = 0.997), while the marginal thrust gain decreased. The momentum-based combustion-to-jet conversion indicator ranged from 7.3 to 9.1% and was highest at intermediate hydrogen flows. Chamber pressure and fan supply power were measured in a separate session with paired runs without and with combustion. Measured thrust was 1.7 to 3.5% below the ideal nozzle prediction from chamber pressure. Combustion increased fan supply power from 1.49 kW to between 1.60 and 2.02 kW. As hydrogen flow increased, jet kinetic power fell from 40% to 19% of the combined electrical and hydrogen chemical exergy input.
Keywords: 
;  ;  ;  ;  ;  

1. Introduction

Aircraft propulsion requires high specific energy and power density in a light, compact, and reliable system. Hydrocarbon fuels meet these requirements with mature storage and combustion technology, but their carbon dioxide emissions and dependence on fossil fuels motivate research into alternative aviation energy carriers [1,2]. The constraints are particularly severe for small aircraft and unmanned aerial vehicles (UAVs), where tanks, power electronics, thermal management, and safety hardware account for a large share of the available mass and volume.
Hydrogen has a lower heating value of approximately 120 MJ kg − 1 , compared with about 43 MJ kg − 1 for conventional jet fuel, and produces no carbon dioxide during direct combustion [1,3]. Storage remains difficult: compressed hydrogen commonly requires containment at 350 to 700 bar, while liquid hydrogen requires temperatures near − 253 ∘ C, insulation, and boil-off management [4,5,6,7]. These storage penalties become more restrictive as vehicle size decreases.
On-demand generation avoids carrying hydrogen in a prefilled gas tank. Aluminum is a stable solid precursor with a high theoretical hydrogen yield through hydrolysis. An alkaline solution disrupts its native oxide layer and allows the reaction to release hydrogen [4,8,9,10,11]. Propulsion use also requires control of the production rate and delivery of the gas to the thrust-producing flow.
The prototype combines an alkaline aluminum-water reactor, two EDFs, a static combustion chamber, and an interchangeable nozzle. The EDFs supply the baseline thrust and combustion airflow. Rod immersion depth regulates hydrogen generation, and the gas is injected downstream of the fans. The experiments test whether this arrangement produces repeatable static thrust augmentation under controlled bench conditions without prefilled high-pressure or cryogenic hydrogen storage.

1.1. Related Work

Relevant research covers hydrogen combustion for aviation, aluminum-water hydrogen production, and hybrid or hydrogen-electric propulsion. Experimental integration of these processes into a single thrust-producing system remains limited.
Hydrogen combustion research has concentrated on gas-turbine and turbofan systems. Energy and exergy studies of small turbojets have found favorable thermodynamic performance relative to conventional jet fuel [3]. Experiments and reviews have examined flashback, nitrogen oxide formation, flame stabilization, injector design, and material compatibility [1,12,13,14]. Small combustors must achieve adequate mixing and residence time over a short flow path, with a greater influence of wall heat transfer [15]. Hydrogen-powered UAV experiments have demonstrated operation on small platforms, generally using gaseous or liquid hydrogen stored onboard [6]. These studies establish the combustion constraints but do not examine direct coupling of an on-demand chemical hydrogen source to combustion in standalone EDF airflow.
Aluminum-water reactions provide hydrogen without compressed or cryogenic gas storage. Aluminum is stable under normal conditions and releases hydrogen when its protective oxide layer is disrupted chemically or mechanically [8,11,16,17]. Aviation experiments have supplied fuel cells with hydrogen from activated aluminum powder [9]. Research on recyclable metal fuels treats aluminum as a solid energy carrier and distinguishes low-temperature hydrogen generation from direct high-temperature metal combustion [18,19]. Recent studies address activation methods, hydrolysis kinetics, and recyclable aluminum-based materials [10,20]. Propulsion also requires the hydrogen supply to follow the operating demand. Mutlu proposed continuous aluminum feeding to regulate fuel-cell hydrogen supply [21]. In the present reactor, rod immersion depth sets the exposed aluminum area, and the generated hydrogen goes directly to a combustor.
Electric, hybrid-electric, and turboelectric aircraft offer flexible control and integration, although endurance depends strongly on the specific energy and power of the onboard energy system [22]. Hydrogen-electric systems normally use a fuel cell to supply electricity to a motor-driven propeller, rotor, or ducted fan [23]. Fuel-cell and battery hybrid propulsion has been studied for lift+cruise eVTOL UAVs to extend battery-only range and endurance [24]. The NederDrone demonstrator uses stored hydrogen and a complete fuel-cell electrical conversion chain [25]. Parallel electric boosting has produced thrust increases on the order of 70% in small-aircraft e-boost configurations [26]. Other studies examine aircraft feasibility, energy management, and fuel-system control [27,28]. Fuel-cell specific power and electrical-powertrain mass and volume affect system performance [23,29]. A related turbineless concept combined a fuel-cell-powered fan and compressor with downstream afterburners; it was assessed thermodynamically, without experimental static-thrust characterization [30].
The literature reviewed here contained no peer-reviewed experiment coupling on-demand aluminum-water hydrogen generation to combustion in standalone EDF airflow with matched EDF-only baselines, controlled hydrogen-flow variation, and static-thrust measurements. Table 1 places the present work within these research areas.

1.2. Contribution and Objectives

This work experimentally couples immersion-depth control of aluminum-water hydrogen generation with direct combustion in standalone EDF airflow and measures static thrust against matched EDF-only baselines. The generated hydrogen burns without prior conversion to electricity, and the flow path contains no turbine. Varying the exposed aluminum area regulates fuel generation for the combustor. This arrangement avoids prefilled high-pressure or cryogenic hydrogen storage.
The experiments cover prototype construction, reactor calibration, matched EDF-only and hybrid thrust measurements, and the response to changes in hydrogen production. A control-volume framework defines the derived performance quantities and their uncertainties. The assessment is limited to static bench operation.

2. System Design and Performance Framework

2.1. Overall Architecture and Operating Principle

The system has an alkaline aluminum-water hydrogen generator, two 70 mm EDFs, a cylindrical combustion chamber, and a converging nozzle. Hydrogen passes through a buffer and regulation line to the chamber; the EDFs provide cold-flow thrust and combustion air. Figure 1 shows the flow paths, Table 2 gives the dimensions and operating parameters, and Figure 2 shows the assembled geometry.
The EDF-only thrust is F EDF , 0 . In hybrid operation, hydrogen is injected downstream of the fans and ignited in the chamber, giving
F tot = F EDF , 0 + Δ F comb
where Δ F comb is the combustion-induced increment. Direct measurements of F EDF , 0 and F tot at matched throttle settings give Δ F comb as their paired difference.

2.2. Hydrogen Generation Subsystem

Hydrogen is generated by reacting aluminum with water in a 5.0 mol L − 1 NaOH solution:
2 Al + 2 NaOH + 6 H 2 O → 2 Na [ Al ( OH ) 4 ] + 3 H 2
NaOH removes the protective oxide layer and sustains the reaction at the aluminum surface [8,10,17]. Solution temperature was not actively controlled and ranged from T sol = 287 to 302 K. Drift during each stable averaging window was below 2 K and was included in the uncertainty calculation.
The theoretical hydrogen-to-aluminum mass ratio from Eq. (2) is
m H 2 m Al th = 3 M H 2 2 M Al ≈ 0.112
with M H 2 = 2.016 g mol − 1 and M Al = 26.98 g mol − 1 . Aluminum consumption and hydrogen production are related through the effective hydrogen-yield factor:
m ˙ H 2 = η r 3 M H 2 2 M Al m ˙ Al
where η r is the measured reaction-yield factor. It accounts for the difference between stoichiometric yield and hydrogen measured downstream of gas conditioning and regulation, including incomplete reaction and losses in the reactor and delivery line. Section 3.4 describes its calibration.
Fuel rate was controlled by the immersed length of 23 cylindrical rods of aluminum alloy 2030, each 0.7 cm in diameter. Each of the 16 calibration runs used a new rod set (Section 3.4). During propulsion tests, the rods and NaOH solution were replaced every 10 runs. The rods stayed fixed in the vessel; adjusting the volume of 5.0 mol L − 1 NaOH solution set the liquid level and immersion depth. Figure 2 shows the reactor arrangement. At immersion depth ℓ imm , the exposed area is
A Al ≈ N π d rod ℓ imm + d rod 4
where the second term accounts for the submerged end face. A Al uses the nominal rod diameter. Propulsion tests covered depths from 2 to 15 cm, corresponding to A Al = 110.0 to 767.5 cm 2 . Although the vessel allowed deeper immersion, preliminary tests above this range produced increasingly vigorous gas evolution and liquid carry-over. The propulsion tests were therefore limited to 15 cm.
Within the measured range, the reactor calibration is
m ˙ H 2 ≈ k A A Al + b
This empirical relation applies to the tested reactor and range. Section 4.2 gives the fitted coefficients.

