DEVELOPMENT OF A ROCKET ENGINE IGNITER USING THE CATALYTIC DECOMPOSITION OF HYDROGEN PEROXIDE

Transcription

1 DEVELOPMENT OF A ROCKET ENGINE IGNITER USING THE CATALYTIC DECOMPOSITION OF HYDROGEN PEROXIDE Wouter A. Jonker (1), Alfons E.H.J. Mayer (2), Barry T.C. Zandbergen (3) (1) TNO Science&Industry, P.O.Box 155, 2600AD, Delft, the Netherlands, Wouter.Jonker@tno.nl (2) TNO Defense,Security&Safety, P.O.Box 45, 2280AA, Rijswijk, the Netherlands, Alfons.Mayer@tno.nl (3) Delft University, Aerospace Engineering, P.O.Box 5058, 2600GB, Delft, the Netherlands, B.T.C.Zandbergen@lr.tudelft.nl ABSTRACT TNO has defined a new concept for a restartable rocket engine igniter which facilitates the catalytic decomposition of hydrogen peroxide monopropellant to generate hot, oxygen-rich gas. Such an igniter can offer advantages over existing igniter concepts: the reactivity of the hot oxygen aids in the ignition while the use of a monopropellant and catalyst allow a simple design. To demonstrate the concept, a propellant feed system and an igniter prototype were constructed, based on requirements derived from the new European Vinci rocket engine. Test results indicated complete decomposition of 87.5% concentrated hydrogen peroxide at a mass flow of 45 g/s, corresponding to a power output of 105 kw, one quarter of that of the Vinci igniter. It was demonstrated that it is possible for a full-scale igniter to meet realistic requirements for restartability and starting transient, while thermal analysis shows losses remain acceptable. This paper presents the design of the igniter prototype and the feed system, and the results of testing and analysis. The work was executed at TNO as part of a student Masters Graduation assignment from Delft University of Technology. 1. INTRODUCTION Most modern restartable rocket engines use either a spark-torch igniter or a pyrophoric (hypergolic) igniter. Spark-torch igniters burn a bipropellant mixture, which is obtained from the main propellant tanks (e.g. RL10) or from a separate feed system (Vinci). This requires many valves, tubes, and relatively complex and expensive control electronics. Pyrophoric igniters have been used to ignite restartable LOX/Kerosene engines (e.g. RD-180) and LOX/LH2 engines (EADS cryogenic 300N engine), and operate by injecting a third chemical into the combustion chamber, which reacts spontaneously with oxygen. The chemical used is usually a mixture of the substances triethylborane and triethylaluminium. These chemicals are highly toxic and ignite spontaneously on contact with air, making them difficult to produce, store and handle. An igniter based on the catalytic decomposition of HP requires less complex hardware than a spark-torch igniter since it needs only a single feed line and valve, and compared to spark-torch igniter with separate feed system it needs only a single storage tank. The use of a monopropellant and a catalyst also removes the need for electronics for numerous valves and a spark exciter, which is complex and therefore expensive to design and qualify. Unlike hydrogen or oxygen, HP can be stored at high density under ambient temperature and pressure, while unlike triethylborane and triethylaluminium, HP is a clean liquid of low toxicity and has completely nontoxic reaction products[1]. The adiabatic decomposition of 90% concentrated HP produces gas of 1033K, well above the auto-ignition temperature of e.g. LOX/LH2 or LOX/Kerosene mixtures[2], but low enough to not require active cooling. Contrary to existing igniter concepts, an igniter based on HP decomposition produces an exhaust containing hot oxygen; for 90% concentrated HP this is about 40% by mass. The reactivity of the hot oxygen results in an easier ignition than the inert flame alone could achieve, especially since bipropellant mixtures in rocket engines are usually fuel-rich. It is expected that a complete ignition system based on HP decomposition can be made lighter, safer, more reliable and at lower cost than current systems. An important step early in the development process is the design and construction of an igniter prototype and a ground-based propellant feed system. To provide a realistic background for the development, the performance requirements for the Vinci igniter are considered typical for modern engines; design and testing of the HP igniter prototype must indicate whether it is possible for a full-scale HP igniter to meet similar requirements. 2. IGNITER AND FEED SYSTEM DESIGN 2.1 Requirements The difficulty with igniter design in general is that it is not very well understood what exactly is required to ignite a rocket engine. The approach usually taken is to

