- Email: [email protected]
The impact of advanced platform and ion propulsion technologies on small, low-cost interplanetary spacecraft
Acta Astronautica 59 (2006) 899 – 910 www.elsevier.com/locate/actaastro
The impact of advanced platform and ion propulsion technologies on small, low-cost interplanetary spacecraft Stephen D. Clarka,∗ , David G. Fearnb a Space Department, Y72 Building, QinetiQ, Farnborough, Hants, GU14 0LX, UK b EP Solutions, 23 Bowenhurst Road, Church Crookham, Fleet, Hants, GU52 6HS, UK
Available online 19 September 2005
Abstract This paper assesses the application of small ion-propelled spacecraft of about 200 kg launch mass to low-cost interplanetary missions, for which very high velocity increments (V ) are required. Here, the “low-cost” criterion eliminates the utilisation of the launcher for the attainment of escape velocity, thereby implying the use of an ion propulsion system for orbit-raising from an initial low Earth orbit or from the geostationary transfer orbit. The sample missions considered were to rendezvous with a wide range of asteroids; they require a total V of the order of 10–12 km/s. Although the performance of the solar array is critical, recent advances in solar cell technology have enabled considerable progress to be made in this area. Other relevant advances include improved batteries, attitude and orbit control sensors, and communication systems. Gridded ion thrusters are essential, owing to the need for very high values of specific impulse. They allow a very large number of target bodies to be accessible within a launch mass of 200–250 kg. © 2005 Elsevier Ltd. All rights reserved.
1. Introduction Deep space missions with significant scientific objectives usually require a very large velocity increment V ). Although some of this can often be provided by planetary swing-bys, until recently, the only options available to achieve this were to launch on a very large and costly rocket or to supply the necessary V by use of an on-board propulsion system. The latter approach led to a large propellant load, and thus ∗ Corresponding author.
E-mail addresses: [email protected] (S.D. Clark), [email protected] (D.G. Fearn). 0094-5765/$ - see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.actaastro.2005.07.022
to a heavy spacecraft and the need for an expensive launcher. The advent of the operational ion propulsion system (IPS), as demonstrated so convincingly by the Deep Space 1 (DS-1) [1] mission, has permitted a complete re-evaluation of this situation. Moreover, a more ambitious mission is now underway; this is Muses-C [2], which is to take a sample from an asteroid and return it to Earth for laboratory examination. Owing to the high specific impulse (SI) available from an IPS, which can be a factor of 10–20 greater than that from a conventional chemical propulsion system, the propellant load required in the second option mentioned above can be reduced by the same factor. This very large gain can be further enhanced by
900
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
utilising an even less costly launch vehicle to deploy the spacecraft initially into a low Earth orbit (LEO) or to geostationary transfer orbit (GTO). Earth escape can then be achieved via a spiral orbit-raising manoeuvre, as recently undertaken by the Artemis communications satellite [3]. A similar transfer from GTO will soon be demonstrated during the SMART-1 [4] lunar mission. The potential mass reductions and associated financial gains available from the use of an IPS has led to several studies which have explored the minimum spacecraft mass that could provide a viable deep space mission [5–8]. Such an objective is assisted substantially by significant advances in other technological fields [9], notably lightweight solar arrays and batteries and high temperature semiconductors. This progress recently led to the conclusion [5] that a 10 km/s mission to an asteroid could be launched as an Ariane 5 auxiliary payload into GTO, which has a mass limit of 120 kg. This mission would commence with an orbit-raising manoeuvre to escape from the Earth’s gravitational field, but could not accommodate redundancy. This earlier work has thus been extended to examine the impact of increasing the launch mass to about 200 kg, with the aim of carrying redundant thrusters and a larger payload, whilst simultaneously raising the total velocity increment available.
2. Asteroid missions There has, for many years, been great interest in deep space missions, heightened by the need to understand better the origin and evolution of the solar system, and also by the possibility of finding evidence of past or present life elsewhere. This interest extends to the asteroids and comets orbiting the Sun, mainly because they are considered to be remnants of the material from which the solar system was formed. However, a major problem associated with missions to these bodies is the substantial cost involved. This is due largely to the very high V required for escape from the Earth’s gravitational field and then to reach the target in a reasonable time. A rendezvous with the target adds further to this requirement, which totals many km/s. The initial objective of the present study was to select suitable targets, taking into account the need to minimise launch cost. Although a very large V
was deemed mandatory, preliminary calculations suggested that 12–14 km/s might be achieved by a 200 kg spacecraft carrying a reasonable payload. A list of accessible asteroids [10] was examined, together with their orbital parameters and the values of V required to rendezvous with them following Earth escape. These were between 4.5 and 8 km/s, so are feasible using the spacecraft envisaged in this study. Indeed, it was found to be possible to target up to three such bodies in a single mission. Three asteroids can be visited for a total V of below 10 km/s in at least 14 combinations and all multiple-targets studied require less than 12 km/s. From this examination of possible targets, it was clear that most represent desirable scientific objectives. It was also evident that special expertise and extensive discussion within the science community will be required to select the asteroids giving the best scientific return. Since none of them require a total value of V exceeding 12 km/s, they are all accessible and there was thus no need to be specific at this stage.
3. Propulsion requirements 3.1. Primary propulsion A preliminary assessment indicated that it is necessary to achieve values of SI of the order of 5000 s to realise the above objectives. Thus, gridded ion thrusters are essential, since no other developed technologies provide an adequate SI, coupled with a thrust of the order of tens of mN and long life. While the absolute performance of such thrusters has almost reached a plateau, significant improvements have been reported in life expectancy, partly through the use of carbon ion extraction grids and also through the attainment of higher values of propellant utilisation efficiency. As regards the overall mass of an IPS, important advances are being made in the propellant feed and power conditioning fields. In the former, there have been moves towards microminiaturisation, although the early promise of micro-electronic mechanical systems (MEMS) has not yet been realised. In the latter, SiC semiconductor technology will permit hightemperature operation of power conditioning circuits, allowing thermal design constraints to be relaxed and leading to substantial mass and volume reductions.
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Table 1 M and thrusting time as functions of SI, with two T5 thrusters providing a V of 12 km/s SI (s)
M (kg)
Thrusting time (h)
3000 4000 5000 6000
67.0 52.7 43.4 36.9
11,445 11,491 11,831 12,059
901
Thus, the primary propulsion requirements assume that the total V needed is 12 km/s and that two T5 thrusters operate in parallel at an SI of 5000 s. The propellant mass is then 43.4 kg and the thrusting time 12,000 h. 3.2. Secondary propulsion
Table 2 Typical operating parameters of a T5 ion engine at 25 mN thrust and a SI of 3500 s Thruster parameter
Value
Thrust Ion beam current Beam accelerating potential Beam power Ion velocity Total mass flow rate Neutraliser mass flow rate Specific impulse Propellant utilisation efficiency Anode current Solenoid power Keeper discharge power Maximum discharge power Neutraliser discharge power Maximum total power Electrical efficiency Total efficiency Power to thrust ratio
25 mN 457 mA 1100 V 503 W 40.2 km/s 0.725 mg/s 0.04 mg/s 3515 s 85.7% 3A <3W < 15 W 138 W < 16 W 657 W 76.5% 65.6% 26.3 W/mN
The “rocket equation” shows that reasonable values of the propellant mass required, M, can be attained only if the SI is large; within limits, the higher the value achieved the better. This is illustrated by the data in Table 1, which assumes an initial deployed spacecraft mass of 200 kg and the use of two T5 ion thrusters [11] operating at 25 mN to produce a V of 12 km/s. For reference purposes, typical T5 thruster parameters are given in Table 2 at an SI of 3500 s. The estimated thruster lifetime is in excess of 10,000 h with molybdenum grids, so all values of SI listed are probably achievable employing this technology. However, an improvement by a factor of 3–5 is immediately available using carbon.
It is likely that the optimum method of achieving three-axis attitude control will involve a combination of thrust vectoring of the IPS and control moment gyros (CMGs). The effectiveness of the former technique was demonstrated during the DS-1 mission [1], for which the NSTAR ion thruster was mounted on a gimbal platform. The CMGs will be mainly exercised when the IPS is not operating, but will also be required continuously for control about the roll axis. The main thrusters will also be suitable for momentum dumping about two axes after periods when no thrust is required. Auxiliary thrusters will be needed for momentum dumping about the roll axis and for attitude control when the IPS is inactive. It is recommended that this capability be provided by the hollow cathode arcjet (HCA) [12], small field-emission electric propulsion (FEEP) devices [13], or miniature colloid thrusters [14]. Of these, the HCA is the currently preferred solution, despite its rather low predicted SI of perhaps 500–600 s. This is because of its commonality with the primary propulsion system; it is able to share the same propellant, tank, regulator, and power supplies. Thus, the attitude control system (ACS) requires four CMGs arranged in a tetrahedron configuration and providing redundancy against the failure of one wheel. Momentum offloading needs eight HCAs, although degraded control could be maintained with six, but this would incur significant additional fuel and reorientation manoeuvres. Taking the T5 cathode as an example, this has a mass of 70 g, to which must be added flow control valves at 100 g each, mounting brackets, pipework, and other items, giving a total of 2 kg. Although it might be assumed that the thruster power conditioning units (PCUs) could be used to operate the HCAs, this will only be possible if the thrusters are not in use. Thus, an allowance must be added for power units for two HCAs to operate simultaneously, plus a switch system to interconnect the
902
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
eight. These supplies will have a mass of about 0.5 kg. If the switch system weighs 0.3 kg, the total is 1.3 kg. The pointing stability required will depend on the payload selected. However, at a range of 10 km from the target body, an estimate of 200 mrad/s is considered adequate. Commercially available CMGs of this standard weigh of the order of 1–3 kg; the mean figure of 2 kg is adopted in this case. This corresponds well with the 2 kg system flown on Clementine [15].