2.3. Combustion Chamber and Nozzle

The stainless-steel chamber has an inner diameter of 76 mm, a length of 520 mm, and an internal volume of approximately 2.36 L. Hydrogen enters 280 mm downstream of the EDF exit, with the spark source 62 mm farther downstream (Figure 3). The small nozzle exit relative to the chamber cross-section leaves the chamber flow nearly stagnant. Its cold-flow bulk velocity is approximately 7.6 m s − 1 , giving a characteristic residence time of about 70 ms.
The nozzle contracts from the 76 mm chamber diameter over an axial length of 85 mm, with a cone half-angle of approximately 19.3 ∘ . Caliper readings at six angles across the slightly elliptical exit ranged from 16.2 to 16.7 mm, giving
A e = π D max D min 4 = 212.5 mm 2
with D max = 16.7 mm and D min = 16.2 mm. This corresponds to a circular diameter of 16.45 mm and A ch / A e ≈ 21.3 . Throughout the tested range, the nozzle pressure ratio remains below the critical value (Section 4.6), so the exit static pressure equals ambient and thrust reduces to the momentum term. A separate pressure-thrust component is therefore unnecessary.
The fixed spark source ignited the hydrogen and was switched off 5 s after a stable flame formed. In valid runs, the flame stayed downstream of the injection region, with no observed propagation toward the EDFs.

2.4. Control-Volume Performance Quantities

The control volume extends from the quiescent far-upstream ambient to the nozzle exit, with the geometry shown in Figure 3. EDF-only thrust is the reference. Section 4.6 checks thrust against chamber pressure and reports fan supply power with and without combustion.
Gas mass flow is calculated from chamber pressure using the isentropic relations for a converging nozzle discharging to ambient. For π = ( p amb + p ch ) / p amb and γ = 1.4 , the exit Mach number, temperature, and mass flow are
M e = 2 γ − 1 π γ − 1 γ − 1 , T e = T 0 1 + γ − 1 2 M e 2 , m ˙ = p amb R air T e A e M e γ R air T e
with R air = 287 J kg − 1   K − 1 , c p = 1010 J kg − 1   K − 1 , and chamber total temperature T 0 . In hybrid operation, T 0 is the chamber-exit thermocouple reading at each condition. In EDF-only operation, the chamber adds no combustion enthalpy; T 0 is taken from the post-EDF temperature trend extrapolated to m ˙ H 2 = 0 . The pressure ratio π remains below 1.26 , compared with a critical value of 1.893 . Flow is therefore subsonic throughout, with exit static pressure equal to ambient.
The hydrogen chemical-power input is
Q ˙ chem = m ˙ H 2 LHV H 2 , LHV H 2 = 120 MJ kg − 1
For the static, momentum-dominated flow, F ≃ m ˙ U e and P jet = 1 2 m ˙ U e 2 . Eliminating U e gives P jet = F 2 / ( 2 m ˙ ) , used below to calculate the jet-power difference.
Vehicle speed and useful propulsive power are zero in a static test. A momentum-based combustion-to-jet indicator is therefore defined as
η cj = Δ P jet Q ˙ chem , Δ P jet ≈ F tot 2 2 m ˙ g − F EDF , 0 2 2 m ˙ air
where m ˙ air and m ˙ g are the EDF-only and hybrid nozzle mass flows. Both are calculated from Eq. (8) at the corresponding pressure and total temperature. Equation (8) already includes hydrogen in the total gas flow. The numerator covers the whole control volume, including any change in fan work as combustion raises chamber pressure against the fixed nozzle area. The denominator contains only the hydrogen chemical input. Thus, η cj measures incremental jet kinetic power per unit hydrogen chemical power. Ideal constant-pressure heat addition in a cycle with pressure ratio π converts at most 1 − π − ( γ − 1 ) / γ of the heat to jet kinetic power [31]. At the measured ratios of 1.22 to 1.26, this limit is 5.6 to 6.4%. Fan supply power with combustion, reported in Section 4.6, enters the exergy calculation in Section 4.7.
The paired thrust measurements also give the thrust ratio and incremental hydrogen thrust-specific fuel consumption (TSFC):
R F = F tot F EDF , 0 = 1 + Δ F comb F EDF , 0
TSFC H 2 , inc = m ˙ H 2 Δ F comb

2.4.1. Uncertainty Framework

For y = f ( x 1 , … , x n ) with uncorrelated inputs, the combined standard uncertainty is
u ( y ) = ∑ i = 1 n ∂ y ∂ x i u ( x i ) 2 1 / 2
following the Guide to the Expression of Uncertainty in Measurement [32]. Table 3 lists the main contributions. Condition standard deviations and paired-bootstrap confidence intervals describe between-run variability. For a mean of n independent runs, the Type-A standard uncertainty is u A = s / n , where s is the between-run standard deviation. Instrument and calibration contributions in Table 3 are Type-B inputs. Pressure-mean uncertainty combines repeated measurements from the paired session with the adopted Type-B contribution.
The combined standard uncertainty of a condition mean was then
u c = u A 2 + u B 2 1 / 2
The tabled standard deviations describe run-to-run variability; u c describes uncertainty in a condition mean. The separately reported paired-bootstrap confidence intervals exclude instrument-related Type-B effects.
For the directly measured hydrogen flow, the flow-meter specification gives
u B ( m ˙ H 2 ) = 0.008 m ˙ H 2 + 0.002 m ˙ FS , m ˙ FS = 75 mg s − 1
Combined with the standard error of the 76 run-level measurements, this gives relative standard uncertainties of 5.8% at the lowest hydrogen flow and 1.4% at the highest.
Thrust increments were calculated for each matched pair before averaging. Their Type-A uncertainty therefore uses the distribution of 76 pair-level differences. Including the load-cell calibration residual gives
u c ( Δ F comb ) = s Δ F 76 2 + 2 ( 0.004 N ) 2 1 / 2
The relative uncertainty of the mean thrust increment decreases from 5.6% at the lowest flow to 1.9% at the highest.
For the incremental hydrogen TSFC, propagation through Eq. (12) gives
u c ( TSFC H 2 , inc ) TSFC H 2 , inc 2 = u c ( m ˙ H 2 ) m ˙ H 2 2 + u c ( Δ F comb ) Δ F comb 2
The relative standard uncertainty ranges from 2.4% to 8.1% across the ten conditions. At minimum TSFC ( m ˙ H 2 = 9.90 mg s − 1 ), it is approximately 4.5%.
Reaction-yield uncertainty was propagated separately through Eqs. (23) and (25), treating the molar masses as constants:
u ( η r ) η r 2 = u ( m ˙ H 2 , meas ) m ˙ H 2 , meas 2 + u ( Δ m Al ) Δ m Al 2 + u ( Δ t ) Δ t 2
The hydrogen term uses the same calibrated mass-flow meter. Balance and drying/recovery contributions were combined in quadrature. Across the 16 calibration runs, instrument-related relative uncertainty in an individual yield estimate ranged from approximately 1.5% to 11.1%, with the largest values at the two lowest flows.
Pressure Type-A uncertainty uses the n = 4 or 6 measurements per condition in the paired session (Section 3.2):
u c ( p ¯ ch ) = ( s p / n ) 2 + ( 0.50 kPa / 3 ) 2
Combined standard uncertainty was 0.29 to 0.31 kPa for pressure means with combustion and 0.29 kPa without combustion. Propagation to the derived quantities treats the Type-B pressure contribution as common to both states. Its contribution to η cj is 0.47 percentage points at the lowest hydrogen flow and 0.02 to 0.14 percentage points elsewhere.

3. Experimental Methodology

3.1. Prototype Construction

The modular prototype used a bolted flange to join the counter-rotating EDF pair to the upstream end of the stainless-steel chamber, with the converging nozzle at the downstream end. Hydrogen first entered a cylindrical buffer vessel with the same nominal internal dimensions as the reactor: 10.3 cm diameter and 25.4 cm height. Bubbling the gas through dilute acetic-acid solution neutralized entrained sodium-hydroxide aerosol. The gas then passed through an anhydrous CaCl 2 drying column, an adjustable pressure regulator, and the hydrogen mass-flow meter. Approximately 1.2 m of polyurethane tubing with an internal diameter of 8 mm and push-in pneumatic fittings connected the meter to a single-orifice injector of nominal diameter 0.5 mm. The regulator, buffer connection, and fuel line were leak-checked before each test sequence.
At each immersion depth, the regulator was adjusted until hydrogen delivery was stable against chamber back pressure. Delivery pressure therefore varied between conditions. Reported hydrogen flow was measured by the hydrogen-calibrated mass-flow meter; it was not inferred from regulator settings, pressure differences, or injector geometry.
Shutdown controls and fire-extinguishing equipment remained accessible during testing. Personnel stayed clear of the nozzle axis during ignition and operation. Handling of sodium hydroxide and post-reaction liquid followed containment and neutralization procedures. Hydrogen was generated only during testing, with the gas inventory limited by buffer volume and operating pressure.

