Bepicolombo - Solar Electric Propulsion System Test and Qualification Approach

BepiColombo - Solar Electric Propulsion System Test and Qualification Approach

IEPC-2019-586

Presented at the 36th International Electric Propulsion Conference University of Vienna • Vienna, Austria September 15-20, 2019

S. Clark1, P. Randall2, R. Lewis3, D. Marangone4 QinetiQ, Cody Technology Park, Farnborough, GU14 0LX, UK

D. Goebel5, V. Chaplin6 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109, USA

H. Gray7 Airbus Defence and Space Ltd., Portsmouth, Hants, PO3 5PU, UK

K Kempkens8 Airbus Defence and Space GmbH, 88039 Friedrichshafen, Germany

N. Wallace9 European Space Agency, ESTEC, Noordwijk, 2200AG, Netherlands

Abstract: In support of the BepiColombo mission to Mercury, a qualification programme for the Solar Electric Propulsion Thruster (SEPT) and Solar Electric Propulsion System (SEPS) was conducted by QinetiQ at its facilities in Farnborough, UK. The paper discusses the qualification of the thruster lifetime against the mission requirements and the test approach and modelling activities that form the validation. Excellent correlation between ground test data, physical measurements and model predictions have been demonstrated during the project, in which some detail is provided.

QinetiQ also undertook the qualification of the Solar Electric Propulsion System (SEPS) chain, involving a complex setup of the SEPS Units (SEPT, PPU, FCU and SEPH/P). This was a unique opportunity to test the hardware in a representative chain in full beam mode, which was not possible during System level AIV activities or final spacecraft integration. The Unit level qualification of the other SEPS units (PPU and FCU) are not addressed in this paper, but have been considered in the frame of the Coupling Test Activities.

1 BepiColombo SEPS AIV lead engineer, QinetiQ, Space UK, [email protected] 2 BepiColombo SEPS system lead engineer, QinetiQ, Space UK, [email protected] 3 BepiColombo SEPS thruster lead engineer, QinetiQ, Space UK, [email protected] 4 BepiColombo SEPS test engineer, QinetiQ, Space UK, [email protected] 5 JPL Fellow, Propulsion, Thermal, & Materials Engineering, Dan.M.Goebel@jpl.nasa.gov 6 Technologist, Electric Propulsion Group, [email protected] 7 BepiColombo EP Architect, Propulsion Department, [email protected] 8 System Engineering Manager, Department TESNG, [email protected] 9 Senior Electric Propulsion Engineer, TEC-MPE, [email protected] 1

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 Nomenclature BB = Bread Board BOL = Beginning of Life CMM = Coordinate Measurement Machine DANS = Discharge, Accel and Neutraliser Supply EBS = Electron Backstreaming EGSE = Electrical Ground Support Equipment EM = Engineering Model EMC = Electromagnetic Compatibility EOL = End Of Life EOM = End Of Mission EQM = Engineering Qualification Model FCU = Flow Control Unit FGSE = Fluidic Ground Support Equipment FDIR = Failure Detection, Isolation and Recovery FCU = Flow Control Unit FM = Flight Model HPEPS = High Power Electric Propulsion System ISP = Specific Impulse JAXA = Japan Aerospace Exploration Agency LEEP = Large European Electric Propulsion MEPS = MTM Electric Propulsion System MMO = Mercury Magnetospheric Orbiter MPO = Mercury Planetary Orbiter PFM = Proto Flight Model PPU = Power Processing Unit SEPH = Solar Electric Propulsion Harness SEPP = Solar Electric Propulsion Pipework SEPS = Solar Electric Propulsion System SEPT = Solar Electric Propulsion Thruster TCF = Thrust Correction Factor Xe = Xenon

I.Introduction epiColombo is an ESA cornerstone mission to Mercury executed in collaboration with its partner the Japan B Aerospace Exploration Agency (JAXA), designed to place 2 spacecraft into orbit around Mercury. Airbus Defence and Space is the overall System prime and is additionally responsible for the development and provision of the Mercury Planetary Orbiter (MPO) and the Mercury Transfer Module (MTM). JAXA is responsible for the development and provision of the Mercury Magnetospheric Orbiter (MMO). The BepiColombo spacecraft has a dedicated transfer module, which provides all the propulsion and power needed to deliver the 2 orbiters to Mercury7. This makes extensive use of electric propulsion through a number of thrust arcs. The use of electric propulsion for orbital transfer between bodies within the solar system is now well established, with its use on the Deep Space 11, Smart-12, Hayabusa Explorer3, Dawn4,5 and Hayabusa26 programmes. This follows the increasing use of electric propulsion for large telecommunications satellites. The BepiColombo mission exploits this heritage. A substantial energy reduction is needed to reach Mercury. The high specific impulses which can be provided by electric propulsion systems offer the ability to achieve this large velocity increment with significantly reduced propellant mass when compared with chemical propulsion systems. As a mission to Mercury will also inherently require flight trajectories with reducing sun-earth distances, the power available from the solar arrays will increase as Mercury is approached, ensuring adequate power for electric propulsion operations. The design, development and qualification of the BepiColombo electric propulsion system is based on previous developments and flight heritage of similar electric propulsion technologies, from a variety of applications.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 II.The BepiColombo Mission

