I ..I INVESTIGATION INTO LOW-COST PROPULSION SYSTEMS FOR SMALL SATELLITE MISSIONS ~I .lerry .Ion Sellers. Alalcolm Paul, ,Maarten Meerman & Robert Wood Umversity of Surrey. Guildford. Surrey_ U.K. ., Abstract introduction LO\\ -cost satellites need low-cost propulsion The UniverSity of Surrey satellite research .. systems. The research sUl1U11arised in this paper group (UoSA T) is a world leader in 100\-cost. , small satellite engineering. To achievl.: high has focused on investigating low-cost propulsion system options for small satellites reliability at an affordable price. the UoSA T ~I \\Ith specific application to the upcoming philosophy has emphasised simple. flexibk UoSA T -12 minisatellite mission. The research designs with maximum use of off-the-shelf :._, began b~ looking at available propulsion terrestrial hardwarc Evolved over II system wchnology Low-cost spacecraft missions. the 50 kg UoSAT. gravity-gradient cngineenng techl1Iques were then explored to stabilised microsatellites have performed a identify specific systcm cost dnvers for further variety of miSSIOns in LEO. However. for .1 investigation This led to parallel research these small satellites to continue to evolve efforts aimed at ( I ) basic research & economically beyond the niche of LEO and development into hybrid rockets to characterise exploit emerging new opportunities such as -I their applicability to 100v-cost spacecraft. and more ambitious communication. remote sensing (2) applied research into low-cost spacecraft and science missions. a larger. more flexible systems engineering-to design and implement bus with a propulsion capability is essential. a system for UoSA T -12. The experimental To pursue this goaL UoSAT researchers. I results from a 400-Newton thrust hydrogen peroxide and polyethylene hybrid motor are sponsored by Surrey Satellite Technology Ltd. (SSTL). han: undertaken the deSign of a presented. Initial results indicate that >90~(1 :.1 modular. multi-miSSIOn minisatellite bus The combustion efficiency is achievable and an experimental hybrid mission could feasibly be UoSA T -12 mission IS the maiden flight of thiS developed over the next fe\\ years. Additional ne\\ platfom1 Intended as a technolog~ I . research into low-cost propulSIOn is also demonstratIon mission. it will bl.: launched from Baikonur Cosmodrome in October 1996 aboard discussed mcluding the application of a lo\\,­ thrust (20-Ne\\10n) bl-propellant rocket a Rockot booster. The UoSA T and SSTL I engine-the LEROS-20. devdoped by British minisatellite bas an approximate mass of 250 Aerospace. Royal Ordnance Rocket Motors kg. The minisatellite structural design builds on DiviSIOn. The combination of innovative the modular approach used in UoSAT I manufacturing techmques along with low-cost microsatellites in a wa~' that allO\vs maximum procurement practices makes this engine an re-use of subs:vstems bet\\een the t\\O attractive. low-cost option. Finally. research platforms.

~ I lI1to decreasmg the cost of support subsystems Along. with enhancements to the has lead to a simple. low-cost design which IS communications. data handling and attitude bell1g Implemented Oil UoSA T -12 at a fraction control systems. UoSAT -12 \vill include an , of the cost of that predicted by industry­ experimental low-cost propulsion system standard models. derived from our resl;:arch. A diagram of UoSA T -12 is ShO\VI1 in Figure 1. The ilV I budgets for the UoSA T -12 propulsion system include ~200 111/S for orbit COrrl;:ctlOI1 and maintenance in addition to ~200 Nms for I supplementary attitude controL t spacecraft systems engineering, going beyond the "paper study" phase to determine the actual, achievable cost savings. I

Propulsion System Options ';.,- I From a conceptual standpoint. propu)sion system technology can be divided roughly-into four categories based on tlte method used to I accelerate the reaction mass. These- are summarised below along with the approximate specific impulse, Isp, for each. Isp is a useful , measure of engine efficiency. much ~Iike miles-per-gallon for cars. (Here we focus only on current or near-tenn technology. For a list , of more far-out ideas-fusion, anti-matter, negative matter, etc.-the reader is referred to l Humble ). I • Cold-gas systems-use the energy of a gas stored at high pressure to accelerate the. gas to high velocity through a nozzle. Isp 70 , seconds. Figure 1: The UoSA T-12 spacecraft based on • Chemical systems:"'-"use the energy inherent the new UoSATISSTL modular minisatellite in chemical bonds released through I bus. catalytic action or combustion to produce high temperature exhaust products which The addition of propulsion represents a giant are then expanded out a nozzle to?igh I step in tenns of UoSA T capability. Of course, velocity. Isp - 300 seconds. . propulsion systems are an integral part of most • Nuclear systems-use the intense heat large spacecraft. They enable orbit generated by nuclear fission (or fusion) to I manoeuvres, station keeping, constellation heat up an inert reaction mass (sud; as maintenance and attitude control. However, hydrogen) to high temperature. The mass because of their prohibitive cost and is then expanded out a nozzle to ,high , complexity, they are rarely found on very low­ velocity. Isp - 1000 seconds. cost missions such as UoSAT-12. The • Electric systems-use electrical energy to research summarised in this paper has focused accelerate a reaction mass either through I on investigating low-cost propulsion system electrothermal, electromagnetic . or options and identifying specific applications electrostatic means to high velocity. Isp- for this technology on small satellites. The 10,000 seconds. approach taken in the research-and I We can begin by examining the ele<:tric summarised in this paper-begins by looking options. At first glance, the incredible at available propUlsion system options. The efficiency these systems offer-which I concept of low-cost spacecraft engineering is translates into lower propellant rilass r"- then explored to understand why current requirements-seems ideal for small sate-Ilite systems are so expensive and identify specific appliaations. However. there are two inherent cost drivers for further investigation. This I drawbacks to electrodynamics systems. leads to a discussion of parallel research efforts aimed at (1) basic research & • Very low thrust , development into alternative systems­ • Very high power requirements specifically hybrid rocket motor technology­ The first drawback necessitates that for:any to characterise its applicability to low-cost significant manoeuvre. the engine must. be spacecraft (2) applied research into low-cost I kept on for a very long time. Because small

