Applications in Energy and Combustion Science 19 (2024) 100282

Contents lists available at ScienceDirect

Applications in Energy and Combustion Science

journal homepage: www.sciencedirect.com/journal/applications-in-energy-and-combustion-science

Scale effects on rotating detonation rocket engine operation

Tyler Mundt , Carl Knowlen * , Mitsuru Kurosaka
University of Washington, Seattle, WA, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Rotating detonation rocket engines are propulsive devices employing detonation waves moving circumferentially
Rotating-detonation around an annular channel that consume axially fed propellants. Theoretically, this provides benefits with
Pressure-gain-combustion respect to combustion pressure gain and thermodynamic efficiency when compared to deflagration-based
RDRE-combustor-scaling
combustors. To facilitate size scaling of these devices, the relationships between geometric parameters, perfor-
mance, and wave dynamics have been investigated with gaseous methane-oxygen propellant. Empirical relations
were derived between combustor geometry, fueling conditions, and engine operation, as well as correlation to
thermodynamic parameters calculated with chemical kinetics codes. The radius of curvature effects were
explored in annular combustors having outer diameters of 25 mm, 51 mm, and 76 mm with a fixed gap width of
5 mm. The injectors were scaled to have same oxidizer-to-fuel injector port area ratio, impingement distance, and
injector-to-gap area ratio. Larger combustors had higher wave counts during operation at a given mass flux and
equivalence ratio. Combustor axial pressures were found to be more dependent on propellant mass flux and
equivalence ratio than geometry. Mass flux and the inner-to-outer radius ratio, the latter of which was related to
other geometric ratios, dictated the operating mode transition thresholds and the number of resulting waves,
respectively.

1. Introduction and fundamentals In an effort to advance understanding of governing principles of

RDRE design, an experimental investigation utilizing three RDRE con-
1.1. Rotating detonation rocket engine figurations at three diameter scales was conducted to explore scaling
laws associated with cylindrical annulus sizing. This research sought
Rotating detonation rocket engines (RDRE) utilize continuously- new insights into operational limits and dynamics at small diameter
spinning detonation waves within an annular channel to combust re- scales, with potential application to satellite propulsion systems. Radius
actants fed in from one end [1]. A depiction of the annular flow field is of curvature effects were investigated by varying the combustor outer
shown in Fig. 1 [2]. A RDRE is considered a pressure gain combustor diameter while holding annular gap width and injection characteristics
(PGC), which, unlike traditional deflagration combustor designs, theo- constant. The RDRE combustor characteristics and wave dynamics
retically exhibits a total pressure increase within the combustor. If arising from variations in geometry, fueling conditions, and thermody-
realized, a PGC would lead to a reduction in the demands, and therefore namic properties such as Chapman-Jouguet (CJ) detonation speed and
size and weight, of supporting hardware in propulsion systems, cell size from 286 tests were reported here.
including turbopumps and compressors, and holds potential efficiency
advantages [3]. Recent advances in RDRE development include
demonstration of pulsed operation for attitude control [4], the first 1.2. RDRE scaling and design review
successful in-flight demonstration within the environment of space [5]
and testing of a 30 kN thrust class engine at NASA Marshall Space Flight The results from prior experimental and numerical efforts to deter-
Center [6]. Even with this progress, the benefits of RDREs expressed by mine the influence of design characteristics on RDRE operation,
theory have not been realized, and the fundamentals of design and including injection design, combustor geometry, and exit conditions are
performance are not well identified [7]. presented here. Prior publications that have presented some of the data
provided herein are identified and the novelty of this body of work is

* Correspondence author at: Research Associate Professor University of Washington W.E.B. Dept. of Aeronautics & Astronautics 3940 Benton Lane, UW Mail
352400 Seattle, WA 98195.
E-mail address: [email protected] (C. Knowlen).

https://doi.org/10.1016/j.jaecs.2024.100282
Received 2 April 2024; Received in revised form 22 July 2024; Accepted 24 July 2024
Available online 27 July 2024
2666-352X/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

the range of mass flux changes tested here.

1.2.2. Combustor geometry

The effects of radius of curvature on the operation of RDREs have
been investigated numerically and experimentally. A computational
study in two-dimensional channels by Kim and Choi [18] showed higher
wave stability at increased radius of curvature. Experiments by Ishii
et al. [19] demonstrated a methane-oxygen engine with a larger outer
diameter, but the same channel width and mass flux, mflux, operated at a
higher wave count. Furthermore, the mflux at which wave count transi-
tions occurred was different across the two outer diameter scales. The
smaller diameter combustor also exhibited a larger increase in wave
speed with mflux across the same wave count than the larger engine.
Zhdan et al. [20] discussed the effect of combustor length on an
RDRE and found it had little impact on operation, as long as the
combustor length was greater than a minimum required for the trans-
Fig. 1. Rotating detonation rocket engine combustor with impinging in- verse detonation wave to be self-sustaining. Testing with a
jectors. [2]. methane-oxygen RDRE with convergent nozzles by Bennewitz et al. [21]
at two different lengths showed no change in operability, but small in-
highlighted. creases in thrust and wave speeds were seen in the shorter configuration.
In addition, the shorter configuration saw a reduction in
1.2.1. Injector and plenum considerations counter-rotating-wave activity. These results were confirmed in a
Propellant injection plays a strong role in RDRE performance as consequent numerical study by Ross et al. [22], where wave dynamics
illustrated by Wyatt et al. [8] with experiments in an ethylene-nitrous varied with chamber length in combustors with constricted exits.
oxide engine. Using different injection designs with the same The impacts of alterations to the annular gap width have been
combustor, differences were seen in both wave spin speed and detonable explored experimentally by Bigler et al. [23] with a 76-mm-diameter
mass flow rate range. These results were attributed to mixing quality, methane-oxygen RDRE. At constant mass flow rate, the smallest gap
which concured with numerical simulations by Yoshidomi et al. [9] that width tested showed the highest performance, but had the lowest
showed wave spin speed and pressure gain increased with increasing operating range with respect to ER. When scaled by mflux, the largest gap
number of injector pairs and better mixing. An experimental investiga- width had the highest chamber pressure and comparable thrust to the
tion by Bigler et al. [10] showed the effects of mixing on thrust per- baseline configuration, whereas the smallest gap width showed uni-
formance when utilizing impinging and misaligned injectors in the same versally lower performance. Loss mechanisms for the smallest width,
combustor. The misaligned injector with inferior mixing produced fewer including heat loss due to increased combustor surface area and higher
waves traveling at lower speeds that resulted in lower thrust and specific plenum pressures that demonstrated increased injector-plenum
impulse, but exhibited no changes in detonative operating limits with coupling, were discussed. Experiments with variation of annular gap
respect to equivalence ratio. Interaction of a detonation wave with a width in a hydrogen-air annular combustor at constant mflux by Brophy
particular RDRE injector at different levels of mixing has been explored et al. [24] also found the largest width to have the highest pressure gain.
in detonation channel tests by Redhal et al. [11]. Mixing was found to Heat transfer losses were again discussed, with the combustor
affect the structure of the passing detonation wave and its connections to volume-to-surface area ratio per unit length highlighted as an important
counter rotating-rotating waves. The detonation wave standoff distance value.
seen during nominal RDRE operation was also discussed.
An investigation of characteristic timescales in RDREs by Bennewitz 1.2.3. Related research publications
et al. [12] found that there was a coupling between detonation perfor- The 76-mm-diameter engine configuration investigated here has the
mance and the recovery time of the injectors. Interactions between the main parameters of the RDRE test rig used in a model validation for
plena and detonation wave have also been seen experimentally, e.g., propulsion program organized by the Air Force Research Laboratory
such as by Feleo et al. [13], even with nominally choked injectors. (AFRL) at Edwards AFB [25]. Data from both engines have been pro-
Paxson and Miki [14] explored the influence of backflow through the vided for a Computational Fluid Dynamics (CFD) validation effort being
injectors numerically. They found that injectors with low diodicity, carried out at several universities. Example CFD results compared to the
which had more backflow, exhibited lower performance. A further generated data were presented by Strakey and Ferguson [26], as well as
exploration of this result was performed numerically by Keller et al. [15] Pal et al. [27], with results of the experimental effort reported in Ben-
on an additively manufactured, high-diodicity impinging injector newitz et al. [28]. This AFRL program used a specific test matrix span-
design, with production of a cold flow test article underway. ning a range of mflux at a specified ER and a range of ER at constant mflux.
A numerical investigation by Schwer and Kailasanath [16] showed The same AFRL test matrix was used for the 76-mm-RDRE investigated
that performance increased as the ratio between total injection area and here and was scaled on a mflux basis for the smaller engine
annulus area increased. This simulation did not include the effects of configurations.
backflow through the injectors. The effect of the injector area ratio on Prior articles that reported some of the data presented herein include
RDRE operation was later studied experimentally by Koch et al. [17] a gap-width investigation in the 25-mm-RDRE that utilized inner cores
using radial injectors in a 76-mm-diameter methane-oxygen engine. with different diameters in [29], the AFRL test matrix data for
Reported performance metrics included specific impulse based on 76-mm-RDRE experiments with emphasis on axial pressure distributions
thrusting pressure, taken from the upstream wall of the combustor, and and back pressure influence in [30], and a subset of the available data
the pressure ratio across the injectors. As the injector area ratio was for 25-mm and 51-mm RDRE experiments with emphasis on the facility
increased, the pressure ratio across the injectors decreased, leading to description and data processing approach in [31]. The plena pressuri-
enhanced injector-plenum coupling and reduction in specific impulse. zation and injector pressure drops for these three RDREs when operating
The wave dynamics were found to be strongly influenced by the injector under the same ER and mflux fueling conditions were compared in [32].
area ratio and equivalence ratio, ER, but were relatively insensitive to The current paper collates an expanded data set from the 25-mm,
51-mm, and 76-mm diameter combustors that includes the AFRL test

