Experimental research on rotating detonation with liquid hypergolic propellants

Shuaijie XUE,Hongjun LIU,Lixin ZHOU,Weidong YANG,Hongbo HU,Yu YAN

Science and Technology on Liquid Rocket Engine Laboratory,Xi’an Aerospace Propulsion Institute,710100 Xi’an,China

KEYWORDS Hypergolic;Initiation;Liquid propellant;Nozzle;Rotating detonation

AbstractThis paper describes experimental research into the initiation and propagation of rotating detonation for liquid Nitrogen TetrOxide(NTO)and liquid MonoMethylHydrazine(MMH).An annular rocket-type combustor without nozzle was designed to investigate detonation combustion.The propellants were injected through unlike impingement injectors.The combustion fiame fronts and pressure waves were detected using optical diagnostics and dynamic pressure sensors,respectively.The propagation of rotating detonation was established spontaneously by increasing the mass fiow rate of propellants.The velocity of propagation of the fiame fronts and pressure waves was nearly equal and reaches supersonic speed.Two different detonation combustion patterns are present,single wave mode and double waves mode.And in double waves mode,the two detonation waves are always counter-rotating.The possibility of rotating detonation initiation in a combustor with nozzle was also checked.Stable rotating detonation can be initialized and sustained at similar operating conditions.

1.Introduction

Continuous rotating detonation is a hot spot in the field of combustion research.Due to the high thermal cycle efficiency and heat release speed,rotating detonation combustion could provide significant advantages in thrust performance and combustor size.1In pioneering research,Voitsekhovskii2and Wolanshi3firstly stabilized rotating detonation in a cylindrical chamber.The feasibility of a rocket motor utilizing a detonation wave rotating in an annular combustion chamber was then possible.4,5In recent years,great efforts have been made in research on rotating detonation.A popular way of initializing continuous rotating detonation is to use gaseous fuel and oxidizer,6–12to achieve highly efficient mixture.However,for the application of detonation engines in the future,it is necessary to employ liquid propellants due to advantages in energy density,safety,and storage.Bykovskii et al.13,14presented the possibility of continuous spin detonation of liquid fuels with gaseous oxygen.Kindracki15described experimental research on the initiation and propagation of rotation detonation for liquid kerosene and air mixtures.For the application of rotating detonation in rocket engines,both fuel and oxidizer should be liquid.Until now,only Bykovskii et al.14have reported the successful initiation of continuous rotating detonation with liquid kerosene and liquid oxygen.

Satellite orbit-control engines mainly use liquid hypergolic propellant,which has the merits of being storable and hypergolic,and has specific-impulse characteristics.If it was possible to initialize and stabilize rotating detonation with liquid hypergolic propellants,the specific-impulse of orbit-control engines would be improved dramatically,which would advance service life and lower the cost of satellites significantly.Therefore,is it possible to initialize rotating detonation with liquid hypergolic propellant?

Clayton et al.16,17described liquid rocket engine resonant combustion with liquid hypergolic propellants of nitric acid and aniline/furfuryl alcohol.A steep-fronted,high-amplitude pressure wave spinning around the combustion chamber was found duringresonant combustion and described asa‘‘detonation-like” wave.Kan and Heister18,19explored a pulsed detonation rocket engine concept through the use of non-toxic liquid hypergolic propellant in a fuel-centered pintle injector combustor.Based on this research,we developed experiments to initialize rotating detonation with liquid hypergolic propellant and explore the characteristics of the detonation waves.The propellants used were liquid Nitrogen TetrOxide(NTO)and liquid MonoMethylHydrazine(MMH),which are popular for satellite orbit-control engines.Development of a detonation chamber with liquid hypergolic propellants is an attractive method to improve the performances of satellite orbit-control thrusters.