A. Mission Objectives Mercury is an extreme of our planetary system. Since its formation, it has been subjected to the highest temperature and has experienced the largest diurnal temperature variation of any object in the solar system. It is the closest planet to the Sun and has the highest uncompressed density of all planets. Solar tides have influenced its rotational state. Its surface has been altered during the initial cooling phase and its chemical composition may have been modified by bombardment in its early history. Mercury therefore plays an important role in constraining and testing dynamical and compositional theories of planetary system formation. To date, only the American probes Mariner 10 and Messenger have returned significant data from Mercury. Although these data have been fully exploited, a lot of gross features remain unexplained. Many conclusions are still speculative and have evoked a great number of new questions. The main scientific objectives of the BepiColombo mission are:  Investigation of the origin and evolution of a planet close to its parent star  Investigation of Mercury’s figure, interior structure, and composition  Investigation of the interior dynamics and origin of its magnetic field  Investigation of the exogenic and endogenic surface modifications, cratering, tectonics, and volcanism  Investigation of the composition, origin and dynamics of Mercury’s exosphere and polar deposits  Investigation of the structure and dynamics of Mercury’s magnetosphere  Test of Einstein’s theory of general relativity The mission will achieve these objectives by delivering 2 separate spacecraft into orbit around Mercury, namely the Mercury Magnetospheric Orbiter (MMO) and the Mercury Planetary Orbiter (MPO). These 2 spacecraft will be placed into different orbits around Mercury. The MPO has an initial polar elliptic orbit, with an altitude of between 400 and 1508 km; the MMO also has an initial polar elliptic orbit, with an altitude of between 400 and 11824 km.

B. Mission Analysis and Propulsion System Usage The spacecraft was launched by Ariane 5 from Kourou at 03:45 CEST on 20th October 2018; Mercury arrival and orbit insertion will be in December 2025, giving an overall transfer duration of 7.2 years. Both chemical and electric propulsion systems are required to achieve the mission requirements. The selected mission profile uses the launcher for direct injection into an interplanetary trajectory. The electric propulsion system is then used over a number of thrust arcs, with intermediate coast phases and planetary fly-bys. A gravitational capture, via the Mercury-Sun Lagrange point, is finally used to place the spacecraft into a weakly bound Mercury orbit. A bipropellant chemical propulsion system on the transfer module is used for attitude and orbit control during the transfer, where higher thrust levels are required. A dual mode chemical propulsion on the MPO is used in bipropellant mode for lowering into the final operational orbits; this is then used in monopropellant mode for orbit and attitude control of the MPO during the operational phase of the mission. T6 thrusters and C. MTM Structure pointing mechanisms High pressure The MTM structure design is driven by the following main Flow control regulator requirements: units  Spacecraft stiffness to avoid coupling with low frequency excitations from the launch vehicle  Propellant and Xenon tanks and two solar arrays (SA) constrain the structure height  Centre of mass constraints  Compatibility with thermal environment temperature range of -30°C to +90°C  Compatibility with the launch vehicle adaptor interface Power  Loads resulting from launch environment and processing units lifting / handling / transportation cases Tanks  Mass minimisation Figure 1. MEPS Layout. The MTM structure consists of a central cone connected to four shear panels and to two main tank floor panels. Three radiator panels are connected to the tank floor, shear panels

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 and struts to the base of the cone, as well as the +Y Sunshield Support Structure form the complete MTM structure. The central cone contains a thruster floor panel, with 4 SEPS thrusters and pointing mechanisms. The SEPS is a key sub-system of the MEPS (MTM Electric Propulsion System7), and comprises the four T6 thrusters (SEPTs), two Power Processing Units (PPUs), four Flow Control Units (FCUs) and interconnecting harness and pipework (SEPH and SEPP). The MEPS layout is shown in Figure 1.

III.Solar Electric Propulsion Thruster (SEPT) Qualification

A. Overview of SEPT Qualification Programme The purpose of the SEPT qualification programme was to verify the T6 thruster capability against the mission requirements. The main programme of work consisted of a series of tests using an Engineering Qualification Model (EQM) T6 thruster covering Beginning of Life (BOL) performance characterisation followed by environmental testing and lifetime verification testing. This was supplemented by additional testing using other thruster models to characterise the thruster’s EMC performance, magnetic moment and plume characteristics. The test programme is shown in Figure 2.

Results of the initial qualification testing, the EMC testing and the magnetic moment testing are presented in the Qualification of the T6 Thruster for BepiColombo8, whilst results of plume characterisation testing performed at NASA JPL are presented in performance evaluation of the T6 Ion Engine9.

Figure 2. SEPT qualification test sequence

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 B. SEPT Lifetime Verification 1. Lifetime Assessment and Verification Approach The SEPT lifetime verification approach was underpinned by a thorough assessment of the wear mechanisms and potential failure mechanisms, which are comparable to other gridded ion thrusters10. These are summarised in Table 1. The verification activities performed during the BepiColombo programme are also summarised in the table.

Table 1. SEPT wear mechanisms

Sub- Wear mechanism Potential failure Criticality for overall Verification activity assembly mechanism thruster lifetime Grid Screen grid erosion due Loss of structural Current driver of T6 Periodically remove grid assembly to ion bombardment integrity of screen grid: thruster lifetime assembly and measure inability to sustain screen grid thickness (via x- beam ray technique) and upstream chamfering Accel grid erosion due Loss of structural Substantial margin11 Periodically remove grid to ion bombardment integrity of accel grid: assembly and measure inability to sustain accel bore profiles and beam downstream surface profile Reduction to electron Electron backstreaming: Substantial margin Periodically measure EBS backstreaming (EBS) reduction in real thrust margin during test margin due to screen and/or damage to programme grid and accel grid discharge chamber erosion Cathode Depletion of low work Inability to emit Substantial margin12,13, Monitor electrical assemblies function material from required discharge does not drive overall performance during test dispenser current thruster life programme Contamination of Inability to emit Substantial margin14, Monitor electrical dispenser by propellant required discharge does not drive overall performance during test impurities current thruster life programme Tip erosion due to ion Transition to plume Erosion of the discharge Separate modelling and test bombardment mode leading to cathode tip is negligible programme to verify EOM unstable performance / in T6, hence does not neutraliser performance and rapid erosion drive lifetime; erosion confirm flow rate of the neutraliser cathode tip does not Periodically perform drive overall life if neutraliser spot to plume suitable flow rate transition measurements selected during thruster endurance test programme Embrittlement of the Heater fracture; cannot Substantial margin14, 180 heater cycles heaters initiate discharge does not drive overall performed prior to (primarily a risk prior to life qualification vibration to launch due to cover ground acceptance mechanical loads) operations prior to launch