2
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

matrix results as a subset. It also numerically determined the thermo- in the combustor annulus. Each reactant delivery line had an indepen-
dynamic properties of the propellants and compared them to experi- dent nitrogen supply for purging. The mass flow rate metering was
mental results, which provided unique insights into RDRE operability as enabled by Flow-Dyne critical flow nozzles with precision throat areas,
the combustor size scale was varied. ACN. These contoured sonic nozzles were insensitive to downstream
pressure changes with pressure ratios as low as 1.2 [35]. The throat
2. Experiment design diameters of the oxidizer nozzles were 4.20 mm and 2.18 mm and the
fuel nozzles were 2.79 mm and 1.27 mm. Based on calibrations provided
The main goals of this geometric scaling investigation were to by Flow-Dyne, the uncertainty in the product of discharge coefficient,
determine scaling laws for RDREs using empirical data and explore the Cd, was at most ±0.5 % at Reynolds numbers greater than 5 × 105 that
operability and characteristics of RDREs at diameter scales of 76 mm were typical in operation. Pressure and temperature in the calibrated
and below. Because the capabilities of the test facility and the details of approach tubes to the critical flow nozzle were monitored by NoShok
the RDRE geometries have been reported in detail elsewhere [31,33], 615–2000–2–1–2–8-CC pressure transducers (±0.125 % uncertainty)
only a summary of the experimental apparatus is provided here. and low-noise OMEGA Type-K exposed bead thermocouples (±0.75 %
uncertainty), respectively. Control of the upstream pressures, and thus
the flow rates, were governed by TESCOM 44–4000 Series air load
2.1. Facility description regulators controlled by ER5000 electro-pneumatic actuator units. The
ER5000s utilized the building air supply and a hand-tuned PID loop to
The University of Washington Rotating Detonation Engine Lab has maintain predetermined set points for the pressures upstream of the
successfully conducted testing with cylindrical RDREs ranging in outer- critical nozzles during engine operation and minimize their settling time
diameters from 10 mm to 154 mm using both gaseous methane-oxygen during engine startup.
and hydrogen-oxygen mixtures [33,34]. The indoor facility captures Detonation initiation was accomplished by a spark-ignited pre-
combustion effluent during experiments and can set sub-atmospheric detonator system. Auxiliary 1000 cc cylinders were filled to nominal
exhaust conditions. While the test stand does not provide thrust mea- levels of 1380 kPa for methane and 4140 kPa for oxygen. During engine
surements, the RDRE combustor exit is optically accessible by a startup, the reactants were injected by matching Parker Series 9 mini-
high-speed camera. The facility is also equipped with both low- and ature calibrant valves into opposing ends of a stainless steel cross. The
high-speed data acquisition systems (DAQ) and its propellant delivery cross was connected with a 6.4-mm-diameter stainless steel tube, having
system can provide flow rates between 8 g/s and 450 g/s of stoichio- a length-to-inner diameter ratio of 50, to one of the radial instrument
metric methane-oxygen. The test stand with a 154-mm-RDRE installed is ports in the combustor. To ignite the engine, the calibrant valves were
shown in Fig. 2. left open for 200 ms to fill the cross and attachment tube. Then, 5 ms
The RDRE test rigs were mounted to a 0.25-m-inner diameter after closing the calibrant valves, a co-annular gap spark plug was fired
exhaust duct that routed combustion products into a 4 m3 dump tank at 500 Hz for 100 ms. Upon firing the pre-detonator, the initial transient
containing 500 kg of aluminum plates to cool and condense combustion wave behavior settled within 200 ms into one of several potential steady
products. After each experiment, the dump tank contents were purged state operating modes. The typical operating modes were single rotating
by a Kinney KT-300 vacuum pump, which was also used to set sub- wave, multiple co-rotating waves, co-rotating with counter rotating-
atmospheric back pressure as low as 2 kPa. The exhaust duct featured rotating waves, deflagration without coherent rotating waves, axial
a wye section with a straight branch having a 12.7-mm-thick poly- pulsations, and no combustion. Images of these wave systems were in
carbonate window that was 2.5 m from the test rig mounting plane. This Ref. [36].
enabled observation of combustor wave dynamics through the exit of
the RDRE annulus with a Phantom v1211 high-speed camera using a
Nikon 70 mm to 200 mm zoom lens.
Fuel and oxidizer were handled independently with reactants mixing

Fig. 2. Rotating Detonation Engine Lab test stand with 154-mm-RDRE.

3
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

2.2. Engine design Table 1

Combustor dimensions for each RDRE configuration.
Testing with the three test rigs scaled by the outer diameter of the RDRE Combustor Feature 76-FI-5 51-FI-5 25-FI-5
combustor annulus determined the influence of radius of curvature. An
Annulus Outer Diameter (mm) 76.2 50.8 25.4
overview of these geometries were provided here, with RDRE design Annular Gap (mm) 5.00 5.00 5.00
details presented in Ref. [36]. Tested combustors with diameters of Annulus Length (mm) 76.2 50.8 25.4
76.2-mm, 50.8-mm, and 25.4-mm and inner core lengths equaling their
outer diameters were shown in Fig. 3. Annulus width was kept at 5 mm
for all configurations with the intent to provide a constant ratio between 2.2.3. 25-mm-RDRE design
gap and detonation cell width. Injector jets impinged at an angle of 60◦ The 25-mm-RDRE utilized the same front endplate, oxidizer ring,
and were separated radially at the same distance to keep impingement and mounting flange as the 51-mm-RDRE. The test rig was designed with
height constant in all RDRE configurations. The net injector-to-annulus the same exterior diameter and one more combustor pressure tap than
area ratios were kept constant so that the coupling between the plena the 51-mm-RDRE, as shown in Fig. 3c. Reusing the front endplate led to
and combustor would be similar [17]. The port area ratios of the in- modifications of the injector design, prompting the need to bolt the
jectors were nominally the same so that tests conducted at identical mflux injector to the internal fuel distributer section of the endplate. Another
conditions would have similar plena pressures for all configurations. The departure from previous engines was removal of the sealing surface on
test rigs were identified by their outer combustor diameter (mm), flat the inner core, which permitted core changes without the need for leak
face impinging (FI), and gap width (mm) as indicated in Table 1. testing. The tolerances of the injector and inner cores resulted in a these
components being centered to within ±0.013 mm and ±0.025 mm,
2.2.1. 76-mm-RDRE design respectively, with respect to the outer core. More details of this 25-mm-
The installation of the injector and inner core in the 76-mm-RDRE is RDRE design are provided in Ref. [29].
shown in Fig. 3a. To minimize non-uniformity of pressure in the plena,
they were axially fed through four distributor lines each (see Fig. 2). This 2.2.4. Combustor instrumentation
combustor contained 24 instrument ports in its 152-mm-long outer wall, The quantity and placement of instrument ports along the outer core
however, not all ports were used. The relatively thin outer wall of the differed with each engine. Their locations in terms of axial distance from
combustor (12.7 mm) necessitated the use of a stainless steel shell to injector face, clocking with respect to the top of the engine, and port
provide additional strength. More details of this RDRE design are pro- diameters are listed in
vided in Ref. [30]. This RDRE configuration was used as the basis for Table 2. All standoff tubes used for static pressure stations were 1.40-
smaller scale combustor designs. mm-ID with a length-to-inner diameter ratio of 145.
Combustor pressures, including that on the front wall of the 76-mm-
2.2.2. 51-mm-RDRE design RDRE, were monitored with OMEGA PX319–500A5V pressure trans-
The 51-mm-RDRE oxidizer plenum was radially-fed with four ducers (±0.25 % uncertainty for all PX319s). Plena pressures for all
distributor lines (9.5-mm-diameter) and its annular fuel plenum was configurations were monitored with OMEGA PX319–1KA5V trans-
axially-fed through a single 25-mm weld-port fitting, as shown in ducers. Plenum gas temperatures were measured with OMEGA Type-K
Fig. 3b. A brass bushing aligned the centerlines of the injector and inner exposed bead thermocouples, and a surface-mount Type-K was used
core with the outer wall of the combustor to within ±0.025 mm. This for exterior engine temperature. Static pressure and temperature sensors
particular RDRE was designed with only three axial combustor pressure with amplified signals were sampled at 2 kHz by a National Instruments
taps to reduce the influence of ports on wave dynamics. The injector and DAQ.
outer core were sandwiched between the front endplate having the A pair of PCB 111A24 piezoelectric dynamic pressure sensors were
plena and adapter flange at its rear. It had thick enough walls (minimum used. One PCB was located in the oxidizer plenum of all engines. The
19 mm) that a stainless steel sleeve on the outer core was not required. second PCB was used in the fuel plenum for the 76-FI-5 configuration,
More details of this 51-mm-RDRE design are provided in Ref. [31]. but was either unused or installed in a combustor instrument port in the

Fig. 3. Section views of geometrically scaled RDREs.