3.2. Test Setup, Operating Modes, and Airflow Measurement

The propulsion unit was mounted on a load-cell thrust stand. The hydrogen generator sat outside the exhaust path and connected to the chamber through the buffer and regulation line. A BME280 sensor recorded ambient pressure and temperature before each run for the air-density calculation, with stated uncertainties of 100 Pa and 0.5 K. During hybrid runs, a DS18B20 sensor measured temperature downstream of the EDF pair, and a DFRobot DFR0558 K-type thermocouple with MAX31855K conversion measured gas temperature on the flow centerline near the exhaust outlet. Supply voltage and combined EDF current were measured at the DC supply upstream of the two ESCs, giving fan supply power P elec = V I . The reported electrical measurements come from the paired session described below. Figure 4 and Figure 5 show the acquisition interface and test arrangement.
EDF-only runs used no hydrogen injection or ignition and defined the baseline F EDF , 0 . Hybrid runs added hydrogen generation, injection, and combustion at the same EDF command. All operating points used 100% PWM and a nominal 25 V supply. Each hybrid run was paired with an EDF-only run at the same throttle and supply condition.
At each of ten immersion depths, 19 measurement sets of four replicates gave 76 paired EDF-only and hybrid runs. Each run contributed one mean from a 30 s steady interval. The 10 Hz samples within that interval were used to calculate the run mean and were not counted as independent replicates.
Air mass flow is calculated from chamber gauge pressure, total temperature, and nozzle geometry using Eq. (8). The wall pressure tap is perpendicular to the flow and 60 mm upstream of the nozzle contraction. Because chamber flow is nearly stagnant, wall static pressure approximates total pressure. Open-inlet vane-anemometer readings depended on position and orientation and did not provide an area-mean inlet velocity; they were excluded from the flow calculation.
Chamber pressure and fan supply power were measured separately on 17 October 2026 using the same rig, nozzle, and operating point. One ascending pass covered all immersion depths, with six pairs of runs without and with combustion at 2 and 3.5 cm and four pairs at each remaining depth. Five runs without combustion preceded the pairs and three followed them. Each run recorded 30 s means of thrust, chamber pressure, hydrogen flow, chamber-exit temperature, supply voltage, and combined current. Post-EDF temperature was not recorded in this session. All channels were checked against independent instruments beforehand; the current channel agreed with a reference ammeter within its reading resolution. Combustion remained continuous in every run but was unsteady at 2 and 3.5 cm. Supply voltage ranged from 24.8 to 25.2 V. Across the 52 runs without combustion, chamber gauge pressure was 21.71 ± 0.05 kPa, with less than 0.07 kPa difference between the start and end of the session.
For each depth, hybrid gas flow m ˙ g combines the paired-session mean chamber pressure with the mean chamber-exit temperature from the 76 runs. EDF-only air flow m ˙ air uses the paired-session mean pressure without combustion and the post-EDF temperature extrapolated to zero hydrogen flow, 321.6 K. The resulting flows are m ˙ air = 0.0459 kg s − 1 and m ˙ g = 0.0448 to 0.0461 kg s − 1 . Current and power values cover both EDFs together.

3.3. Thrust Measurement and Signal Processing

A 20 N load cell with an HX711 amplifier measured static thrust at 10 Hz. Linear calibration used four reference loads from 5.8 to 12.6 N:
F raw = a S + b
where S is the raw sensor output. Calibration residuals were below 0.004 N. The zero offset was recorded immediately before EDF start-up, and calibration was checked before and after each run and verified with a reference mass after the campaign. Bench mounting fixed the thrust axis, which was leveled with a spirit level. Two control experiments checked for forces unrelated to combustion. Hydrogen injection at the highest flow without ignition produced no change above 0.01 N over five repeats, bounding the contribution from the pressurized fuel line and injection momentum. With the fans off, external chamber heating produced load-cell readings below 0.02 N at body temperatures up to 253 °C. Thermal distortion of the stand therefore made no measurable contribution to the reported increments.
A 10-sample moving average, equivalent to 1.0 s, filtered the calibrated signal. Steady-state values used a 30 s interval after the ignition transient, accepted when filtered thrust varied by less than 5% over 10 s. Peak values were excluded.
For each matched pair i, the thrust increment was calculated before averaging across the condition:
Δ F comb , i = F tot , i − F EDF , 0 , i
The mean and standard deviation of Δ F comb use the 76 pair-level differences. EDF-only and hybrid means use their respective run values. The standard uncertainty of a mean is
u F = σ F n
with σ F taken from the relevant run-level or pair-level distribution.

3.4. Hydrogen Production Measurement

Hydrogen flow was measured in every propulsion test under steady generation and delivery conditions. Buffer accumulation was negligible, so flow to the combustor represented the production rate. Aluminum consumption was measured in 16 calibration runs, each with new rods, to compare metered hydrogen with the stoichiometric rate. Rods were weighed before and after reaction on a balance with 0.1 g resolution. At the end of the prescribed interval, they were quenched in clean water, dried with compressed air, and weighed immediately. Repeated washing, drying, and weighing of reference rods differed by less than 0.1 g, which was included as recovery uncertainty. The dip-to-quench interval Δ t averaged 88.9 s and ranged from 85.8 to 92.1 s. Repeated flow measurements showed an approximately constant production rate after the initial transient for the rest of this interval. Aluminum consumption rate was calculated as
m ˙ Al = Δ m Al Δ t
The 5.0 mol L − 1 solution contained 200 g of NaOH per liter of final solution. A waterproof DS18B20 probe (SONOFF WTS01) recorded solution temperature at the start and end of each run; drift within stable windows was below 2 K. Each calibration run used fresh solution, and the vessel was cleaned between runs.
The 16 independent gas-production calibration runs determined the reaction-yield factor. Gas conditioning comprised bubbling through dilute acetic-acid solution to neutralize entrained sodium-hydroxide aerosol, followed by an anhydrous calcium-chloride column to remove water vapor. Flow was measured downstream of these stages and the regulator, with continuous buffer-pressure monitoring. The Omega FMA-1600A laminar differential-pressure mass-flow meter used its hydrogen calibration and had a nominal full-scale range of 0 to 75 mg s − 1 . The manufacturer specifies measurable flow up to 125% of full scale, accuracy of ± ( 0.8 % of reading + 0.2 % of full scale), repeatability of ± 0.2 % of full scale, and a 200:1 turndown ratio. At its reference conditions of 25 °C and 101.3 kPa, hydrogen density is consistent with the 82.4 mg L − 1 implied by the recorded volumes. Propulsion-test flows covered 4.5 to 44% of full scale. Combined instrument uncertainty ranged from 0.177 mg s − 1 ( 5.2 % ) at the lowest flow to 0.412 mg s − 1 ( 1.3 % ) at the highest. All flows exceeded the 0.375 mg s − 1 lower limit set by the turndown ratio. Metered hydrogen flow was compared with the stoichiometric rate from aluminum loss,
m ˙ H 2 , st = 3 M H 2 2 M Al m ˙ Al
and
η r = m ˙ H 2 , meas m ˙ H 2 , st
Reported hydrogen rates are the flow-meter readings, logged with thrust and exhaust temperature on a common time base. Equation (4) was used to characterize yield, not to calculate those rates. The effective yield factor and its correlation with aluminum consumption come from the same 16 calibration runs; no separate validation dataset was used.

3.5. Validity Criteria

In the 760-run campaign, a pair was accepted when the load-cell record was complete, throttle commands matched, ignition succeeded, the flame remained stable during averaging, and no upstream propagation was observed. The hydrogen line also had to pass its leak check, and thrust had to satisfy the steady-window criterion. Exclusions were limited to technical or procedural failures; low thrust augmentation alone did not exclude a run.

4. Results

The steady-state dataset comprises 760 paired EDF-only and hybrid runs, with 76 pairs at each of ten immersion depths. Reactor calibration used 16 independent runs. Five repeated sequences formed the ignition-aligned transient dataset, and the pressure and power session contained 44 pairs plus eight runs without combustion. One transient run (run 3) was excluded from quantitative analysis as an outlier: its thrust increment (1.98 N) differed from those of the other four runs (1.49–1.61 N) at the 95% level (Dixon’s Q test, Q = 0.77 > 0.71 for n = 5 ). Its trace remains in the ensemble in Figure 6. Ambient conditions were p amb = 100.8 ± 0.2 kPa and T amb = 297.2 ± 0.5 K.

4.1. Empirical Calibration of the Effective Hydrogen Yield

The 16 calibration runs gave
η r = 0.720 ± 0.035 , n = 16
where ± 0.035 is the run-to-run standard deviation. The standard uncertainty of the mean is 0.0088, with a two-sided 95% t-interval of approximately 0.701 to 0.739. The mean lies within the reported range of about 0.65 to 0.80 for comparable alkaline aluminum-water systems [10]. Regression of metered hydrogen flow against the stoichiometric rate from aluminum consumption,
m ˙ H 2 , meas = s r m ˙ H 2 , st + b r
gave s r = 0.724 ± 0.012 , b r = − 0.153 mg s − 1 , and R 2 = 0.996 . This fit describes the calibration data. The effective hydrogen-yield factor in Eq. (26) is the mean of the per-run ratios.

4.2. Hydrogen Production as a Function of Immersion Depth

Across ten immersion depths from 2 to 15 cm, measured hydrogen flow increased from 3.41 ± 0.74 to 32.72 ± 1.89 mg s − 1 . The fitted relation is
m ˙ H 2 = ( 2.356 ± 0.047 ) ℓ imm − ( 1.555 ± 0.444 ) , R 2 = 0.997
with ℓ imm in centimeters. In terms of exposed aluminum area,
m ˙ H 2 = k A A Al + b , k A = 0.0466 ± 0.0009 mg s − 1 cm − 2 , b = − 1.968 ± 0.444 mg s − 1
The calibration applies over the measured 2 to 15 cm range, which provided a hydrogen-flow range of approximately 9.6:1. Table 4 gives the calibration results.