2 satellites are generally not in a hurry, this is spacecraft must be designed around the not necessarily a problem. However, to take system. This integration complexity increases I advantage of this engine efficiency the thrust the cost of these systems significantly for vector needs to be continually pointed in an first-time use on a new satellite bus. optimum direction, requiring a more complex, Furthermore, the toxicity of the propellants I continuous 3-axis control system. Again. this make them a significant safety risk (which may not be overly constraining for a low-cost further increases the cost). Despite this, liquid system. Unfortunately, it is the second bi-propellant engines using storable t drawback of electrodynamic systems that is propellants are the "work horses" of the the hardest to overcome. Discussion of satellite industry. They are used in every electric propulsion typically focuses on thrust aspect of propulsion-launch. orbit insertion, level per of input power. Even a orbit manoeuvre, orbit correction and attitude modest svstem such as the augmented control, They have been to the surface of the 2 hydrazine 'resistojet built by Olin Ae;ospace Moon and are on their way to the stars. in the U.S.A. produces relatively high th~ust Because they are such an industry standard, a (0.27 N) at an impressive 300 sec Isp but with vast infrastructure already exists to support a power requirement of SOO W. With a total their operations. Thus. they \\ arrant serious available power in the UoSAT minisatellite consideration in any analysis of small satellite system of less than 100 W. electric propulsion application. The challenge is to drive down systems had to be excluded from their cost. , consideration due to these excessivc power requirements. Similarly, nuclear systems were The only remaining chemical system to excludcd from consideration for practical consider is the hybrid. There are currently no I reasons. commercially available hybrid systems. So far, hybrids have only seen service in a limited Thus. the potential shopping list quickly role on sounding rockets. No hybrid rocket comes down to chemical (solid, liquid and has ever been used in space. However. as we I hybrid) and cold-gas systems. Solid motors will present, hybrid rockets offer many can only be used once, therefore. for multiple inherent advantages which make them a prime bums, multiple motors-with all the candidate for low-cost application on small I additional overhead that entails-must be satellites. A large part of our research has employed. Furthermore, you are slaved to only focused on hybrid systems, in basic and a few potential suppliers and their "off-the­ applied research into finding ways to realise I shelf" motors. Costs of custom-designed their potential. motors go up dramatically. For these reasons, and because of the requirement for a flexible Finally, there are cold-gas systems. Because I of their very low lsp. cold-gas systems are '" orbital manoeuvring and orbit correction system for the minisatellite, solid motors were really only practical in the attitude control and not considered in our research. minor orbit correction role. However, both liquid and hybrid systems require pressurised The most widely available liquid systems rely gas to deliver the propellant to the combustion on storable propellants, primarily hydrazine chamber. Therefore. when either of these two 1 used as a mono-propellant at an lsp of 21S- systems are employed for primary orbit ... 240 seconds or hydrazine (or its derivatives manoeuvring, the cold-gas system can be had such as monomethyl hydrazine (MMH), "for free" for attitude control applications. I UDMH. Aerozine-SO, etc.) with nitrogen - tetroxide (NTO) used in a bi-propellant . Defining a Low-cost SYstem system with an lsp of around 300 seconds. , What is a "low-cost propulsion system?" As a satellite designer, its important to keep in There is an expression which goes-"you get mind that even with "off-the-shelf' liquid what you pay for." Large space programmes systems, they can not be just "bolted-on," for multinational communications or manned • ready to fly. In fact, to a large extent, the space flight can usually justify every cent of I 3 I their enormous expenditure on system quality gas (6.3 litre @ Ne, 200 bar, 24.87 cm and product assurance. The potential price for diameter). failure is extremely high: billions of dollars or • Orbit Correction !J. V = 200 m/s. I human lives. However, propUlsion systems • Bi-propellant engine Isp = 290 sec for small satellites, especially on low-cost, • Standard system architecture higher risk missions, would not be feasi~le if I implemented on the same basis. One of the gives an overall propulsion system cost objectives of our research was to deter.mine estimate as summarised in Table I. (They how best to design, implement and ope(pte a worked in accounting units where lAU == t propulsion system on a small satellite platform $0.98US. For simplicity we will assume lAU while carefully trading off cost, risk' and $IUS). performance. We can conclude from this analysis that. at I' Svstem Cost Models first glance, a small satellite designer is faced with propulsion system costs of >$600k in The most important question to be answered non-recurring mission design costs in addition I before designing a "low-cost" system is '-'what to around $700k per mission. To make does a high-cost system cost?" Unfortunptely, matters worse, by the nature of the small because of the complexity involved~ith satellite industry, "production runs" of i adapting an existing system to a new mission, spacecraft are virtually unheard of. All the answer to this question is not necessarily technology development costs must be funded 3 straight-forward. In 1984, Smith and Horton "in-house" and charged against the first I did a comprehensive survey of available mission. There is no external customer to pass propulsion systems and developed· cost on these costs to and no easy way to amortise functions to allow us to make basic system these costs over several missions. Thus. there I cost estimates for engines, tanksandcither is a very sensitive "enabling cost" level for components. new technology systems, meaning that below this enabling level the mission can proceed. Applying their cost relationships to: the 'I Above this level the mission is not feasible. following mission scenario: • Spacecraft deployed mass == 250 kg . I • Shell tanks for oxidiser, fuel (10 litre.each @ 20 bar, 27.18 cm diameter, 1.65: I oxidiserlfuel (O/F) ratio) and pressurant I Table 1: Smith & HorlOn 3 cost model applied to a 250 kg spacecraft with a 200 mls !J.V requirement along with cold-gas thrusters based on standard :.ystem architecture. , Component Qty. Unit.Mass Total . Recurring Non- Total First I I (kg) Mass (kg) Cost Recurring spacecraft Cost cost ,j Liquid Engine I 3.36 3.36 S200,000 $150.000 $350.000 Propellant Tank 2 1.00 2.00 S14,775 S]08,720 S138.270 Nitrogen Tank 1 3:36 3.36 S12,370 S99,480 Sl1 1,850 Regulator 1 0.84 0.84 S30,000 $100,000 S130.000 I Pressure Transducer 4 0:23 0.92 S6,500 20000 S21,495 Pyro Valve -'" 0.15 0.44 ., S2,500 S50,000 S57,500 J Relief Valve I 0.45 0.45 S10,000 S80,OOO $90.000 Filter 3 0.13 0.68 $4,000 S30,000 $42,000 Fil\!Drain Valve 5 0.15 0.73 S2,700 SlO,OOO $23,500 Cold-Gas Thruster 12 0.100 1.20 S20,OOO S100,000 $340.000 I Thruster Iso Valve (pyro) 2 0.15 0.29 S2.500 $50.000 S55.000

.' !Total ~,', 12.77 S711,415 $648,200 S1.359.615 I

4 I I I Therefore, the challenge we faced was to find (usually provided in a several inches thick ways. within the context of "faster, better. document e.g. NASA, ESA or DoD cheaper" paradigm for which the small programme specifications). , satellite community has always been known, 2. Individual components at minin:1Um to substantially reduce the overall system cost. possible cost as the result of relaxatioh of In the next subsections we will break apart requirements as addressed in the attached I these high costs to determine the primary cost RFQ and modification of your nQ[mal drivers and then pull together basic system production process. engineering approaches for reducing them. Unfortunately, response to this RFQ was .Iess , than overwhelming. Only 12 compa.nies Market Analysis provided serious responses to any components. While these cost models are useful to gIve a These responses are summarised in Tab1e 2 I starting point for a paper-study mission, a (because most companies only providecr the "reality check" of the industry is necessary in second type of bid asked for, minimum c-ost. order to proceed with a real mission. To only those results are reported). I quantify price ranges and availability of propulsion system components for UoSA T-12, Table 2: Results of market anaZysis. we prepared a formal Request for Quotation Represents first-unit cost (including 110n­ , (RFQ) which was sent to nearly 30 different recurring cost). aerospace companies. The RFQ component Component Bid Range. requirements were based on a conceptual N2 Tank $15,000 - $208.000 propulsion system design using the 20-Nev\'ton • I Propellant Tank • $15,000 - $27.200 bi-prope l1ant thruster as described later in this N2 Fill/drain valve • $4,000 - $25.600 paper. Propellant FilJldrain • $4,000 - $16,000 1\ valve In the cover letter, we tried to establish a Pyrotechnic valve $12.000 (one bid)~ rapport with the supplier to get them to work • N2 or Propellant Filter $25,600 (one bid) \vith us to lower costs. We wrote: "Our • Pressure Transducers $350 (one bid) I experience in building other satellite systems • Pressure Relief valve • $675 (terrestrial model) has demonstrated that there are ways to lower - $30.000 (space component costs by relaxing performance model) I requirements and judiciously modifying the Pressure regulator • $3.950 - $30.000·(one procurement and parts production process. bid) We know that a $50,000 component cannot Cold-gas thruster • $3.100 - $1 1.117 I simply be changed into a $20,000 component isolation valve by changing the amount of paperwork that gets Cold-gas thruster valve • $2.525 - $12.560 . shuffled. We recognise that the less expensive Propellant check valves • $24.000 (one bid) 1 component can not have the same assured quality as its more expensive counterpart, Using the lowest priced components, we. can however, in many cases it may be sufficient develop a modified-more realistic-budget I for our higher-risk, low-cost missions. We are for the small satellite propulsion system priced - asking Yill! to tell us how best to reduce the above. As \ve did not get a separate bid for the cost of each component with understandable engine, we use the same value (combi'~ing I consequences." recu~~jng and non-recurring costs) to get a revised total system price for the. first Companies were given an abbreviated set of spacecraft of $588,125. While this is a component requirements and asked to provide significant savings from the one predicted by I two separate quotations: the cost modeL it is still quite dear. For il I. Individual components in compliance with minisatellite with a total budget of $4M, the I all typical industry standard performance propulsion system alone would cons'-!,me requirements and acceptance testing nearly 15%. But to get the cost of "the I a 5 propulsion system down any more requires a promlsmg low-cost terrestrial hardware must more basic understanding of the fundamental be rejected due to materials or other aspects I cost drivers for space hardware. We will that are incompatible with space. In this way. explore these cost drivers in the next section. the space environment serves as a hard driver , of the technical approach that must be used.