4
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

Table 2
Annular instrumentation locations by configuration.
RDRE Location Distance Azimuth Diameter Sensor
(mm) (deg) (mm)

76-FI- Front 0.00 357.5 0.8 Pressure

5 Wall
Annulus 22.2 330 3.2 Pressure
Annulus 34.9 30 3.2 Pressure
Annulus 47.6 330 3.2 Pressure
Annulus 60.3 30 3.2 Pressure
Annulus 73.0 330 3.2 Pressure
Annulus 73.0 210 3.2 Pre-
detonator
76 mm 85.7 30 3.2 Pressure
Duct
76 mm 98.4 330 3.2 Pressure
Duct

51-FI- Annulus 26.7 150 3.2 Pressure

5
Annulus 33.7 330 3.2 Pressure
Annulus 33.7 210 3.2 Pre-
detonator
Annulus 40.6 30 3.2 Pressure

25-FI- Annulus 16.5 270 2.4 Pressure

5
Annulus 18.5 330 2.4 Pressure Fig. 4. RDRE injector section views: (left) 76-mm-FI-5, (middle) 51-mm-FI-5,
Annulus 20.3 150 2.4 Pressure (right) 25-mm-FI-5.
Annulus 22.4 30 2.4 Pressure
Annulus 22.4 210 3.2 Pre-
detonator 2.2.6. Operating procedure
76 mm 38.1 90 3.2 Pressure Prior to each run the dump tank was evacuated to the desired back
Duct
pressure whenever sub atmospheric exhaust conditions were desired.
Sensor input and output to the DAQ were controlled with MATLAB. The
smaller configurations due to fuel plenum access restrictions. These Phantom camera and ER5000 were controlled with software provided
PCBs were sampled by a secondary National Instruments DAQ at a rate by their manufacturers. After arming the DAQ, valve actuation, spark
of 1.25 MHz. ignition, camera, and DAQ recording events were synchronized by the
DAQ according to a pre-determined timing matrix. Burn durations for
2.2.5. Injector design hot fires were always at least 700-ms-long, with steady plenum pressure
The number of injector pairs and port diameters for each configu- traces required for the last 100 ms of burn duration for inclusion in the
ration are summarized in Table 3. In each injector, the nominal oxidizer- data set. While data were being saved after the run, the contents of the
to-fuel area ratio of 2.50 was chosen to produce identical plena pressures dump tank were evacuated, and the engine was convectively cooled by
under stoichiometric conditions for the planned propellants of methane passing air through the oxidizer plenum. The next run was not con-
and oxygen. The total injection area was scaled to maintain a ratio of ducted until the exterior combustor wall temperature dropped below
0.110 between it and the annular combustor exit area for the 5-mm-gap. 300 K. More details of the operation procedure are provided in Ref. [31].
In all configurations, oxidizer and fuel injectors were inclined 30◦ from
the combustor axis and their impingement height was 2.16 mm from the
2.3. Data processing and uncertainty
injector face. The impingement diameters of the injectors were 5 mm
smaller than their annulus outer diameter so that the jets always
Calibration factors for the pressure transducers (3.45 MPa range)
impinged at the gap centerline.
were provided by their manufacturers. The dump tank pressure was
Section views of the injectors are shown in Fig. 4. Both the 76-mm
measured with an OMEGA PX319–050A5V absolute pressure transducer
and 51-mm oxidizer injectors were fed with 32 radial blind holes that
with a range of 345 kPa. Because of its higher resolution, the initial
each connected with two injector ports. Scaling the injector design down
values of all pressure traces were synchronized to the dump tank pres-
from 51 mm to 25 mm required a reduction in the number of injectors,
sure at the start of each run.
but radial spacing between injector pairs was kept at 2.34 mm to
Thermocouple signals passed through an amplifier (Model TCA-MS-
maintain the same impingement standoff distance. In order to adapt the
K-8-A4 from The Sensor Connection) before being read by the DAQ. This
51-mm-RDRE front endplate, the 25-mm-RDRE fuel injectors were fed
unit converted all thermocouple input signals to linearized, independent
axially and its oxidizer injectors were fed radially by annular plena.
analog outputs. For the purposes of the following uncertainty analysis,
the errors arising from the thermocouple signal conversion are assumed
Table 3 to be negligible.
Injector parameters for each RDRE configuration. Mass flow rate for each reactant through a choked orifice was
Reference ID 76-FI-5 51-FI-5 25-FI-5 calculated with Eq. (1) [37]:

Injector Pairs 72 72 48 ( )− γ+1

ACN P √̅̅̅ γ+1 2(γ− 1) ACN P
Inclination Fuel / Ox 30◦ / 30◦ 30◦ / 30◦ 30◦ / 30◦ ṁ˙ = Cd √̅̅̅̅̅̅̅̅̅̅̅̅̅ γ = Cd √̅̅̅̅̅̅̅̅̅̅̅̅̅ CFF (1)
Fuel Injector Diameter (mm) 0.787 0.635 0.508 T Rgas 2 T Rgas
Oxidizer Injector Diameter (mm) 1.245 0.991 0.813
Oxidizer-to-Fuel Area Ratio 2.50 2.43 2.56 Here, pressure, P, and temperature, T, were measured in the
Injection-to-Annulus Area Ratio 0.110 0.109 0.108 approach tube to the critical nozzle, and Rgas represents the species gas
Fuel Injector Feed Radial Radial Axial constant. Two pairs of critical nozzles were used. The discharge co-
Oxidizer Injector Feed Radial Radial Radial
efficients, Cd, for the largest critical nozzle set (13.8 mm2 and 6.13 mm2)

5
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

were provided by the manufacturer (Flow-Dyne) with 0.5 % uncertainty 3. Experimental results
at Reynolds numbers greater than 5 × 105 that were typical in operation,
which also accounts for uncertainty in throat area, ACN. This set was For each configuration, test matrices of ER sweeps at mflux ≈ 243 kg/
then used to calibrate the smaller critical nozzles (3.75 mm2 and 1.27 s/m2 and mflux sweeps at ER ≈ 1.15 were conducted with gaseous
mm2) using nitrogen. The variable CFF represents the critical flow factor methane and oxygen. Scaling test points in this manner led to similar
that included real gas effects based on the measured P and T and a plena pressures for each RDRE at each fueling condition. As part of this
lookup table generated by REFPROP [38]. investigation, additional ER sweeps at different mflux were conducted in
The average values from the last 100 ms of the run were used for all the smaller RDREs. The initial back pressures were reduced from at-
data processing, resulting in N = 200 samples per sensor for the time-of- mospheric to ~50 kPa in the 25-mm and 51-mm combustors to enable
record. This imposed a relative uncertainty, dxi/xi, in the measurement back pressure insensitive operation with lower mass fluxes; i.e., mflux <
of xi due to sample averaging for 95 % confidence given by Eq. (2) [28]: 200 kg/s/m2 [40].
dxi sx sx
= zα/2 √̅̅̅̅ = 1.96 √̅̅̅̅̅̅̅̅̅ (2)
xi N 200 3.1. Full data set by configuration