Include on-off cycles in thruster endurance test programme, monitor heater electrical performance Discharge Erosion of discharge Failure to meet Substantial margin, Periodically remove grid chamber chamber components performance does not drive overall assembly and measure assembly (baffle and magnetic requirements due to life dimensions of discharge pole pieces) due to ion reduced discharge chamber components bombardment efficiency Electrical Isolation degradation Inability to apply Primarily a problem for Periodically measure isolators due to sputter operating voltages ground testing, not in- electrical isolation during contamination flight operation test programme

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 Prior to the main BepiColombo SEPS programme, the total impulse capability of the T6 grid set had been assessed11, by means of a technology readiness programme and technology demonstration activity. As part of these programmes a single grid set was tested on a T6 thruster, accumulating a total impulse of 3.6MNs at a constant thrust level of ~168mN, with a total operating time of 6,056 hours. The accel grid erosion was measured and extrapolated, and it was concluded that the total impulse capability of the accel grid was far in excess of the BepiColombo requirement. However, significant erosion of the screen grid was noted, and this became the subject of a separate investigation. It was concluded from this work that the anode current and xenon main flow rate needed to be slightly increased and the grid gap reduced in order to limit the rate of screen grid erosion to an acceptable level, and this was verified by means of a 2,000 hour test. These changes were set as the new baseline for the BepiColombo mission and were applied during all qualification tests preceding the endurance test (Figure 2). Grid erosion represents the primary life-limiting wear mechanism for T6, therefore the main lifetime verification activity for BepiColombo was an endurance test designed to accumulate impulse on the grid set in conditions that were as flight-representative as practicable, supplemented with periodic erosion measurements and ion optics modelling to predict the trends and the overall lifetime. Originally it was planned to perform a total of 8,000 hours of endurance testing at 145mN thrust level. The detailed planning has evolved over the course of the programme due to a number of factors, including an evolving understanding of the true need as the mission planning progressed. The periodic grid inspections during the endurance test also provided an opportunity to inspect other parts of the thruster which experience wear, and confirm that wear rates were not higher than expected. The T6 thruster design features two hollow cathodes with an identical core design: the discharge cathode is used to produce the main discharge and the neutraliser cathode is used to neutralise the ion beam. Prior to the main BepiColombo SEPS programme a number of tests had been performed to investigate the lifetime of the T6 cathodes. These tests concluded that the cathode wear mechanisms do not drive the overall thruster lifetime. One possible exception is erosion of the neutraliser tip which could result in operation in the undesirable plume mode if the selected flow rate is not suitable. Neutraliser tip erosion was addressed via a separate programme of work, in order to determine an appropriate neutraliser flow regime to apply over the course of the mission and verify that the neutraliser would remain in spot mode. Erosion modelling was used to predict the End of Mission (EOM) eroded profile, and these results were used to derive an EOM neutraliser geometry which was fitted to a thruster and characterised to verify acceptable operation. Neutraliser performance and erosion was also monitored over the course of the main thruster endurance test.

2. BepiColombo Endurance Testing Following the completion of environmental qualification tests, the SEPT EQM unit commenced its endurance test programme in the LEEP3 facility. For the majority of the test programme the thruster was coupled to flight- representative FCU and PPU. Ultimately the endurance testing was split into two main parts termed “ET1” and “ET2”. ET1 and ET2 each consisted of a number of test phases separated by periodic inspection points. Figure 3 illustrates the actual sequence of test phases performed. Testing has continued up to and beyond launch. Key events are labelled on the figure.

Figure 3. Endurance test sequence

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 Early in the test programme it was discovered that the screen grid erosion rate was higher than the earlier 2,000 hour pre-development test had suggested. This was addressed via two changes to the operating conditions:  The main flow rate into the discharge chamber was increased by ~12% across the thrust range. This significantly reduces the discharge voltage and Xe++ to Xe+ ratio, on which screen grid erosion rate is heavily dependent. Effectively the selection of main flow rate is a trade-off between ISP performance and screen grid lifetime. A review of the mission requirements concluded that the corresponding reduction to ISP could be accommodated.  The grid spacing was increased by 0.2mm, returning it to the previous spacing employed during pre- development tests. The grid spacing had been reduced in order to reduce the discharge voltage, but with the increase to the main flow rate this was no longer necessary. The grid spacing was increased to alleviate concerns that the grids could touch or be unable to sustain the high voltage conditions needed for beam operation; a concern that at the time was compounded by a lack of experimental data and hence high uncertainty regarding the operational grid spacing. During the endurance test programme the main thrust level was reduced from 145mN to 125mN. This lower thrust level was sufficient to meet the revised mission requirements and had the benefit of reducing the beam out rate during ground testing. Note that early flight operations15 have since confirmed a further beam out rate reduction in the flight environment. The 125mN thrust level was also found to be a worst case for screen grid erosion rate (as a function of accumulated impulse), attributed to a higher Xe++ to Xe+ ratio at this thrust level, hence selection of the 125mN thrust level for ground endurance testing ensures that the worst case for thruster lifetime is tested. Figure 4 summarises the approximate total impulse accumulated on each of the main sub-assemblies at the end of ET2 phase 5. The same thruster body was used throughout ET1 and ET2. The cathodes were replaced early in ET2 following a facility anomaly (due to concerns that the performance of the original devices may have been compromised by this anomaly). At the start of ET2 grid spacing was increased by 0.2 mm. The figure also shows how the mission requirement of 5.75MNs is broken down, illustrating that all the main sub-assemblies have exceeded the total impulse of the ideal mission cruise, without contingencies. Note that all four flight thrusters survived launch15.