4. Solar cells and arrays In all studies, the performance of the solar array is critical to the achievement of a significant payload mass. In addition to the power generated per unit mass and per unit area, which can now reach 125 W/kg and 250 W/m2 , the degradation during transit through the Earth’s radiation belts is of concern. However, recent progress has enabled the degradation during transfer from GTO to escape to be reduced to less than 5%. 4.1. Solar cells There has been considerable success in improving the performance of solar cells and thus of complete arrays. For example, certain high-eta Si cells have initial efficiencies ranging from 15.5% to 17.3%, depending on whether the cell is optimised for beginning-of-life (BOL) efficiency, radiation resistance or cell mass. The main advantages of these cells over other candidates are that they are lighter, with approximately one-third of the mass and less than 50% of the cost. GaAs/Ge cells, which are available from commercial sources, have a typical BOL efficiency of 19%, with superior radiation resistance compared to Si cells. The temperature coefficients are also better, so that there is less reduction of efficiency with increasing temperature. However, dual and triple junction (TJ) cells are the current state-of-the-art devices, although the former are already being superseded by the latter. Dual junction cells are typically 24–25% efficient, and the TJ types reach 27–28% and are forecast to be close to 30% in the next few years. They all have a superior radiation resistance compared to GaAs/Ge technology. These cells have a relatively high operating voltage, which simplifies array construction; they
Table 3 Performance of the X4 solar array with modern solar cells Cell type
Cell efficiency Mass (kg) Watts (W) Watts/kg
Original X4 7.7 High eta 16 GaAs/Ge 19 Triple junction 28
6.25 6.85 8.04 8.26
310 644 724 1127
49.6 94.0 90.0 136.4
provide 2.2 V, as compared to 0.85 and 0.45 V for GaAs/Ge and Si, respectively. 4.2. Solar arrays Rigid panel arrays are the most common type in use. Carbon-fibre skinned aluminium panels are hinged together and deployed out from the sides of the spacecraft. Although most applications use two wings, some employ one wing. The mass per unit area is rather high, so a flexible blanket array is more attractive for the present application. This consists of a reinforced Kapton blanket, supporting the solar cells and their interconnections, which is folded up, concertina fashion, into a box and then deployed using an appropriate boom or mast. In between each fold is a “leaf” of Kapton to protect the cells. An early example of this type was the RAE (Royal Aerospace Establishment) array [16] used on the Miranda (X4) spacecraft flown in 1974. The Miranda array produced 310 W from 7440 2 cm × 2 cm Si cells with an efficiency of 7.7% for a mass of 6.25 kg, i.e. about 50 W/kg. If modern cells are transposed onto this array, the resulting performance figures are shown in Table 3. These improvements are very significant indeed. Another method for enhancing the power from an array is to increase the light intensity on the cells, using a concentrator technique. Several variants have been proposed, including cassegrainian reflectors, reflective troughs and fresnel lenses, although the simplest is the trough concentrator. This gives typically a 50% rise in power for 33% less cells, compared to a planar array. However, all concentrators require more stringent pointing accuracy in one axis compared to planar arrays. This leads to important ACS requirements. A more rigid assembly than the flexible blanket type is provided by the hybrid array. This usually consists of
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
903
Panel current (A)
1.05 1 Predicted Degradation
0.95 0.9 1 year Pm dose 2.1E14
0.85
Fig. 1. Partially deployed QinetiQ hybrid array. 0.8 0
100
200
300 400
500 600
700 800
900
Days since launch
blankets or reinforced skins in a carbon fibre frame, allowing the use of simpler deployment methods, while minimising the amount of structure required and the mass. As an example, QinetiQ have developed a hybrid array specifically aimed at small satellite applications. The current design provides 1–2 kW, will fit onto a 120 kg minisatellite, and is depicted in Fig. 1. The frame is constructed from carbon fibre reinforced tubing joined with moulded corner connectors. The cells are supported on a stretched membrane, with the panel interconnects attached to the rear of the membrane.
Fig. 2. Degradation of STRV-1b GaAs/Ge solar panel in GTO.
Table 4 Comparison between solar cell technologies after 15 years in GEO Cell type
Normalised EOL power
High-eta silicon GaAs/Ge Dual junction Triple junction
0.66 0.80 1.00 1.16
4.4. Summary and recommendations 4.3. Radiation considerations For orbit-raising through the Earth’s radiation belts, the radiation dose received by the array is the main design driver. This will depend upon the initial orbit and how quickly the radiation belts can be traversed. Ideally, this transfer requires radiation hard technology and a thick coverglass to shield the cell, thus a TJ type is preferable. For example, the degradation of GaAs/Ge cells using 500 m coverglasses, flown on the QinetiQ STRV1-b satellite [17] in GTO, is illustrated in Fig. 2. This predicts a degradation of below 7% for a transfer from GTO to Earth escape in 1 year, and it should be noted that TJ cells will provide a better performance than this. To illustrate the long-term performance of solar arrays in a radiation environment, Table 4 compares the relative performances of various cell types after a geostationary Earth orbit (GEO) mission of 15 years duration, where EOL is the end of life. This shows that a high-eta array would be 75% larger than the TJ equivalent if it had to provide the same EOL power.
The above discussion leads to the array parameters summarised in Table 5. From this it is clear that TJ cells provide major advantages and that they should be utilised unless strong reasons are available to dictate otherwise. Similarly, the flexible array configuration provides significant benefits and was therefore selected for the interplanetary spacecraft discussed later.
5. Other technologies 5.1. Power electronics The problem of reducing the mass and volume of the PCU of the IPS, and of other electronics systems, can be distilled to the single issue of heat rejection. If a system could be produced where the dissipation was lower, the operating temperature was higher, and the thermal conductance to the spacecraft was better, major improvements could be realised.
904
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
Table 5 Summary of solar array characteristics and recommendations Array type
Cell type
Efficiency
Mass/area (kg/m2 )
Power/mass (W/kg)
Power/area (W/m2 )
Muses-C (EOL data) Upgraded X4/X5 flexible type
TJ GaAs/Ge TJ GaInP2 /GaAs/Ge TJ TJ TJ TJ TJ
– 19 28 22.5 25 25 25 28 28
4.45 2.01 1.96 4.44 3.00 1.64 6.7–5.0 3.00 2.00
53 90 136 45 77 100 45–60 83 125
236 181 267 200 230 164 300 250 250
Concentrator (DS-1) QinetiQ hybrid QinetiQ X4 type QinetiQ hybrid; prediction Future performance—hybrid Future performance—flexible
The temperature of any electronic system is limited to that of the maximum safe operating temperature of the semiconductor components. This is currently determined by the upper junction temperature of typical Si-based semiconductors, which is 110 ◦ C. In analysing this situation, it is assumed that increasing the operating temperature of the PCU would allow the desired reduction in the physical size of the unit, if all other aspects are kept constant. As the power is largely dissipated within the junction, doubling this limit from ∼ 100 to 200 ◦ C would double the power density, leading to either half the mass, half the volume or twice the power capability for the same power conversion efficiency and conductive thermal resistance. The recent availability of SiC diodes has demonstrated the advantages of SiC over Si for power converter applications [18]. This technology, in the form of diodes and switching transistors, yields savings of 50% and more in the losses in power converter systems [19], and can therefore provide the benefits discussed above. A further advantage is the ability of SiC to operate at high switching frequencies without performance degradation. It also has very good radiation tolerance. It should also be noted that these advanced power electronics concepts are equally applicable to other associated systems, including battery charge and discharge regulators, and solar array shunt regulators.