4.3. Static Thrust Response

Transient tests used the maximum immersion depth (15 cm) and steady hydrogen flow at the highest tested level. Figure 6 includes all five traces; the quantitative analysis uses the four retained runs. The figure’s window mean is therefore 10.12 N, compared with 10.06 N for the retained runs. Flame establishment occurred on average 1.28 s after the ignition command, and the spark was switched off 5.00 s later. In the four retained runs, mean thrust rose from 8.51 ± 0.03 N before ignition to 10.06 ± 0.06 N during the common 23.4 to 42.9 s interval. The increment was 1.55 ± 0.05 N, or 18.2%. With all five runs included, the increment is 1.64 ± 0.20 N. A descriptive two-exponential fit gives faster and slower response times of approximately 2.5 and 9 s. The dashed first-order curve is a separate visual guide.
Across the 760 steady-state runs, EDF-only thrust averaged 8.46 ± 0.15 N. The paired thrust increment rose from 0.175 ± 0.070 N at 3.41 ± 0.74 mg s − 1 hydrogen flow to 1.450 ± 0.231 N at 32.72 ± 1.89 mg s − 1 . At the highest flow, matched EDF-only and hybrid thrusts were 8.49 ± 0.15 N and 9.94 ± 0.21 N, respectively, an increase of 17.1% (Table 5). The descriptive quadratic curve in Figure 7(b) was fitted to the ten condition means ( R 2 = 0.995 ).
The derivative of the quadratic fit decreased monotonically from approximately 0.063 to 0.021 N per (mg s − 1 ) across the measured range. Higher-order polynomials gave unstable slopes near the range limits, so the derivative uses the quadratic fit. The declining slope indicates diminishing marginal thrust gain.
Figure 7. (a) Measured hydrogen mass-flow rate versus aluminum-rod immersion depth. (b) Combustion-induced thrust increment versus measured hydrogen mass-flow rate. Markers are condition means from 76 independent paired runs, and error bars are between-run standard deviations. The curves are descriptive least-squares fits of the condition means: linear in panel (a) and quadratic in panel (b).
Figure 7. (a) Measured hydrogen mass-flow rate versus aluminum-rod immersion depth. (b) Combustion-induced thrust increment versus measured hydrogen mass-flow rate. Markers are condition means from 76 independent paired runs, and error bars are between-run standard deviations. The curves are descriptive least-squares fits of the condition means: linear in panel (a) and quadratic in panel (b).
Preprints 239931 g007
Figure 8. Momentum-based combustion-to-jet conversion indicator versus measured hydrogen mass-flow rate. Markers are condition values calculated as the ratio of the mean jet-power increment to the mean chemical-power input from 76 independent paired runs. Error bars and the shaded maximum-location band show paired-bootstrap 95% intervals conditional on fixed pressure inputs. The cubic trend, selected by AICc among orders 2 to 4, has a fitted maximum near 14 mg s − 1 . The right-hand axis shows the marginal thrust gain, obtained as the derivative of a quadratic fit of the thrust means.
Figure 8. Momentum-based combustion-to-jet conversion indicator versus measured hydrogen mass-flow rate. Markers are condition values calculated as the ratio of the mean jet-power increment to the mean chemical-power input from 76 independent paired runs. Error bars and the shaded maximum-location band show paired-bootstrap 95% intervals conditional on fixed pressure inputs. The cubic trend, selected by AICc among orders 2 to 4, has a fitted maximum near 14 mg s − 1 . The right-hand axis shows the marginal thrust gain, obtained as the derivative of a quadratic fit of the thrust means.
Preprints 239931 g008
The thrust ratio R F reached 1.171 at the highest flow. Minimum incremental hydrogen TSFC was 65.2 kg kN − 1   h − 1 near 9.90 mg s − 1 , so the condition giving maximum thrust differed from that using the least hydrogen per unit of added thrust.

4.4. Momentum-Based Conversion Indicator

At each condition, the conversion indicator is the mean jet-power increment divided by the mean chemical-power input. The calculation uses an EDF-only air flow of 0.0459 kg s − 1 and hybrid gas flows from 0.0448 to 0.0461 kg s − 1 , with 0.0454 kg s − 1 at the highest hydrogen flow (Section 3.2). The indicator was 7.29 % at the lowest flow, ranged from 8.50 to 9.07 % between 6.73 and 20.80 mg s − 1 , and fell to between 7.55 and 8.16 % at the four highest flows.
Among polynomial fits of order 2 to 4, the corrected Akaike information criterion (AICc) selects a cubic. It gives R 2 ≈ 0.97 and a maximum of 9.18% near 13.8 mg s − 1 , with a paired-bootstrap 95% interval of 12.7 to 15.4 mg s − 1 for the maximum’s location. This preference depends mainly on the lowest-flow point; excluding it makes AICc favor a quadratic. At 2 cm, the thrust increment was comparable with its run-to-run scatter, giving wider relative uncertainty.
All indicator values exceed the ideal constant-pressure limit of 5.6 to 6.4% (Section 2.4), consistent with the additional fan work in Section 4.6. The global equivalence ratio, calculated from measured hydrogen flow and total gas flow, ranges from 0.0025 to 0.025 . This very lean chamber average is consistent with combustion in locally richer regions near the injector.

4.5. Chemical Input and Estimated Jet-Power Increment

Across the ten conditions, hydrogen chemical input rose from 0.409 ± 0.088 to 3.926 ± 0.226 kW, and the estimated jet-power increment rose from 29.8 ± 13.1 to 302.6 ± 48.1 W. The slower increase in jet power accounts for the lower conversion indicator at high hydrogen flow.
Extrapolating post-EDF temperature to zero hydrogen flow gave T 0 = 321.6 K. Linear and quadratic extrapolations differed by 0.4 K, changing EDF-only mass flow by less than 0.1 % . Combined with the paired-session chamber pressure without combustion, this gives m ˙ air = 0.0459 kg s − 1 . The quantities in Table 6 use the static-bench assumptions of Eq. (10). Measured chamber-exit total temperature increased from 331.7 to 392.4 K; the table includes pressure and temperature so that m ˙ g can be calculated from Eq. (8).
Table 6 also gives f H = ( T 0 − T EDF , out ) / [ Q ˙ chem / ( m ˙ g c p ) ] , the fraction of hydrogen heating value appearing as sensible enthalpy across the chamber. It divides the measured total-temperature rise from fan exit to chamber exit by the rise that full hydrogen heat release would produce in the same gas flow. From 9.90 to 27.49 mg s − 1 , this fraction ranged from 0.861 to 0.891 , then fell to 0.796 and 0.745 at the two highest conditions. At the two lowest conditions, the temperature rise was only 10 to 17 K. Temperature and flow uncertainty then dominate the estimate, so f H is omitted.

4.6. Pressure-Based Check of the Thrust Increment

Chamber gauge pressure reaches 26.0 kPa, keeping the nozzle pressure ratio below 1.26 against a critical value of 1.893 . Exit Mach number ranges from 0.54 to 0.58 . The nozzle remains unchoked, with exit pressure equal to ambient, so thrust is entirely the momentum term. Chamber bulk velocity is 7.6 m s − 1 , giving a dynamic pressure of 38 Pa compared with about 22 kPa static pressure. The wall tap therefore approximates total pressure to within 0.2 %.
The pressure-based thrust estimate follows the one-dimensional nozzle momentum relation
F tot p = m ˙ U e + ( p e − p amb ) A e
For unchoked discharge to ambient, p e = p amb and F tot p = m ˙ U e . Equation (8) gives m ˙ and U e from measured chamber pressure, total temperature, ambient pressure, and the fixed nozzle geometry.
Measured hybrid thrust is 1.7 to 3.5 % below the pressure-based prediction from Eq. (8) (Table 6); EDF-only thrust is 1.1 to 2.1 % below its prediction. In the paired session, which recorded pressure and thrust simultaneously, the measured-to-predicted ratio ranged from 0.957 to 0.994 across 96 runs, with a mean of 0.980 . A real converging nozzle is expected to produce slightly less thrust than the ideal prediction because its discharge and velocity coefficients are below unity.
From 6.73 to 32.72 mg s − 1 , the measured thrust increment is 0.88 to 1.01 of the pressure-predicted increment. At the lowest flow, the ratio is 0.56 and the increment is comparable with its run-to-run scatter. Combustion at 2 cm was unsteady during the paired session, which gave an increment of 0.39 N compared with 0.175 N for the 76-run mean. From 6.73 mg s − 1 upward, paired-session increments agreed with the 76-run results within their scatter. These comparisons support using chamber pressure as a consistency check; the load cell remains the primary thrust measurement.
At fixed nozzle pressure ratio and geometry, Eq. (8) gives m ˙ ∝ T 0 − 1 / 2 and U e ∝ T 0 1 / 2 . Their product m ˙ U e is independent of T 0 , so heating at fixed pressure ratio leaves the modeled momentum thrust unchanged. The increase in chamber pressure links the hybrid state to higher nozzle momentum flux.
In the paired session, combined EDF supply power without combustion averaged 1491 ± 26 W over 52 runs. With combustion, it rose from 1604 W at 2 cm to 2023 W at 15 cm, an increase of 9 to 34% over the paired runs without combustion. Ideal compression work estimated from the measured pressure ratio rose by only 31 to 133 W. Most of the extra electrical power therefore did not become flow work, consistent with reduced fan efficiency as combustion raised back pressure. At every condition, the electrical-power increase exceeded Δ P jet .