.', Cost Drivers In addition, there is an equally constraining Why are propulsion systems specifically. and political environment which is just as hard to I spacecraft in general, so expensive? Some, ignore. It is an unfortunate but true fact of life like space economist Chris Elliot4 argue that that the optimum engineering solution is often space missions are so' expensive simply irrelevant in the face of political demands. because we expect them to be. When one I Any discussion of options for buying or assumes missions are high-cost, then decisions producing low-cost propulsion system are made which turn this into a self-fulfilling hardware can only be effective once this basic prophecy e.g. higher cost == fewer missions I context is understood. While the purely higher reliability demands = higher cost. To technical side of this political environment break out of this vicious circle, you must first determines the "acceptable risk" of a given accept that space missions can be inexpensive. I engineering solution, the political dimension Only then can decisions be made which to determines "allowable risk." allow this to happen. , What is "acceptable risk?" This is perhaps As part of our research, we defined a generic, one of the most important, but least politically systematic process for space hardware correct questions that can be asked for any selection in an attempt to identify the primary miSSIOn. Risk. or its inverse reliability. is I cost drivers. We can divide the evolution of a something that should be budgeted along with mission into three phases: power, mass and every other precious • Mission Definition commodity on a spacecraft. In the traditional • Mission Design 'vicious circle' of high-cost missions. risk was • Procurement something that had to be minimised through­ out the entire system. But, of course. trying to We can then isolate significant cost drivers operate inside a very tight 'risk margin' can be within each phase. very expensive. Mission Definition Phase I Obviously. if designers were willing to accept Despite the best cost-cutting attention of any higher risk. mission cost could be reduced. engineer or programme manager, there are But 100:v-cost and high-risk are relative terms. I certain harsh realities of planning and While a large government programme may operating space missions that cannot be simply view a $lIvl microsatellite mission as budgeted away. During the Mission inexpensive. they may not be prepared to Definition Phase, the "who, what, where, why accept additional risk for political reasons. On & when" of the mission is addressed. The the other hand. if a University is paying for it, • answers to these questions determine the entire $1 M is viewed as a significant portion of the I mission environment. research budget. yet they may be more willing to accept risk. So, the trick in low-cost There are two aspects of this environment that engiI]~ering is to effectively spend the risk I are among the biggest cost drivers for space budget to achieye the best return. One way to hardware. Perhaps the most obvious of these is do this is to look for areas where you are the punishing physical environment of space: paying for the same risk reduction more than I radiation, vacuum, etc. Any potential hardware once. And to best understand that. the options must be screened and tested to ensure majority of risk should be centred \\ithin the they will function in this harsh environment. Because of this hard constraint, some I I 6 I prime contractor or rationally shared with technical options that can be considered. For I subcontractors, example, mission ~ V requirements. coupled with mass constraints, can limit you to ,., Regardless of how cleverly you can manage to consider only propulsion options with a cut corners on a particular engineering sufficiently high Isp. Unfortunately. solution and how confident you are it will requirements cost money. Furthermore. for actually work, if your customer or the launch most sYstems there is an exponential "1 authority cannot be convinced the risk is relationship between performance (whether acceptable then it will not be allowed. measured in reliability. specific impulse or Fortunately. within the. context of small bits/sec) and cost. This simple point can not I missions, especially those done by University be overemphasised to the entire design team. researchers. the customer and the designer are By carefully considering the impact of each sometimes one and the same. When you are requirement on the total system in some cases ~.• your own customer, you are more likely to it may be possible that by simply relaxing a fully. appreciate all the mission trade-offs and single requirement by 10%. the overall savings therefore more willing to raise the threshold of could be many times that. ,I acceptable risk. Another point to consider is the performance For propulsion systems, an important margin. Margin is the engineering tolerance "I controller of acceptable risk is the launch given to a particular specifications. For authority. In addition to requirements and ~xample, specifications for an engine inlet constraints specifically spelled out by the pressure may be 250 psi psi. The smaller launch agency, the launch site management the gap between a system's performance and I may have other rules and regulations to which its specification, the more it will cost to ensure you must comply. For example, propulsion compliance. Over-specifying the margin can ,I, systems to be used \vith launch operations out needlessly drive up the cost. Low cost of Vandenberg AFB m California are spacecraft engineering focuses on increasing governed bv WRR 127-1 Range Safety such margins by controlling requirements. Requirement~5. In addition to this regulation, This can lead to a wider range of possible I Military Standards dictate requirements for solutions as well as less testing and analysis. specific components, for example MIL-STD- By changing the margin for inlet pressure 6 1522A covers "Standard General specification in the above example to 0 psi. I Requirements for Safe Design and Operation very little is lost in overall performance but it of Press uri sed Missile and Space Systems." becomes much easier and cheaper to ensure compliance and less sensitive to small changes Mission Design Phase I in configuration or performance of related Once the mission is defined, the top-down support systems, Mission Design Phase can begin in earnest. During this phase many key designs and Required quality level can also be considered as part of hardw'are requirements. " philosophies are established which serve as • important cost drivers for spacecraft hardware. Unfortunately, there is no industry accepted Among the most critical of these for definition of what makes a particular component "space qualified." The most " 'propulsion systems are fundamental performance requirements and constraints, obvious justification, of course. would be if a with corresponding margins, and the basic comp6nent had actually been used in space for system architecture definition. some considerable length of time under similar operating conditions without failure. As any :1 The cost impact of specific mission engineer knows, no amount of testing or performance requirements and their associated analvsis can completely guarantee that a margins can ripple through an entire system. particular component will never fail. But I To begin with, these determine the range of I 7 I attempting to manage this uncertainty is what may never have actually flown in space rather the whole "space qualified" process is about. than use one that is subjectively qualified but A simple bolt is a classic example of how the lacks all the necessary paperwork. most basic of components can attract System architecture is a broad term used to expensive precautions. In all probability, a describe the selection, deployment and bolt bought at the local hardware store would interaction of components within a system. It perform adequately on a spacecraft. However, is another cost driver established during parts used for space missions are typically Mission Design. For propulsion systems, this procured to specifications requiring particular process determines the "plumbing" lay-out of titanium aJloys, traceability of raw materials, the tanks, valves and engines and their concept I precision threads and high torque heads. The for operations. This is fundamentally a design cost of such a bolt may be hundreds of times exercise that ultimately determines the number more expensive than the hardware store and type of components used. During this t equivalent. This is in addition to the process, cost vs. performance trade-offs are administration cost of writing and maintaining identified. The decisions made will directly the specifications, ensuring the high-quality corresponds to system cost. I parts remain traceable and not forgetting that specialised heads require specialised tools and Several important philosophies are established specialised suppliers. Even so, this is often during the system architecture process which considered necessary in the effort to manage also directly drive cost. One of these is the I the uncertainty of mission success: one rogue approach to redundancy. In the realm of low­ bolt may make all the difference. cost, small satellites, true I: I hardware redundancy is virtually impossible. The I To understand this uncertainty management additional mass, power and volume process for specific pieces of hardware, let us requirements of flying complete back-up imagine a quality spectrum for components. systems, in addition to the additional overhead This "quality spectrum" can be defined in required to manage this redundancy, is terms of increasing confidence or reliability as prohibitive. Yet, even on low-cost missions, • you start at one end with commercial or single-point failures, while inevitable, must be I industrial grade components and then move up dealt with. to High REL and finally Space Qualified at the top end. These are listed below in increasing Thus, systems design engineers must balance t order of "quality": the trade-off between two competing philosophies: (I) "all single point failures are I. Commercial/industrial grade components bad", and (2) "if it ain't there, it can't fai I." , 2. Military Spec or military standard or On the one hand, you would like multiple "MiISPEC/MiIStd" paths (e.g. valves) to ensure gas or propellant 3. High reliability or "HiRel" flow in case of failure. On the other hand, I 4. Space Qualified or "SpaceQual" each additional valve leads to additional leak paths, wiring, software, mass and cost. It is important to realise that there are Designers must carefully review the objective measures of "space qualified," e.g. acceptable risk associated with the entire I formal certification through rigorous testing, system and try to establish an approach which as well as subjective measures, e.g. you have balances this risk against cost and complexity . flown it in space before and it worked. .- I Ironically, the political environment may be such that the subjective criteria is not enough. A customer or launch authority my demand I you use a component with an objective certification of "space quality" even though it I