5.2. Batteries Although Ni-H2 batteries were, until recently, regarded as the best available technology, Li-ion batter-
ies are now much superior [20], and their use was assumed in the present study. Bearing in mind that only a modest number of discharge cycles will be needed in the orbit-raising phase of the mission, it has been assumed that the depth of discharge (DoD) will be limited to 80%, giving a performance of about 150 Wh/kg. At this DoD, the limit is 30,000 cycles. 5.3. Data handling and storage Ideally, a single data handling unit (DHU) is required to perform all of the spacecraft and payload functions, i.e. housekeeping, autonomous navigation and control, operation of the IPS and of the payload. In general, the ERC-32 (Sparc) processor is likely to be the best candidate to meet these requirements, as it is commercially available in a radiation-tolerant form. An additional digital signal processor (DSP) could be accommodated for the data processing of the payload. The main design driver is the radiation dose expected during the mission. The aggregate mass of the onboard computer (OBC) are given in Table 6 for a variety of memory capacities; the option in the last line of the table was selected for this study. 5.4. Propellant feed system Provided that the problems which face MEMS systems in new applications can be successfully overcome, there is no reason why selected propellant feed components could not benefit from these technologies in the near future. The limiting factor is the technical performance achievable when compared to operational requirements. Clearly, the benefits include
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
905
Table 6 Mass of OBC as a function of shielding thickness and memory capacity Shielding thickness (mm)
Mass memory (Mbits)
Computer mass (g)
Mass excluding box (g)
Total mass (g)
4 4 7 7
500 1000 500 1000
477 477 477 477
645 674 645 674
1552 1609 2321 2400
reduced mass and volume and, eventually, an improvement in reliability. In the longer term, a significant additional advantage is the degree of component integration that can be achieved into a single device, allowing production of a complete system “on a chip” or the inclusion of multiple redundancy with little extra mass or volume penalty. Two possible examples at opposite ends of the integration spectrum include a single microvalve chip, containing ten or more identical valves, and a complete propellant feed system integrated at wafer level into a single unit. However, it was not felt that MEMS capabilities are sufficiently developed for their use to be assumed in the present study. 5.5. Communications The frequency bands of choice are either Ka- or X-band and, to minimise cost, it is assumed that 12 m diameter ground antennas will be used. As a baseline, the system was flown very successfully on DS-1 [1] operated in Ka-band and had a mass of 3 kg. Allowing 5 kg for the high-gain antenna (HGA) and cabling, the total would be about 8 kg. However, at X-band the latest power module technology could be used; these are smaller and lighter than those of the Ka-band system. Thus, an X-band system was selected, with a 0.5 m diameter HGA reflector, a horn feed, a continuously powered low-noise amplifier, an X-band receiver, and a transmitter. Small omni-directional antennas are also provided to allow for emergency access to the spacecraft and for low-rate housekeeping data to be downloaded. It is estimated that this system, operating at 8.5 GHz, will provide a data rate of 600 bps at a distance of 2 AU from the Earth. The assumed 9.6 dB margin (equivalent to a bit error rate of 1 : 105 ) is exceeded by 1.7 db by the use of standard coding techniques. The
beamwidth of the HGA will be less than 1◦ and will therefore require accurate Earth-pointing. The transmitter RF output of 180 W requires about 400 W of DC power. Since the antenna is assumed to be fixed to the spacecraft, so that the complete vehicle must be orientated correctly during communications sessions, a constraint on the mission is that the transmitter is not operated whilst thrusting. A mass of 7 kg was assumed for the complete system. 5.6. Attitude and orbit determination Attitude determination during orbit-raising will utilise digital sun and Earth sensors, based on MEMS technology, augmented as necessary by data from a star sensor/tracker. A number of MEMS devices currently in development weigh only 10’s of grammes each, so a total allocation of 1 kg for sun and Earth sensors is generous. The star tracker can be used during both orbit-raising and the cruise phase and provides an essential autonomous navigation capability. This is vital owing to the large distances from Earth and the need to minimise ground segment costs. Excellent results from DS-1 [1] suggest that this concept is entirely feasible. Indeed, this mission provided en-route accuracy of better than 400 km and 0.2 m/s, which improved to better than 3 km at the encounter. The star sensor will be aligned along the solar array axis, which places the sun at 90◦ , avoiding blinding. This allows the device to be small and reduces the requirement for a baffle around the head. The star tracker on Clementine [15] weighed only 0.6 kg, and consumed 4.5 W; this is suitable in the present case, although an allocation of 1.5 kg provides redundancy. Orbit determination whilst in Earth orbit will be by standard spacecraft ranging through the transponder. This is likely to be augmented by use of a GPS system, although the utility of this in and beyond GTO has yet
906
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
to be demonstrated. As commercial units for terrestrial use are available with a mass of a few hundred grammes, 1 kg for this would appear to be reasonable. 5.7. Structure, launch adaptor, thermal control and harness The assumed structural configuration employs both carbon fibre reinforced plastic struts and honeycomb panels. A single toroidal filament-wound fuel tank is situated around a central thrust/torsion cylinder, which also houses the ion thrusters and their flow control units. Subsystems and payload units are attached to floors bonded above the tank, and external shear walls carry the loads through to the launcher interface ring. A preliminary assessment suggests that 12% of the launch mass is a reasonable allocation for this. To this must be added about 3% to allow for the components of the launch adaptor fixed to the spacecraft. Removal of heat from both the communications equipment and each ion thruster PCU will require additional mass. A heat shield will also be required between the ion thrusters and the adjacent spacecraft structure. An additional 3% of the launch mass (6 kg), is allocated for these purposes. The harness is likely to require about 1% of the launch mass, plus the separate allocation of 2 kg to cover the parts of the harness connecting the thrusters, switch system and PCUs. 5.8. Spacecraft autonomy To minimise ground segment costs, the omnidirectional low-gain antenna on the spacecraft radiates continuously at one of four frequencies. These indicate the health of the spacecraft, as determined by the OBC, and the need for intervention from the ground, and the urgency with which action is required. This is very effective in reducing ground segment costs, and it is recommended that a similar approach to autonomy is considered for most future missions.
together with a variety of US piggy-back alternatives. In addition, a viable possibility is a launch to GTO as an Ariane 5 auxiliary payload within the main thrust cone, for which the mass limit is 350 kg. However, many of these vehicles could launch 200 kg to a much higher altitude. For example, the Rockot could provide an initial circular altitude of several 10,000 km and probably an escape capability. Thus, a very wide spectrum of performance is available, ranging from LEO to attaining more than escape velocity. So the total requirements for the IPS vary from more than 10 km/s to a few km/s, depending upon the target. As one objective of this mission is to demonstrate the capabilities of ion propulsion, it is appropriate to choose a modest launcher, thereby also minimising cost. If Vega is disregarded owing to its relatively high performance, the remaining European option is an Ariane 5 auxiliary payload launch into GTO, which requires 4 km/s to achieve escape velocity.
7. Payload An idea of the type of instrument which would be suitable for an asteroid mission can be obtained from the NEAR spacecraft [21], which acquired extensive data from and eventually landed on Eros. Relevant instruments were also carried by DS-1 [1] and the Clementine mission to the moon [15]. In the latter case the payload consisted of two star tracker cameras, other cameras operating in the ultra-violet (UV)/visible, near infra-red (IR) and long-wave IR regions, a high-resolution camera, and a laser transmitter. The total mass was 7.37 kg and the maximum power consumption 97.3 W. More ambitious is the Muses-C asteroid sample return mission [2], and relevant future missions include ESA’s SMART-1 lunar spacecraft [4]. Additional capabilities can be provided by combinations of individual instruments and by the communication systems. A realistic aim for the total mass of the payload is 15 kg, as indicated in Table 7.
6. Launch options To achieve minimum cost, an attractive option is a launch to LEO using a relatively inexpensive vehicle, such as the Pegasus-XL, Taurus, Rockot and Vega. Several other low-cost Russian options also exist,
8. Mission profile In this mission, the dominating constraints are the initial mass of 200 kg and the use of an Ariane
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Table 7 List of possible payload instruments, with masses and power consumptions Experiment
Origin
Multispectral camera X-ray spectrometer Radio science Magnetometer IR spectrometer Laser rangefinder Penetrator
NEAR SMART-1 SMART-1 NEAR SMART-1 Muses-C –
Total
Mass (kg)
Power (W)
5 3 0 1 2 2 2
10 5 0 1.5 2 17 5
15
40.5
50
Propellant Mass (kg)
From 300 km From 2000 km
40
From GTO
30
20
10 3000
3500
4000
4500
5000
5500
Specific Impulse (s) Fig. 3. Propellant mass as a function of SI for the orbit-raising phase of the mission.
5 auxiliary payload launch to GTO. Target selection, which is best left to later scientific assessments, is represented by a worst case V of 8 km/s. To this must be added the 4 km/s needed for Earth escape, giving a total of 12 km/s. It is proposed that the orbit-raising phase be based on thrusting over a wide angle about the apogee to raise the perigee to geostationary altitude, followed by continuous thrusting to achieve escape. For comparison purposes, the Pegasus-XL, SHTIL2/Volna and Ariane 5 launch options were analysed to establish what can be accomplished using small ion thrusters for orbit-raising. The altitudes to which the first two will reach were assumed to be 2000 and 300 km, respectively. The results (Fig. 3) assume that T5 ion thrusters [11] operating at a thrust of 25 mN are employed. The range of SI covered was determined by limiting the ion accelerating potential to 2.3 kV, which is representative of present grid systems.
907
All three options are viable, although the launches into LEO require the greatest M. For the GTO launch, there is a mass advantage of about 8 kg in going from an SI of 3200–5000 s. Assuming the use of carbon grid technology, the operating times from GTO are acceptable, even if only one thruster is operated. However, it is likely, for single thruster operation, that a spare will be needed to achieve the total impulse required, with a switching system to permit a single PCU to be connected to either device. In general, an optimum SI can be derived by adding the total mass of the propulsion and power systems, since a change in SI at constant thrust alters both M and the power needed by the thrusters. The Earth escape manoeuvre will be followed by the interplanetary trajectory, represented by the V required and the maximum distance from the sun. Thrusting with the IPS to maximum velocity (to minimise trip time) will then be followed by a retardation phase, again using the IPS, with the aim of a rendezvous with the target. The V required is between 4.6 and 8.0 km/s, giving a total of 8.6–12.0 km/s, including the initial orbit-raising manoeuvre. Bearing in mind the low gravitational force exerted by the target, the thrust required for capture into orbit is well within the capabilities of the IPS. A more significant challenge involves terminal navigation, since the target will almost certainly have a low albedo. It may therefore be difficult to acquire it visually at a large distance, so a very slow approach, with the possibility of needing to thrust perpendicular to the velocity vector, may be necessary. However, the value of V required is negligible compared to 12 km/s. As the power consumption increases with SI, the mass of the solar array rises as M falls, and there will normally be an optimum value of SI at which the total mass is a minimum. To determine this minimum, performance data for a flexible array using TJ cells were adopted; these are 2 kg/m2 , 125 W/kg, and 250 W/m2 (Table 5). An allowance was also made for degradation due to the Earth’s radiation belts during orbit-raising. Using two thrusters, the transfer time is about 6 months and the worst case degradation is about 3% (Fig. 2), assuming that the whole of this period is spent in GTO; a value of 5% was taken to provide a suitable margin. Approximate analytical expressions were derived for the power consumed by the thrusters as a function
908
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Total, with margin Total
100
Propellant
Mass (kg)
Array and PCU
80
60
40
20 2000
3000
4000
5000
6000
Specific Impulse (s)
To summarise, the total velocity increment needed is 12 km/s and two T5 thrusters are used in parallel, operating at an SI of 5000 s and a total thrust of 50 mN. M is then 47.7 kg if a margin of 10% is included. The thruster on-time is less than 12,000 h, which is well within the capabilities of carbon grids. Of this, about 4300 h are required for the orbit-raising phase of the mission. For redundancy, three thrusters will be flown, mounted on individual gimbal platforms, and supplied by two PCUs. A switching system permits either PCU to operate any thruster. The total mission duration will be at least 2.5 years, assuming an operational stay of about 1 year at the target asteroid.
Fig. 4. Propellant, power system and aggregate masses as functions of SI.