4.7. Exergy Assessment of the Propulsion Subsystem

The exergy assessment compares the thrust increase with use of the combined electrical and hydrogen chemical inputs. Its boundary begins at the measured DC supply upstream of the two ESCs and includes the electrical drive, both EDFs, chamber, and nozzle. It excludes the aluminum-water reactor and the production or regeneration of aluminum and other reactants.
The accounted input is
E ˙ in , acc = P elec + m ˙ H 2 e ch , H 2 , e ch , H 2 = 117 MJ kg − 1 .
Electrical power is electrical exergy. The hydrogen chemical term uses the tabulated value adopted by Rakopoulos et al. [33], with a standard reference environment of 298.15 K and 101.325 kPa and mole fractions of 0.7567 N 2 , 0.2035 O 2 , 0.0303 H 2 O, 0.0003 CO 2 , and 0.0092 trace gases. This value differs from the 120 MJ kg − 1 LHV. Physical stream exergy uses the measured ambient state, T a = 297.2 K and p a = 100.8 kPa; the chemical value is retained without an ambient-state correction.
With constant-property ideal-gas relations, the non-kinetic physical outlet exergy is
E ˙ ph , out = m ˙ g c p T e − T a − T a ln T e T a + R T a ln p e p a
This expression follows from ( h − h a ) − T a ( s − s a ) . The modeled exit has p e = p a , so the pressure term is zero. Static exit temperature T e comes from Eq. (8). Using chamber total temperature together with the full kinetic term would count kinetic energy twice. Jet kinetic power and its fraction of the accounted input are
P jet ≈ F 2 2 m ˙ g , χ jet = P jet E ˙ in , acc
The ratio expresses jet kinetic power as a fraction of accounted exergy input.
Electrical input uses the mean supply power with combustion from the paired session. As hydrogen flow increases from 3.41 to 32.72 mg s − 1 , accounted input rises from 2.00 to 5.85 kW, jet power from 0.80 to 1.09 kW, and non-kinetic physical outlet exergy from 0.02 to 0.33 kW. Static exit temperatures at the ends of the range are approximately 313 and 368 K. Despite increasing thrust, χ jet falls at every step, from 40.0% to 18.6%.
Pressure uncertainty was propagated jointly through m ˙ g , T e , P jet , and E ˙ ph , out . It contributes approximately 0.27 and 0.10 percentage points of standard uncertainty to χ jet at the lowest and highest hydrogen flows. The corresponding unallocated balances are about 1.18 and 4.43 kW. The unallocated balance comprises exhaust chemical exergy, heat loss, and exergy destruction.

5. Discussion

5.1. Thrust Response and Operating Limits

Absolute thrust increased throughout the tested hydrogen-flow range, with decreasing marginal gain. The highest flows gave lower η cj values, but the tests did not reach a hard thrust limit.
Chamber pressure in the paired session rose almost linearly with hydrogen flow. A descriptive quadratic fit gives slopes of about 0.14 kPa per mg s − 1 at the lowest flows and 0.11 kPa per mg s − 1 at the highest. Fan supply power rose by 9 to 34% (Section 4.6), so additional fan work contributed to the thrust increment.
The enthalpy fraction f H in Table 6 stays between 0.86 and 0.89 from 9.90 to 27.49 mg s − 1 , then falls to 0.80 and 0.75 . At the two highest flows, a smaller fraction of hydrogen chemical input therefore appears as estimated sensible-enthalpy rise. Mixing, combustion completeness, wall heat transfer, and the representativeness of the local temperature reading may contribute. Compact hydrogen combustors face similar mixing and residence-time constraints [14,15].
During valid runs, the flame stayed downstream of the injector after ignition and persisted after the spark was switched off at 5 s, with no observed propagation toward the EDFs. These observations establish repeatable flame persistence in this geometry.
Maximum thrust occurred at the highest tested flow, while minimum incremental hydrogen TSFC occurred near 10 mg s − 1 . This trade-off was measured at 100% EDF throttle. Other throttle settings would change air mass flow, chamber pressure, hydrogen-to-air ratio, residence time, and the fan’s response to combustion back pressure, so the optimum hydrogen flow will differ.

5.2. Immersion-Depth Control of Hydrogen Production

Rod immersion depth controlled fuel rate reproducibly within the calibration range. Mutlu’s mechanically regulated aluminum-feeding system [21] regulates hydrogen supply through feed rate; this reactor varies exposed surface area.
The measured yield factor, η r = 0.720 ± 0.035 , falls within the reported range of about 0.65 to 0.80 for comparable alkaline aluminum-water systems [10]. The calibration is specific to this reactor and tested range. Longer operation may change reaction rate through heating, product accumulation, or solution aging. Practical use also requires reaction-product management and regeneration [18,19].

5.3. Comparison with Existing Propulsion Approaches

The present system generates hydrogen on demand and burns it in EDF airflow. Other architectures in Table 7 store hydrogen for turbine combustion or convert it to electricity before propulsion. Avoiding hydrogen storage and electrochemical conversion introduces requirements for reactor hardware, reactants, product handling, and thermal management.
The measured 17 to 18% static-thrust augmentation is below the cited parallel-electric boost. The systems are not normalized by installed mass or stored energy. A vehicle comparison must include total installed mass, usable energy, endurance, and balance-of-plant requirements.

5.4. Limitations

All tests were static. Gas mass flow was inferred from chamber pressure, total temperature, and nozzle geometry using isentropic relations, with uncertainty from exit area and the approximation of total pressure by the wall tap. Because pressure and fan supply power were measured separately from the 76-run thrust dataset, the derived flows combine two sessions (Section 3.2). Hybrid total temperature was measured at the chamber exit, while the EDF-only value of 321.6 K was extrapolated from post-EDF temperature at zero hydrogen flow. No independent area-mean airflow measurement, direct EDF-only temperature measurement, nozzle-exit pressure measurement, or spatial exit velocity profile was available. These limits leave η cj as a momentum-based indicator and prevent a complete energy balance.
Forward flight introduces inlet velocity and ram pressure, which can change the EDF operating point, chamber pressure, nozzle pressure ratio, and mass flow. Useful propulsive power also depends on flight speed. The static increments therefore cannot be converted directly into in-flight thrust gains or conventional propulsive efficiency.
Local hydrogen concentration, flame structure, exhaust composition, and NO x were not measured. Run durations were too short to establish long-term reactor stability or component durability. EDF-only thrust also drifted from 8.41 to 8.51 N during the sequence.

5.5. Future Work

A full campaign should measure thrust, chamber pressure, electrical power, and gas temperatures simultaneously, together with nozzle-exit pressure and velocity. Varying EDF throttle and hydrogen flow would test the pressure-based interpretation and locate efficient operating conditions for a given thrust demand using both electrical and chemical inputs. Transient tests should examine flame stability and closed-loop immersion-depth control, accounting for the measured response times and decreasing marginal thrust gain.
An energy balance should quantify exhaust enthalpy, wall heat transfer, and incomplete heat release alongside electrical and hydrogen chemical inputs. It could then assess recovery of exhaust or reactor heat for hydrolysis-temperature control or air preheating, subject to pressure-loss constraints. Vehicle assessment requires longer tests, flight-representative inlet conditions, and exhaust measurements of unburned hydrogen and NO x , as well as installed mass and balance-of-plant requirements. It should also include the life-cycle effects of aluminum production, sodium-hydroxide regeneration, and reaction-product recovery [18,19].

6. Conclusions

The prototype produced repeatable static thrust augmentation by generating hydrogen on demand from aluminum and water and burning it in EDF airflow, without prefilled high-pressure or cryogenic hydrogen storage. At the highest tested flow, thrust rose from 8.49 ± 0.15 to 9.94 ± 0.21 N, an increase of 17.1% over the matched EDF-only baseline. Rod immersion depth controlled hydrogen supply, with an approximately linear relation between exposed aluminum area and flow ( R 2 = 0.997 ). More hydrogen increased absolute thrust but reduced marginal gain. The momentum-based conversion indicator ranged from 7.3 to 9.1% and was highest at intermediate flows.
In the paired pressure and power session, measured thrust was 1.7 to 3.5% below the ideal nozzle prediction, and combustion increased fan supply power by 9 to 34%. As hydrogen flow rose, jet kinetic power fell from 40% to 19% of the combined electrical and hydrogen chemical input within the stated subsystem boundary. Vehicle-level performance still requires assessment of durability, installed mass, and flight operation.
Declaration of Prior Presentation: An earlier version of this work was presented by Ermis V. Tatsis at the 36th European Union Contest for Young Scientists (EUCYS), held in Riga, Latvia, from 15 to 20 September 2025, under the title “Hybrid turbineless propulsion engine with on-board aluminum-based hydrogen production,” where it received a second prize in physics.1 The contest entry was not published in proceedings or in any other archival form. The present manuscript substantially extends that work through additional measurements, an uncertainty analysis, and quantitative characterization of the coupled hydrogen-generation and propulsion system.

Funding

This research received no external funding.

Data Availability Statement

The measurement data of this study (thrust, hydrogen flow, and temperatures of the 760 paired runs; the five transient thrust records; and chamber pressure, thrust, and fan supply power of the paired session) and the analysis script are openly available in Zenodo at https://doi.org/10.5281/zenodo.23541391.

Conflicts of Interest

The author declares no competing financial or non-financial interests.

Use of Artificial Intelligence

Claude (Anthropic) were used for language editing. The author checked all results and takes full responsibility for the content of the article.