2 determine the thermal power required to raise the temperature of the bipropellant mixture from the inner ring of injector elements to the auto-ignition temperature. It is assumed that once the inner ring has been ignited, the rest of the engine will follow. The thermal power is delivered by heat transfer from the hot igniter exhaust to the cold bipropellant mixture, and so takes the form of Eq. 1: ( H ) P = m & (1) c H 0 where H c is the enthalpy of the exhaust at the chamber temperature and H 0 is the enthalpy at reference temperature. During the initial development of the Vinci igniter, a reference temperature of 298K was used; during newer studies within the frame of ESA s GSTP program the reference temperature is taken as the autoignition temperature of the main propellants. Table 1 shows assumed performance requirements for the fullscale HP igniter; the prototype will have to demonstrate whether these requirements can be met. The number of ignitions, ignition duration and output power are based on the requirements of the Vinci igniter[3]. Table 1. Assumed HP igniter requirements Number of ignitions 5 minimum Duration of ignition 2 seconds minimum Output power > 440 kw at the start of the last ignition for a reference temperature of 298K Starting transient P c 0.95 P nominal in <100ms Envelope Must fit existing interface equipment 2.2 The effect of hot oxygen The definition of Eq. 1 takes only the energy of the inert flame into account, while one of the key advantages of the HTP igniter is the fact that it produces an oxygenrich exhaust. This hot oxygen can react immediately with hydrogen in the combustion chamber and contribute strongly to the output power. Consider the case where the decomposition of hydrogen peroxide is used to ignite a LOX/LH2 mixture with an auto-ignition temperature of 840K[2]. From an initial temperature of 298K, the adiabatic decomposition of 1kg of 87.5% concentrated HP yields kg H2O and kg of O2 at 969K. The output power from equation 1 for a reference temperature of 840K is 225kW so 440kW requires a mass flow of 1.96kg/s. If it is assumed all the oxygen in the igniter exhaust is used to burn hydrogen from the inner injector ring, an additional 5.57MW is available, which lowers the required mass flow to kg/s. Most modern engines with regenerative cooling have injector elements that inject gaseous hydrogen around a core of liquid oxygen[4]. In such a case, the igniter exhaust first comes into contact with the hydrogen, and the advantage of hot oxygen is used to its fullest. 2.3 Design of the igniter For comparison with the Vinci igniter, the reference temperature for the HP igniter prototype is set at 298K and only the energy of the inert flame is considered. An 87.5% concentrated HP solution was selected as it could be easily obtained; relevant physical parameters of this solution and its decomposition products have been summarized in table 2. The enthalpy difference H c - H 0 is ~2520 kj/kg so a power output of 440kW would require a mass flow of kg/s. Table 2 Properties of 87.5% HP and its decomposition products[5] H 2 O 2 H 2 O O 2 Density, kg/m K Heat of kj/mol formation Molar mass g/mol The HP igniter prototype consists of a main body which houses a catalyst bed, a plug containing the HP inlet, and a tube/nozzle (fig.1, fig.2). The distance from the inlet to the catalyst pack is minimized to achieve a short starting transient. The plug is movable so that the voids below and above the catalyst bed can be varied should the minimum void prove too small. The tube is required to transport the hot gas from the space reserved for the igniter to the injection area of the main engine. Stainless steel 316L was selected for the prototype housing because of its easy machinability, good corrosion resistance and low cost. The prototype was designed using basic engineering equations to withstand a pressure of 75 bar at the reduced material strength that occurs at a temperature of 1033K, with a safety margin of ~2. The prototype fits the envelope of the VINCI igniter except for a ring of material on which sensors are mounted. The available space for the catalyst pack is limited by the need for 6 M5 bolt holes, which are used to mount the igniter in the engine. 2.4 Design of the catalyst pack The catalyst pack is designed along guidelines from Davis&McCormick[6] (fig.3). Silver mesh screen is selected as the catalyst material, based on the high reactivity and easy machinability. The catalyst pack starts with an injector plate, containing some 200 holes of 1.5mm diameter, resulting in an open area of about 20%. Along with the injector plate, 4 mesh screens made of 304 stainless steel act as a diffusor to spread the peroxide evenly over the available area. Stacks of