9. Spacecraft design of SI and thrust, taking into account the losses in the PCU. This then provided the array mass. The PCU mass, which also depends on the power, was calculated and doubled to take into account dual thruster operation. The result was then halved to allow for advanced power processing technologies [18,19]. The resulting mass was added to that of the array to give the total dependent upon SI, recognising that the thruster itself and the propellant feed system are independent of this parameter. The results are given in Fig. 4, in which M, the total mass of the solar array and PCU, and the aggregate of these are plotted against SI. Although the upward trend of the mass of the power systems is evident, this is less rapid than the rate at which M falls with increase of SI, so there is no minimum in the curve for the total mass, despite extending the range covered to 6000 s. Thus, there is no optimum value of SI within this range. To check the sensitivity of this conclusion to the assumptions made about the power systems, 20% was added to their mass, giving the upper curve; this also has no minimum. In order to proceed, it was decided to limit the SI to 5000 s, for which M is 43.4 kg. At this value, the ion thruster grid operating point is acceptable, with a beam accelerating potential of 2.25 kV. If an allowance of 300 W is made for platform systems during thruster operation, together with a 5% degradation due to radiation damage, an array power to 2.4 kW at the start of the mission is a reasonable objective. The array mass is then 19.2 kg and the area 9.6 m2 .
The proposed configuration of the spacecraft is shown in Fig. 5. It is conventional, with solar arrays on either side of the platform, which contains all service systems and payload instruments and is dominated by the HGA. This is fixed in position and thus requires the spacecraft to point it towards the Earth for the transmission of data. Two omni-directional antennas are also included for low data rate communications. For simplicity, only a single T5 ion thruster is indicated, mounted on the lower face of the platform, although it is recommended that three thrusters be employed. These can be mounted readily onto one face of a platform with major dimensions of about 1 m. The platform is modular, with the IPS situated partly within the launcher interface adapter. The masses of the components of the IPS were deduced largely from the Artemis programme [3], taking into account expected advances in electronics and
Fig. 5. Close-up, partially cut-away view of the underside of the spacecraft, showing one thruster only.
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Table 8 Preliminary mass budget Item
Mass (kg)
Margin (%)
Payload (assumed) Propellant (Xe) Tank Pressure regulator system T5 thrusters (3) Gimbal mounts (3) PCUs (2) Switch unit Propellant control units (3) IPS harness IPS pipework and fittings Solar array Array rotation mechanism Array electronics Battery OBC Communications system Sun and Earth sensors Star tracker (redundant) GPS system ACS thrusters ACS supplies, switches CMGs Structure Launch adaptor Thermal control Harness Totals
15.0 44.9 5.0 3.0 4.8 4.5 12.4 1.0 1.5 2.0 1.0 19.2 2.5 1.8 3.4 2.4 7.0 1.0 1.5 1.0 2.0 1.3 2.0 24.0 6.0 6.0 2.0 178.2
(kg)
a
10 20 10 5 20 20 20 10 20 20 20 10 20
4.5 1.0 0.3 0.3 0.9 2.5 0.2 0.2 0.4 0.2 3.9 0.3 0.4
a
10 10 10 10 10 10 10 10 10 20 20 20
0.3 0.7 0.1 0.2 0.1 0.2 0.2 0.2 2.4 1.2 1.2 0.4 22.3
a Mass defined by mission.
propellant feed technologies [9]. The latter assumed the use of MEMS components for certain functions, such as temperature and pressure sensors. The individual items are listed in the mass budget in Table 8, together with their estimated masses and appropriate margins. To the total of 30.2 kg must be added to the mass of the propellant tank. The preferred tank is a recent QinetiQ development and is a filament-wound toroid, which makes the best use of the volume available and has a mass of about 10% of M. Assuming that the ACS will also use xenon and that it requires 1.5 kg, the total M will be 44.9 kg for a 12 km/s capability. This becomes 49.4 kg with a margin of 10%. The tank will therefore weigh 5 kg and the total IPS dry mass will be 35.2 kg.
909
A T5 thruster operating at 25 mN thrust and an SI of 5000 s consumes about 830 W. With a PCU efficiency of 87%, the input power becomes 954 W. To this must be added the 20 W consumed by the propellant feed system, giving a total of 974 W. Thus, an array power of 2 kW is adequate for two thrusters. If 300 W covers the needs of the platform systems and payload, the total requirement becomes 2248 W. To this was added an allowance for array degradation; the worst case figure of 5% gives a required capability of 2366 W, which was rounded up to 2400 W. Assuming flexible array technology, the mass is 19.2 kg. If the array must be packaged within the outline of the spacecraft body, with a side of length 1 m, it is likely to have dimensions of about 0.9 m × 5.5 m. Small, commercially available, array drive systems weigh about 1.1 kg, so a dual redundant unit will require perhaps 2.5 kg. In estimating the mass of the unregulated power bus system, the maximum power to be handled by the battery charge and discharge regulators is limited to the peak demands on the battery, as determined by the sum of the payload and platform consumptions. This is less than 400 W. The total mass using present technology was estimated to be 2.5 kg. The shunt regulator must be able to handle the full design output power of 2400 W. Assuming a bus potential of 50 V, the maximum current is 48 A and the predicted mass is 1.1 kg. Thus, the total system mass is 3.6 kg, which can be reduced to 1.8 kg using advanced electronics technologies [18,19]. The battery will be lithium-ion to capitalise on the high Wh/kg that this technology offers [20]; 150 Wh/kg should be achievable with a DoD of 80%. A high capacity is not needed, as the only eclipses will be during orbit-raising from GTO, when the IPS will not be operated. It was assumed that 500 Whr will be fully adequate, so the mass becomes 3.4 kg. 9.1. Mass budget Table 8 gives the preliminary mass budget, including estimated margins which have been selected in the range 5–20% according to the maturity of the technology involved. It is clear that the spacecraft can be constructed within the 200 kg limit, including the estimated margins, and that the 12 km/s velocity increment can be achieved. However, this budget does not include any scope for an overall 20% margin, as often
910
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
required at the start of many projects; this amounts to 31 kg in this case, since the propellant mass would not be included in the calculation. 10. Conclusions This study aimed to minimise the mass of an interplanetary mission to an asteroid, to meet a 200 kg requirement while using an economical launch vehicle. After consideration of a range of targets, a velocity increment of 8 km/s was assumed for the interplanetary phase of the mission to ensure that a very wide range of asteroids will be accessible. The assumption of an Ariane 5 auxiliary launch into GTO increased V to 12 km/s, to provide an orbit-raising manoeuvre to achieve escape from the Earth’s gravitational field. Assuming the use of two 25 mN gridded ion thrusters operating simultaneously with an SI of 5000 s, together with one spare thruster, the total predicted mass of the proposed spacecraft is 178 kg. This includes a solar array with a 2.4 kW capability at the BOL and a payload of 15 kg, and enables 12 km/s to be achieved. If appropriate margins are included for individual components and systems, the 200 kg limit remains feasible. However, a 20% global margin cannot be accommodated.
Acknowledgement This study was funded by ESA/ESTEC. References [1] M.D. Rayman, The Deep Space 1 extended mission: challenges in preparing for an encounter with comet Borrelly, IAF Paper IAF-01-Q.5.01, October 2001. [2] J. Kawaguchi, K.T.K. Uesugi, Technology development status of the Muses-C sample and return project, IAF Paper IAF99-IAA.11.2.02, October 1999. [3] R. Killinger, et al., ARTEMIS orbit-raising inflight experience with ion propulsion, IEPC Paper 03-096, March 2003. [4] G.D. Racca, B.H. Foing, P. Rathsman, An overview on the status of the SMART-1 mission, IAF Paper IAA-99IAA.11.2.09, October 1999.
[5] N. Wells, D. Fearn, Minimising the size and mass of interplanetary spacecraft, IAF Paper IAF-01-Q.5.07, October 2001. [6] P. Bianco, C. Moratto, L. Tucci, FEEP propulsion system for small spacecraft interplanetary space missions, IAF Paper IAF-01-S.4.05, October 2001. [7] A.R. Martin, W.E.F.L. Moulford, D.G. Fearn, Low-cost interplanetary missions using electric propulsion, J. Br. Interplan Soc. 49 (1996) 447–454. [8] D. Rodgers, J. Brophy, Ion propulsion technology for fast missions to Pluto, IEPC Paper 01-1716, October 2001. [9] S.D. Clark, D.G. Fearn, F. Marchandise, A study into the techniques for miniaturised electric propulsion systems, and mission categories, for small spacecraft, IEPC Paper 03-216, March 2003. [10] R.L.S. Taylor, Probability Research Group, private communication, August 2002. [11] D.G. Fearn, P. Smith, A review of UK ion propulsion— a maturing technology, IAF Paper IAF-98-S.4.01, September/ October 1998. [12] P. Gessini, S.B. Gabriel, D.G. Fearn, The hollow cathode as a micro-ion thruster, IEPC Paper 01-233, October 2001. [13] S. Marcuccio, A. Genovese, M. Andrenucci, FEEP microthruster technology status and potential application, IAF Paper IAF-97-S.3.04, October 1997. [14] M. Paine, S. Gabriel, A micro-fabricated colloidal thruster array, AIAA Paper 01-3329, July 2001. [15] P.L. Rustan, Clementine: mining new uses for SDI technology, Aerospace Am. 32 (1) (1994) 38. [16] Staff of Space Department, RAE, A proposal for the X4 and X5 spacecraft in the Black Arrow programme, RAE Tech Report 68144, June 1968. [17] R.J. Blott, N.S. Wells, J. Eves, The STRV1 microsatellite series: exploiting the geosynchronous transfer orbit, Acta Astronaut. 41 (1997) 4–10, 481–491. [18] R. Pierobon, et al., Characterization of Schottky SiC diodes for power applications, Report by Department of Electronics and Informatics—University of Padova, via Gradenigo 6/a, 35131 Padova, Italy. [19] Anon, The Role of SiC in low loss power electronics for energy management—Sceptre, http://eee.ncl.ac.uk/ HOMEPAGE.NSF/. [20] M. Slimm, R. Spurrett, C. Thwaite, D. Lizius, Lithium -ion batteries for space, Proceedings of the Sixth European Space Power Conference, Oporto, Portugal, 6–10 May 2002. [21] A.G. Santo, S.C. Lee, R.E. Gold, NEAR spacecraft and instrumentation, J. Aeronaut. Sci. 43 (4) (1995) 373–397.
Further reading [22] S.D. Clark, D.G. Fearn, F. Marchandise, Miniaturised EP Systems Study—Final Report, March 2003.