References

  1. Verstraete, D. Long range transport aircraft using hydrogen fuel. Int. J. Hydrogen Energy 2013, 38, 14824–14831. [Google Scholar] [CrossRef]
  2. Hoelzen, J.; Silberhorn, D.; Zill, T.; Bensmann, B.; Hanke-Rauschenbach, R. Hydrogen-powered aviation and its reliance on green hydrogen infrastructure—Review and research gaps. Int. J. Hydrogen Energy 2022, 47, 3108–3130. [Google Scholar] [CrossRef]
  3. Akdeniz, H.Y.; Balli, O. Energetic and exergetic assessment of operating biofuel, hydrogen and conventional JP-8 in a J69 type of aircraft turbojet engine. J. Therm. Anal. Calorim. 2021, 146, 1709–1721. [Google Scholar] [CrossRef]
  4. Akbar, D.; Unal, D. On-demand hydrogen production and storage via the aluminum–water reaction: a strategic evaluation. Chem. Pap. 2025, 79, 6975–6983. [Google Scholar] [CrossRef]
  5. Eberle, U.; Mueller, B.; von Helmolt, R. Fuel cell electric vehicles and hydrogen infrastructure: status 2012. Energy Environ. Sci. 2012, 5, 8780–8798. [Google Scholar] [CrossRef]
  6. Gavrilovic, N.; Carbonneau, X.; Moschetta, J.M.; Prakash, R.; Schimpf, J.; Kim, S.Y. Ground testing and analysis of a liquid-hydrogen propulsion system for UAVs. Int. J. Hydrogen Energy 2025, 187, 151960. [Google Scholar] [CrossRef]
  7. Ebrahimi, A.; Rolt, A.; Jafari, S.; Huete Anton, J. A review on liquid hydrogen fuel systems in aircraft applications for gas turbine engines. Int. J. Hydrogen Energy 2024, 91, 88–105. [Google Scholar] [CrossRef]
  8. Petrovic, J.; Thomas, G. Reaction of Aluminum with Water to Produce Hydrogen: A Study of Issues Related to the Use of Aluminum for On-Board Vehicular Hydrogen Storage (Version 2 - 2010 Update); Technical report; US Department of Energy (USDOE): Washington, DC (United States). Office of Energy Efficiency and Renewable Energy (EERE), Office of Sustainable Transportation; Hydrogen Fuel Cell Technologies Office (HFTO), 2011. [Google Scholar] [CrossRef] [PubMed]
  9. Elitzur, S.; Rosenband, V.; Gany, A. On-board hydrogen production for auxiliary power in passenger aircraft. Int. J. Hydrogen Energy 2017, 42, 14003–14009. [Google Scholar] [CrossRef]
  10. Gai, W.Z.; Deng, Z.Y. Enhanced hydrogen production from Al-water reaction: Strategies, performances, mechanisms and applications. Renew. Energy 2024, 226, 120397. [Google Scholar] [CrossRef]
  11. Gao, X.; Wang, C.; Hou, Y.; Zhao, L.; Bai, W.; Che, D. Experimental study on characteristics and mechanism of hydrogen production from aluminium–water reaction for carbon-free power generation. Fuel 2025, 380, 133206. [Google Scholar] [CrossRef]
  12. Manigandan, S.; Praveenkumar, T.R.; Ryu, J.I.; Verma, T.N.; Pugazhendhi, A. Role of hydrogen on aviation sector: A review on hydrogen storage, fuel flexibility, flame stability, and emissions reduction on gas turbines engines. Fuel 2023, 352, 129064. [Google Scholar] [CrossRef]
  13. Cameretti, M.C.; De Robbio, R.; Ferrara, V.; Tuccillo, R. Performance and Emissions Evaluation of a Turbofan Burner with Hydrogen Fuel. Aerospace 2025, 12, 231. [Google Scholar] [CrossRef]
  14. Clemen, C.; Ravikanti, M.; La Bianca, N.; Eggels, R.; Wurm, B.; Young, K. Considerations for Hydrogen Fueled Aerospace Gas Turbine Combustion Sub-System Design. In Proceedings of the ASME Turbo Expo 2024: Turbomachinery Technical Conference and Exposition, London, United Kingdom, June 2024; Volume 3A: Combustion, Fuels, and Emissions, p. V03AT04A016. [Google Scholar] [CrossRef]
  15. Shehata, A.M.; Schoen, M.P. Comprehensive Survey of Micro Turbojet Experimentation: Sensor Technologies, Methodologies, and Research Trends. Machines 2025, 13, 989. [Google Scholar] [CrossRef]
  16. Parmuzina, A.V.; Kravchenko, O.V. Activation of aluminium metal to evolve hydrogen from water. Int. J. Hydrogen Energy 2008, 33, 3073–3076. [Google Scholar] [CrossRef]
  17. Kravchenko, O.V.; Semenenko, K.N.; Bulychev, B.M.; Kalmykov, K.B. Activation of aluminum metal and its reaction with water. J. Alloys Compd. 2005, 397, 58–62. [Google Scholar] [CrossRef]
  18. Bergthorson, J.; Goroshin, S.; Soo, M.; Julien, P.; Palecka, J.; Frost, D.; Jarvis, D. Direct combustion of recyclable metal fuels for zero-carbon heat and power. Appl. Energy 2015, 160, 368–382. [Google Scholar] [CrossRef]
  19. Bergthorson, J.M. Recyclable metal fuels for clean and compact zero-carbon power. Prog. Energy Combust. Sci. 2018, 68, 169–196. [Google Scholar] [CrossRef]
  20. Wang, H.; Ma, H.; Jin, Z.; Shi, J.; Zhang, Z.; Wei, C.; Gao, Q.; Hou, G. Sn-free Al-based on-demand hydrogen production materials for easy recycling. Int. J. Hydrogen Energy 2024, 49, 1332–1343. [Google Scholar] [CrossRef]
  21. Mutlu, R.N.; Kandasamy, J.; Kiymaz, T.B.; Guleryuz, D.; Boncu, E.; Eroglu, E.; Gokalp, I. Optimization of aluminum hydrolysis reactions and reactor design for continuous hydrogen production using aluminum wire feeding. Int. J. Hydrogen Energy 2024, 52, 1390–1403. [Google Scholar] [CrossRef]
  22. Brelje, B.J.; Martins, J.R.R.A. Electric, hybrid, and turboelectric fixed-wing aircraft: A review of concepts, models, and design approaches. Prog. Aerosp. Sci. 2019, 104, 1–19. [Google Scholar] [CrossRef]
  23. Ji, Z.; Qin, J.; Cheng, K.; Zhang, S.; Wang, Z. A comprehensive evaluation of ducted fan hybrid engines integrated with fuel cells for sustainable aviation. Renew. Sustain. Energy Rev. 2023, 185, 113567. [Google Scholar] [CrossRef]
  24. An, J.H.; Kwon, D.Y.; Jeon, K.S.; Tyan, M.; Lee, J.W. Advanced Sizing Methodology for a Multi-Mode eVTOL UAV Powered by a Hydrogen Fuel Cell and Battery. Aerospace 2022, 9, 71. [Google Scholar] [CrossRef]
  25. De Wagter, C.; Remes, B.; Smeur, E.; van Tienen, F.; Ruijsink, R.; van Hecke, K.; van der Horst, E. The NederDrone: A hybrid lift, hybrid energy hydrogen UAV. Int. J. Hydrogen Energy 2021, 46, 16003–16018. [Google Scholar] [CrossRef]
  26. Weintraub, D.; Koppelberg, J.; Köhler, J.; Jeschke, P. Ducted fans for hybrid electric propulsion of small aircraft. CEAS Aeronaut. J. 2022, 13, 471–485. [Google Scholar] [CrossRef]
  27. Fuel Cells and Hydrogen 2 Joint Undertaking; Clean Sky 2 Joint Undertaking. Hydrogen-powered aviation: A fact-based study of hydrogen technology, economics, and climate impact by 2050; Technical report; Publications Office of the European Union: Luxembourg, Study prepared by McKinsey & Company for the Clean Sky 2 Joint Undertaking and the Fuel Cells and Hydrogen 2 Joint Undertaking; 2020. [Google Scholar] [CrossRef]
  28. Plummer, A.; Adeyemi, D.; Sell, N.; Sciatti, F.; Tamburrano, P.; Amirante, R. Fuel System Control for Hydrogen-Powered Aircraft. In Proceedings of the 2024 UKACC 14th International Conference on Control (CONTROL 2024); IEEE, 2024; pp. 201–202. [Google Scholar] [CrossRef]
  29. Soleymani, M.; Mostafavi, V.; Hébert, M.; Kelouwani, S.; Boulon, L. Hydrogen propulsion systems for aircraft: A review on recent advances and ongoing challenges. Int. J. Hydrogen Energy 2024, 91, 137–171. [Google Scholar] [CrossRef]
  30. Ji, Z.; Qin, J.; Cheng, K.; Zhang, S.; Dong, P. Performance assessment of a solid oxide fuel cell turbine-less jet hybrid engine integrated with a fan and afterburners. Aerosp. Sci. Technol. 2021, 116, 106800. [Google Scholar] [CrossRef]
  31. Hill, P.G.; Peterson, C.R. Mechanics and Thermodynamics of Propulsion, 2nd ed.; Addison-Wesley: Reading, MA, USA, 1992. [Google Scholar]
  32. Joint Committee for Guides in Metrology. Evaluation of measurement data—Guide to the expression of uncertainty in measurement, 2008. GUM 1995 with minor corrections. corrected version 2010. [CrossRef]
  33. Rakopoulos, D.C.; Rakopoulos, C.D.; Kosmadakis, G.M.; Giakoumis, E.G.; Kyritsis, D.C. Exergy Analysis in Highly Hydrogen-Enriched Methane Fueled Spark-Ignition Engine at Diverse Equivalence Ratios via Two-Zone Quasi-Dimensional Modeling. Energies 2024, 17, 3964. [Google Scholar] [CrossRef]
1
Contest results: https://www.eucys2025.eu/en/participants/eucys-2025-winners/ (accessed 8 October 2026).
Figure 1. System layout and principal air, hydrogen, electrical, and combustion-energy paths.
Figure 1. System layout and principal air, hydrogen, electrical, and combustion-energy paths.
Preprints 239931 g001
Figure 2. CAD representation of the integrated propulsion system.
Figure 2. CAD representation of the integrated propulsion system.
Preprints 239931 g002
Figure 3. Combustion-chamber geometry (section view).
Figure 3. Combustion-chamber geometry (section view).
Preprints 239931 g003
Figure 4. Acquisition and control interface showing the reactor pump controls, liquid-detection interlock, the two EDF channels at 100% PWM, and readings of exhaust temperature, hydrogen mass-flow rate, and static thrust.