8 Table 3: Summary of advantages and disadvantages associated with buying vs. buildingf!ight I hardware. Option Advantages Disadvantages Buying Hardware YOU ... YOU, .. :~~ - I • Control sharing of risk with • Have less control over ' . supplier specifications • Get tried and tested hardware • Have less control over schedule , • Have no development costs and cost '., • Learn from subcontractors • Spend overhead outside your organisation Building Hardware YOU ... YOU ... I ~ Carryall the risk • Control the specifications and • .. performance • Need the in-house expertise to design the entire component • Control the design and .. ..I interfaces and manufacture it -. • Control the schedule • Need to acceptance test it • Control the cost • Need to space qualify it • Spend overhead in-house • Gain expertise that should make •t: it even cheaper next time Procurement Phase satellite builder and subcontractors. the method for handling system specifications, Once you know what components you want, type and amount of documentation and overall you have to procure them. This is not a trivial project management. At each step of the • exerc ise. The procurement practices and process, there are opportunities to decrease (or , process used have a profound cost impact increase) the final cost of the hardware. ,The giving you a drastically reduced (or increased) challenge is to develop a working relation-ship price for essentially identical hardware. with suppliers that allows you to conduct business with the minimum of documentation, Where can you get low-cost space hardware? . , meetings and overhead. Some low-cost missions have the luxury of • being able to "borrow" donated hardware or When it comes to building your own , purchase flight spares for a much reduced hardware, the primary advantages you gain are price. Unfortunately these situations, while 111 increased control over a particular they can be found (if you are willing to look component. This of course creates. the I hard enough, hat-in-hand) do not offer a potential for cost savings, but by no means repeatable system. Thus, to build a series of guarantees it. The disadvantages flow from low-cost satellites using the same basic ;he hassle caused by having to do it all propulsion system you are left with two yourself. Once you have made the decision to options: build, the basic steps in the manufacturing, • integration and testing processes differ very 1. Buying flight hardware through normal little from those used throughout industry.' -.a procurement channels, or Whether you choose to buy or build 'your ') Building your own flight hardware. hardware, somehow you must test it sO'you 1 Each of these options has their advantages and will' flave some confidence it will work in disadvantages as well as corresponding cost space. On one hand. testing is essential to implications as summarised in Table 3. ensure compliance with requirements. On the One very important cost driver associated with other hand, it can be a tremendous drain on '\ resources-time. money and manpower:--to buying hardware deserves special attention­ the procurement process itself. This has to do conduct. review and document the testing. I For a low-cost system. you only want· to , with division of risk between the primary 9 I perform the testing that is absolutely no hybrid rocket has ever been used or even necessary and avoid those that are frivolous or tested in space. However, as we will see. redundant. hybrid rockets offer several exciting features I that made them the logical candidate for a . But how much testing is enough? To answer build-your-own, low-cost rocket. Additional this question, it is important to distinguish background work on our research can be I between qualification and acceptance testing. found in the March 1995 Journal of the British Qualification tests are used to check the InterplanetGlY Sociel),16. . manufacturing process and determine whether I the product complies with specifications. A typical hybrid rocket motor (by convention. Once a component has passed this hurdle. hybrids are called motors whi Ie all-liquid much less demanding acceptance tests are systems are called engines) consists of a solid I done to individual components to screen them fuel and a liquid oxidiser. Start-up involves for basic functionality. Unfortunately, many openmg a single valve. allowing the spacecraft components do not come off pressurised oxidiser to flow into the , production lines and so satellite builders are fuel/combustion chamber where combustion forced to do both qualification and acceptance takes place either spontaneously or with an testing which drives up the cost. However. ignitor. components are often subjected to a variety of tests following manufacture e.g. vibration, Laying somewhere between a traditional • and then. once fitted into a subassembly are liquid engine and solid motors. it is often said , tested again and then, yet again when the that a hvbrid "combines the best of both 7 subassembly goes into the spacecraft. While worlds ,,- As such. hybrid rockets have these multi-levels of testing do help to screen several inherent advantages. Hybrids: out faulty parts, they create additional costs • Can be easily turned off by stopping the which small satellite designers can ill-afford. flow of oxidiser and then restarted at a later • Our Approach time. • Can be throttled. With this background understood, we set • Have a comparable Isp to storable bi­ • about designing a low-cost propulsion system propellants, and, with oxidiser to fuel for the UoSA T -12 spacecraft, mindful that this t combinations at high ratios (8: I or greater), was to be an in-orbit demonstration mission of their density (the product of and a generic bus to be used for several follow-on Isp Isp average specific gravity of the propellants) mIssIons. Thus, we needed to focus on I is well above liquid bi-propellant systems. repeatable solutions. From a cost standpoint. we split the system into two separate pieces­ • ese readily available fuels such as HTPB the main engine and the support components (rubber) or polyethylene (plastic) and I (tanks, valves, etc.). In the interest of basic oxidisers such as liquid oxygen (LOX) or engineering research. we chose to investigate hydrogen peroxide. the option of building our own main engine • Have environmentally friendly exhaust while pursuing in parallel the application low­ products. cost system design and procurement • Are extremely safe-it is virtually • techniques to bought-in engines and support impossible for a hybrid to explode. , components. A discussion of the preliminary A hybrid rocket combines the storability of a results of these efforts will be the focus of the solid system with the restart and throttle ,I following sections. capability of a liquid system. Perhaps most Hybrid Rocket Development important, the inherent simplicity and design robustness of hybrids make them ideal for safe II The Hybrid Option experimentation and development within a university environment. Thus. they promise a As discussed above, there are currentlv no low-cost, easy-to-use propulsion option "off-the-shelf' hybrid rocket systems worthy of further study. I available for use on small satellites. In fact,

10 I I Therefore, one of the primary objectives of 2. Test-Using this infrastructure and motor. our research was to establish whether or not conduct basic research into hybrid I hybrid rockets really live up to their combustion characteristics expectations. After all. it is often the case in 3. Assess-Along the way, gain sufficient engineering that a particular technology practical experience with hybrid r2C ket I application looks great but in design and operations to enable a crliical practice is sorely lacking. assessment of cost schedule and Other development requirements for a near~.ierm Beginning in April 1994. we undertook an I space mission. ambitious hybrid rocket research and development programme. The fundamental Since then. a proof-of-concept motor has been I objectives were to assess the feasibility of designed, fabricated and tested using ~5% using hybrid rockets on small satellites. high test hydrogen peroxide (HTP) as .. the oxidiser and polyethylene (PE) as the Iuel. This feasibility was to be assessed using three Figure 2 shows a cut-away drawing of the I basic criteria: motor attached to its support plate along with a photograph of the entire test rig. 1. Simplicity-defined in terms of ease of I design. fabrication and operations. In operation, the HTP is injected into a ") Peljormance-defined in terms of catalyst bed where it chemically decomposes measured vs. theoretical combustion into superheated (>600°C) oxygen and st¢am. I efficiency. These decomposition products are then vented Cost-defined in terms of development and into the combustion chamber ~here operational cost of a flight motor as compared spontaneous ignition of the PE takes place. I to other low-cost options.From these basic Onboard a satellite. the HTP offers the added objectives, we developed a research advantage that it can also be used as a mono­ programme that encompassed the following propellant with an Isp of around 140 seconds I tasks: for attitude control and minor orbit corrections. 1. Design-Starting from "scratch," establish I a hybrid development infrastructure and proof-of-concept motor. I

Figure 2: CUI-away drawing oj the University oj Surrey proof-oj-concept hybrid motor attached to I its support plate. Photograph shows test stand with support plumbing installed in test cell at Royal Ordnance test Jacility, Westcott. u.K.