Figure 4. SEPT accumulated total impulse

During the BepiColombo endurance test programme a number of new diagnostic capabilities were introduced in order to inform a better understanding of the T6 performance:  A co-ordinate measurement machine (CMM) was used to perform high precision measurements of the accel grid aperture profiles and downstream surface profile, and the neutraliser orifice profile.  An x-ray measurement technique was used to measure the screen grid thickness without needing to disassemble the grid set. This method calculates the thickness of the molybdenum screen grid by measuring the attenuation of an x-ray beam, noting that the carbon accel grid is virtually transparent to x-rays.  A spectroscope was introduced to observe optical emissions from the thruster. It was used to measure the relative quantity of molybdenum under different operating conditions, thus providing an instantaneous relative measure of screen grid erosion rate. In particular it was used to experimentally establish the relationship between screen grid erosion rate and main flow rate. 7

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019  A new diagnostic instrument was developed by the UK’s National Physical Laboratory which uses a frequency scanning interferometry technique to measure the operational grid spacing. The instrument is still in development but some preliminary measurements were taken and used as an input into the ion optics modelling conducted by NASA JPL (see following section).

3. Ion Optics Lifetime Modelling Ion optics modelling was used to support the grid lifetime verification programme, in order to inform a theoretical understanding of the wear mechanisms, to predict the trends over life and to inform an understanding of any differences between the wear rates in-flight and on the ground. Some initial simulations were performed using the FFX ion optics code. Further simulations were performed by NASA JPL, which are summarised in this paper. JPL’s simulations utilised the CEX2D16 and CEX3D17,18 ion optics codes. CEX2D was used to simulate the accel grid bore erosion and was also used to confirm the screen grid erosion trend based on the initial measured erosion rate from endurance testing. The code also estimated the electron back-streaming margin over the lifetime. CEX3D was used to simulate the pits and grooves erosion on the downstream face of the accel grid. The CEX models were first benchmarked using the ground test data from both ET1 and ET2 for the erosion pattern and electron back-streaming based on the operating conditions used during the endurance tests. There was good agreement between barrel erosion patterns of the accel grid simulated in CEX2D and the measured profiles during the endurance test (Figure 11). However, the depth of the barrel erosion predicted by CEX was higher, up to double that measured during the endurance test. This discrepancy is thought to be a consequence of the way re-deposition of the eroded material is addressed in CEX2D i.e. in reality more material is re-deposited than was predicted. Based on these results the erosion trends predicted by CEX2D are considered a worst case. CEX3D was also benchmarked using the measured pits and grooves erosion profile of the accel grid (Figure 6). Again the profile of the simulation was found to agree closely with the measured profile, but with CEX3D over predicting the magnitude of the erosion, similarly attributed to treatment of re-deposition.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019

Figure 5. Benchmarking of accel grid barrel erosion: a) endurance test measurements, b) CEX2D prediction

Figure 6. Benchmarking of accel grid pits and grooves erosion: Left - CEX3D prediction, Right - endurance test measurements (pink trace depicted as model prediction for hole A)

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 After benchmarking, CEX2D flight simulations (with zero background gas pressure) were performed at thrust level of 125 mN over the full mission lifetime with and without additional 50% margin, noting that the mission total impulse requirement equates to 12.76 khr at 125mN. Figure 7 shows the 125 mN predicted erosion profiles of the accel grid, showing an acceptable level of erosion, especially given that benchmarking indicated that it is likely to be overestimated. CEX2D is also able to simulate the screen grid erosion. Whilst the accel grid erosion is driven by Figure 7. Predicted accel grid barrel erosion at 125mN charge exchange ions in the vicinity of and for mission life, with and without margin downstream of the accel grid bores, screen grid erosion is driven by the plasma conditions in the discharge, in particular the discharge voltage and the doubly charged ion content. Since the latter was unknown for the revised flow rate, this input was adjusted in JPL’s discharge model such that good agreement to QinetiQ’s measured screen grid erosion rate was achieved at the start of the simulations. The simulations were then used to show that the screen grid erosion rate over life would remain approximately linear; this trend was corroborated by QinetiQ’s measurements. CEX3D flight predictions demonstrated that in flight the pits and grooves erosion profile would be less focused than during ground testing. This effect, combined with a significant reduction to the downstream ion current due to fewer charge exchange interactions when there is no background gas, results in a significantly lower rate of pits and grooves growth rate in flight. In any case the ground test results show that pits and grooves erosion is not a life-limiting wear mechanism for T6. The CEX2D code was used to predict the EBS trend over life. Figure 8 shows the predicted EBS trend in flight. It has been calculated at four different radial distances across the grid, corresponding to different screen grid zones11, with the overall trend assumed to follow whichever trend line is most negative i.e. has the lowest margin to EBS. These results suggest that the outer zones of the grids (zones 3 & 4) drive the EBS trend rather than the central zones (1 &2). The nominal accel voltage is -265V, hence there is substantial margin to EBS throughout the simulation. A number of sensitivity studies were performed to investigate the influence of uncertainties within the input conditions. In particular the sensitivity of the EBS trend to the operating grid spacing was investigated. It was found that a uniform change to the spacing effectively results in a uniform offset to the EBS trend, whilst a (non-uniform) change to the spacing profile will effectively result in a different offset for each zone. To account for the uncertainty of the operating grid spacing profile as a function of radius, QinetiQ therefore assumes that in the worst case the EBS trend could actually follow the gradient of the zone 1 trend line shown in Figure 8. The comparison of the ET2 measured EBS data to the JPL nominal and worst case flight predictions (with y-offsets adjusted to match the measured data) is shown in Figure 9. The measured trend shows some scatter but is otherwise consistent with the predicted trend. In the worst case, the EBS limit after accumulation of 1.5x the mission total impulse is still approximately 50V above the nominal operating voltage, confirming that EBS is not a life-limiting mechanism for BepiColombo.