Figure 4. Acquisition and control interface showing the reactor pump controls, liquid-detection interlock, the two EDF channels at 100% PWM, and readings of exhaust temperature, hydrogen mass-flow rate, and static thrust.
Preprints 239931 g004
Figure 5. Static-thrust test setup. The propulsion unit is carried on two low-friction linear rails and reacts against the load cell, so that only the axial component is transmitted and thermal growth of the chamber is not restrained. Air enters through the inlet of the counter-rotating EDF pair, driven by the two ESCs from the 25 V supply. Hydrogen is delivered through the injector 280 mm downstream of the EDF exit and ignited by the spark plug fed from the high-voltage module; a second, unused spark-plug port is provided further downstream. The downstream chamber pressure tap and the high-temperature thermocouple used for the reported measurements are visible near the downstream end. The interchangeable nozzle is shown detached, so the exhaust exit appears here as the bare chamber opening. Load-cell, pressure, thermocouple, and hydrogen-flow signals are collected by the Arduino unit and logged on a common time base through the acquisition interface of Figure 4.
Figure 5. Static-thrust test setup. The propulsion unit is carried on two low-friction linear rails and reacts against the load cell, so that only the axial component is transmitted and thermal growth of the chamber is not restrained. Air enters through the inlet of the counter-rotating EDF pair, driven by the two ESCs from the 25 V supply. Hydrogen is delivered through the injector 280 mm downstream of the EDF exit and ignited by the spark plug fed from the high-voltage module; a second, unused spark-plug port is provided further downstream. The downstream chamber pressure tap and the high-temperature thermocouple used for the reported measurements are visible near the downstream end. The interchangeable nozzle is shown detached, so the exhaust exit appears here as the bare chamber opening. Load-cell, pressure, thermocouple, and hydrogen-flow signals are collected by the Arduino unit and logged on a common time base through the acquisition interface of Figure 4.
Preprints 239931 g005
Figure 6. Ignition-aligned mean thrust from repeated runs. The shaded region is ± 1 standard deviation across all five recorded traces; the quantitative values in the text are taken from the four runs retained after the outlier test. Vertical markers identify ignition, flame establishment, and spark deactivation. The dashed line is a descriptive guide to the transition and is not used in the quantitative analysis.
Figure 6. Ignition-aligned mean thrust from repeated runs. The shaded region is ± 1 standard deviation across all five recorded traces; the quantitative values in the text are taken from the four runs retained after the outlier test. Vertical markers identify ignition, flame establishment, and spark deactivation. The dashed line is a descriptive guide to the transition and is not used in the quantitative analysis.
Preprints 239931 g006
Table 1. Position of the present work relative to the main related research areas.
Table 1. Position of the present work relative to the main related research areas.
Research stream Representative focus Scope limitation Role in the present study
Hydrogen combustion for aviation Hydrogen combustion in gas turbines, turbofan burners, and compact combustors [3,13,14] Generally assumes turbine-based propulsion and stored hydrogen; does not couple chemical generation to EDF-supplied airflow Provides combustion context and flame-stability constraints
Aluminum-water hydrogen generation Activated aluminum, alkaline hydrolysis, recyclable metal fuels, and mechanically regulated hydrogen production [8,9,10,11,18,19,21] Focuses mainly on hydrogen generation, storage replacement, or fuel-cell supply; no direct combustion thrust measurement in an air-breathing prototype Provides onboard hydrogen source and fuel-rate control mechanism
Hydrogen-electric and hybrid-electric propulsion Fuel-cell electric propulsion, EDF/propeller systems, UAV demonstrators, and hybrid-electric architectures [22,23,24,25,26] Hydrogen is typically converted to electricity before propulsion; thrust is limited by fuel-cell and electrical powertrain characteristics Provides electric baseline thrust and controllable airflow
Direct combustion augmentation of EDF flow Related studies have modeled fuel-cell-powered fan systems with downstream afterburners [30] and examined hydrogen combustion in conventional turbofan burners [13] No peer-reviewed experimental characterization was identified for direct hydrogen combustion in standalone EDF airflow, including EDF-only baselines, hydrogen-flow variation, thrust response, and diminishing-gain behavior Defines the immediate experimental gap addressed here
Present work Turbineless hybrid electric-hydrogen propulsion system with on-demand aluminum-water hydrogen generation and static thrust characterization Bench-scale and static conditions only; no flight testing or full propulsion-cycle optimization Integrates generation, fuel-rate control, EDF airflow, combustion, and thrust measurement
Table 2. Main system parameters and subsystem functions.
Table 2. Main system parameters and subsystem functions.
Subsystem Main parameters Role in the system
Hydrogen generation subsystem Cylindrical alkaline aluminum-water reactor (10.3 cm internal diameter, 25.4 cm internal height); 23 aluminum rods; tested immersion depth 2–15 cm; two-stage gas conditioning: buffer vessel containing dilute acetic-acid solution and anhydrous CaCl 2 drying column; regulation line; exposed aluminum surface area A Al [ cm 2 ] Produces hydrogen on demand; A Al , set through the rod immersion depth, is used as the fuel-rate control variable
Electric ducted fans Two EDF units (70 mm, 12-blade, counter-rotating) mounted in series in a common duct; supply voltage 25 V; motor constant: 2100 rpm  V − 1 ; open annular inlet area A eff = 25.6   cm 2 (single annulus, hub subtracted); throttle fixed at 100% PWM for all reported tests Provide the EDF-only static thrust baseline F EDF , 0 and the forced airflow m ˙ air supplied to the combustion chamber
Combustion chamber Static cylindrical chamber; inner diameter: 76 mm; length: 520 mm; injection port location: 280 mm downstream of EDF exit; spark igniter location: 62 mm downstream of injection port Receives EDF-supplied airflow and injected hydrogen; provides the control volume where combustion increases gas temperature and exhaust momentum
Interchangeable nozzle Reported-test geometry: simple converging cone; chamber-side diameter 76 mm; slightly elliptical exit, 16.2 mm × 16.7 mm (exit area 212.5  mm 2 ); axial converging length 85 mm; cone half-angle approximately 19.3 ∘ ; material: stainless steel; interchangeable mounting Expands and accelerates the heated exhaust flow; defines the exit-flow area used in the reported performance quantities
Table 3. Primary measurement channels and uncertainty inputs used in the experimental and derived quantities. Instrument specifications and observed run-to-run scatter are stated separately where applicable.
Table 3. Primary measurement channels and uncertainty inputs used in the experimental and derived quantities. Instrument specifications and observed run-to-run scatter are stated separately where applicable.
Quantity Instrument or method Uncertainty / specification Used in
Static thrust F 20 N load cell with HX711 amplifier; linear calibration with 5.8–12.6 N reference loads Calibration residual below 0.004 N; run-to-run SD 0.14–0.23 N; standard error of condition means ≤ 0.027  N for n = 76 F EDF , 0 , F tot , Δ F comb , η cj , TSFC H 2 , inc
Chamber gauge pressure p ch AMS 5915-0500-D differential-pressure sensor at the chamber wall tap; paired session: 4 or 6 runs without and with combustion per condition 0–50 kPa differential-pressure range; adopted Type-B standard uncertainty u B ( p ) = 0.289  kPa; dynamic-pressure correction below 0.2% of reading m ˙ air , m ˙ g , F tot p , η cj
Nozzle exit area A e Caliper measurement of the elliptical exit, 16.2 mm × 16.7 mm Area uncertainty: 1.8 × 10 − 6   m 2 ( ± 0.1  mm on each axis) m ˙ air , m ˙ g
Ambient pressure and temperature BME280 pressure/temperature sensor p amb : 100 Pa; T amb : 0.5 K ρ air , m ˙
Hydrogen mass flow m ˙ H 2 Omega FMA-1600A laminar differential-pressure mass-flow meter, nominal full-scale range 0–75 mg  s − 1 ; measurable flow up to 125% of full scale; hydrogen calibration ± ( 0.8 % of reading + 0.2 % of full scale); 5.2% at the lowest and 1.3% at the highest tested flow Q ˙ chem , η cj , TSFC H 2 , inc