2.Experimental apparatus

A rocket-type annular combustor for initializing rotating detonation was developed,as shown in Fig.1.The annular combustor has a copper alloy heat sink chamber,with outer and inner diameters of 60 mm and 20 mm respectively.The inner column is a changeable module element.The length of the chamber is 150 mm.The liquid propellants were injected and atomized through 24 pairs of unlike impinging injectors distributed uniformly around the injector plate.The diameter of the fuel injecting orifices is 0.31 mm,and that of the oxidizer injecting orifices 0.37 mm.By applying the Phase Doppler Particle Analyzer(PDPA)technology,20the Sauter Mean Diameter(SMD)of droplets from the injectors was measured with water,based on hot tests operating conditions.The SMD is about 88–95 lm in atmosphere conditions.The distance between the measuring plane and the impinging point is about 30 mm,beyond which the SMD changes slightly.Due to hypergolic propellant,the combustor does not need an igniter or a hotshot tube.7

Nitrogen-compressing fiuid supply systems were used to supply the propellants to the combustor.Combustion products were exhausted into a tail gas absorber and vented to the atmosphere.The mass fiow rate of propellants was controlled through regulating the pressure of the tank and the fiow resistance of the supply systems.The fiow resistance was calibrated through a series of cold tests conducted with NTO and MMH in the same test bench as that of the hot tests.Two Coriolis force mass fiowmeters(Micro Motion,model F050)were used to monitor the mass fiow rate in the cold tests.Piezoresistance gauge-pressure sensors were installed in the manifolds and the combustor to measure the static pressure in the cold and hot tests.The sensors for capturing the static pressure before injecting were installed in the manifolds,and the sensor for measuring the static pressure of the combustor was installed in the middle of the chamber.The installation plane was 75 mm away from the injector plate.Two dynamic pressure sensors(Kistler,type 6052C)were installed in the combustor to obtain the propagating pressure waves in the hot tests and labeled C-D(as shown in Fig.1).The distance between the installation plane of Sensors C and D and the injector plate was 10 mm,and the angle between Sensors C and D was 45?.The sampling rates of static and dynamic pressure sensors were 1 and 102.4 kHz,respectively.The piezoresistance pressure sensors showed errors of less than±0.025 bar(1 bar=105Pa)in the area of interest.The mass fiowmeters were calibrated with an error of 0.5%of the reading.All channels of pressure data started to sample simultaneously through the test bench.

A Phantom V12.1 high-speed camera was used to detect the combustion fiame directly from the combustor exit in the hot tests,as shown in Fig.2.The camera was equipped with a CMOS(Complementary Metal-Oxide-Semiconductor)sensor.The fiame was detected at a frame rate of 50 kHz with a resulting optical resolution of 256 pixel?256 pixel,covering the full exit area of 75 mm?75 mm.Each pixel represented an area of 0.3 mm?0.3 mm.The exposure time was chosen through a series of hot pre-tests and fixed at 9.6 ls,which balanced the brightness and contrast of fiame images.A large quartz glass was placed between the camera and the combustor to protect the camera from the hot exhaust.

Fig.1Experimental combustor.

Fig.2Optical setup for capturing fiame images.

3.Experimental results

Table 1 lists some of the operating conditions and test results of the hot tests,where OFR is the mixing ratio of oxidizer and fuel.The temperature of NTO and MMH before being injected into the combustor was room-temperature.Due to the heat sink combustor,the duration of a typical hot test was about 1.0 s,which is shorter than the response time of the massflowmeter.However,the mass fiow rate of propellant in the hot tests could be calculated using the pressure drop and the lf ow coefficient of the injectors.20The fiow coefficient of the injectors was also obtained through cold tests conducted with NTO and MMH.For Tests 1–4,the experiments were conducted with the combustor as shown in Fig.1.For Tests 5–6,a converging nozzle was installed in the combustor by switching the inner column.

Fig.3 shows a series of fiame images detected by the highspeed camera during Test 3(in Fig.3,t is the time).There are two combustion fiame fronts rotating along the outer wall,and the fiame fronts propagate in opposite directions.The light intensity at the point of fiame fronts collision is especially heightened,and the fiame intensity near the inner wall is especially weak.Based on statistical analysis of fiame images,the propagation velocity of the fiame fronts could be calculated using the following formula:

Table 1Operating conditions and results.