-120

-140

-160

-180

-200

EBS EBS limit,V -220 EBS data (ET2) JPL prediction (125 mN flight) -240 JPL prediction (125 mN flight - WC) -260 Nominal operating voltage -280 0 2 4 6 8 10 Total Impulse, MNs Figure 8. Predicted EBS trend Figure 9. EBS comparison to ET2 measured data (taken when operated with FCU)

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 4. Neutraliser Lifetime Erosion modelling was performed by NASA JPL19 to predict the end of mission condition of the neutraliser cathode orifice. This analysis indicated that the tip erosion would effectively stop after attaining a ‘barrel’ shape after approximately 10,000 hours of operation. A neutraliser with an orifice representing this End of Mission (EOM) condition was manufactured by QinetiQ and tested for just over 1,000 hours. Orifice profile measurements confirmed that no additional erosion had occurred over this period, whilst spot-to-plume transition testing confirmed adequate xenon flow and keeper current margins at end of mission. During the endurance testing, measurements of the neutraliser orifice to a maximum depth of 0.9mm were taken periodically (Figure 10). Over a depth range of 0.5mm to 0.9mm the measurements show excellent agreement to the model: both the shape of the erosion profile and the erosion rate as a function of time (the measured erosion rate is slightly slower, noting that the modelling assumed operation at 145mN whilst a significant proportion of the testing was actually performed at 125mN). At a depth less than 0.5mm the measurements show development of a chamfer not predicted by the model; the chamfer stopped growing after approximately 4,000h. Despite this, the as- manufactured profile of the EOM neutraliser previously tested still provides a good approximation to the real erosion profile. The endurance test neutraliser still has margin to plume mode having accumulated about 75% of the mission total impulse, with the tip erosion measurements demonstrating that the erosion rate has significantly reduced over time, in line with expectation.

0.45

Manufactured End of Mission Orifice 0.4

0.35

0.3

0.25

0.2

JPL model 200 hr JPL model 500 hr Orifice radius (mm) Orificeradius 0.15 JPL model 1000 hr JPL model 2000 hr

JPL model 4000 hr JPL model 8000 hr 0.1 Post-ET1 phase 2 (550hrs on T6/Neut/021) Post-ET1 phase 3 (1100hrs on T6/Neut/021)

Post-ET1 phase 4 (2700hrs on T6/Neut/021) Post-ET2 phase 1 (4100hrs on T6/Neut/021) 0.05

Post-ET2 Phase 3 (6600 hrs on T6/Neut/021) Post-ET2 Phase 5 (8500 hrs on T6/Neut/021) 0 -2.1 -1.9 -1.7 -1.5 -1.3 -1.1 -0.9 -0.7 -0.5 -0.3 -0.1 Depth from down stream face (mm)

Figure 10. Neutraliser tip erosion, comparing measurements to model results

5. T6 Lifetime Capability On the basis of the lifetime verification activities described in this paper the total impulse capability of the BepiColombo SEPT is predicted to be 8.0MNs, which provides comfortable margin above the mission requirement of 5.75MNs. The extensive testing and modelling activities conducted during the BepiColombo programme have resulted in a very good understanding of the T6 life-limiting wear mechanisms. As a result of this work QinetiQ has very high confidence in the predicted lifetime of the T6.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 The T6 lifetime is currently limited by screen grid erosion. The lifetime can be significantly extended by reducing the screen grid erosion rate. There are a number of potential means by which this may be achieved; this is being actively pursued by QinetiQ as part of its product development strategy.

6. SEPT Updated Performance The SEPT performance is presented in Table 2. The values are derived from a SEPT performance model which uses data from ground testing of the four SEPT FM units in order to calculate typical performance. The performance is calculated at the thruster interface. Note that during early BepiColombo flight operations it was decided to implement an additional slight increase to the nominal main flow rate settings in order to accommodate the system flow uncertainty and guarantee that the actual flow rate in flight is at least as high as the flow rate used during endurance testing15. This approach ensures that the screen grid erosion rate during ground endurance testing, on which the lifetime prediction is primarily based, is a worst case compared to the in-flight performance. The values in Table 2 are based on the endurance test main flow rate.

Table 2. SEPT updated performance

Nominal Operating Point 75 100 125 145 Actual Thrust, mN 76.5 102.1 127.7 147.8 Thrust Correction Factor (TCF) 0.951 0.957 0.958 0.961 Total Xenon Flow, mg/s 2.10 2.69 3.27 3.78 Specific Impulse, s 3,720 3,880 3,990 3,980 Beam Power, W 2,100 2,780 3,480 4,010 Total Power, W 2,520 3,280 4,040 4,620 Power to Thrust Ratio, W/mN 33 32 32 31 Thruster Total Efficiency 55% 59% 62% 62%

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 IV.Solar Electric Propulsion System (SEPS) Qualification

A. Overview of SEPS Qualification Programme QinetiQ executed a test/qualification programme of the SEPS to allow a staged build-up of the hardware, and hence, logical test sequences to allow direct comparison between ground-support equipment (i.e. Electrical Ground Support Equipment (EGSE) and Fluidic Ground Support Equipment (FGSE)) to the real BepiColombo hardware (i.e. Power Processing Unit (PPU) and Flow Control Unit (FCU)). Specially developed diagnostics were introduced when the hardware was located outside of the vacuum chamber, which allowed characterization of key interfaces that were not possible at spacecraft level due to the limitations in the vacuum chamber performance and access during AIV activities. This proved to be a key test logic in the development of the SEPS hardware. Figure 11 shows a test logic overview of the Coupling Tests conducted by QinetiQ for the SEPS and the associated hardware list that comprised of the designated test activity. The subsequent sections provide an overview for each test iteration (red highlighted text indicates equipment installed inside of the vacuum chamber) finally concluding with the detail of the final coupling test (Coupling Test Part 2b) where a full end-to-end chain was configured inside of the QinetiQ vacuum chamber. The previous results from the SEPS Coupling Tests have been previously reported21.