The impact of advanced platform and ion propulsion technologies on small, low-cost interplanetary spacecraft Stephen D. Clarka,∗ , David G. Fearnb a Space Department, Y72 Building, QinetiQ, Farnborough, Hants, GU14 0LX, UK b EP Solutions, 23 Bowenhurst Road, Church Crookham, Fleet, Hants, GU52 6HS, UK
Available online 19 September 2005
Abstract This paper assesses the application of small ion-propelled spacecraft of about 200 kg launch mass to low-cost interplanetary missions, for which very high velocity increments (V ) are required. Here, the “low-cost” criterion eliminates the utilisation of the launcher for the attainment of escape velocity, thereby implying the use of an ion propulsion system for orbit-raising from an initial low Earth orbit or from the geostationary transfer orbit. The sample missions considered were to rendezvous with a wide range of asteroids; they require a total V of the order of 10–12 km/s. Although the performance of the solar array is critical, recent advances in solar cell technology have enabled considerable progress to be made in this area. Other relevant advances include improved batteries, attitude and orbit control sensors, and communication systems. Gridded ion thrusters are essential, owing to the need for very high values of specific impulse. They allow a very large number of target bodies to be accessible within a launch mass of 200–250 kg. © 2005 Elsevier Ltd. All rights reserved.
1. Introduction Deep space missions with significant scientific objectives usually require a very large velocity increment V ). Although some of this can often be provided by planetary swing-bys, until recently, the only options available to achieve this were to launch on a very large and costly rocket or to supply the necessary V by use of an on-board propulsion system. The latter approach led to a large propellant load, and thus ∗ Corresponding author.
E-mail addresses: [email protected] (S.D. Clark), [email protected] (D.G. Fearn). 0094-5765/$ - see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.actaastro.2005.07.022
to a heavy spacecraft and the need for an expensive launcher. The advent of the operational ion propulsion system (IPS), as demonstrated so convincingly by the Deep Space 1 (DS-1) [1] mission, has permitted a complete re-evaluation of this situation. Moreover, a more ambitious mission is now underway; this is Muses-C [2], which is to take a sample from an asteroid and return it to Earth for laboratory examination. Owing to the high specific impulse (SI) available from an IPS, which can be a factor of 10–20 greater than that from a conventional chemical propulsion system, the propellant load required in the second option mentioned above can be reduced by the same factor. This very large gain can be further enhanced by
900
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
utilising an even less costly launch vehicle to deploy the spacecraft initially into a low Earth orbit (LEO) or to geostationary transfer orbit (GTO). Earth escape can then be achieved via a spiral orbit-raising manoeuvre, as recently undertaken by the Artemis communications satellite [3]. A similar transfer from GTO will soon be demonstrated during the SMART-1 [4] lunar mission. The potential mass reductions and associated financial gains available from the use of an IPS has led to several studies which have explored the minimum spacecraft mass that could provide a viable deep space mission [5–8]. Such an objective is assisted substantially by significant advances in other technological fields [9], notably lightweight solar arrays and batteries and high temperature semiconductors. This progress recently led to the conclusion [5] that a 10 km/s mission to an asteroid could be launched as an Ariane 5 auxiliary payload into GTO, which has a mass limit of 120 kg. This mission would commence with an orbit-raising manoeuvre to escape from the Earth’s gravitational field, but could not accommodate redundancy. This earlier work has thus been extended to examine the impact of increasing the launch mass to about 200 kg, with the aim of carrying redundant thrusters and a larger payload, whilst simultaneously raising the total velocity increment available.
2. Asteroid missions There has, for many years, been great interest in deep space missions, heightened by the need to understand better the origin and evolution of the solar system, and also by the possibility of finding evidence of past or present life elsewhere. This interest extends to the asteroids and comets orbiting the Sun, mainly because they are considered to be remnants of the material from which the solar system was formed. However, a major problem associated with missions to these bodies is the substantial cost involved. This is due largely to the very high V required for escape from the Earth’s gravitational field and then to reach the target in a reasonable time. A rendezvous with the target adds further to this requirement, which totals many km/s. The initial objective of the present study was to select suitable targets, taking into account the need to minimise launch cost. Although a very large V
was deemed mandatory, preliminary calculations suggested that 12–14 km/s might be achieved by a 200 kg spacecraft carrying a reasonable payload. A list of accessible asteroids [10] was examined, together with their orbital parameters and the values of V required to rendezvous with them following Earth escape. These were between 4.5 and 8 km/s, so are feasible using the spacecraft envisaged in this study. Indeed, it was found to be possible to target up to three such bodies in a single mission. Three asteroids can be visited for a total V of below 10 km/s in at least 14 combinations and all multiple-targets studied require less than 12 km/s. From this examination of possible targets, it was clear that most represent desirable scientific objectives. It was also evident that special expertise and extensive discussion within the science community will be required to select the asteroids giving the best scientific return. Since none of them require a total value of V exceeding 12 km/s, they are all accessible and there was thus no need to be specific at this stage.
3. Propulsion requirements 3.1. Primary propulsion A preliminary assessment indicated that it is necessary to achieve values of SI of the order of 5000 s to realise the above objectives. Thus, gridded ion thrusters are essential, since no other developed technologies provide an adequate SI, coupled with a thrust of the order of tens of mN and long life. While the absolute performance of such thrusters has almost reached a plateau, significant improvements have been reported in life expectancy, partly through the use of carbon ion extraction grids and also through the attainment of higher values of propellant utilisation efficiency. As regards the overall mass of an IPS, important advances are being made in the propellant feed and power conditioning fields. In the former, there have been moves towards microminiaturisation, although the early promise of micro-electronic mechanical systems (MEMS) has not yet been realised. In the latter, SiC semiconductor technology will permit hightemperature operation of power conditioning circuits, allowing thermal design constraints to be relaxed and leading to substantial mass and volume reductions.
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Table 1 M and thrusting time as functions of SI, with two T5 thrusters providing a V of 12 km/s SI (s)
M (kg)
Thrusting time (h)
3000 4000 5000 6000
67.0 52.7 43.4 36.9
11,445 11,491 11,831 12,059
901
Thus, the primary propulsion requirements assume that the total V needed is 12 km/s and that two T5 thrusters operate in parallel at an SI of 5000 s. The propellant mass is then 43.4 kg and the thrusting time 12,000 h. 3.2. Secondary propulsion
Table 2 Typical operating parameters of a T5 ion engine at 25 mN thrust and a SI of 3500 s Thruster parameter
Value
Thrust Ion beam current Beam accelerating potential Beam power Ion velocity Total mass flow rate Neutraliser mass flow rate Specific impulse Propellant utilisation efficiency Anode current Solenoid power Keeper discharge power Maximum discharge power Neutraliser discharge power Maximum total power Electrical efficiency Total efficiency Power to thrust ratio
25 mN 457 mA 1100 V 503 W 40.2 km/s 0.725 mg/s 0.04 mg/s 3515 s 85.7% 3A <3W < 15 W 138 W < 16 W 657 W 76.5% 65.6% 26.3 W/mN
The “rocket equation” shows that reasonable values of the propellant mass required, M, can be attained only if the SI is large; within limits, the higher the value achieved the better. This is illustrated by the data in Table 1, which assumes an initial deployed spacecraft mass of 200 kg and the use of two T5 ion thrusters [11] operating at 25 mN to produce a V of 12 km/s. For reference purposes, typical T5 thruster parameters are given in Table 2 at an SI of 3500 s. The estimated thruster lifetime is in excess of 10,000 h with molybdenum grids, so all values of SI listed are probably achievable employing this technology. However, an improvement by a factor of 3–5 is immediately available using carbon.
It is likely that the optimum method of achieving three-axis attitude control will involve a combination of thrust vectoring of the IPS and control moment gyros (CMGs). The effectiveness of the former technique was demonstrated during the DS-1 mission [1], for which the NSTAR ion thruster was mounted on a gimbal platform. The CMGs will be mainly exercised when the IPS is not operating, but will also be required continuously for control about the roll axis. The main thrusters will also be suitable for momentum dumping about two axes after periods when no thrust is required. Auxiliary thrusters will be needed for momentum dumping about the roll axis and for attitude control when the IPS is inactive. It is recommended that this capability be provided by the hollow cathode arcjet (HCA) [12], small field-emission electric propulsion (FEEP) devices [13], or miniature colloid thrusters [14]. Of these, the HCA is the currently preferred solution, despite its rather low predicted SI of perhaps 500–600 s. This is because of its commonality with the primary propulsion system; it is able to share the same propellant, tank, regulator, and power supplies. Thus, the attitude control system (ACS) requires four CMGs arranged in a tetrahedron configuration and providing redundancy against the failure of one wheel. Momentum offloading needs eight HCAs, although degraded control could be maintained with six, but this would incur significant additional fuel and reorientation manoeuvres. Taking the T5 cathode as an example, this has a mass of 70 g, to which must be added flow control valves at 100 g each, mounting brackets, pipework, and other items, giving a total of 2 kg. Although it might be assumed that the thruster power conditioning units (PCUs) could be used to operate the HCAs, this will only be possible if the thrusters are not in use. Thus, an allowance must be added for power units for two HCAs to operate simultaneously, plus a switch system to interconnect the
902
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
eight. These supplies will have a mass of about 0.5 kg. If the switch system weighs 0.3 kg, the total is 1.3 kg. The pointing stability required will depend on the payload selected. However, at a range of 10 km from the target body, an estimate of 200 mrad/s is considered adequate. Commercially available CMGs of this standard weigh of the order of 1–3 kg; the mean figure of 2 kg is adopted in this case. This corresponds well with the 2 kg system flown on Clementine [15].