I Catalyst Pack Load Cell Nozzle .~ I \

I -.:::'-"cj__ \ ~combustion I ~ Chamber I Test Stand Support Plate I 11 I Throughout the programme, we were able to 4. Efficient-as characterised by measured vs. gain access to various information which had theoretical chamber temperature. I been unavailable at the start which helped to 5. Minimum pressure drop-higher pressure steer the selection of HTP catalyst options. In drop requires higher delivery pressure. addition. as is often the case in applied Unfortunately. our research revealed that,;these I engineering research, adjustments were made characteristi~s are much easier stated:~ than to the test apparatus and instrumentation along achieved. The search for the "holy grail:' of a the way. Throughout, this was very much a reliable and effective catalyst pac~, has I learning process and we suffered a variety of consumed the bulk of our rocket development labour pains in the process of starting from and experimentation thus far. This work was nothing and evolving an entire rocket research somewhat frustrating because we knew that I and development infrastructure. reliable catalyst packs had been,-fully developed and widely used in the 1950s and The following is a list of just some of the 60s. However. when HTP fell out of favour in technical problems we had to overcome during the early 60s. this technology was abandoned. I this effort. Thus, during our research at times w~ felt • Obtaining a reliable supply of HTP more like archaeologists, trying to reconstruct I • Locating HTP-compatible valves an ancient art from a few vague and dusty • Designing a nozzle insert capable of scrolls, than a true rocket scientists. withstanding the +2300°C temperature of We started out by looking for any I hybrid combustion and finding a company commercially available HTP calalysts. to manufacture it. Unfortunately. none were available. • Selecting, procuring, integrating and Therefore, we had to develop our own. ,When I testing the hardware and software for data we began our system design, we had only a collection and experimental control. sketchy description of the catalyst packs used 89 • Designing, fabricating, installing 'and in earlier work . From these we were able to I testing a cavitating venturi to overcome determine some general guidelines for the catalyst pack "wet-start" problems. catalyst pack configuration and fix the basic Results pack geometry (diameter and depth). I Catalvst Pack Deyelopment What remained was finding a suitable catalyst material to use. Unfortunately, descriptjons of The hybrid combustion process actually catalyst material were even more vague. I begins in the catalyst pack. In fact, it is Davis8 describes using samarium oxide treated perhaps the w'eakest link in the chain. If the silver gauze. The Hydrogen Peroxide catalyst pack does not work, the hybrid will Handbo~k9 describes a variety of catalyst I not ignite. Thus. as part of over-all system, a types including calcium permanganate reliable catalyst pack needs to initiate smooth, "stones," silver gauze and silver-plated brass, rapid and complete decomposition of HTP nickel and stainless steel gauze, 16 - 20 ,mesh. I with minimum pressure drop across the pack. Four specific catalyst pack design parameters With only this to go on we started our emerge from this description: research. During the past year, we have I conducted over 50 catalyst pack firings. using 1. Reliable-relatively long lifetime with 8 different catalysts with additional variations. little degradation in performance, working .' "first-time, every-time." • 7),pe-J: Maximum density silver-plated I 2. Smooth-no radical surges or oscillations nickel gauze in \vorking pressure. • Type-2: Maximum surface area silver­ I 3. Rapid--decomposition time should be on plated nickel gauze the order of tenths of seconds. "Wet starts" • Type-3: KMn04 crystals should not be observed. • Type-2A: Type-2 catalyst. sintered Jor 30 I min at 800°C I 12 I Measured Steady-state Temperature vs. Theoretical Value for 5 different catalyst I conti gurati ons 700 I 680 660 ___ Type-2 640 __ Type-2A E 620 e: -.Ir- Type-4 I 600 il e --*- Type-7 .. __ Type-8 ~ 580 E ___ 100% ~ 560 ' I 540 520 500 ' I 0 5000 10000 15000 20000 25000 Cummulat"tive HTP Throughput (kglm2)

I " Figure 3: Summary of cata(vst pack test results. Graph shows the relationship between cumulative HTP throughput and average steady-state temperature for various catalyst options. . p I • 7)pe-.J: LCH 212 (an industry-standard thermochemistry computer programme - r hydrazine catalyst) . which assumes equilibrium flow). • -rlpe-5: Silver-plated gauze provided by ,J As Figure 3 indicates, the best overall catalyst the U.S. Air Force Academy (USAFA) we tested was the -rvpe-8 pure silver gauze. • Type-.JA: LCH 212 plus Shell 405 This is a 40-mesh gauze of99.9%+ pure silver I' • -rvpe-6: Permanganate stones provided by treated in a 5% solution of samarium nitrate the USAF Rocket Lab. Edwards AFB. (soaked for 25 minutes and then vacuum • -rvpe- '7: Silver-plated nickel gauze plated to dried). In practice we found that this very fine I a modified specification provided by mesh packed much more densely than the British Aerospace, Royal Ordnance!o._ silver-plated nickel gauze. As a result, we • -rlpe-8: Pure silver gauze treated with ended lip with a catalyst pack which was , samarium nitrate!!. effectively shorter. However, this did not noticeably reduce its performance. As of this For hybrid development work. we wer~ most writing, the same catalyst pack containing the interested in catalyst performance' and silver gauze has been used for 8 30-second I lifetime. Performance can be defined in terms firings at an HTP loading rate of ~ 75 of measured decomposition temperature v~, a kghn2/sec plus an additional 3 20-second theoretical value. Lifetime is defined by the hybrid firings at a similar load rate with no I total throughput of HTP for a given pack with appreciable decrease in performance. Longer units of kg/m::. term research will try to determine the Unfortunately. there is a practical limit to the effective lifetime of this catalyst I lifetime issue due to finite HTP resources. configuration . This was especially true early in the Hybrid Development programme when we only had 60 litres with 'which to experiment. Nevertheless. we were The. hybrid combustion process is actually able to characterise a definite trade-off quite similar to that of a log in a fireplace. • The log of course is the fuel. Oxygen from between performance and lifetime. This C'ln I be seen in Figure 3 which plots cumulative the air interacts with the fuel surface causing it . throughput for the most reactive cataly~ts vs. to vaporise and burn. If more oxygen is steady-state decomposition temperature. The added-the basic function of a bellows-then i theoretical value of 634°C is includ-e,d for combustion takes place more rapidly. Hybrid combustion theory attempts to describe and I comparison (derived using the USAF Isp 13 I predict this process within the combustion of the oxidiser which. in turn. determines the chamber. The simplest hybrid combustion rate of heat generated in the combustion zone. chamber is a solid rod of fuel with a In practice, this combustion or flame zone cylindrical port through the centre. Just like exists within the boundary layer above the the fireplace analogy. oxygen is added U.tJ- the solid fuel. Oxidiser enters the flame zone by form of a suitable oxidiser) into the port:and, diffusion from the centre of the port. Fuel in the presence of an ignitor (or with.pre­ enters as a result of vaporisation from the heated oxidiser) the fuel vaporises and starts .solid surface. to burn. Prediction of regression rate is fundamental to The principle metric used for discussing the hybrid rocket motor design. Unfortunately. hybrid combustion process is the rate at which the regression rate constants can only be the fuel burns. This speed of fuel determined empirically. Therefore, the consumption is called the regression rate. In primary goal of all basic hybrid rocket harmony with common sense, theory desii"ribes combustion research is determining the the regression rate as a function of the regression rate constants for a given propellant concentration of fuel and oxidiser. The combination. With these numbers in hand, the standard regression rate formula is no~ally designer can then tailor a hybrid motor to given as support specific mission objectives. The definitive paper which pre-dates our work . (\) on HTPfPE hybrids was published by George where Moore and Kurt Berman in the November 1956 issue of Jet Propulsion 13. Their paper r fuel regression rate (m Is) describes work conducted several years before (presumably kept secret until 1956) at the G total = total propellant (fuel plus oxidiser) General Electric Company as part of an Army 2 mass flux (kg I /11 s) Ordnance contract. Moore and Berman used X = distance down the port (m) the same HTPfPE combination we used to a. n. m = regression rate constants, (although they had access 90% HTP while we have to get by on 85%). At the outset, characteristic ofthe propellants they list several desirable characteristics of Hybrid combustion differs from liquid and this particular hybrid combination: solid combustion in that the process occurs as 1. Good theoretical specific impulse. Isp. a macroscopic diffusion flame. That is, combustion occurs in a region near the fuel "') High average density. surface where the fuel and oxidiser 3. Spontaneous ignition. usually after < 0.5 concentrations allow it. Typically, fue( is seconds .. In this way the fuel combustion concentrated near the solid fuel wall while can be thought of augmenting the HTP oxidiser is concentrated near the centre of the mono-propellant performance. "Ignition is port. The two mix through the diffusion generally reliable and smooth over a \vide process. While oxidiser to fuel ratio, OfF, is range of oxidiser/fueI ratios." basically constant for liquid and solid rockets, 4. With proper design of the engine and fuel I for hybrids this ratio gradually decreases charge "hard" starts should never occur: do\yn the length of the port. That is the reason the peroxide decomposition product is a regression rate is shown as a function of x in gm;eous oxidiser and therefore cannot Eqn. (I). accumulate in the combustion chamber prior to ignition. • Theory, backed up by experimental evidence, 5. Intermittent operation and throttling can be indicates that the rate of heat transfer to the I accomplished by means of a single valve in fuel and its heat of decomposition are the the peroxide line. controlling factors in combustion. The tate of M 6. The system has the simplicity of a mono­ heat transfer is controlled by the rate of flow I propellant but with a safety, at a given I 14 I performance, probably unattainable with We achieved our first successful hybrid firing I most liquid mono-propellants. in November 1994-just 7 months after our 7. The design and construction of the fuel programme began. Our experimental results charge is not very critical in regard to the are quite encouraging and compare favouabJy I presence of cracks or voids: there is verv with the ground-breaking work of Moore and little possibility of an explosion of th~ Berman as well as the preliminary results of entire propellant, and in these respects the Wernimont and Meyer. Table 4 gives specific I system is more desirable than most solid performance data for each test. Notice that propellants. characteristic exhaust velocities (C*. an effective measure of combustion efficiency) At the same time, thev noted onlv a few .. . exceed 87% and in most cases are greater than I disadvantages: 90%. Figure 4 plots the results our 6 most 1. Relatively high freezing point and potential successful hybrid tests conducted to-date (we I instability ofHTP. had several cases of failed ignition due to ') Difficult to vary the fuel burning rate by catalyst problems early in the programme) more than a factor two which can against those of Moore/Berman and I complicate the fuel charge design for some Wernimont/Meyer. Our data appears to agree applications. more closely with the classic Moore/Berman results with essentially the same slope to the The remainder of their paper outlines the curve. I results from over 300 tests! They used three basic chamber designs-straight tube, rod & Conclusions tube and straight tubes with alternating inside Our on-going investigation into hybrid rockets diameters to generate eddy currents for better I as a low-cost alternative for small satellites mixing. They used both extruded tube as well has proven the basic feasibility of the concept. as stacks of 112" plastic wafers (the fact that Overall. the hybrid system is safe and easy to the gaps between the wafers had no effect on I work with. Good performance is readily burning underscores how insensitive hybrids achievable. Estimates of development costs are to cracks or voids in the fuel grain unlike and manpower, based on our experience, solid motors). I indicates a proto-flight hybrid motor could be With the exception of some work published in ready for in-orbit demonstration in under 2 1970 by M. Pugibet and H. Moutetl4 on years. Much work remains. Future I additives to hybrid fuel to initiate hypergolic publications will expand on our experimental combustion with HTP, little was done with results and discuss the future objectives of this HTP/PE hybrids until 1994. In that vear, we research in greater detail. I began our work and E. Wernimon; and S. Meyer at Purdue University in the U.S. published some initial results from their I5 I \vork .