Post-EDF and chamber-exit temperature DS18B20 post-EDF sensor; DFRobot DFR0558 K-type thermocouple with MAX31855K conversion for chamber-exit temperature DS18B20 accuracy: ± 0.5 ∘ C from − 10 to 85 ∘ C; K-type probe range: 0– 800 ∘ C; converter error within ± 2 ∘ C below 700 ∘ C; probe error within ± 2.5 ∘ C; chamber-exit run-to-run SD: 3.5–12.2 K m ˙ air , m ˙ g , η cj , f H
EDF supply voltage Waveshare Current/Power Monitor HAT (INA219); voltage channel only 0–26 V measurement range; 12-bit conversion P elec
Combined EDF current LEM HAS 100-S Hall-effect current transducer 100 A nominal current; ± 4  V output; manufacturer accuracy: 1% P elec
Aluminum mass change Δ m Al Pre/post-reaction balance measurement Balance resolution: 0.1 g; drying/recovery uncertainty: 0.1 g η r calibration
Reaction time Δ t Recorded dip-to-quench interval Timing resolution: 1 s η r calibration
Immersion depth and exposed Al area NaOH liquid-level setting and measured rod geometry Immersion-depth resolution: 0.1 cm; rod-diameter uncertainty: 0.01 cm A Al calibration
Table 4. Hydrogen-generation calibration results.
Table 4. Hydrogen-generation calibration results.
Calibration item n Estimate Intercept R 2
Per-run reaction-yield ratio 16 η r = 0.720 ± 0.035 - -
Measured versus stoichiometric H 2 rate 16 s r = 0.724 ± 0.012 b r = − 0.153  mg s s − 1 0.996
H 2 rate versus immersion depth 10 2.356 ± 0.047 mg  s − 1   cm − 1 − 1.555 ± 0.444 mg  s − 1 0.997
H 2 rate versus exposed Al area 10 k A = 0.0466 ± 0.0009 mg  s − 1   cm − 2 b = − 1.968 ± 0.444 mg  s − 1 0.997
Table 5. Static thrust and derived performance quantities for 76 independent paired EDF-only and hybrid runs per operating point. η cj is reported as the ratio of means with a paired-bootstrap 95% confidence interval in brackets, conditional on fixed pressure inputs, as in Figure 8. All other quantities are condition means ± between-run standard deviations.
Table 5. Static thrust and derived performance quantities for 76 independent paired EDF-only and hybrid runs per operating point. η cj is reported as the ratio of means with a paired-bootstrap 95% confidence interval in brackets, conditional on fixed pressure inputs, as in Figure 8. All other quantities are condition means ± between-run standard deviations.
ℓ imm
[cm]
m ˙ H 2
[mg  s − 1 ]
n F EDF , 0
[N]
F tot
[N]
Δ F comb
[N]
η cj [%]
(95% CI)
R F
[-]
TSFC H 2 , inc
[kg  kN − 1   h − 1 ]
2 3.41 ± 0.74 76 8.411 ± 0.149 8.585 ± 0.174 0.175 ± 0.070 7.29 [ 6.74 , 7.84 ] 1.021 70.2
3.5 6.73 ± 0.62 76 8.418 ± 0.153 8.758 ± 0.159 0.341 ± 0.116 8.50 [ 7.90 , 9.12 ] 1.040 71.1
5 9.90 ± 0.84 76 8.424 ± 0.165 8.971 ± 0.167 0.546 ± 0.173 9.05 [ 8.44 , 9.65 ] 1.065 65.2
6.5 13.32 ± 0.75 76 8.455 ± 0.153 9.178 ± 0.176 0.723 ± 0.194 8.98 [ 8.43 , 9.52 ] 1.085 66.3
8 16.94 ± 0.81 76 8.456 ± 0.152 9.376 ± 0.175 0.920 ± 0.206 9.07 [ 8.61 , 9.53 ] 1.109 66.3
9.5 20.80 ± 0.90 76 8.458 ± 0.148 9.528 ± 0.182 1.070 ± 0.204 8.81 [ 8.44 , 9.18 ] 1.126 70.0
11 25.11 ± 0.84 76 8.457 ± 0.152 9.630 ± 0.173 1.173 ± 0.199 8.16 [ 7.87 , 8.45 ] 1.139 77.0
12 27.49 ± 1.03 76 8.476 ± 0.154 9.701 ± 0.183 1.225 ± 0.225 8.10 [ 7.79 , 8.43 ] 1.145 80.8
13.5 30.66 ± 1.58 76 8.507 ± 0.136 9.818 ± 0.200 1.311 ± 0.223 7.55 [ 7.25 , 7.84 ] 1.154 84.2
15 32.72 ± 1.89 76 8.485 ± 0.155 9.935 ± 0.206 1.450 ± 0.231 7.71 [ 7.41 , 8.00 ] 1.171 81.2
Table 6. Chemical input and momentum-based jet-power increment for the ten hybrid operating points. Hydrogen flow, chemical input, and jet-power increment are condition means ± between-run standard deviations. η cj is the ratio of means; its paired-bootstrap confidence intervals are reported in Table 5 and Figure 8. p ch is the mean chamber gauge pressure with combustion measured in the paired session, reported as mean ± sample SD ( n = 6 at the two lowest conditions, n = 4 at the others). F tot p is the hybrid thrust predicted from that pressure through Eq. (8), independently of the load cell; the last column is the ratio of the measured to the predicted value. T 0 is the measured chamber-exit total temperature and f H is the fraction of the hydrogen heating value appearing as sensible enthalpy across the chamber, with a paired-bootstrap 95% confidence interval in brackets, conditional on fixed pressure inputs; f H is not reported at the two lowest conditions (Section 4.5).
Table 6. Chemical input and momentum-based jet-power increment for the ten hybrid operating points. Hydrogen flow, chemical input, and jet-power increment are condition means ± between-run standard deviations. η cj is the ratio of means; its paired-bootstrap confidence intervals are reported in Table 5 and Figure 8. p ch is the mean chamber gauge pressure with combustion measured in the paired session, reported as mean ± sample SD ( n = 6 at the two lowest conditions, n = 4 at the others). F tot p is the hybrid thrust predicted from that pressure through Eq. (8), independently of the load cell; the last column is the ratio of the measured to the predicted value. T 0 is the measured chamber-exit total temperature and f H is the fraction of the hydrogen heating value appearing as sensible enthalpy across the chamber, with a paired-bootstrap 95% confidence interval in brackets, conditional on fixed pressure inputs; f H is not reported at the two lowest conditions (Section 4.5).
m ˙ H 2
[mg  s − 1 ]
Q ˙ chem
[kW]
p ch
[kPa]
T 0
[K]
Δ P jet
[W]
η cj
[%]
f H [-]
(95% CI)
F tot p
[N]
F tot / F tot p
[-]
3.41 ± 0.74 0.409 ± 0.088 22.54 ± 0.27 331.7 29.8 ± 13.1 7.29 – 8.899 0.965
6.73 ± 0.62 0.808 ± 0.074 22.74 ± 0.15 340.9 68.7 ± 22.0 8.50 – 8.973 0.976
9.90 ± 0.84 1.188 ± 0.101 23.16 ± 0.10 346.1 107.5 ± 32.8 9.05 0.867 [ 0.807 , 0.926 ] 9.125 0.983
13.32 ± 0.75 1.598 ± 0.090 23.77 ± 0.05 355.4 143.5 ± 37.8 8.98 0.889 [ 0.845 , 0.932 ] 9.350 0.982
16.94 ± 0.81 2.033 ± 0.098 24.32 ± 0.04 364.0 184.3 ± 40.5 9.07 0.875 [ 0.824 , 0.925 ] 9.550 0.982
20.80 ± 0.90 2.496 ± 0.108 24.79 ± 0.14 374.2 219.8 ± 40.9 8.81 0.881 [ 0.850 , 0.913 ] 9.723 0.980
25.11 ± 0.84 3.013 ± 0.101 25.25 ± 0.06 384.2 246.0 ± 40.1 8.16 0.861 [ 0.826 , 0.897 ] 9.889 0.974
27.49 ± 1.03 3.299 ± 0.123 25.36 ± 0.04 392.9 267.4 ± 45.8 8.10 0.891 [ 0.863 , 0.917 ] 9.927 0.977
30.66 ± 1.58 3.679 ± 0.190 25.81 ± 0.05 392.6 277.7 ± 46.4 7.55 0.796 [ 0.768 , 0.824 ] 10.091 0.973
32.72 ± 1.89 3.926 ± 0.226 26.03 ± 0.09 392.4 302.6 ± 48.1 7.71 0.745 [ 0.716 , 0.775 ] 10.169 0.977
Table 7. Architectural comparison with selected hydrogen and hybrid-electric propulsion approaches. Values are representative of different system scales and are not directly mass-normalized.
Table 7. Architectural comparison with selected hydrogen and hybrid-electric propulsion approaches. Values are representative of different system scales and are not directly mass-normalized.
Approach Energy or hydrogen pathway Representative reported metric Principal comparison with the present work
Hydrogen turbine or turbofan combustion [3,13,14] Stored hydrogen; direct combustion in gas-turbine or turbofan flow Hydrogen LHV approximately 120 MJ  kg − 1 ; aircraft-engine combustion and power scale High absolute power and established turbomachinery framework, but requires hydrogen storage and a turbine-based propulsion cycle
Hydrogen fuel-cell UAV [23,25] Stored hydrogen converted to electricity before motor-driven propulsion Multi-hour UAV endurance has been demonstrated; the NederDrone reported approximately 3.5 h endurance High electrical conversion efficiency and endurance, but retains hydrogen storage and the complete fuel-cell electrical conversion chain
Parallel electric e-boost [26] Battery electrical energy supplied to an auxiliary or boosted electric propulsor Reported thrust increment on the order of 70% Larger reported boost fraction, obtained through additional battery and electrical-system mass rather than removal of fuel storage
Mechanically regulated aluminum-water hydrogen generation [10,21] Solid aluminum and water converted to hydrogen on demand Reported conversion factors approximately 0.65–0.80; mechanical fuel-rate control demonstrated for hydrogen supply Avoids stored hydrogen, but most reported systems supply a fuel cell or external hydrogen user rather than a directly coupled thrust-producing combustor
Present work Aluminum-NaOH hydrogen generation coupled to direct combustion in EDF airflow m ˙ H 2 = 3.41 –32.72 mg s s − 1 ; maximum Δ F comb ≈ 1.45  N (approximately 17–18%); maximum η cj ≈ 9.1 % Bench-scale static integration without prefilled high-pressure or cryogenic hydrogen storage; lower augmentation than e-boost, but with a different mass and energy-carrier trade-off
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.