3 ~15 catalyst screens are alternated with anti-channel baffles which prevent HP from flowing around the catalyst material. Four stainless steel screens provide mechanical support between the relatively weak silver screens and the support plate on which the whole pack rests. The total available height for the catalyst pack is 40mm; with this configuration 27.5mm is available for silver mesh screens. The silver is cut from 40-mesh 12 x24 sheets of 0.5mm thickness into 46mm-diameter disks using a pneumatic punch. Silver screen catalyst has an initially low reactivity which only picks up after some time of use, but the reactivity can be increased through a samarium nitrate coating, nitric acid wash or heat treatment[7]. The silver screens for the HP igniter prototype are rinsed with nitric acid for 2 minutes. After being cleaned with distilled water, the screens are placed in a tube oven and heated at 10K per minute from room temperature to 900K. The treatment results in a clean and porous silver surface and should thereby increase the reactivity of the catalyst material. The nitric acid wash dissolved part of the silver: screen mass had reduced from gram to gram, thickness had reduced to 0.32 mm and the gap size had increased. After the heat treatment the silver screens had turned from grey into bright white, had lost elasticity and could easily be deformed. Despite the parameters of the surface treatment being chosen rather arbitrarily, testing at Delft University proved the reactivity of the treated screens had greatly increased compared to the untreated screens. Bengtsson[7] has performed some more extensive reactivity tests for silver screen catalyst; his first order approximation is that some 200 kg HP can be decomposed per minute per liter of catalyst material. Based on these findings we expect the HP igniter prototype to be able to fully decompose a mass flow of 125g/s. 2.5 Design of the feed system The feed system is designed for a maximum operating pressure of 12.5 MPa. The main components are a propellant tank which can be pressurized using nitrogen, and a solenoid valve which controls the HP flow to the igniter (fig.4). There are provisions for filling/draining the tank, emergency depressurization and for purging the propellant lines. Sensors monitor temperature and pressure at critical points. Mass flow rate is measured using a custom-built venturi fitted with a differential pressure transducer. Component specifications are dictated by the expected pressure and mass flow, and by the required chemical compatibility. Recommended materials[8,9] are aluminum or stainless steel, and plastics such as PTFE (Teflon), PCTFE (Kel-F), PVDF, Viton and polycarbonate. All critical valves of the feed system are computer operated and automatically revert to their safe position in the event of a power failure. 3. TEST RESULTS The igniter firing tests took place at TNO in December 2004 in Test Facility 3 (TF3) for igniters and small thrusters. The igniter is fitted with a pressure sensor and thermocouple to measure chamber conditions; two video cameras and a thermographic camera are used to monitor the igniter during operation. Due to the cold weather the peroxide had a temperature of 275K leading to a maximum expected chamber temperature of 938K. The initial cold-start ignition test was unsuccessful: chamber pressure remained constant, chamber temperature increased only some 50K and liquid HP was observed gushing from the igniter tube. During several small tests, short pulses were given to pre-heat the catalyst bed. The heating proved most effective for 0.2-second pulses at 5-second intervals. In follow-up tests five of these pulses were given prior to the main ignition pulse. Four successful 8-second igniter tests were conducted with chamber temperatures reaching 865K. During one 8-second test loud oscillations were heard, and were visible in the exhaust jet on the video monitor. Fig.5 shows the temperature and pressure measurement of one of the 8-second igniter firing tests. The warm-up pulses raise the catalyst bed temperature to some 450K. During the main ignition the chamber temperature is about constant at 8.63 bar while the temperature increases quickly to ~820K in 2.5 seconds, then continues to rise slowly for another 5.5 seconds eventually reaching 840K. The mass flow during the 8- second ignition is on average 45 g/s, resulting in an output power of 105kW (fig.6). The chamber pressure rises to >95% of the nominal value of 8.63 bar in about 90 ms, then oscillates a few times before finally settling above the 95% line after some 300 ms (fig.7). Fig.8 shows the igniter itself during a warm-up pulse, producing a white jet of water vapor and oxygen. 4. ANALYSIS To get an idea of the losses in the igniter, convective heat transfer is modeled in a FE analysis. Relevant gas properties throughout the igniter are calculated using NASA s CEA2000 program for chemical equilibrium calculations, assuming a mass flow of 45g/s, a chamber pressure of 8.63 bar and complete decomposition. The heat flux is then calculated for every point on the igniter wall for all time from t = 0 to 8 seconds. Like the thermographic camera, the analysis shows a quick heating of the igniter tube while the main body stays relatively cool. Convective heat transfer from the gas to the igniter wall is largest in the tube section: initially 5.2kW or some 5% of the output power, causing a drop in gas temperature of about 65K. Within 2 seconds the heat transfer has decreased to <1kW, corresponding to a