where sx is the relative standard deviation of the sample and the z-score, After individual run processing was completed, data from multiple
Zα/2, is 1.96. Assuming that uncertainties from sampling and the runs having fueling conditions within ±1.2 % of the desired mflux and ER
manufacture sensor specifications were random, the net sensor uncer- that exhibited identical wave dynamics (modes) in terms of primary
tainty for each time-of-record datum point was the square root of the wave count and the presence of counter-rotating wave(s) were averaged
sum of the squares of these two quantities. All of the relative un- together. For the AFRL test matrix data set, the uncertainty due to
certainties for the inputs to Eq. (1) were summed as shown in Eq. (3) to averaging mass flow rates and ERs was less than ±0.03 %, which was
determine the corresponding uncertainty of mass flow rate, dmi/mi, of much less than the sensor accuracy limits. The standard deviations in the
species i through each orifice with 95 % confidence [28]. averaged measurements of combustor pressure were less than ±0.1 %
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2 ( )2 ( ) ( )2 ( )2̅ and ±0.3 %, respectively. These latter results indicated the observed
dṁ˙i dACN dCd dCFFi 2 dPi dTi phenomena were very repeatable. Because the averaging uncertainties
= + + + + (3)
ṁ˙i ACN Cd CFFi Pi 2Ti were smaller than the plotting symbols, they were not included in the
Assuming the uncertainties in CFF and ACN were negligible and that plots.
the others were random resulted in the mass flow rate uncertainty for
both the oxidizer and fuel being dmi/mi ≈ 0.8 %. The corresponding 3.1.1. 76-FI-5 results
uncertainty for the net mass flow rate was then ±1.1 %. Applying the Data from 63 runs reduced to 21 averaged points for the 76-FI-5
mass flow rate uncertainties to the mixture mole fractions results in an configuration presented in Fig. 5 used different plotting symbols to
average uncertainty in ER of ±0.008 with less than 1 % standard devi- indicate the wave count. This configuration did not generate counter-
ation over the range of test conditions considered here. The un- rotating wave activity in any of the test conditions considered here.
certainties in mass flow rate and ER were smaller than the plotting Wave count increased as mflux increased with a transition from two to
symbols used here and thus they were not included in the plots. three waves occurring at 250 ≤ mflux ≤ 300 kg/s/m2. Within this tran-
The wave dynamics within the RDRE combustors were recorded with sition region, both two-wave and three-wave modes were established
128 × 128 pixel images at 240,000 FPS and exposure of 2 μs by a with similar test conditions. At constant mflux, the wave count was
Phantom v1211 camera. In a manner similar to that described in usually three at low fuel-lean and high fuel-rich conditions with two
Ref. [39], the resulting video images of the annulus were overlaid on a waves occurring when 1.15 ≤ ER ≤ 1.40. The maximum number of co-
polar mesh with 180 bins to track the rotation of combustion luminosity rotating primary waves was four, which was repeatedly recorded at ER
for the duration of the experiment. In the 25-mm RDRE, the polar mesh ≈ 0.26 with mflux ≈ 243 kg/s/m2.
had 60 bins due to the fewer pixels in the luminous regions of the smaller
combustor. Detonation surface plots tracking the angular wave motion 3.1.2. 51-FI-5 results
versus time were generated from these data. Through the Data obtained for the 51-FI-5 configuration, shown in Fig. 6,
two-dimensional Fast Fourier Transform of the detonation surface, the encompass the largest set of off-cross test conditions, as well as the most
number of waves and operational spin frequency were determined. variety in wave dynamics. Like with the 76-FI-5 configuration, the
Because the spinning detonation waves appeared to be in rigid body highest wave counts were seen at high flux and low ER conditions. The
rotation and their spin speed varied significantly with radius in the onset of deflagration occurred at fuel-rich ER thresholds that increased
smallest diameter combustor, the wave speeds reported here were based with increasing mflux. Counter-rotating waves appeared near transitions
on outer annulus circumference for comparison purposes.
Utilizing the procedure described in Ref. [31], uncertainties based on
frequency distribution taken from time-of-record samples were evalu-
ated with a statistical analysis for each measurement that accounted for
frequency fluctuations around the mean with 95 % confidence. Com-
bined with video framing rate resolution, this approach resulted in the
worst-case relative uncertainty in frequency and spin velocity to be ±2.5
%. The corresponding error bars were included in the plots.
The most luminous waves under stable operating conditions always
rotated in one direction for the duration of the test firing and were
designated as the “primary” waves. Under some mflux and ER fueling
conditions, however, less luminous waves observed spinning in the
opposite direction of the primary waves would, in some cases, fade away
and then regularly reappear. The primary wave counts were indicated
with “Wv” and the presence of counter-rotating waves was denoted with
“CW” in the data plots presented here.
Fig. 5. Mode data for 76-FI-5 experiments.

6
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

3.2. Experimental scaling comparisons

The results grouped by the combustor outer annulus diameter were

presented here. This data set only included matching test conditions
along sweeps of ER at mflux ≈ 243 kg/s/m2 or sweeps of mflux at ER ≈
1.15. The influence of mass flow rate, combustor pressure, wave fre-
quency, and wave speed on wave dynamics were presented. All con-
figurations had a pressure station near 65 % of the combustor length,
whose data were used for combustor pressure comparisons. Wave speeds
were compared to CJ speeds, DCJ, determined at the pre-combustion
pressure for each test condition.

3.2.1. Mass flux sweeps

Data from mflux sweeps at ER ≈ 1.1were plotted in Figs. 8 to 11. Wave
count, in general, increased with radius of curvature (Fig. 8) at any given
mflux. Wave count transitions to a higher number and/or cessation of
counter-rotating waves were observed in all combustors when mflux was
greater than ~250 kg/s/m2. Only the 25-FI-5 configuration experienced
pulsing, which occurred whenever mflux ≤ 100 kg/s/m2.
The 76-mm, 51-mm, and 25-mm combustors had pressure stations
located at 62.5 %, 66.3 %, and 65.0 % of their respective inner core
lengths. At these stations, pressure data were insensitive to the wave
count (Fig. 9) in all combustors. The 25-mm pressure data, however,
Fig. 6. Mode data for 51-FI-5 experiments. were about 14 % higher than the larger combustors at mflux ≤ 320 kg/s/
m2, which was likely due to the closer proximity of the detonation zone
to deflagration and changes in wave count. Similar to the 76-mm-RDRE, to the pressure port under these conditions. All pressures increased
the 51-mm-RDRE had a wave count transition from one to two waves linearly with mflux and were of similar amplitude at mflux > 350 kg/s/m2
between mflux of 243 kg/s/m2 and 280 kg/s/m2 at ER ≈ 1.15. conditions. This implied that the detonation zone moved closer to the
injector face in the 25-FI-5 configuration with increasing mflux in a
3.1.3. 25-FI-5 results manner like that reported in Ref. [41], which reduced its influence on
In contrast to the 51-FI-5 and 76-FI-5 configurations having their the 65 % pressure port at higher mflux.
highest wave counts at low ER and high mflux, the 25-FI-5 configuration At ER ≈ 1.15, the wave frequency in the 51-mm and 76-mm com-
demonstrated steady one-wave operation within the same upper-left bustors steadily increased with increasing mflux until it abruptly
quadrant, as shown in Fig. 7. For the ER sweep at mflux ≈ 243 kg/s/ increased to around 20 kHz when additional waves appeared at mflux >
m2, one-wave operation continued up to ER ≈ 1.8. Thereafter increasing 250 kg/s/m2 (Fig. 10). In the 25-mm combustor, however, the frequency
the excess methane resulted in deflagration and ultimately no combus- smoothly increased from ~17 kHz to ~22 kHz with increasing mflux even
tion. The ER sweep at mflux ≈ 160 kg/s/m2 resulted in either one spin- though the operating mode made a transition from one primary wave
ning wave with a single counter-rotating wave, deflagration, or no with a counter-rotating wave to only one primary wave at the same mflux
combustion. Tests at ER ≈ 1.15 with mflux ≤ 100 kg/s/m2 experienced a where the wave counts suddenly increased in the larger combustors. The
repetitive pulsing mode that was characterized by the annular wave speed, Dspin, in the 51-mm and 76-mm combustors abruptly
combustor suddenly becoming brightly luminous, followed by a period decreased from around 2 km/s to 1.7 km/s as additional waves
of complete darkening that was assumed to be the propellant refill appeared, and then steadily increased with increasing mflux (Fig. 11). For
period, akin to the longitudinal pulsing operation of a pulse detonation the 25-mm combustor, Dspin ≈ 1.6 km/s near the mode transition con-
engine. The average frequency of the pulsing was near 16.6 kHz for the dition at mflux ≈ 250 kg/s/m2. At higher mass flux, the wave speeds of
included runs. Similar phenomena were reported at low mflux in Ref. [1]. the 25-mm and 51-mm combustors were in very good agreement. Ac-
counting for pressure effects on DCJ, the ratios of Dspin/DCJ for the mflux
sweep data at ER ≈ 1.15 were within the range of 0.60 to 0.80 for the 51-

Fig. 7. Mode data for 25-FI-5 experiments. Fig. 8. Mode data from mflux sweeps at ER ≈ 1.15.

7
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

Fig. 12. Mode data from ER sweeps at mflux ≈ 243 kg/s/m2.

Fig. 9. Pressure port data at ~65 % inner core length for ER ≈ 1.15.
appeared during fuel-lean conditions in any of the combustors.
Combustor pressures at the nominal 65 % point monotonically
increased from fuel-lean to fuel-rich conditions until the maxima were
reached for each configuration around 1.3 < ER < 1.7 (Fig. 13). This is
consistent with the highest Chapman-Jouguet detonation pressure ratio
determined by chemical equilibria calculation being at ER ≈ 1.6 for
CH4–O2 propellant [47]. When deflagration occurred at ER > 2.0, the
combustor pressures dropped by more than 25 %. For the extremely
fuel-rich tests in the 25-mm-RDRE, no combustion occurred and the
pressure was less than 10 % of that generated with spinning detonation
waves.
Wave frequencies were similar at all size scales in the fuel-lean and
fuel-rich regions of operation (Fig. 14). When the wave count decreased
by one in the near stoichiometric region, however, the spin frequencies
suddenly decreased in the 51-mm (~12.5 kHz) and 76-mm (~16 kHz)
combustors. Local maxima in wave frequency (~21 kHz) for all con-
figurations occurred at ER of ~0.6.
Fig. 10. Spin frequency at ER ≈ 1.15. The highest recorded wave spin speeds of ~2 km/s and ~1.9 km/s in
the 51-mm and 76-mm RDREs, respectively, occurred in the near stoi-
chiometric region. (Fig. 15). Otherwise, aside from the very fuel-lean
test having four primary waves, Dspin values for all RDREs were of
similar amplitude (1.5 km/s to 1.7 km/s) throughout the remaining ER
range.
In all cases, Dspin/DCJ values decreased with increasing ER in the fuel-
lean region from their local maxima of 0.75–0.8 at an ER ≈ 0.4 (Fig. 16).
Near stoichiometric, Dspin/DCJ values were again 0.75–0.8 for the 51-
mm and 76-mm combustors, yet only ~0.6 in the 25-mm combustor.
All combustors experienced their lowest Dspin/DCJ values of ~0.55 under
their most fuel rich conditions, which coincided with the highest DCJ
values for this propellant.