4. Solar cells and arrays In all studies, the performance of the solar array is critical to the achievement of a significant payload mass. In addition to the power generated per unit mass and per unit area, which can now reach 125 W/kg and 250 W/m2 , the degradation during transit through the Earth’s radiation belts is of concern. However, recent progress has enabled the degradation during transfer from GTO to escape to be reduced to less than 5%. 4.1. Solar cells There has been considerable success in improving the performance of solar cells and thus of complete arrays. For example, certain high-eta Si cells have initial efficiencies ranging from 15.5% to 17.3%, depending on whether the cell is optimised for beginning-of-life (BOL) efficiency, radiation resistance or cell mass. The main advantages of these cells over other candidates are that they are lighter, with approximately one-third of the mass and less than 50% of the cost. GaAs/Ge cells, which are available from commercial sources, have a typical BOL efficiency of 19%, with superior radiation resistance compared to Si cells. The temperature coefficients are also better, so that there is less reduction of efficiency with increasing temperature. However, dual and triple junction (TJ) cells are the current state-of-the-art devices, although the former are already being superseded by the latter. Dual junction cells are typically 24–25% efficient, and the TJ types reach 27–28% and are forecast to be close to 30% in the next few years. They all have a superior radiation resistance compared to GaAs/Ge technology. These cells have a relatively high operating voltage, which simplifies array construction; they
Table 3 Performance of the X4 solar array with modern solar cells Cell type
Cell efficiency Mass (kg) Watts (W) Watts/kg
Original X4 7.7 High eta 16 GaAs/Ge 19 Triple junction 28
6.25 6.85 8.04 8.26
310 644 724 1127
49.6 94.0 90.0 136.4
provide 2.2 V, as compared to 0.85 and 0.45 V for GaAs/Ge and Si, respectively. 4.2. Solar arrays Rigid panel arrays are the most common type in use. Carbon-fibre skinned aluminium panels are hinged together and deployed out from the sides of the spacecraft. Although most applications use two wings, some employ one wing. The mass per unit area is rather high, so a flexible blanket array is more attractive for the present application. This consists of a reinforced Kapton blanket, supporting the solar cells and their interconnections, which is folded up, concertina fashion, into a box and then deployed using an appropriate boom or mast. In between each fold is a “leaf” of Kapton to protect the cells. An early example of this type was the RAE (Royal Aerospace Establishment) array [16] used on the Miranda (X4) spacecraft flown in 1974. The Miranda array produced 310 W from 7440 2 cm × 2 cm Si cells with an efficiency of 7.7% for a mass of 6.25 kg, i.e. about 50 W/kg. If modern cells are transposed onto this array, the resulting performance figures are shown in Table 3. These improvements are very significant indeed. Another method for enhancing the power from an array is to increase the light intensity on the cells, using a concentrator technique. Several variants have been proposed, including cassegrainian reflectors, reflective troughs and fresnel lenses, although the simplest is the trough concentrator. This gives typically a 50% rise in power for 33% less cells, compared to a planar array. However, all concentrators require more stringent pointing accuracy in one axis compared to planar arrays. This leads to important ACS requirements. A more rigid assembly than the flexible blanket type is provided by the hybrid array. This usually consists of
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
903
Panel current (A)
1.05 1 Predicted Degradation
0.95 0.9 1 year Pm dose 2.1E14
0.85
Fig. 1. Partially deployed QinetiQ hybrid array. 0.8 0
100
200
300 400
500 600
700 800
900
Days since launch
blankets or reinforced skins in a carbon fibre frame, allowing the use of simpler deployment methods, while minimising the amount of structure required and the mass. As an example, QinetiQ have developed a hybrid array specifically aimed at small satellite applications. The current design provides 1–2 kW, will fit onto a 120 kg minisatellite, and is depicted in Fig. 1. The frame is constructed from carbon fibre reinforced tubing joined with moulded corner connectors. The cells are supported on a stretched membrane, with the panel interconnects attached to the rear of the membrane.
Fig. 2. Degradation of STRV-1b GaAs/Ge solar panel in GTO.
Table 4 Comparison between solar cell technologies after 15 years in GEO Cell type
Normalised EOL power
High-eta silicon GaAs/Ge Dual junction Triple junction
0.66 0.80 1.00 1.16
4.4. Summary and recommendations 4.3. Radiation considerations For orbit-raising through the Earth’s radiation belts, the radiation dose received by the array is the main design driver. This will depend upon the initial orbit and how quickly the radiation belts can be traversed. Ideally, this transfer requires radiation hard technology and a thick coverglass to shield the cell, thus a TJ type is preferable. For example, the degradation of GaAs/Ge cells using 500 m coverglasses, flown on the QinetiQ STRV1-b satellite [17] in GTO, is illustrated in Fig. 2. This predicts a degradation of below 7% for a transfer from GTO to Earth escape in 1 year, and it should be noted that TJ cells will provide a better performance than this. To illustrate the long-term performance of solar arrays in a radiation environment, Table 4 compares the relative performances of various cell types after a geostationary Earth orbit (GEO) mission of 15 years duration, where EOL is the end of life. This shows that a high-eta array would be 75% larger than the TJ equivalent if it had to provide the same EOL power.
The above discussion leads to the array parameters summarised in Table 5. From this it is clear that TJ cells provide major advantages and that they should be utilised unless strong reasons are available to dictate otherwise. Similarly, the flexible array configuration provides significant benefits and was therefore selected for the interplanetary spacecraft discussed later.
5. Other technologies 5.1. Power electronics The problem of reducing the mass and volume of the PCU of the IPS, and of other electronics systems, can be distilled to the single issue of heat rejection. If a system could be produced where the dissipation was lower, the operating temperature was higher, and the thermal conductance to the spacecraft was better, major improvements could be realised.
904
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
Table 5 Summary of solar array characteristics and recommendations Array type
Cell type
Efficiency
Mass/area (kg/m2 )
Power/mass (W/kg)
Power/area (W/m2 )
Muses-C (EOL data) Upgraded X4/X5 flexible type
TJ GaAs/Ge TJ GaInP2 /GaAs/Ge TJ TJ TJ TJ TJ
– 19 28 22.5 25 25 25 28 28
4.45 2.01 1.96 4.44 3.00 1.64 6.7–5.0 3.00 2.00
53 90 136 45 77 100 45–60 83 125
236 181 267 200 230 164 300 250 250
Concentrator (DS-1) QinetiQ hybrid QinetiQ X4 type QinetiQ hybrid; prediction Future performance—hybrid Future performance—flexible
The temperature of any electronic system is limited to that of the maximum safe operating temperature of the semiconductor components. This is currently determined by the upper junction temperature of typical Si-based semiconductors, which is 110 ◦ C. In analysing this situation, it is assumed that increasing the operating temperature of the PCU would allow the desired reduction in the physical size of the unit, if all other aspects are kept constant. As the power is largely dissipated within the junction, doubling this limit from ∼ 100 to 200 ◦ C would double the power density, leading to either half the mass, half the volume or twice the power capability for the same power conversion efficiency and conductive thermal resistance. The recent availability of SiC diodes has demonstrated the advantages of SiC over Si for power converter applications [18]. This technology, in the form of diodes and switching transistors, yields savings of 50% and more in the losses in power converter systems [19], and can therefore provide the benefits discussed above. A further advantage is the ability of SiC to operate at high switching frequencies without performance degradation. It also has very good radiation tolerance. It should also be noted that these advanced power electronics concepts are equally applicable to other associated systems, including battery charge and discharge regulators, and solar array shunt regulators.