Table 4: Summary of tests results for University ofSurrey hybrid motor. I Ave, C*'hcory I 0/0 ... I Test I Burn Steady I Oxidiser Fuel I Total G ox Gtotal : Average IActual Time -state. mdot mdot mdot (kg/m 21 (kg/m2/i regression. OIF C* err, ~ : (sec) Pc (kg/s) (kg/s) (kgls) sec) sec) rate (10' meas,. I I (bar) i . m/s) 1 17.8 18.1 0.144 0.017 0.162 159.0 178.21 4.84 8.3 11482 1575 94% 2 16.9 10.2 0.155 0.015 0.170 243.3 267.11 5.04 10.2 1320 1522 87% I 3 13.2 18.9 0.162 0.0]8 0.180 319.8 354.91 6.62 9.1 1395 1554 90%

7 18.6 16.1 I 0.142 0.014 0.156 166.7 183.3 4.05 10.1 • 1372 1522 90% 8 18.6 17.5 0.145 0.161 ! 256.1 283.7 5.51 9.3 1447 1554 93% I I 0.016 9 18.2 18.2 0.143 I 6.58 6.6 1471 1568 94% I 0.022 • 0.165 223.1 I 256.7 I 15 I I Regression Rate Data from Moore/Berman (M/B) and WeinimontfMeyer (W/M) vs. Hybrid tests 1-3, 7-9 (SIP) I CJ 09 Qj til E 0.8 E- 0.7 I Qj C!I -+-M'B c:::- 0.6 t: 0.5 _W/M 0 I 'iii S/P til 0.4 -a- ...Qj en Qj 0.3 c::: Qj en 0.2 I f! Qj 0.1 >

Table 5 gives the basic specification for a Table 5: System parameters for a hybrid hybrid motor which could fulfil the motor to fulfil the UoSAT-12 missioll I performance requirements for the UoSA T -12 requirements. mission compared to the LEROS-20, 20- Newton bi-propellant engine discussed in the I Parameter Hybrid LER~ I next section. Notice that while the hybrid Option Opti rocket is slightly less efficient than the liquid Propellant(s) HTP/PE MMH/MON system, it has an advantage in density isp. OIF i 8 1.65 I I meaning you can package more deliverable Nozzle Expansion I 80:1 180: I Ratio impulse for a given volume, Figure 5 shows Nominal Thrust (N) 500 22 how such a system could be packaged within Specific Impulse I 280* 290 I the minisatellite structure. (sec) I Unfortunately, given the short time-line for Chamber Pressure I 18 8.88 the UoSAT-12 mission it was not feasible to (bar) I consider the hybrid as a realistic option. Total thrust time 97 sec 40 min (sec) Luckily, parallel research had revealed i .... another attractive low-cost option . Total propellant 17.6 17 I mass (kg) Total propellant 13.5 14.6 volume (/) I I Density lsp 364 337 *Assumes 90% of theoretical vacuum Isp achievable. I I I 16 I wavs to increase the reliability and pe;formance of the engine while slashing the production cost. By focusing on design and manufacturing techniques the Royal Ordnance team has been able to develop a production engine at very low cost. LEROS-20 Engine ~.. I Figure 5: Cut-away diagram showing possible configuration of a hybrid motor and support I tanks within the minisatellite structure. I Low-Thrust Bi-propellant Engin~s In parallel with longer-term research into the viability of hybrid rocket systems. we Figure 6: Diagram of LEROS-20 engine continued to look for other low-cost chi?mical 'I integrated onto the attach fitting plate of the propulsion system options. Traditional large UoSAT-J2 minisatellite. Propellant tanks are spacecraft use large bi-propellant engines shown on the other side of the attach plate. I (400+N) for major orbit correction. 'These engines come with a comparably large price­ Why are these engines a low~cost option for tag (>$200,000) as indicated above. Therefore, small satellites? To begin with, these small I we had to look for a less expensive option. We engines are produced in much larger numbers found one by taking the approach of adapting than the 400-N version, thus we were more mission requirements to fit an engine rather likely to be able to take advantage of existing I than the traditional approach which IS the production runs. In addition, because this is other way around. the same engine repeatedly tested for large expensive programs, direct evidence of basic British Aerospace, Royal Ordnance Rocket qualification was available without resort to Motors Division (RO) has a long hist~ry in I expensive, and in this case redundant, developing, testing and manufacturing bi­ qualification testing. As a result, 20-N engines propellant engines. RO. along with several are generally 75% less expensive than their other manufacturers, make an industry­ I larger, 400-N thrust big brothers. standard 20-Newton thrust engine burning MMH/MON. These engines are used on large On top of these basic savings, RO have been I satellite platforms for attitude control. Use on able to reduce the cost of these engines by a small spacecraft for primary propulsion nearlv an additional 50% through careful represents an innovative application of atten;ion to design and manufacturing issues. existing technology, one that would offer a For any liquid engine. the two most important relatively inexpensive engine that was already design areas that effect performance are the " supported by the industry. injection and the cooling. In the LEROS-20, I these" problems are handled together in a A diagram of the LEROS-20 engine is shown synergistic way that achieves good reliable in Figure 6 integrated into the satellite attach performance at low cost. The engine is film fitting base-plate. The engine itself is about I and radiation cooled yielding a simple, cost­ the size of a wine glass with a total mass of effective design. Furthermore, the film cooling 0.45 kg. Propellant mass flow rate is on the barrier allows the core combustion process to order of 7.5 gm/sec. This 20-Newton -;:thrust be optimised for maximum performance. engine represents years of research' and development at Royal Ordnance in finding