Fig. 11. Wave speed at ER ≈ 1.15.

mm and 76-mm combustors and as low as 0.50 for the 25-mm combustor
[31].

3.2.2. Equivalence ratio sweeps

Operating mode data for ER sweeps at mflux ≈ 243 kg/s/m2 were
shown in Figs. 12 to 16. In general, the wave count was either at a
minimum or exhibited a single primary wave with a counter-rotating
wave when 1.0 < ER < 1.4 (Fig. 12). Deflagration occurred when the
25-mm and 51-mm combustors operated extremely fuel-rich (i.e., ER >
2) and combustion could not be stabilized at ER ≈ 2.6 in the 25-mm
combustor. At the mflux of these ER sweeps, no counter-rotating waves
Fig. 13. Pressure port data at ~65 % inner core length for mflux ≈ 243 kg/s/m2.

8
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

ahead of the detonation zone (through the injector face), a basis for
computing the pre-detonation pressure for the other configurations was
needed. When spinning detonation waves were present, combustor
pressure data were predominantly influenced by the mflux and ER con-
ditions, rather than either the wave mode or geometry (see Fig. 13). In
all configurations, the combustor pressures were proportional to the
mflux. Furthermore, by geometrically scaling the injectors and main-
taining a 5-mm-annular gap, the RDRE plena experienced similar pres-
sure levels when operated under the same mflux and ER conditions.
Because the injection and mixing processes were similar in the scaled
combustors, their detonation zone standoff distances from the injector
face were expected to be similar. Nonetheless, variations in pressure
between the 25-FI-5 and two larger RDREs ranged from 6 % to 16 % at
near stoichiometric conditions. This is likely primarily due the smaller
inner core having a much larger fraction of its length exposed to the
Fig. 14. Spin frequency data at mflux ≈ 243 kg/s/m2. detonation zone. Thus, the 65 % pressure port for the 25-mm-RDRE
would be closest to the detonation zone and would therefore have
higher pressure than in the larger RDREs at a given fueling condition, as
evident in Figs. 10 and 14.
Motivated by the observation of comparable pressures for all test
conditions in the 51-mm and 76-mm combustors, a pressure map from
the 65 % length station to the injector face pressure, p0, for the 76-FI-5
configuration was generated based on inputs of mflux and ER. Utilizing a
second-order polynomial fit to ER and mflux (kg/s/m2) data, an R2 value
of 0.998 was achieved for p0 (kPa) with the following:

p0 = − 531.2 + 2.965 mflux + 526.6 ϕ + .0009336 mflux 2 − 1.418 mflux ϕ

− 54.73 ϕ2
(4)
This relationship was then used to estimate injector face pressure in
the two smaller combustors.

4.1.2. Detonation cell size

Fig. 15. Wave speed at mflux ≈ 243 kg/s/m2. Cantera chemical kinetics software [42] and Shock and Detonation
Toolbox from the Caltech Explosion Dynamics Laboratory [43] were
used to calculate detonation cell width. The inputs were pre-detonation
pressure, temperature, gas mixture, and reaction mechanism. As both
fuel and oxidizer plena temperatures averaged ~300 K across the full
data set, this was the temperature input for all cases. The pre-mixed
methane-oxygen ER was altered over the range of test conditions
covered in experiments with the pressure for each case determined from
Eq. (4). The GRI-Mech 3.0 combustion mechanism was selected due to
its accuracy for methane-oxygen mixtures [44].
Calculations for the detonation cell width, λ, utilized the Zel’dovich-
von Neumann-Döring model to relate induction time through the
methods described by Westbrook [45]. The induction time was defined
as the point of maximum rate of temperature increase, which typically
translated to the time it took the temperature to complete half of its total
increase. After converting this to an induction length based on the
post-shock gas velocity, the result was multiplied by a factor of 29,
which represents the average relation between detonation cell width
Fig. 16. Dspin/DCJ at mflux ≈ 243 kg/s/m2. and induction length. To reach this factor, the average for unconfined
hemispherical detonations over various fuel-air and fuel-oxygen mix-
4. Scaling analysis tures based on experiments relating critical tube diameter and induction
length was computed, then combined with the number of transverse
4.1. Thermodynamic calculations waves needed to initiate such a detonation at the critical tube diameter
[45].
Processes for computing thermodynamic properties were presented The resulting theoretical λ values at four different initial pressures
for determination of pre-detonation pressure, detonation cell width, were shown in Fig. 17 along with some experimental data points from
heat release, and injection fill height. Ref. [46]. Minimum λ values were at fuel rich conditions and further
decreased with increasing initial pressure. The calculated λ values
4.1.1. Pre-detonation pressure trended with experimental results for the methane-oxygen experiments
In order to calculate detonation cell size and other thermodynamic at elevated pressures, however, the theory tended to under predict the
properties, pressure information is required in addition to propellant cell widths.
composition. As only the 76-FI-5 configuration possessed a pressure tap

9
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

by uniformly injected, premixed reactants were determined based on

their axial velocity, u, and arrival time, t, of the next detonation wave.
Axial velocity was found by assuming mass conservation with ideal gas
law as follows:

ṁ˙Rgas T0
u= (6)
p 0 Ae

where ṁ˙ represented mass flow rate and Ae represents annulus exit area,
from which mflux was determined, and Rgas was the mixture gas constant.
The next detonation wave arrival time was:
2πr0 1
t= ≡ (7)
N Dspin f

where r0 represents outer radius and N is wave count of the primary

system. Because Dspin values were based on annulus perimeter, use of
outer radius was appropriate here. In doing so, t became equivalent to
the inverse of operating frequency, f. Combining Eqs. (6) and 7 allowed
Fig. 17. Detonation cell size vs. ER for CH4–O2 propellant (data from
fill height, h, to be written as:
Ref. [46] included).
mflux Rgas T0 1
h=u×t = (8)
4.1.3. Heat release p0 f
For calculation of heat release, the NASA Chemical Equilibrium This fill height relation was similar to those in other publications,
Applications (CEA) program was used to generate the mass fractions at such as Ikema et al. [49]. With p0 and T0 specified, fill height was a
equilibrium [47]. Properties of CJ detonation waves were calculated for function of experimental mflux, ER, and f. The resulting h vs. ER values
sweeps of pressure and ER at an initial temperature of 300 K. Query for this data set were plotted in Fig. 19. Fill height ranged from 2 mm to
points, particularly in ER, were selected near the test conditions used in 7 mm as mflux increased and from 2 mm to 4 mm during the ER sweep.
experiments. To calculate heat release, q, the following equation was Single wave operation in general had smaller h than when a
applied for reactant and product species: counter-rotating wave was present. This indicated that the
∑ Xi Δhf,i ∑ Xi Δhf,i counter-rotating wave consumed reactants not processed by the primary
q= − (5) wave, which was consistent with counter-rotating wave(s) arising near
Reactants
MWi Products
MWi
transition points. In such cases, there was not enough propellant ϕ low to
Heat release units were MJ/kg of propellant, Xi was species mass support the presence of an additional co-rotating wave, leading to a
fraction, Δhf,i was standard enthalpy of formation, and MWi was mo- relatively weak counter-rotating wave. Mixing effects at off-design
lecular weight. Standard enthalpy of formation and MWi for each injector conditions may also have contributed to the appearance of the
product species were retrieved from NIST Chemistry Webbook [48]. counter-rotating wave(s).
Heat release values based on ER and pressure shown in Fig. 18 were
interpolated for experimental conditions. The largest pressure effects
were near ER ≈ 1.4, however, depending on pressure, maximum heat 4.2. Thermodynamic and geometric comparisons
release occurred within the range of 1.6 < ER < 1.8. After computing q
for each test condition, results were multiplied by mflux to get chemical Pertinent RDRE geometric features influenced by thermodynamic
power flux.

4.1.4. Fill height

After detonation wave passage, fresh reactants flowed again into the
combustor (blue region in Fig. 1). The maximum fill heights achievable

Fig. 18. Heat release at different initial pressures for CH4–O2 propellant. Fig. 19. Mode dependence on fill height and ER for all combustors.

10
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

properties were listed in Table 4. Note, both the gap-to-outer radius

(Gap/ro) and annulus volume-to-surface area (V/{SAxL}) ratios multi-
plied by the annulus length, L, decreased linearly as the ratio between
inner and outer radius (ri/ro) increased.