5.2. Batteries Although Ni-H2 batteries were, until recently, regarded as the best available technology, Li-ion batter-
ies are now much superior [20], and their use was assumed in the present study. Bearing in mind that only a modest number of discharge cycles will be needed in the orbit-raising phase of the mission, it has been assumed that the depth of discharge (DoD) will be limited to 80%, giving a performance of about 150 Wh/kg. At this DoD, the limit is 30,000 cycles. 5.3. Data handling and storage Ideally, a single data handling unit (DHU) is required to perform all of the spacecraft and payload functions, i.e. housekeeping, autonomous navigation and control, operation of the IPS and of the payload. In general, the ERC-32 (Sparc) processor is likely to be the best candidate to meet these requirements, as it is commercially available in a radiation-tolerant form. An additional digital signal processor (DSP) could be accommodated for the data processing of the payload. The main design driver is the radiation dose expected during the mission. The aggregate mass of the onboard computer (OBC) are given in Table 6 for a variety of memory capacities; the option in the last line of the table was selected for this study. 5.4. Propellant feed system Provided that the problems which face MEMS systems in new applications can be successfully overcome, there is no reason why selected propellant feed components could not benefit from these technologies in the near future. The limiting factor is the technical performance achievable when compared to operational requirements. Clearly, the benefits include
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
905
Table 6 Mass of OBC as a function of shielding thickness and memory capacity Shielding thickness (mm)
Mass memory (Mbits)
Computer mass (g)
Mass excluding box (g)
Total mass (g)
4 4 7 7
500 1000 500 1000
477 477 477 477
645 674 645 674
1552 1609 2321 2400
reduced mass and volume and, eventually, an improvement in reliability. In the longer term, a significant additional advantage is the degree of component integration that can be achieved into a single device, allowing production of a complete system “on a chip” or the inclusion of multiple redundancy with little extra mass or volume penalty. Two possible examples at opposite ends of the integration spectrum include a single microvalve chip, containing ten or more identical valves, and a complete propellant feed system integrated at wafer level into a single unit. However, it was not felt that MEMS capabilities are sufficiently developed for their use to be assumed in the present study. 5.5. Communications The frequency bands of choice are either Ka- or X-band and, to minimise cost, it is assumed that 12 m diameter ground antennas will be used. As a baseline, the system was flown very successfully on DS-1 [1] operated in Ka-band and had a mass of 3 kg. Allowing 5 kg for the high-gain antenna (HGA) and cabling, the total would be about 8 kg. However, at X-band the latest power module technology could be used; these are smaller and lighter than those of the Ka-band system. Thus, an X-band system was selected, with a 0.5 m diameter HGA reflector, a horn feed, a continuously powered low-noise amplifier, an X-band receiver, and a transmitter. Small omni-directional antennas are also provided to allow for emergency access to the spacecraft and for low-rate housekeeping data to be downloaded. It is estimated that this system, operating at 8.5 GHz, will provide a data rate of 600 bps at a distance of 2 AU from the Earth. The assumed 9.6 dB margin (equivalent to a bit error rate of 1 : 105 ) is exceeded by 1.7 db by the use of standard coding techniques. The
beamwidth of the HGA will be less than 1◦ and will therefore require accurate Earth-pointing. The transmitter RF output of 180 W requires about 400 W of DC power. Since the antenna is assumed to be fixed to the spacecraft, so that the complete vehicle must be orientated correctly during communications sessions, a constraint on the mission is that the transmitter is not operated whilst thrusting. A mass of 7 kg was assumed for the complete system. 5.6. Attitude and orbit determination Attitude determination during orbit-raising will utilise digital sun and Earth sensors, based on MEMS technology, augmented as necessary by data from a star sensor/tracker. A number of MEMS devices currently in development weigh only 10’s of grammes each, so a total allocation of 1 kg for sun and Earth sensors is generous. The star tracker can be used during both orbit-raising and the cruise phase and provides an essential autonomous navigation capability. This is vital owing to the large distances from Earth and the need to minimise ground segment costs. Excellent results from DS-1 [1] suggest that this concept is entirely feasible. Indeed, this mission provided en-route accuracy of better than 400 km and 0.2 m/s, which improved to better than 3 km at the encounter. The star sensor will be aligned along the solar array axis, which places the sun at 90◦ , avoiding blinding. This allows the device to be small and reduces the requirement for a baffle around the head. The star tracker on Clementine [15] weighed only 0.6 kg, and consumed 4.5 W; this is suitable in the present case, although an allocation of 1.5 kg provides redundancy. Orbit determination whilst in Earth orbit will be by standard spacecraft ranging through the transponder. This is likely to be augmented by use of a GPS system, although the utility of this in and beyond GTO has yet
906
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
to be demonstrated. As commercial units for terrestrial use are available with a mass of a few hundred grammes, 1 kg for this would appear to be reasonable. 5.7. Structure, launch adaptor, thermal control and harness The assumed structural configuration employs both carbon fibre reinforced plastic struts and honeycomb panels. A single toroidal filament-wound fuel tank is situated around a central thrust/torsion cylinder, which also houses the ion thrusters and their flow control units. Subsystems and payload units are attached to floors bonded above the tank, and external shear walls carry the loads through to the launcher interface ring. A preliminary assessment suggests that 12% of the launch mass is a reasonable allocation for this. To this must be added about 3% to allow for the components of the launch adaptor fixed to the spacecraft. Removal of heat from both the communications equipment and each ion thruster PCU will require additional mass. A heat shield will also be required between the ion thrusters and the adjacent spacecraft structure. An additional 3% of the launch mass (6 kg), is allocated for these purposes. The harness is likely to require about 1% of the launch mass, plus the separate allocation of 2 kg to cover the parts of the harness connecting the thrusters, switch system and PCUs. 5.8. Spacecraft autonomy To minimise ground segment costs, the omnidirectional low-gain antenna on the spacecraft radiates continuously at one of four frequencies. These indicate the health of the spacecraft, as determined by the OBC, and the need for intervention from the ground, and the urgency with which action is required. This is very effective in reducing ground segment costs, and it is recommended that a similar approach to autonomy is considered for most future missions.
together with a variety of US piggy-back alternatives. In addition, a viable possibility is a launch to GTO as an Ariane 5 auxiliary payload within the main thrust cone, for which the mass limit is 350 kg. However, many of these vehicles could launch 200 kg to a much higher altitude. For example, the Rockot could provide an initial circular altitude of several 10,000 km and probably an escape capability. Thus, a very wide spectrum of performance is available, ranging from LEO to attaining more than escape velocity. So the total requirements for the IPS vary from more than 10 km/s to a few km/s, depending upon the target. As one objective of this mission is to demonstrate the capabilities of ion propulsion, it is appropriate to choose a modest launcher, thereby also minimising cost. If Vega is disregarded owing to its relatively high performance, the remaining European option is an Ariane 5 auxiliary payload launch into GTO, which requires 4 km/s to achieve escape velocity.
7. Payload An idea of the type of instrument which would be suitable for an asteroid mission can be obtained from the NEAR spacecraft [21], which acquired extensive data from and eventually landed on Eros. Relevant instruments were also carried by DS-1 [1] and the Clementine mission to the moon [15]. In the latter case the payload consisted of two star tracker cameras, other cameras operating in the ultra-violet (UV)/visible, near infra-red (IR) and long-wave IR regions, a high-resolution camera, and a laser transmitter. The total mass was 7.37 kg and the maximum power consumption 97.3 W. More ambitious is the Muses-C asteroid sample return mission [2], and relevant future missions include ESA’s SMART-1 lunar spacecraft [4]. Additional capabilities can be provided by combinations of individual instruments and by the communication systems. A realistic aim for the total mass of the payload is 15 kg, as indicated in Table 7.
6. Launch options To achieve minimum cost, an attractive option is a launch to LEO using a relatively inexpensive vehicle, such as the Pegasus-XL, Taurus, Rockot and Vega. Several other low-cost Russian options also exist,
8. Mission profile In this mission, the dominating constraints are the initial mass of 200 kg and the use of an Ariane
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Table 7 List of possible payload instruments, with masses and power consumptions Experiment
Origin
Multispectral camera X-ray spectrometer Radio science Magnetometer IR spectrometer Laser rangefinder Penetrator
NEAR SMART-1 SMART-1 NEAR SMART-1 Muses-C –
Total
Mass (kg)
Power (W)
5 3 0 1 2 2 2
10 5 0 1.5 2 17 5
15
40.5
50
Propellant Mass (kg)
From 300 km From 2000 km
40
From GTO
30
20
10 3000
3500
4000
4500
5000
5500
Specific Impulse (s) Fig. 3. Propellant mass as a function of SI for the orbit-raising phase of the mission.
5 auxiliary payload launch to GTO. Target selection, which is best left to later scientific assessments, is represented by a worst case V of 8 km/s. To this must be added the 4 km/s needed for Earth escape, giving a total of 12 km/s. It is proposed that the orbit-raising phase be based on thrusting over a wide angle about the apogee to raise the perigee to geostationary altitude, followed by continuous thrusting to achieve escape. For comparison purposes, the Pegasus-XL, SHTIL2/Volna and Ariane 5 launch options were analysed to establish what can be accomplished using small ion thrusters for orbit-raising. The altitudes to which the first two will reach were assumed to be 2000 and 300 km, respectively. The results (Fig. 3) assume that T5 ion thrusters [11] operating at a thrust of 25 mN are employed. The range of SI covered was determined by limiting the ion accelerating potential to 2.3 kV, which is representative of present grid systems.
907
All three options are viable, although the launches into LEO require the greatest M. For the GTO launch, there is a mass advantage of about 8 kg in going from an SI of 3200–5000 s. Assuming the use of carbon grid technology, the operating times from GTO are acceptable, even if only one thruster is operated. However, it is likely, for single thruster operation, that a spare will be needed to achieve the total impulse required, with a switching system to permit a single PCU to be connected to either device. In general, an optimum SI can be derived by adding the total mass of the propulsion and power systems, since a change in SI at constant thrust alters both M and the power needed by the thrusters. The Earth escape manoeuvre will be followed by the interplanetary trajectory, represented by the V required and the maximum distance from the sun. Thrusting with the IPS to maximum velocity (to minimise trip time) will then be followed by a retardation phase, again using the IPS, with the aim of a rendezvous with the target. The V required is between 4.6 and 8.0 km/s, giving a total of 8.6–12.0 km/s, including the initial orbit-raising manoeuvre. Bearing in mind the low gravitational force exerted by the target, the thrust required for capture into orbit is well within the capabilities of the IPS. A more significant challenge involves terminal navigation, since the target will almost certainly have a low albedo. It may therefore be difficult to acquire it visually at a large distance, so a very slow approach, with the possibility of needing to thrust perpendicular to the velocity vector, may be necessary. However, the value of V required is negligible compared to 12 km/s. As the power consumption increases with SI, the mass of the solar array rises as M falls, and there will normally be an optimum value of SI at which the total mass is a minimum. To determine this minimum, performance data for a flexible array using TJ cells were adopted; these are 2 kg/m2 , 125 W/kg, and 250 W/m2 (Table 5). An allowance was also made for degradation due to the Earth’s radiation belts during orbit-raising. Using two thrusters, the transfer time is about 6 months and the worst case degradation is about 3% (Fig. 2), assuming that the whole of this period is spent in GTO; a value of 5% was taken to provide a suitable margin. Approximate analytical expressions were derived for the power consumed by the thrusters as a function
908
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Total, with margin Total
100
Propellant
Mass (kg)
Array and PCU
80
60
40
20 2000
3000
4000
5000
6000
Specific Impulse (s)
To summarise, the total velocity increment needed is 12 km/s and two T5 thrusters are used in parallel, operating at an SI of 5000 s and a total thrust of 50 mN. M is then 47.7 kg if a margin of 10% is included. The thruster on-time is less than 12,000 h, which is well within the capabilities of carbon grids. Of this, about 4300 h are required for the orbit-raising phase of the mission. For redundancy, three thrusters will be flown, mounted on individual gimbal platforms, and supplied by two PCUs. A switching system permits either PCU to operate any thruster. The total mission duration will be at least 2.5 years, assuming an operational stay of about 1 year at the target asteroid.
Fig. 4. Propellant, power system and aggregate masses as functions of SI.
9. Spacecraft design of SI and thrust, taking into account the losses in the PCU. This then provided the array mass. The PCU mass, which also depends on the power, was calculated and doubled to take into account dual thruster operation. The result was then halved to allow for advanced power processing technologies [18,19]. The resulting mass was added to that of the array to give the total dependent upon SI, recognising that the thruster itself and the propellant feed system are independent of this parameter. The results are given in Fig. 4, in which M, the total mass of the solar array and PCU, and the aggregate of these are plotted against SI. Although the upward trend of the mass of the power systems is evident, this is less rapid than the rate at which M falls with increase of SI, so there is no minimum in the curve for the total mass, despite extending the range covered to 6000 s. Thus, there is no optimum value of SI within this range. To check the sensitivity of this conclusion to the assumptions made about the power systems, 20% was added to their mass, giving the upper curve; this also has no minimum. In order to proceed, it was decided to limit the SI to 5000 s, for which M is 43.4 kg. At this value, the ion thruster grid operating point is acceptable, with a beam accelerating potential of 2.25 kV. If an allowance of 300 W is made for platform systems during thruster operation, together with a 5% degradation due to radiation damage, an array power to 2.4 kW at the start of the mission is a reasonable objective. The array mass is then 19.2 kg and the area 9.6 m2 .