17 In addition, the LEROS-20 design Again, we focused on the cost drivers incorporates several features which reduce identified earlier. manufacturing costs: Mission Definition • Low total parts count-the entire thruster assembly has only 5 parts. ...;' During this phase we concentrated on understanding the "political" environment for • Small number of manufacturing the mission and how it would affect our operations-there are only three.'welds mission design. We recognised that any required on the entire thruster assemb'ly. technical solution would have to be acceptable • iVa special manufacturing proceses to the launch authority, At the outset we did requiring specialised tools or fixtures­ not know what launch vehicle we would be injector holes drilled with standard micro using so we chose Vandenberg AFB launch I drills rather than electro-discharge. '.' requirements as a baseline to work from • .lvfinimum number of hot fire tt}sts- knowing that these would be representative of extensive use of water flo\v to qualitatively those applied at other launch sites and I and quantitatively evaluate injector probably more constraining. This imposed performance., two important requirements that impacted our Finally, by working with RO on the' basic mission architecture: I procurement process for the engine we. have • At least 3 physical and electrical breaks devised a flexible set of requirements for between propellant tanks and the I acceptance testing which gives us confi'dence combustion chamber. in the engine's performance while reducing • At least 2 physical and electrical breaks the final cost an additional 10%. Working between high pressure gas and the outside. together we have devised the following I approach: Mission Design • Take maximum advantage of qualification Beginning with system specifications and , by similari(v-the LEROS-20 already has a margins, based on lessons learned in low-cost strong development heritage. This goes for spacecraft hardware design, ample mass the engine valves which also have a proven margins for the propulsion system was flight heritage. " allowed. Total deployed mass of the satellite was targeted to be < 250 kg with < 50 kg for • Reduce acceptance tests·-thruster and • the "wet" mass of the propulsion system. valve acceptance tests would include However, in every case, we were allowed to I leakage and functional tests . only. trade-off mass for cost. The importance of this Minimum number of hot-fire tests to verify mass flexibility can not be overstated. Having performance over blow-down range. the luxury of not having to count every gram 1 • Reduce performance requirements-accept had a significant knock-on effect which a nominal Isp of 290 sec vs. 293 sec. ultimately lowered the system cost. Operate over established engine blow­ Specifically, this allowed the use of: I down range of 20 to lObar inlet pressure (total performance loss -7%). • A "bang-bang" pressure control scheme in place of a much more expensive regulator. Best of aIL this is a long-term solution which II • Cold-gas thrusters and nitrogen control can be adapted to a variety of future missions, valves that could be selected based on For these reasons, the LEROS-20 was chosen a"q.ilable, cost and reliability rather than as the main engine for the UoSAT-12 mission. mass, I The UoSAT-12 Propulsion System • Tanks that could be manufactured using off-the-shelf material to a higher than I With the LEROS-20 selected as the 'main necessary operating pressure reducing cost, engine, \ve turned our attention to reducing the delivery time and the additional worry of cost of the rest of the system for UoSkT-12. ensuring pressure compliance with tight margins.

18 • Standard aerospace 37° flared fittings sufficient nitrogen to operate over the I instead of an all-welded construction. This LEROS-20 blow down range of 20 to lObar. also greatly simplifies system integration. Traditional designs would pressurise both • Standard stainless steel line (pressure rated tanks from a common source. requiring the I to >5000 psi) throughout the~ystem addition of redundant check valves upstream. instead of lighter (but much: more By decoupling the tanks upstream. we expensive) titanium line. eliminate the possibility of propellant leaking past check valves causing catastrophic I The complete system mass break-down in reaction. shown in Table 6. While the actuaj"-mass impact of our system engineering approach I easilty met our design goals, the overall system mass exceeds the earlier estimate Table 6: Mass breakdown/or UoSAT-12 (based on a more conventional apgroach propulsion system. I shown in Table 1) by about 38%. Ho!Vever, this increase represents only about 2%. of the Component 1 Qty. Unit Total I i deployed spacecraft mass. As we will see, this I Mass Mass (kg) , (kg) I overall approach has lead to a significa.nt cost savings making it well worth it for our type of Liquid Engine 1 0.75 0.75 low-cost programme. Propellant Tank 2 1.00 2.00 I Nitrogen Tank I 3.36 3.36 The complete system architecture is sh0wn in Accumulator/Ullage 6 0.5 3.0 Figure 7. The system includes several 'unique "Bang-bang" valves 2 0.9 1.8 features designed to make it simple, saf~, easy I Pressure Transducer 6 0.23 1.38 to operate and low-cost. ... Flow Restrictors -' 0.15 0.45 To begin with, the cold-gas system is Relief Valve 1 0.45 0.45 I completely separate from the liquid system Filter 3 0.23 0.69 (conventional systems would use a c6mmon FillfDrain Valve 5 0.15 0.75 reservoir of pressurant gas). Pressure Cold-Gas Thruster 8 0.15 1.2 I regulation is achieved using two "bang-bang" IUlbge tank 1,0 Valve 2 0.9 1.8 valves, controlled by feedback from a pressure transducer. This gives the added advantage of Total 17.63 I selectable pressure which allows us to vary the cold-gas thruster torque on-demand. . Simple. one-shot burst disks are used On the liquid side. each propellant tank is downstream of the tanks as an additional I individually pressurised at the beginning of isolator prior to tank pressurisation rather than the mission to an initial blow-down ptessure. 'more con1plex and expensive pyrotechnic No further tank pressurisation control or I va·lves. operations are needed. In addition. we evolved the following This architecture decouples the propellant redundancy philosophy which guided the tanks from the main N2 supply and includes I design of our propulsion system architecture: only enough nitrogen and ullage to initially pressurise the tanks to blow-down pressure. • Inc lude the minimum redundancy to I This has the advantage of not needing any ~pmply with range safety requirements. type of pressure regulation device for the • Keep it simple-the fewer things there are liquid system. to go wrong, the less redundancy you need. I • Do not put redundant components The two ullage tanks. operating at upstream of single-point failures. approximately 45 bar along with additional I ullage in each tank of around 2 litres, \ViII give I I 19 I