Fig.3Flame images of Test 3.

where Doutis the outer diameter of the combustor,ncirclethe number of a fiame front’s rotating circles,n the statistical magnitude of the images(generally n>5000,to make sure the time scale>100 ms),and Dt the sampling time of the camera(equal to 1/50000 s).The mean propagation velocity of theflame fronts is about 1450 m/s.

Fig.4 shows the static and dynamic pressures during Test 3.The static combustor pressure is about 0.4 bar(gauge pressure).The oscillation of the static pressure when the combustion starts is due to the water hammer phenomenon of the injectors.The raw dynamic pressure data was high pass filtered at 1 kHz to eliminate zero shift of Kistler sensors.Near t=4.85 s,where the chamber pressure starts to level off,dynamic pressure oscillation grows spontaneously.The peakto-peak value of pressure waves is about 4.0–12.0 bar.The sharp-fronted characteristics of the pressure waves(as shown in Fig.4(b))are similar to those of the waveform in Ref.15.The typical rise-time of the wave front is about 9.8 ls,which is typical for detonation waves.10The collision point of theflame fronts during the time scale of Fig.4(b)is close to the position of Sensor D,and the peak of the dynamic pressure of Sensor D is especially high.When the fiame fronts pass away from the collision point,the peak of the dynamic pressure becomes lower,as shown in the data for Sensor C in Fig.4(b).

The propagating velocity of pressure waves can be calculated using two methods.7,15,21The first method has been shown in Refs.6,7,15;based on the two time values of the occurrences of pressure peaks on the same sensor and the circumference of the chamber,it is possible to calculate one value of velocity propagation.Since the calculations are carried out for all pressure peaks,all values of the instant velocity are obtained.The second method15,21is to use a Fast Fourier Transform(FFT)determining the dominant frequency of the pressure wave.Both methods are valid,21and calculation results for Test 3 are shown in Fig.5.The first algorithm could be influenced by the position of the collision point and the appearances of random pressure peaks,which causes the instant velocity to scatter across a region.

The instant velocity of pressure waves is in a range of 1350–1650 m/s,as shown in Fig.5(a).The mean value of the dominant frequency is 7732 Hz,as shown in Fig.5(b).Based on the dominant frequency,the propagation velocity of the pressure waves is about 1457 m/s,which is very close to the velocity of the fiame fronts.By tracing the fiame fronts,it is found that the dynamic pressure reaches a peak when the fiame front is near a dynamic pressure senor.Therefore,the propagation velocities of pressure waves and fiame fronts could be evaluated as one value,for example,1457 m/s.According to the software NASA CEA,the sonic velocity is 1150 m/s when the total pressure of the combustor is set at 5 bar and the mixing ratio is set at 2.0.The velocity of pressure waves is supersonic(Mach number Ma=1.27),and the pressure waves in the combustor are shock waves.Shock waves could enhance the process of propellant secondary atomization,22,23which may be a propagating mechanism of rotating detonation waves with liquid propellant.

Fig.4Static and dynamic pressure signals(Test 3).

Fig.5Comparison between two calculation methods of propagation velocity during Test 3.

Fig.6Static and dynamic pressure signals(Test 2).

Based on the definition ofrotating detonation in Refs.5,6,15,24,the initialization of stable rotating detonation with liquid hypergolic propellant was successful in Test 3.Throughout the entire course,there were two detonation waves propagating in opposite directions.Only at random times did one detonation wave propagate around the chamber.The combustion pattern with two counter-rotating waves in Test 3 was very similar to that mentioned in Ref.11,which is referred as counter-rotating double-wave mode.

Fig.8Power spectral densities of dynamic pressure based on measured data of Sensor D in Test 2.