Figure 11. BepiColombo SEPS Coupling Test Qualification Flow Logic.

B. SEPS Coupling Test Part 1A The primary purpose of Coupling Test Part 1A was to acquire baseline SEPT performance test data. The SEPT was operated with a Bronkhorst FGSE and SEPT driver EGSE in order to generate a performance data set for later comparison, shown in Figure 12. The Facility Monitoring System /3 performance dataset was used to H:FMS/3-SEPT/DE validate later coupling tests (Part 1B, 1C, &1D) to verify if the SEPT was SEPT/DE/2 LEEP3 H:SEPT/DE/2-FT performing within nominal H:SEPT/DE/2-Brnk H:L3/FT-EQM/TSP/Test Elec FT Support Frame P:Brnk/3-SEPT H:SEPH/EQM/BB/Flex parameters. The baseline performance Xe EQM/TSP/3a P:Reg/3-Brnk/3 Reg/3 EBQBM Bronkhorst/3 data allowed any deviation in the SEPT Xe Fluid FT SEPT

6TCs steady state or temporal parameters PT PT PT PT PT PT

Elec FT from the nominal, to be traced back to H:SEPT/DE/2-FT/SEPT/TC/a H:L3/FT/SEPT/TC/a-SEPT a specific change in the system through the staged coupling and validation Facility Monitoring System /2 process. The tests also included the measurement of backpressures upstream of the SEPP which includes Figure 12. BepiColombo SEPS Coupling Test 1A Configuration the backpressure component of the operating SEPT.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019

C. SEPS Coupling Test Part 1B The primary purpose of Coupling Test Part 1B was to verify that the SEPT performs nominally when propellant flow is supplied by the EM FCU. The test verified control algorithms and ensure FCU electrical compatibility with the Crisa Driver EGSE driver interface (specified and procured to be identical to the PPU FCU driver interfaces) and SEPT. Any incompatibility will be directly observable as steady state or temporal deviations in operating performance/operating parameters of the SEPT from the nominal parameters documented in Part 1A. The most significant of these is a flow accuracy susceptibility to backpressure in the main flow branch and a low neutraliser flow rate when supplied at the nominal inlet pressure. Backpressure measurements were made at the outlet of the FCU. The fluidic connections downstream of the FCU were identical in length to the flight configuration. The Crisa FCU Driver EGSE was not specified or built to communicate with the SEPT Driver EGSE, therefore during this test the FCU was essentially being operated in a closed loop mode in which the FCU flow control settings were operated independently of the SEPT Driver EGSE under operator control. As a result discharge initiation, thrust ramps and beam out recovery were not representative of those expected during operation with Figure 13. FCU fitted with PT’s for CT1B the PSCU/PPU, but these operations were covered in Coupling Test Part 1C/D. However, steady state and temporal performance validation at the nominal thrust levels with the FCU were representative and allowed the system performance to be compared against performance data from Part1A (SEPT Driver EGSE and Bronkhorst FGSE). This data in addition to the Part 1A data served as a baseline comparison for Coupling Tests Part 1C-1D. The preliminary analysis of this data gave confidence to proceed with Coupling Test Part 1C/D. Coupling Test Part 1B test results were very closely matched, in terms of the electrical performance of the SEPT obtained during Part 1A. This suggests that there was no reason to doubt that the EM FCU for SEPS coupling test part 1C/D (in combination with the PPU) would function correctly with the SEPT.

D. SEPS Coupling Test Part 1C/D The primary purpose of Coupling Test Part 1C was to verify electrical compatibility of the HPEPS PSCU (identical interface to the BepiColombo PPU which was currently under development at the time of this test campaign) and demonstrate that the SEPT and FCU perform nominally when coupled with and driven by the PSCU in both the nominal and redundant configuration representing the redundant switching of the PSCU from the primary DANS (Discharge Accel and Neutraliser Supply) terminals to the switched DANS terminals through the Cross-Strap Harness. Parts 1C and 1D involved the use of script files developed for command and control sequencing of the PSCU, which would later be used for in-flight operations. The tests verified that command and control data and telemetry could be uploaded and received from the PSCU, and that the performance of the SEPT remained within acceptable transient and steady state limits. A real-time 1553 Bus Monitor was developed by QinetiQ to allow interpretation of electrical parameters. Initial tests verified electrical compatibility of the various PSCU supplies beginning with static load tests, progressing to plasma loads, beam extraction and thrust Figure 14. PSCU setup for CT1C/D ramping to the initial 75mN set point. The compatibility of

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 the voltage ramps of the accel and beam supply was assessed followed by manual checkout of the initial 30mN to 75mN thrust ramp increments to ensure stability. Once confidence in the start-up regime was established automatic sequences were tested. This included the automatic start-up and ramp to the 4 nominal thrust levels. The primary outcome of the coupling test was that there was no fundamental incompatibility of the PSCU, FCU and SEPT throughout the thrust range with both the nominal and redundant harness. It was however revealed during this early coupling test, that there was not sufficient averaging/filtering of some of the electrical Telemetry (e.g. Anode Voltage) for the FDIR to correctly function. Any one sample below the voltage limit is able to reset the FDIR timer even if all other samples are above the limit. Short-term negative spikes of noise were therefore capable of resetting the counter. The FPGA was also subsequently recoded to include such averaging/filtering functionality, which were implemented onto the Bepi PPU and formed part of the qualification during Coupling Test Part 1E/F. There also exhibited a PSCU reset during a beam recovery cycle. In a beam-out, a piece of sputter or debris (nominally generated during ground testing) intercepts the gap between the screen and accel grid, which causes a short. The short circuit will cause transient currents in the beam supply modules. This reset was determined to be a combination of operating with a grounded PSCU (Bepi PPU is a floating configuration) and inherent inductances and capacitances internal to the PPU that required optimization.