4 drop in gas temperature of 12K. Heat transfer in the main body is some 20 times smaller due to the low flow velocities. Fig.9 shows the heat transfer Q in the tube and the resulting wall temperatures at the tube inlet and outlet. Due to non-conformance of the position of the thermocouple wire, the temperature measurement took place at the tube inlet, in the boundary layer near the igniter wall. In addition, the gas flow approached the thermocouple wire tip from behind so the measured temperature is closer to the static temperature than the total temperature. The resulting measurement does not accurately represent the chamber temperature, but it provides a lower boundary. The measured temperature lies roughly between the film temperature and the static temperature of the flow that were found in the FE analysis assuming complete decomposition (fig.10). The oscillations in pressure and mass flow that were observed during several tests occur due to an insufficient pressure drop over the injector. The measured pressure drop over the injector plate, 94 mesh screens and the support plate combined is 18% while 20-30% over the injector alone is recommended[4]. Insufficient pressure drop allows reaction instabilities and pressure spikes to affect mass flow, causing oscillations such as those that are observed. 5. DISCUSSION Testing of the HP igniter prototype has successfully demonstrated the HP igniter concept. The HP igniter prototype, which is of roughly the same size as the Vinci igniter, has been fired over 15 times with 4 pulses lasting as long as 8 seconds, indicating that requirements for restartability and minimum ignition duration can be met. The exhaust, containing nearly 40% of oxygen, reached a temperature exceeding the auto-ignition temperature of stoichiometric LOX/LH2. The starting transient was some 300ms due to oscillations, but redesign of the injector can prevent these and reduce the transient to <100 ms. Whether a full-scale HP igniter can meet a power requirement will depend on the design envelope and on the exact definition of the requirement. The output power increases linear with mass flow, but the maximum HP mass flow than can be decomposed within a certain space is unknown. The limit of the catalyst pack of the HP igniter was not reached during the tests, and the prototype was not optimized for a maximum catalyst pack size. In addition, the catalyst surface treatment was not optimized. The output power depends even stronger on the effect of the hot oxygen in the exhaust. Further testing is necessary to determine to which extent the hot oxygen contributes to the output power. 4. ACKNOWLEDGEMENTS W.A. Jonker wants to thank all his advisors at Delft University and TNO; in particular A.E.H.J. Mayer and B.T.C. Zandbergen for their support throughout the thesis work and A.G.M. Maree and H.F.R. Schöyer for countless helpful discussions. 7. REFERENCES [1]Iarochenko, N., and Dedic, V., Hydrogen Peroxide as a Monopropellant, Catalysts and Catalyst Beds, Experience from more than Thirty Years of Exploitation, Russian Scientific Center :Applied Chemistry, Russia, St. Petersburg [2]Chemical Engineers Handbook, 3rd edition, by John H. Perry, Ph.D., McGraw-Hill publishing company Ltd, 1950 [3]Frenken, G., Vermeulen, E., Bouquet, F., and Sanders, B., Development Status of the Ignition System for Vinci, 4th international conference on launcher technology space launcher liquid propulsion, december 2002 [4]Huzel, D.K., and Huang, D.H., Modern Engineering for Design of Liquid Propellant Rocket Engines, AIAA Progress in Astronautics and Aeronautics Series, AIAA, Washington, [5]Knacke, O., Kubaschewski, O., and Hesselmann, K., Thermodynamic Properties of Inorganic Substances, 2nd edition, Springer-Verlag, Berlin Heidelberg, ISBN [6]Davis, N.S. and McCormick, J.C., Design of Catalyst Packs for the Decomposition of Hydrogen Peroxide, American Rocket Society paper nr [7]Website: by Erik Bengtsson, June 2005 [8]Whitehead, J.C., Hydrogen Peroxide Propulsion for Smaller Satellites, 12th AIAA/USU Conference on Small Satellites, SS98-VIII-1 [9]Degussa brochure, Hydrogen Peroxide Properties and Handling, Ch NOMENCLATURE p = pressure P = power m& = mass flow H = enthalpy T = temperature 0 = index for reference conditions c = index for chamber conditions DAQ = data acquisition HP = hydrogen peroxide GSTP = General Support Technology Programme TF3 = Test Facility 3 for igniters and small thrusters at TNO Defense, Security & Safety

5 9. FIGURES Figure 1. Igniter schematic Figure 2. Igniter hardware Figure 3. Catalyst bed composition Figure 4. Liquid-propellant feed system schematic

6 E T (K) p T 1.0E E E E E+05 p (Pa) m (kg/s) P m P (kw) E t (s) t (s) Figure 5. Chamber pressure and temperature Figure 6. Mass flow and output power 1.4E+06 P (Pa) 1.2E E E E E E+05 valve opens p=0.95 pnominal 0.0E t (s) Figure 7. Pressure starting transient Figure 8. Igniter firing test Wall temperature (K) T (outlet) T (inlet) Q time (s) Heat transfer (J/s) Temperature (K) Tstatic Tmeasured 750 Tfilm 625 Twall time (s) Figure 9. Tube heat loss and temperature (analysis) Figure 10. Measured / calculated gas temperatures