Results of the other tests are shown in Table 1.In Test 1,detonation combustion was not initialized.In Test 2,rotating detonation combustion was obtained,and there were two different combustion patterns in this test,single-wave mode and counter-rotating double-wave mode,presenting alternately and randomly.No predominant mode was sustained during the test.The combustion pattern of Test 4 was similar to that of Test 3,which was counter-rotating double-wave mode.Overall,with the mass fiow rate increasing,rotating detonation was obtained.

The characteristics of the dynamic pressure and combustionflame of single-wave mode during Test 2 are shown in Figs.6 and 7,respectively.Based on the code used to calculate the propagation velocity of the fiame fronts,the velocity of theflame fronts is about 1352 m/s.The frequency information of pressure waves during Test 2 is shown in Fig.8.The dominant frequency is 7219 Hz.Based on the second method of calculating the pressure wave velocity,the mean velocity is about 1361 m/s.The velocity of pressure waves is supersonic(Ma=1.17).Comparing Tests 2 and 3,the peak value of the dynamic pressure and the velocity of the waves are lower in Test 2,which implies that the intensity of the detonation wave is weaker.Further reducing the mass fiow rate to the operating condition of Test 1,detonation was not obtained.The pressure characteristics and fiame images of Test 1 are shown in Figs.9 and 10,respectively.The fiame is uniform,and the combustion is stable.

Fig.7Flame images of Test 2(detonation wave propagating in counter-clockwise direction).

Fig.9Static pressure(top)of combustor and dynamic pressure(bottom)of Sensor D in Test 1.

Fig.10Flame images of Test 1.

Satellite orbit-control thrusters are usually operated with nozzles for achieving high performance.The possibility of rotating detonation initiation in a combustor with a nozzle is also a major concern.Additional experiments were conducted.A combustor with a nozzle was obtained by switching the inner column of the combustor,as shown in Fig.11.The equivalent diameter of the combustor throat was 35.2 mm.The distance between the throat and the injector plate was 115 mm.The mass fiow rate of the operating condition(named Test 5 and 6)was similar to those of Test 2 and 3,as shown in Table 1.

The pressure characteristics of Test 5 are shown in Figs.12 and 13.The static gauge-pressure of the combustor is about 3.4 bar.The peak-to-peak value of pressure waves is about 8.0–18.0 bar.Based on the dominant frequency(8187 Hz,as shown in Fig.13),the velocity of the pressure wave is about 1543 m/s(Ma=1.34).Stable rotating detonation was initialized and sustained.The pressure waves velocity was higher than those of the cases without a nozzle,implying that the detonation waves were enhanced in Test 5.

Fig.11Experimental combustor with a nozzle for Test 5.

Fig.12Static and dynamic pressure signals(Test 5).

Fig.13Power spectral densities of dynamic pressure based on measured data of Sensor C in Test 5.

Results of Test 6 are shown in Table 1.In Test 6,stable rotating detonation combustion was also obtained.The combustion pattern of Test 6 was similar to that of Test 3,which was counter-rotating double-wave mode.

4.Conclusions

This paper presents an experimental study on rotating detonation with liquid hypergolic propellants,consisting of NTO and MMH at room temperature.The characteristics of pressure waves and fiame fronts in a model combustor were obtained and analyzed based on data from static pressure sensors,dynamic pressure sensors,and optical diagnosis equipment.The results obtained are summarized as follows:

(1)It is experimentally verified that rotating detonation could be spontaneously initialized and sustained with liquid hypergolic propellants in a rocket-type annular combustor.

(2)Two combustion patterns were found and analyzed,namely,single-wave mode and counter-rotating double-wave mode.

(3)Stable rotating detonation combustion could also be initialized and sustained in a combustor with a nozzle,which signals a bright future for the application of rotating detonation to satellite orbit-control thrusters.

The main achievement in this paper is a preliminary demonstration of the feasibility and practicality of rotating detonation with liquid hypergolic propellants.However,several challenges need to be overcome to put the Rotating Detonation Engine(RDE)concept into practice.The next step should focus on the initializing mechanism and the performance of the rotating detonation with liquid hypergolic propellants,conducting experimental,numerical,and theoretical research systematically.