E. SEPS Coupling Test Part 1E/F 1. Objectives Due to a number of issues with the PSCU discovered in Coupling Test Parts 1A-D, Part 1 E-F was devised as an intermediate risk reduction stage prior to formal requirements verification in Coupling Test Part 2. This test was conducted in a similar fashion to Part 1A-D with the PPU and FCU installed outside of the vacuum chamber within a clean tent. LEEP-3 Clean Tent H:FMS/1-PPU/DE H:PPU/EQM-FT The purpose of the Part E-F test campaign was H:EQM/Fix/X/PS.4a Intention is to design a single Single Crisa cable for both DANS #1 EQM/CSP/1 harness that suits both the PSCU & PPU testing isolated PSCU and non- isolated PPU requirements y l Facility Monitoring System /1 p ] p M to perform compatibility between the PPU, SEPT u 1 S S .

U m 3 C a [ F e / B M Q

EMPTY E and FCU as a confidence verification of the : PSCU/DE (Crisa) SLOT H

6TCs H:PPU/DE/1 (Crisa) coupled SEPS hardware. The confidence checks PPU EQM EQM FCU 6TCs

LEEP3 6TCs 6TCs were primarily limited to electrical compatibility, Interface plate HCU/1 P:HCU/1-PPU/TIP (isolated) Support Frame H:TIPC/3-PPU/TIP/TC H:L1/FT-EQM/TSP/1a H:SEPH/EQM/BB/Flex Elec FT EQM/TSP/1a 1

FDIR checks and demonstration of beam-out /

U H:TIPC/3-PPU&FCU/TC

C EQM 6TCs BB SEPT H - P:L1/FCU&Brnk/A-SEPT SEPT 3 /

C H:TIPC/3-FT/SEPT/TC P I

T TC Logger

tolerance/beam-out recovery. : H:L1/FT/SEPT/TC/a-SEPT H Elec FT 1 / U

It was also a primary objective of this test C

TIPC/3 F

- Fluid FT A / S R P

HPRS/DE/A H : campaign to operate the PPU over an extended I/P P H:HPRS/DE/A-HPRS/A PT PT PT HPRS/A A / S voltage and inlet pressure range whilst operating a R Facility Monitoring System /2 P H - 1 / g e R : P real thruster. Xe Reg/1 Bronkhorst/3 Facility Ground Xe During the test campaign, a number of flight Signal Ground representative thrust ramps were performed at a Figure 15. Test Setup for CT1E/F rate of 0.5mN/s following an initial 16 steps from beam on to 75mN. All the ramp profiles demonstrated that the profiles were achievable in terms of thruster electrical data and input power requirements. It was also demonstrated in this test campaign that automatic beam-out recoveries at 75, 100, 125 and 145mN were observed with an input power of 100V and recorded in TM. Every beam-out that was experienced (regardless of the thrust level) was automatically and, from the perspective of an operator and from TM, successfully recovered and managed. It was however noticed that beam recovery sequences at lower input voltages were not successful, which led to the successful conclusion of a modified beam out recovery sequence (sequence timing modification), optimization of clamping network parallel capacitance and further optimization of parasitic inductances inherent within the design. A further end-to-end test20, was conducted at system level, with the SEPS powered by a representative PCDU, and demonstrated compatibility of the PCDU LCLs with beam-out transients.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 F. SEPS Coupling Test Part 2a Following on from SEPS Coupling Test Part 1 E/F (incorporating modifications to the PPU) and providing the test bench for the formal qualification of the EQM PPU, this test campaign was conducted in a similar fashion to Part 1E/F with the PPU and FCU installed outside of the vacuum chamber within a clean tent (see Figure 16). The EQM PPU was fully populated with both DANS, so full verification of direct/cross strapped relays and harnesses could be verified. In addition to the qualification requirements, a number of FDIR/OBC checks and demonstration of beam-out tolerance/beam-out recovery were performed during this campaign. For the first instance, the PPU and FCU were both operated at the extended minimum operating points, and the SEPS continued to operate on both the direct and cross-strapped harness configurations with no observable change in the steady-state operating conditions during the transitions. During this campaign, the verification of the Figure 16. Control Terminals for CT2a inlet pressure ripple to the FCU by the means of the HPRS (High Pressure Regulator Simulator) was conducted, as future test activities included all of the SEPS hardware being mounted inside of the vacuum chamber, whilst the HPRS was not designed to be vacuum compatible, and therefore the pipe volume between the HPRS and FCU inlet would not have been representative. The FCU was tested at the minimum and maximum specified range, and met all of the performance targets.

G. SEPS Coupling Test Part 2b 1. Objectives The final Coupling Test executed by QinetiQ was to place all of the SEPS hardware inside the vacuum chamber and provide, as close as possible, a replica of the flight configuration. Although the harnesses and pipework used were EQM (flight equivalent) the routing and forming were not matched (for example across the TPM’s), so this is why the term ‘as close as possible’ is used. A schematic of the test setup is shown in Figure 17, which shows the SEPT, PPU and FCU all mounted inside the vacuum vessel along with the EQM SEPH and SEPP. It was during this test phase, that the remainder of the requirement verification was achieved and final validation of the in-flight operating procedures was carried out, as the outgassing and commissioning sequences would need to be followed step-by-step to inform the delivery of the SEPS user manual.