4.2.1. Radius-to-gap width

The unique combustor radius-to-gap ratio values provided a means
to compare influences of fueling conditions on operating modes. Shown
in Fig. 20 were mflux and ER sweep data versus Dspin in terms of outer
radius-to-gap ratio. It was evident that in the range of 1.6 km/s < Dspin <
1.8 km/s, the wave count changed by the ratio of combustor outer di-
ameters when there were no counter-rotating waves. Furthermore,
highest wave speeds always occurred when RDREs operated with fewest
waves over the range of mflux tested here.
The influence of λ on operation modes was shown in Fig. 21 with
inner radius-to-gap ratio as a parameter. For most fueling conditions,
regardless of configuration, λ was less than 2 mm. From this plot alone, it
was unclear if λ influenced wave count because modes were inter- Fig. 20. Mode dependence on outer radius-to-gap width ratio and wave
speed, Dspin.
spersed. Nonetheless, larger λ’s at fuel-lean conditions were associated
with maximum wave count and deflagration when fuel rich (Fig. 12),
implying that λ influenced operating range.
The power flux data ranging from 0.1 GW/m2 to 3 GW/m2 were
shown in Fig. 22. Wave count transitions due to ER changes occurred
wherever there were two modes of operation close together at power
fluxes less than 1 GW/m2. The wave count changes due to increasing
mflux occurred at ~1.4 GW/m2.

4.2.2. Inner radius to detonation cell size

Plotting the ratio of inner core radius to cell width, ri/λ, with respect
to ER as in Fig. 23 enabled regions of wave stability to be identified.
Horizontal lines were drawn such that three regions appear; one con-
taining all non-detonative runs (ri/λ < 6), another that includes all
counter-rotating wave activity (ri/λ ≤ 26), and a third region with only
stable co-rotating wave operation (ri/λ > 26). Increasing ri or reducing λ
through fueling conditions; e.g., by raising mflux and thereby fill pres-
sure, provided higher wave stability. Furthermore, these results indi-
cated that even though operating modes were strongly influenced by
inner radius and fueling conditions, other factors were involved.
Fig. 21. Mode dependence on inner radius-to-gap ratio and detonation cell
4.2.3. Fill height to detonation cell size size, λ.
The operational modes for fill height-to-detonation cell size ratio, h/
λ, from both the ER and mflux sweeps were displayed in Fig. 24 along
with the inner radius-to-gap ratio of each combustor. Because the fill
height varied less with ER than the cell size, their ratio decreased with
increasing cell size in non-stoichiometric propellant. Under fuel lean
conditions, the h/λ values in Fig. 24a increased from less than one in
each combustor as the ER approached stoichiometric, and the wave
counts were proportional to the ratio of the diameters. The onset of the
mode transition was at h/λ ≈ 2.5 in the 51-mm and 76-mm combustors,
thereafter the ratio jumped to the first data point indicating a change in
mode, i.e., 6.4 and 5.1 in the 51-mm and 76-mm data sets, respectively.
The h/λ ratio then decreased until another jump occurred with the mode
making a transition back to its original off-stoichiometric condition. In
the 25-mm combustor, the appearance of a counter-rotating wave began
at h/λ ≈ 3.4 and jumped to 4.2. The counter-rotating wave disappeared
with the addition of a more excess fuel while h/λ continued to increase to
its maximum value of ~4.3. Adding more fuel eventually reduced h/λ to
the point where the mixture was too rich to sustain a spinning

Table 4
Geometric ratios of RDRE configurations. Fig. 22. Mode dependence on inner radius-to-gap ratio and chemical
power flux.
Configuration ri (mm) ri / ro Gap / ro ri / Gap ro / Gap V / (SA * L)

76-FI-5 33.1 0.87 0.13 6.62 7.62 0.033

51-FI-5 20.4 0.80 0.20 4.08 5.08 0.049
25-FI-5 7.70 0.61 0.39 1.54 2.54 0.098

11
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

≈ 1.15 had a peak power flux of 3 GW/m2 where wave counts depended
on r0. The wave count transition occurred in the power flux range of 1.4
GW/m2 to 1.7 GW/m2.
Because power flux and λ were computed from fueling conditions,
each test condition had unique values that were independent of both
geometry and operating mode. This was shown in Fig. 26, where power

Fig. 23. Mode dependence on ri/λ and ER.

Fig. 25. Mode dependence on Chemical power flux vs. ER.
detonation.
The values of h/λ changed in a different manner during mode tran-
sitions in the mflux sweeps shown in Fig. 24b. In the 51-mm and 76-mm
combustors, h/λ would increase to the maximum value for its low flux
mode, then make a transition to a higher wave count with a sudden
decrease in h/λ. Thereafter h/λ continued to increase. In the 25-mm
combustor, h/λ monotonically increased as it made a transition to a
single primary wave after the disappearance of its counter-rotating
wave. It is clear from these data that the mode transitions to a
different wave count were accompanied by significant changes in cell
size, which was the primary driver for h/λ. The appearance and disap-
pearance of single counter-rotating wave also resulted in a step increase
in h/λ, however, this relative change was much smaller than that for
wave count transition.

4.2.4. Chemical power flux

Operating conditions in terms of power flux and ER were shown in
Fig. 25. Chemical power flux plateaued at ~1.5 GW/m2 when ER > 1.4,
which coincided with peak combustor pressures. The mflux sweep at ER
Fig. 26. Relationship between chemical power flux and λ.

Fig. 24. Mode dependence on inner radius-to-gap ratio and h/λ. (a) ER sweeps (b) mflux sweeps.

12
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

flux and λ were inversely proportional and test conditions repeated

across configurations overlapped with little scatter. Although there were
some exceptions, in general CJ speed increased as λ decreased and
chemical power flux increased.
Pressure data at the nominal 65 % point in the annuli of the com-
bustors were plotted against chemical power flux in Fig. 27. Trend lines
from both the ER and mflux sweeps in the 25-mm combustor (dash) and
the ER sweeps (dots) and mflux sweeps (dot-dash) for both the 51-mm
and 76-mm combustors were shown. The annulus pressure increased
with increasing power flux in all cases along with the wave count. As
previously mentioned, the offset in 25-mm data was attributed to its
sensor being closer to the detonation zone, although its pressures did
match the 51-mm and 76-mm combustor data from the fuel-lean, low
power flux tests and the high power flux tests at ER ≈ 1.15.

Fig. 28. Mode dependence of off-sweep data on inner radius-to-gap width ratio
4.3. Application of results
and mflux.

The influence of several experimental variables on combustor oper-

counter-rotating wave(s) and deflagration limits [49]. Application of the
ation at all size scales have been discussed, including driving variables
same comparison to off-cross data from Fig. 23 was shown in Fig. 29.
for operating mode, wave speed, and combustor pressure. Utilizing these
The dotted line was the ri/λ = 6 condition, which indicated a minimum
empirical lessons to predict experimental outputs from only geometry
criterion (maximum λ) for detonation to almost always occur. The
and fueling conditions was desirable to guide and optimize future RDRE
dashed line of ri/λ = 26 indicated the boundary below which
designs. In order to test the applicability of the previously discussed
counter-rotating waves would form (minimum λ). The ratio of ri/λ
relations, they were applied to data thus far not utilized, such as the
included geometric and fueling condition influences, and predicted the
deflagration limits and counter-rotating wave results from 25-mm and
presence of deflagration and counter-rotating wave results reasonably
51-mm experiments. This data set also included the results that were not
well. The largest drawback though, as discussed in [48], was the scatter
part of the original mflux and ER sweeps. There were 78 “off-sweep” data
of solely co-rotating wave systems into both regions, meaning this ratio
points for the 51-FI-5 and 25-FI-5 configurations, many of which rep-
only implied a result could be deflagration but not that it necessarily
resented repeated tests that were averaged. Pre-detonation pressure, λ,
would be.
and power flux were calculated for each experiment as described in
The computation of wave speed required knowledge of the wave
Section 4.1.
count, and therefore wave frequency presented itself as a better option
Operating modes were driven by a combination of geometry and
for prediction of wave dynamics. Fill height provided a starting point, as
fueling conditions, notably mflux, as shown in Fig. 28. Wave count and
it is computed from properties based on the initial reactant mixture and
stability increased with both mflux and ratio of inner radius-to-gap width.
operating frequency (see Eq. (8)). Fill heights plotted in Fig. 19 had
Almost all runs with a counter-rotating wave were within the region
minimal variations over large changes in geometry, operating mode, and
having an upper boundary of mflux ≈ 240 kg/s/m2. At the highest mass
fueling conditions, especially when the results associated with counter-
fluxes shown for the 51-mm-RDRE, deflagration occurred at the very
rotating wave modes were not considered. Within the mflux and ER
fuel rich conditions. The deflagrations that were to the right of the upper
sweeps, h had an average value of ~3.25 mm. Holding h constant at this
boundary at mflux ≈ 170 kg/s/m2 tended to be at very fuel-rich condi-
value while solving Eq. (8) for wave frequency provided results
tions, which had cell sizes of 3 mm to 6 mm. Overall, the ratio of inner
remarkably close to the off-sweep data. On average, wave frequency was
radius-to-gap width and mflux were strong indicators of primary wave
over predicted by ~5 %, with the largest deviations seen at extreme ERs.
count, however, the inclusion of ER effects were needed to improve
As the CJ speeds were calculated from fueling conditions using chemical
prediction of operating mode.
equilibrium codes and the majority of wave speeds were within the
For wave stability, which was defined by the absence of a counter-
range of 60 % to 80 % of these values, combining DCJ, outer radius, and
rotating wave, ri/λ values indicated approximate boundaries for
wave frequency provided a path for predicting the wave count. Explo-
ration of the influence of injection and mixing characteristics on fill
height was outside the scope of this investigation, and would be
necessary before applying the correlations presented here to other

Fig. 27. Pressure at ~65 % inner core length vs. chemical power flux from ER
and mflux sweeps in 25-mm (dashed), 51-mm (dots), and 76-mm combustors
(dot-dash). Fig. 29. Relation of off-sweep data with inner radius and λ.