~-, i ' i ' 1Riii.:< -, I ~a"-~;:~::~':,;"'" :;:;c'I-;>"'t~;~

,,- I f'J:rj-FEt,--,"- T 'onduc-::r rT\'(:l,'e ,- I

:: i I I I

Figure 7: UoSAT-12 Propulsion System Architecture. I • Where possible, allow for degraded • Careful selection of components large functional redundancy whereby a single production runs (in some cases these may failure allows for continued, if somewhat already be in place to support larger, more I degraded, mission performance. formal programs). This philosophical approach was applied • No formal configuration management. I throughout propulsion system design, from the • No supplier data items (othe"i' than a circuit board level to the tanks and is also certificate of compliance). reflected in the overall mission design of • Supplier standard practices and pr.ocedures. I UoSAT-12. Together with Arde, we have. carefully evaluated the system requirements and relaxed them when possible to arrive at an effective t All our efforts at reducing the cost in the design solution that could be delivered on design would be for nought if we could not schedule. Low cost was the major driver in actually procure the components in a low-cost the selection of the final approach and specific I way, Fortunately, we were able to establish a components. To the greatest degree possible, working relationship with Arde, Inc, of we are using existing hardware from those Norwood, NJ, who, like us, were keenly suppliers with a known record for good quality I interested in cutting through the normal red­ products to reduce risk during grout:!d testing. tape found in space missions and focusing on and in flight. the essential elements to get a good product at I low cost. Working with Arde, we have • The systems requirements were '~valuated developed a procurement strategy that from the very beginning of the design phase to emphasises: allow the inclusion of selected key space­ grade components. These components would I • Flexible system specification. require little or no analytical effort to • Standard, proven designs for essential determine their structural and thermal modules with simple interfaces. compatibility to the space environment. The I• 20 I 1,1 history of each component was used to has made fundamental progress on hydrogen preclude unnecessary testing and its associated peroxide catalyst and regression rate cost. Actual component selection was made characterisation. We have proven their basic .,., using experienced engineering judgement simplicity and performance and identified key based on traditional space programs. development issues for hybrid rocket upper­ stages. In addition, we have identified a small, Simple and easily manufactured work low-cost bi-propellant engine-the LEROS- :, packages characterise the design of the 20-which can be easily adapted for a variety propulsion system. The entire system has of small satellite missions. Finally, our been broken into subassembly kits containing applied research focused on the propulsion all necessary tanks, valves, ullage/accumulator I system for the upcoming VoSAT-12 mission bottles, pressure transducers, filters, flow has given us practical experience in low-cost restrictors, burst disks and connectors. Each propulsion system design and support kit can be assembled and tested at the factory. component procurement, We hope the lessons ~ Spacecraft integration will require the addition learned on this mission can be adapted to of lines and mechanical brackets to hold the future small satellite missions, giving them a components in place. This overall approach is I low-cost enabling technology to take on a designed to support rapid system assembly and variety of bold, new missions. testing, reducing cost as well as programme risk. I Acknowledgements Conclusions The authors would like to thank: By applying a combination of low-cost t • Dr. Ron Humble of the University of spacecraft system engineering and procurement practices we have been able to Colorado, Colorado Springs for his advice , significantly reduce the cost of the propulsion and support as well as advanced copies of system for VoSAT-12. The total cost for all his excellent text. system hardware is targeted to be -$200,000. • Capt. Mike Lydon at the U.S. Air Force Less than half the cost estimated from our Academy, Mr. Richard Brown of Project market survey and only 15% of that predicted Machinery and Mr. David Andrews for • by typical cost models, their valuable assistance with hybrid rocket development. , Future Work • Arde, Inc. for their on-going involvement Much work remains 011 the VoSA T-12 in the UoSAT-12 project, especially Allan propUlsion system. A critical design review is Fleming for his contributions and advise on I planned for the end of September with Arde to this paper. review the overall system design and technical • British Aerospace, Royal Ordnance, Rocket , solution. Hardware delivery is planned for the Motors Division, for their on-goIng co­ end of January, 1996 with integration and operation on our research. testing to follow to meet the October 1996 References -I launch date. 1. Humble, R., et at Propulsion System Analysis Conclusions and Design, To be published, Fall 1995. 2. Barton, l, Olin Aerospace, phone conversation, Our on-going research into low-cost I r4 November 1994. propulsion systems for small satellites has so 3. Smith, P., Horton, M.A., "Advanced Propulsion far produced several important results. The Systems for Geostationary Spacecraft-Study important cost drivers for space hardware have t· Results," The Marconi Company, AIAA-84- been identified and used to guide our 1230, 1984 . investigation into system design and • procurement. Our basic hybrid rocket research I 21 I 4. Sellers, U., et aL Understanding Space: An Author Biographies Introduction to Astronautics. McGraw-Hill, I New York, N.Y.. 1994. Jerry Sellers is a captain in the U.S. Air Force 5. WRR 127-1 Range Safety Requirements, completing a Ph.D. at the Univerisity of Vandenberg, AFB, USAF. Surrey sponsored by the Air Force Instimie of ., 6. MIL-STD-1522A Standard General Technology. After graduating from the USAF Requirements for Safe Design and Operation of Academy in 1984, Capt. Sellers worked ·in the Pressurised Missile and Space Systems. NASA Johnson Space Center in Space Sbuttle f 7. Goldberg, RE., Wiley. D.R., "Hybrids: Best of Mission Control until 1988. Following Both Worlds," Aerospace America, June 1991. completiong of a Master's Degree at StaN-ford 8. Davis, N.S., jr., McCormick, le.. "Design of University in 1990, he taught in· the Catalyst Packs for the Decomposition of Department of Astronautics at the ttSAF I Hydrogen Peroxide," Liquid Rockets and Academy until beginning his p~esent Propellants, Bollinger, L.E., Goldsmith, M., & assigment in 1993. Lemmon. A.W., jr, (eds.), A Symposium of the ~ American Rocket Society, Ohio State Malcolm Paul began his rocketry career in University, Columbus, Ohio, 18-19 July, 1960. 1963, Jommg the Rocket PropUlsion 9. Hydrogen Peroxide Handbook, Chemical and Establishment, Westcott, U.K., as a liquid ·1 Material Sciences Department. Research engines trials engineer. He worked on ·many Division, Rocketdyne-a division of North liquid engine projects from the Blue Streak American A viation. Inc., Canoga Park, California, Technical Report AFRPL-TR-67- launch vehicle to the LEROS apogee engine. I 144, July 1967. Along the way, he specialised in propulsion ]0. Walder, H .. Spalding, E.G., "Influence of HTP system ground support equipment. He left stabiliser (sodium stannate) on the silver Westcott in 1993 and has since workecj as a catalyst," Technical Memorandum No. RPD 68, consultant to Raufoss Technology AJS, Rocket PropUlsion Department, Westcott, Norway. He began working for SSTL on • January, 1955. hybrid rocket research and minisatellite f 11. Runckel, IF., Willis, e.M., Salters, L.Bjr, propulsion in 1994. "Investigation of Catalyst Beds for 98-Percent­ Concentration Hydrogen Peroxide," NASA Martaan Meerman is a Research Fellow at the Technical Note. D-1808, Langley Research University of Surrey working as a Principle Center, Hampton, Va .. June 1963. Engineer for Surrey Satellite Technology. He • 12. Selph, e., Isp Computer Code, Provided by Dr. has primary responsibility for spacecraft Ron Humble, U.S. Air Force Academy, 1992. mechanical systems and oversees propulsion t 13. Moore, G. E., Berman, K, "A Solid-Liquid system research and development. . He Rocket Propellant System," Jet Propulsion. completed a B.Sc. from Amsterdam November, 1956. Polytechnic in 1988 and has worked at the I 14. Pugibet, M., Moutet, H. "On the Use of University of Surrey in areas of increasing Hydrogen Peroxide as Oxidizer in Hybrid responsibility since then. Systems," NASA Technical Translation, NASA , TT F-13034, May 1970. Robert Wood is the Senior Progr~mme 15. Wernimont, EJ., Meyer, S.E., "Hydrogen Manager of the liquid engine business at Royal Peroxide Hybrid Rocket Engine Performance Ordnance Rocket Motors Division (a division I Investigation," AIAA 94-3147, 30th Joint of British Aerospace Defence. Ltd. ).He is Propulsion Conference, Indianapolis, IN. June responsible for the design and development of 1994. a range of monopropellant and bi-propellant I 16. Sellers. J.J., Meerman, M., Paul, M .. Sweeting, spacecraft thrusters and apogee engines. Prior M., "A Low-Cost Propulsion Option for Small to joining Royal Ordnance in 1987, he worked Satellites," Journal of the British Interplanetary as a development engineer at Pratt & Whitney, Society, Vol. 48, pp. 129-138, March, 1995. t Canada, engaged in turbine development for small turboprop engines. 22 •I t