The proposed configuration of the spacecraft is shown in Fig. 5. It is conventional, with solar arrays on either side of the platform, which contains all service systems and payload instruments and is dominated by the HGA. This is fixed in position and thus requires the spacecraft to point it towards the Earth for the transmission of data. Two omni-directional antennas are also included for low data rate communications. For simplicity, only a single T5 ion thruster is indicated, mounted on the lower face of the platform, although it is recommended that three thrusters be employed. These can be mounted readily onto one face of a platform with major dimensions of about 1 m. The platform is modular, with the IPS situated partly within the launcher interface adapter. The masses of the components of the IPS were deduced largely from the Artemis programme [3], taking into account expected advances in electronics and
Fig. 5. Close-up, partially cut-away view of the underside of the spacecraft, showing one thruster only.
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910 Table 8 Preliminary mass budget Item
Mass (kg)
Margin (%)
Payload (assumed) Propellant (Xe) Tank Pressure regulator system T5 thrusters (3) Gimbal mounts (3) PCUs (2) Switch unit Propellant control units (3) IPS harness IPS pipework and fittings Solar array Array rotation mechanism Array electronics Battery OBC Communications system Sun and Earth sensors Star tracker (redundant) GPS system ACS thrusters ACS supplies, switches CMGs Structure Launch adaptor Thermal control Harness Totals
15.0 44.9 5.0 3.0 4.8 4.5 12.4 1.0 1.5 2.0 1.0 19.2 2.5 1.8 3.4 2.4 7.0 1.0 1.5 1.0 2.0 1.3 2.0 24.0 6.0 6.0 2.0 178.2
(kg)
a
10 20 10 5 20 20 20 10 20 20 20 10 20
4.5 1.0 0.3 0.3 0.9 2.5 0.2 0.2 0.4 0.2 3.9 0.3 0.4
a
10 10 10 10 10 10 10 10 10 20 20 20
0.3 0.7 0.1 0.2 0.1 0.2 0.2 0.2 2.4 1.2 1.2 0.4 22.3
a Mass defined by mission.
propellant feed technologies [9]. The latter assumed the use of MEMS components for certain functions, such as temperature and pressure sensors. The individual items are listed in the mass budget in Table 8, together with their estimated masses and appropriate margins. To the total of 30.2 kg must be added to the mass of the propellant tank. The preferred tank is a recent QinetiQ development and is a filament-wound toroid, which makes the best use of the volume available and has a mass of about 10% of M. Assuming that the ACS will also use xenon and that it requires 1.5 kg, the total M will be 44.9 kg for a 12 km/s capability. This becomes 49.4 kg with a margin of 10%. The tank will therefore weigh 5 kg and the total IPS dry mass will be 35.2 kg.
909
A T5 thruster operating at 25 mN thrust and an SI of 5000 s consumes about 830 W. With a PCU efficiency of 87%, the input power becomes 954 W. To this must be added the 20 W consumed by the propellant feed system, giving a total of 974 W. Thus, an array power of 2 kW is adequate for two thrusters. If 300 W covers the needs of the platform systems and payload, the total requirement becomes 2248 W. To this was added an allowance for array degradation; the worst case figure of 5% gives a required capability of 2366 W, which was rounded up to 2400 W. Assuming flexible array technology, the mass is 19.2 kg. If the array must be packaged within the outline of the spacecraft body, with a side of length 1 m, it is likely to have dimensions of about 0.9 m × 5.5 m. Small, commercially available, array drive systems weigh about 1.1 kg, so a dual redundant unit will require perhaps 2.5 kg. In estimating the mass of the unregulated power bus system, the maximum power to be handled by the battery charge and discharge regulators is limited to the peak demands on the battery, as determined by the sum of the payload and platform consumptions. This is less than 400 W. The total mass using present technology was estimated to be 2.5 kg. The shunt regulator must be able to handle the full design output power of 2400 W. Assuming a bus potential of 50 V, the maximum current is 48 A and the predicted mass is 1.1 kg. Thus, the total system mass is 3.6 kg, which can be reduced to 1.8 kg using advanced electronics technologies [18,19]. The battery will be lithium-ion to capitalise on the high Wh/kg that this technology offers [20]; 150 Wh/kg should be achievable with a DoD of 80%. A high capacity is not needed, as the only eclipses will be during orbit-raising from GTO, when the IPS will not be operated. It was assumed that 500 Whr will be fully adequate, so the mass becomes 3.4 kg. 9.1. Mass budget Table 8 gives the preliminary mass budget, including estimated margins which have been selected in the range 5–20% according to the maturity of the technology involved. It is clear that the spacecraft can be constructed within the 200 kg limit, including the estimated margins, and that the 12 km/s velocity increment can be achieved. However, this budget does not include any scope for an overall 20% margin, as often
910
S.D. Clark, D.G. Fearn / Acta Astronautica 59 (2006) 899 – 910
required at the start of many projects; this amounts to 31 kg in this case, since the propellant mass would not be included in the calculation. 10. Conclusions This study aimed to minimise the mass of an interplanetary mission to an asteroid, to meet a 200 kg requirement while using an economical launch vehicle. After consideration of a range of targets, a velocity increment of 8 km/s was assumed for the interplanetary phase of the mission to ensure that a very wide range of asteroids will be accessible. The assumption of an Ariane 5 auxiliary launch into GTO increased V to 12 km/s, to provide an orbit-raising manoeuvre to achieve escape from the Earth’s gravitational field. Assuming the use of two 25 mN gridded ion thrusters operating simultaneously with an SI of 5000 s, together with one spare thruster, the total predicted mass of the proposed spacecraft is 178 kg. This includes a solar array with a 2.4 kW capability at the BOL and a payload of 15 kg, and enables 12 km/s to be achieved. If appropriate margins are included for individual components and systems, the 200 kg limit remains feasible. However, a 20% global margin cannot be accommodated.
Acknowledgement This study was funded by ESA/ESTEC. References [1] M.D. Rayman, The Deep Space 1 extended mission: challenges in preparing for an encounter with comet Borrelly, IAF Paper IAF-01-Q.5.01, October 2001. [2] J. Kawaguchi, K.T.K. Uesugi, Technology development status of the Muses-C sample and return project, IAF Paper IAF99-IAA.11.2.02, October 1999. [3] R. Killinger, et al., ARTEMIS orbit-raising inflight experience with ion propulsion, IEPC Paper 03-096, March 2003. [4] G.D. Racca, B.H. Foing, P. Rathsman, An overview on the status of the SMART-1 mission, IAF Paper IAA-99IAA.11.2.09, October 1999.
[5] N. Wells, D. Fearn, Minimising the size and mass of interplanetary spacecraft, IAF Paper IAF-01-Q.5.07, October 2001. [6] P. Bianco, C. Moratto, L. Tucci, FEEP propulsion system for small spacecraft interplanetary space missions, IAF Paper IAF-01-S.4.05, October 2001. [7] A.R. Martin, W.E.F.L. Moulford, D.G. Fearn, Low-cost interplanetary missions using electric propulsion, J. Br. Interplan Soc. 49 (1996) 447–454. [8] D. Rodgers, J. Brophy, Ion propulsion technology for fast missions to Pluto, IEPC Paper 01-1716, October 2001. [9] S.D. Clark, D.G. Fearn, F. Marchandise, A study into the techniques for miniaturised electric propulsion systems, and mission categories, for small spacecraft, IEPC Paper 03-216, March 2003. [10] R.L.S. Taylor, Probability Research Group, private communication, August 2002. [11] D.G. Fearn, P. Smith, A review of UK ion propulsion— a maturing technology, IAF Paper IAF-98-S.4.01, September/ October 1998. [12] P. Gessini, S.B. Gabriel, D.G. Fearn, The hollow cathode as a micro-ion thruster, IEPC Paper 01-233, October 2001. [13] S. Marcuccio, A. Genovese, M. Andrenucci, FEEP microthruster technology status and potential application, IAF Paper IAF-97-S.3.04, October 1997. [14] M. Paine, S. Gabriel, A micro-fabricated colloidal thruster array, AIAA Paper 01-3329, July 2001. [15] P.L. Rustan, Clementine: mining new uses for SDI technology, Aerospace Am. 32 (1) (1994) 38. [16] Staff of Space Department, RAE, A proposal for the X4 and X5 spacecraft in the Black Arrow programme, RAE Tech Report 68144, June 1968. [17] R.J. Blott, N.S. Wells, J. Eves, The STRV1 microsatellite series: exploiting the geosynchronous transfer orbit, Acta Astronaut. 41 (1997) 4–10, 481–491. [18] R. Pierobon, et al., Characterization of Schottky SiC diodes for power applications, Report by Department of Electronics and Informatics—University of Padova, via Gradenigo 6/a, 35131 Padova, Italy. [19] Anon, The Role of SiC in low loss power electronics for energy management—Sceptre, http://eee.ncl.ac.uk/ HOMEPAGE.NSF/. [20] M. Slimm, R. Spurrett, C. Thwaite, D. Lizius, Lithium -ion batteries for space, Proceedings of the Sixth European Space Power Conference, Oporto, Portugal, 6–10 May 2002. [21] A.G. Santo, S.C. Lee, R.E. Gold, NEAR spacecraft and instrumentation, J. Aeronaut. Sci. 43 (4) (1995) 373–397.
Further reading [22] S.D. Clark, D.G. Fearn, F. Marchandise, Miniaturised EP Systems Study—Final Report, March 2003.