13
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

configurations. Acknowledgements

5. Conclusion Significant contributions were provided by graduate students J.V.

Koch, L. Chang, and Q. Roberts, and undergraduates M. Ikeda, D. Menn,
Experiments conducted in annular combustors with 25-mm, 51-mm, N. Inahara, D. Nankova, A. Holden, C. Decker, and J. Geldenbott. Con-
and 76-mm outer diameters having flat-faced impinging injectors for sultations with W.A. Hargus, J.W. Bennewitz, and J. Burr from AFRL
gaseous methane-oxygen propellant provided pertinent data on size were very helpful. This work was supported by AFOSR Grant FA
scaling effects. The annulus gap width was 5 mm in all cases. With 9550–18–1–9–0076 with program manager Chiping Li. Video data were
respect to geometry, variations in radius of curvature were associated recorded with a high-speed camera funded through the University of
with changes in wave count and frequency; however, their wave speeds Washington Student Technology Fee.
were nearly the same at any given mass flux and equivalence ratio. Mode
transitions observed when operating with an equivalence ratio near References
stoichiometric occurred in each combustor in the mass flux range of 250
kg/s/m2 to 300 kg/s/m2. [1] Bykovskii FA, Zhdan SA, Vedermikov EF. Continuous spin detonations. J. Propuls.
Power 2006;22(6):1204–16. https://doi.org/10.2514/1.17656.
Chemical equilibrium calculations from Cantera and CEA for the [2] Koch JV. Nonlinear dynamics of rotating detonation waves. Seattle, WA: William E.
experimental fueling conditions provided estimates for detonation cell Boeing Department of Aeronautics and Astronautics, University of Washington;
size and chemical power flux, respectively. Detonation cell size was 2020. PhD Thesis, http://hdl.handle.net/1773/45431.
[3] Lu FK, Braun EM. Rotating detonation wave propulsion: experimental challenges,
associated with detonative limits and operating mode, and related to modeling, and engine concepts. J. Propuls. Power 2014;30(5):1125–42. https://
other variables such as inner radius of the combustor. Chemical power doi.org/10.2514/1.B34802.
flux had a strong relationship with both detonation cell size and [4] Smith RD, Stanley SB. Experimental investigation of rotating detonation rocket
engines for space propulsion. J. Propuls. Power 2021;37(3):463–73. https://doi.
combustor pressure, while also influencing wave speeds when the wave org/10.2514/1.B37959.
count was constant. An increase in wave count and/or cessation of [5] Goto K, Matsuoka K, Matsuyama K, Kawasaki A, Watanabe H, Itouyama N,
counter-rotating waves occurred in the power flux range of 1.4 GW/m2 Ishihara K, Buyakofu V, Noda T, Kasahara J. Flight demonstration of detonation
engine system using sounding rocket S-520-31: performance of rotating detonation
to 1.7 GW/m2 at all combustor size scales.
engine. In: AIAA SciTech 2022 Forum. San Diego, CA: AIAA; 2022. https://doi.org/
Calculation of fill height of fresh propellant showed little variation 10.2514/6.2022-0232. 2022-0232.
with different geometries or fueling conditions. As such, fill height was [6] Teasley TW, Fedotowsky TM, Gradl PR, Austin BL, Heister SD. Current state of
useful in predicting operating frequency, which applied to wave count NASA continuously rotating detonation cycle engine development. In: AIAA
SciTech 2023 Forum. National Harbor, MD: AIAA; 2023. https://doi.org/10.2514/
through the Chapman-Jouguet detonation velocity derived from chem- 6.2023-1873. 2023-1873.
ical equilibrium codes because the majority of observed wave speeds [7] Hargus WA, Schumaker SA, Paulson EJ. Air force research laboratory rotating
were within 60 % to 80 % of these values. detonation rocket engine development. In: 2018 Joint Propulsion Conference.
Cincinnati, OH: AIAA; 2018. p. 2018–4876. https://doi.org/10.2514/6.2018-4876.
Combustor pressure was dominated by adjustments to fueling con- [8] Wyatt JJ, Snow NJ, Fiorino NT, Schauer FR, Polanka MD, Sell B, Cho KY. Injection
ditions, especially mass flux, rather than geometry. Pressure maxima studies on a small-scale rotating detonation engine. In: AIAA SciTech 2022 Forum.
observed in the equivalence ratio range of 1.3 to 1.7 coincided with the San Diego, CA: AIAA; 2022. https://doi.org/10.2514/6.2022-1114. 2022-1114.
[9] Yoshidomi K, Kurita N, Ozawa K, Tsuboi N, Hayashi AK, Kawashima H. Numerical
maximum power flux of the propellant at constant mass flux. Although simulation on rotating detonation engine: effect of number of injection ports in
fueling conditions dominated the operational trends, combustor annulus non-premixed H2-O2 gases. In: AIAA SciTech 2022 Forum. San Diego, CA: AIAA;
size influenced wave stability and operational limits at the most fuel- 2022. https://doi.org/10.2514/6.2022-1112. 2022-1112.
[10] Bigler BR, Bennewitz JW, Schumaker SA, Danczyk SA, Hargus WA. Injector
lean and fuel-rich conditions. Regardless of geometry, however, wave alignment study for variable mixing in rotating detonation rocket engines. In: AIAA
stability, as indicated by elimination of count-rotating waves, always SciTech 2019 Forum. San Diego, CA: AIAA; 2019. https://doi.org/10.2514/
improved through an increase in mflux. 6.2019-2019. 2019-2019.
[11] Redhal SC, Burr JR, Yu KH. Fuel injection dynamics and detonation wave
interaction in rectangular channel. In: AIAA SciTech 2019 Forum. San Diego, CA:
CRediT authorship contribution statement AIAA; 2019. p. 2019–251. https://doi.org/10.2514/6.2019-2251.
[12] Bennewitz JW, Burr JR, Lietz CF. Characteristic timescales for rotating detonation
Tyler Mundt: Data curation, Formal analysis, Investigation, Meth- rocket engines. In: AIAA Propulsion and Energy 2021 Forum. Virtual Event: AIAA;
2021. p. 2021–3671. https://doi.org/10.2514/6.2021-3671.
odology, Software, Writing – original draft, Writing – review & editing. [13] Feleo A, Chacon F, Gamba M. Uncertainties in thrust and EAP measurements of a
Carl Knowlen: Conceptualization, Data curation, Formal analysis, rotating detonation combustor with Axial Air Inlet. In: AIAA Propulsion and
Funding acquisition, Investigation, Methodology, Project administra- Energy 2020 Forum. Virtual Event: AIAA; 2020. p. 2020–3856. https://doi.org/
10.2514/6.2020-3856.
tion, Software, Supervision, Validation, Writing – original draft, Writing [14] Paxson DE, Miki K. Computational assessment of inlet backflow effects on rotating
– review & editing. Mitsuru Kurosaka: Conceptualization, Formal detonation engine performance and operability. In: AIAA SciTech 2022 Forum. San
analysis, Funding acquisition, Investigation, Methodology, Project Diego, CA: AIAA; 2022. https://doi.org/10.2514/6.2022-1263. 2022-1263.
[15] Keller AR, Otomize J, Nair AP, Minesi NQ, Spearrin RM. High-diodicity impinging
administration, Supervision, Validation, Writing – review & editing. injector design for rocket propulsion enabled by additive manufacturing. In: AIAA
SciTech 2022 Forum. San Diego, CA: AIAA; 2022. https://doi.org/10.2514/
Declaration of competing interest 6.2022-1265. 2022-1265.
[16] Schwer D, Kailasanath K. Numerical study of the effects of engine size n rotating
detonation engines. In: 49th AIAA Aerospace Sciences Meeting. Orlando, FL: AIAA;
The authors declare that they have no known competing financial 2011. https://doi.org/10.2514/6.2011-581. 2011-581.
interests or personal relationships that could have appeared to influence [17] Koch JV, Chang L, Upadhye C, Chau K, Kurosaka M, Knowlen C. Influence of
injector-to-annulus area ratio on rotating detonation engine operability. In: AIAA
the work reported in this paper.
Propulsion and Energy 2019 Forum. Indianapolis, IN: AIAA; 2019. p. 2019–4038.
The authors declare that they have no competing financial interests https://doi.org/10.2514/6.2019-4038.
or personal relationships that could have appeared to influence the work [18] Kim T-Y, Choi J. Numerical study of detonation wave propagation in non- circular
reported in this paper. channels for arbitrary shaped rotating detonation engines. In: 50th AIAA/ASME/
SAE/ASEE Joint Propulsion Conference. Cleveland, OH: AIAA; 2014.
p. 2014–3903. https://doi.org/10.2514/6.2014-3903.
Data availability [19] Ishii K, Kurata W, Kawana H, Ohno K, Ikema D. Effects of combustor size on
behavior of rotating detonation waves. In: 27th International Colloquium on the
Dynamics of Explosions and Reactive Systems, Beijing, China; 2019. http://www.
Data will be made available on request. icders.org/ICDERS2019/abstracts/ICDERS2019-251.pdf.
[20] Zhdan SA, Bykovskii FA, Vedernikov EF. Mathematical modeling of a rotating
detonation wave in a hydrogen-oxygen mixture. Combustion, explosion, and shock

14
T. Mundt et al. Applications in Energy and Combustion Science 19 (2024) 100282

waves, 43. US: Springer; 2007. p. 449–59. https://doi.org/10.1007/s10573-007- [35] Critical flow nozzle. Bulletin 2024;101. https://www.flow-dyne.com/Nozzle.htm.
0061-y. Accessed 01/05/.
[21] Bennewitz JW, Bigler B, Danczyk S, Hargus WA, Smith RD. Performance of a [36] Mundt T. Geometric scaling of cylindrical rotating detonation rocket engine
rotating detonation rocket engine with various convergent nozzles. In: AIAA combustors. Seattle, WA: William E. Boeing Department of Aeronautics and
Propulsion and Energy 2019 Forum. Indianapolis, IN: AIAA; 2019. p. 2019–4299. Astronautics, University of Washington; 2023. PhD Thesis, http://hdl.handle.net
https://doi.org/10.2514/6.2019-4299. /1773/50198.
[22] Ross M, Lietz C, Desai Y, Hamilton J, Hargus WA. Constriction-induced counter [37] Rathsack TC, Bigler BR, Bennewitz JW, Danczyk SA, Hargus WA. Laboratory flow
rotating-propagating behavior in RDREs. In: AIAA Propulsion and Energy 2020 measurement in rotating detonation rocket engines. In: AIAA SciTech Forum.
Forum. Virtual Event: AIAA; 2020. p. 2020–3873. https://doi.org/10.2514/ Orlando, FL: AIAA Paper; 2020. https://doi.org/10.2514/6.2020-9401. 2020-
6.2020-3873. 0194.
[23] Bigler BR, Burr JR, Bennewitz JW, Danczyk S, Hargus WA. Rotating detonation [38] Lemmon EW, Bell IH, Huber ML, McLinden MO. NIST standard reference database
rocket engine scaling using variable annular width geometries. In: AIAA Propulsion 23: reference fluid thermodynamic and transport properties-REFPROP, Version
and Energy 2021 Forum. Virtual Event: AIAA; 2021. p. 2021–3686. https://doi. 10.0. In: NIST Standard Reference Data Program, Gaithersburg, MD; 2018. https://
org/10.2514/6.2021-3686. doi.org/10.18434/T4/1502528.
[24] Brophy CM, Codoni JR, Thoeny A. Channel width impact on RDE performance with [39] Bennewitz JW, Bigler BR, Schumaker SA, Hargus WA. Automated image processing
fuel injection parity. In: AIAA SciTech 2022 Forum. San Diego, CA: AIAA; 2022. method to quantify rotating detonation wave behavior. Rev. Sci. Instrum. 2019;90.
https://doi.org/10.2514/6.2022-1877. 2022-1877. https://doi.org/10.1063/1.5067256. 065106.
[25] Comer A, Ihme M, Li C, Lietz C, Oefelein J, Rankin B, Sankaran V. In: Proceedings [40] Koch JV, Washington MR, Kurosaka M, Knowlen C. Operating characteristics of a
of the Fourth Model Validation for Propulsion (MVP 4) Workshop AIAA SciTech CH4/O2 rotating detonation engine in a backpressure controlled facility. In: AIAA
2020 Forum, Orlando, FL; 2020. https://community.apan.org/wg/afrlcg/mvpws/ SciTech 2019 Forum. San Diego, CA: AIAA; 2019. https://doi.org/10.2514/
p/proceedings. 6.2019-0475. 2019-0475.
[26] Strakey P, Ferguson DH. Validation of a computational fluid dynamics model of a [41] Washington MR, Koch JV, Kurosaka M, Knowlen C. Radial injector mixing effects
methane-oxygen rotating detonation engine. In: AIAA SciTech 2022 Forum. San on detonation zone position in rotating detonation engine. In: AIAA Propulsion and
Diego, CA: AIAA; 2022. https://doi.org/10.2514/6.2022-1113. 2022-1113. Energy 2019 Forum. Indianapolis, IN: AIAA; 2019. p. 2019–4131. https://doi.org/
[27] Pal P, Demir S, Kundu P, Som S. Large- e. In: AIAA Propulsion and Energy 2021 10.2514/6.2019-4131.
Forum. Virtual Event: AIAA; 2021. p. 2021–3642. https://doi.org/10.2514/ [42] Goodwin, D.G., Moffat, H.K., Schoegl, I., Speth, R.L., Weber, B.W.: Cantera: an
6.2021-3642. object-oriented software toolkit for chemical kinetics, thermodynamics, and
[28] Bennewitz JW, Burr JR, Bigler BR, Burke RF, Lemcherfi A, Mundt T, Rezzag T, transport processes. https://www.cantera.org, Version 2.6.0, (2022). https://doi.
Plaehn EW, Sosa J, Walters IV, Schumaker SA, Ahmed KA, Slabaugh CD, org/10.5281/zenodo.6387882.
Knowlen C, Hargus WA. Experimental validation of rotating detonation for rocket [43] Kao ST, Ziegler JL, Bitter NP, Schmidt BE, Lawson J, Shepherd JE, (contributors).
propulsion. Sci Rep 2023;13(14204). https://doi.org/10.1038/s41598-023-40156- SDToolbox: numerical tools for shock and detonation wave modeling. Pasadena,
y. CA: Explosion Dynamics Laboratory, California Institute of Technology; 2021.
[29] Mundt T, Chang L, Ikeda M, Menn D, Knowlen C, Kurosaka M. Annular gap width GALCIT Report FM2018.001, http://shepherd.caltech.edu/EDL/PublicResource
variation in 25-mm rotating detonation rocket engine. In: AIAA Paper 2023-1103; s/sdt/.
2023. https://doi.org/10.2514/6.2023-1103. [44] Smith, G.P., Golden, D.M., Frenklach, M., Moriarty, N.W., Eiteneer, B., Goldenberg,
[30] Mundt T, Chang L, Ikeda M, Menn D, Knowlen C, Kurosaka M. Operating M., Bowman, C.T., Hanson, R.K., Song, S., Gardiner Jr., W.C., Lissianski, V.V., Qin,
characteristics of a 76-mm rotating detonation rocket engine. In: AIAA Paper 2023- Z.: GRI-Mech 3.0. Gas research institute (2000). http://combustion.berkeley.edu
1104; 2023. https://doi.org/10.2514/6.2023-1104. /gri-mech/version30/text30.html.
[31] Knowlen C, Mundt T, Kurosaka M. Experimental results for 25-mm and 51-mm [45] Westbrook CK. Chemical kinetics of hydrocarbon oxidation in gaseous detonations.
rotating detonation rocket engine combustors. Shock Waves 2023;33:237–52. Combust Flame 1982;46:191–210. https://doi.org/10.1016/0010-2180(82)90015-
https://doi.org/10.1007/s00193-023-01120-x. 3.
[32] Knowlen C, Mundt T, Kurosaka M. Experimental results for geometrically scaled [46] Schumaker SA, Knisely AM, Hoke JL, Rein KD. Methane-oxygen detonation
rotating detonation rocket engines. In: AIAA Paper 2023-0354; 2023. https://doi. characteristics at elevated pre-detonation pressures. Proc. Combust. Inst. 2021;38
org/10.2514/6.2023-0354. (3):3623–32. https://doi.org/10.1016/j.proci.2020.07.066.
[33] Boening JA, Wheeler EA, Heath JD, Koch JV, Mattick AT, Breidenthal RE, [47] McBride B, Gordon S. Computer program for calculation of complex chemical
Knowlen C, Kurosaka M. Rotating detonation engine using a wave generator and equilibrium compositions and applications II. Users Man. Progr. Descr 1996. http
controlled mixing. J. Propuls. Power 2018;34(6):1364–75. https://doi.org/ s://ntrs.nasa.gov/citations/19960044559.
10.2514/1.B36603. [48] Linstrom, P.J., Mallard, W.G., editors: NIST chemistry WebBook, NIST standard
[34] Knowlen C, Mundt T, Roberts Q, Hamza A, Menn D, Kurosaka M. Operating reference database number 69. National Institute of Standards and Technology,
characteristics of a 10-mm rotating detonation rocket engine. In: AIAA SciTech Gaithersburg, MD, 20899, 2023. https://doi.org/10.18434/T4D303.
2024 Forum. Orlando, FL: AIAA Paper; 2024. p. 2024–610. https://arc.aiaa.org/ [49] Ikema D, Yokota A, Kurata W, Kawana H, Ishii K. Propagation stability of rotating
doi/10.2514/6.2024-2610. detonation waves using hydrogen/oxygen-enriched air mixtures. Trans. Jpn Soc.
Aeronaut. Space Sci 2018;61(6):268–73. https://doi.org/10.2322/tjsass.61.268.

15