Figure 17. SEPS Coupling Test Part 2b Setup

2. Design of the hardware Accommodating all of the SEPS hardware inside the vacuum vessel whilst operating the thruster was a challenge, as the plasma environment directly around the thruster and the variation in facility pressure due to the flow of Xenon gas needed to be decoupled from the PPU to avoid any interactions and of course maintenance of a high vacuum and avoiding any partial pressures inside of the PPU that may cause flashovers. The other obvious concern was shielding the PPU and FCU from facility target sputter. To solve this challenge, QinetiQ designed and built specialist MGSE to house both the PPU and FCU, while still integrated to the thermal interface plate and cooling system that was used throughout the campaigns. The PPU and FCU were integrated into the vacuum compatible trolley, which is manufactured from stainless steel with a maximum lifting load of 200kg. The castors are fabricated from materials compatible with vacuum, including all internal bearings and lubricant. The outer frame and lifting platform is modular, so facilitate integration of the PPU and subsequent AIV activities. 16

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019

As depicted in Figure 18, the PPU was encased in a metallic box with a separate exhaust port to an additional turbo molecular pump, this ensured the PPU box was maintained well below 1x10-5mbar at all times during thruster firing.

Figure 18. Actual and CAD Representation of SEPS Coupling Test Part 2b Hardware

3. Outgassing and Commissioning Sequence To validate the SEPS UM (User Manual) and sequences used for flight operations, CT2b acted as a test bench for a single PPU/FCU and Thruster chain to run through the test logic. As soon as high vacuum was achieved (<1x10- 5mbar) the PPU naturally outgassed before turning on any of the electronics, and in particular the HV powers supplies, to avoid any internal flashovers. A series of FCU venting and purging sequences took place prior to thruster commissioning. The SEPS UM sequences operated flawlessly throughout these commissioning activities.

4. Performance Results The test campaign was operated over various input parameters (BOL and EOL setpoints, minimum bus input voltage, extended bus input voltage and minimum HPRS input to FCU) and thrust levels (75mN to 145mN) in both ‘local’ and ‘remote’ mode to obtain as wide as possible performance data to inform both the SEPS performance model and validate the potential extremes of in-flight operating conditions. During the extended operating inputs (93.7V input to PPU) and operating at 145mN, there were no changes to the thruster operating conditions, and the system maintained the demanded thrust level within specification.

The PPU telemetry collected during the tests were also used to evaluate and compare the electrical thrust stability against the levels defined in the relevant requirement, which stipulated that the stabilised thrust vector magnitude should not vary by more than 0.5mN in any period up to 1 hour. Note, for all of the tests performed, the FCU input pressure was set to minimum and the DANS input voltage of 98V was used and the statistics were calculated over a period of 1 hour. Figure Figure 19. Stabilised thrust vector magnitude @ 145mN 19 illustrates the raw, instantaneous thrust level (sampled at high frequency), and the smoothed average using a simple running average filter. It is seen that the 0.5mN thrust stability requirement is met when averaged over about 4 minutes. The AOCS samples the thrust TM at a lower frequency and is compatible with the measured thrust stability.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 The PPU telemetry collected during the EQM CT2b Remote tests were also used to evaluate and compare the mean electrical thrust against the demand. 75mN 100mN 125mN 145mN This requirement prescribes that the stabilised IB (A) 1.126 1.500 1.876 2.166 VB (V) 1843.18 1842.81 1842.53 1842.29 thrust accuracy for each SEPT (absolute RAW

Value difference between commanded and achieved TACTUAL (mN) 79.772 106.236 132.854 153.387 stabilised thrust levels, as defined) shall not TDEMAND (mN) 79.533 106.045 132.556 153.745 exceed 3mN. |TACTUAL – TDEMAND|(mN) 0.239 0.191 0.298 0.358 To this purpose, the mean actual electrical Req. |TACTUAL – TDEMAND|(mN) <3mN <3mN <3mN <3mN thrust (using the average TCF) is compared to the electrical thrust demand and the absolute Figure 20. Stabilised thrust accuracy (Remote Configuration) difference calculated. Note, all of the tests performed, the FCU input pressure was set to minimum and the DANS input voltage of 98V was used and the statistics were calculated over a period of 1 hour. The results show that the requirement of <3mN has been achieved across the thrust range on both the ‘local’ and ‘remote’ DANS.

A fully representative/automated ramp profile carried out during the test campaign demonstrates the full system capability. The profile was carried out at minimum static FCU input pressure and minimum power bus input voltage (98V). The minimum SEPS power demand, for a given thrust level, usually occurs with the thruster in thermal equilibrium. The peak power conditions, usually arise as the thruster reaches the desired thrust level and therefore the thruster initial conditions are important in determining peak power demand; the colder the engine at the moment of discharge strike, the higher the peak power demand in a ramp. Consequently, to determine the SEPS peak power demand, the engine was subjected to an extended cold soak prior to discharge initiation. The thrust ramps were carried with the system fully instrumented with scope monitoring of the DANS & BS bus currents Figure 21. Demonstration of thrust ramp at SEPS Level and voltages. (Local) Once at the desired 145mN beam current, the thrust control algorithm uses the beam current as the primary feedback parameter. Any thrust level has an associated anode current (determined by the look up table) whereas the solenoid current is variable (between minimum and maximum limits) and is used to ‘trim’ the discharge conditions and therefore the beam current. It can be noted that almost as soon as the desired anode and beam current are achieved, the magnet current begins to ramp down as the engine heats up. This also has an associated effect on the anode voltage, which also begins to drop as the engine heads towards thermal equilibrium. Again, all of the thrust ramps during this campaign were undertaken at various SEPS input conditions to ensure adequate margin over the operational range.

The PPU architecture includes a number of FDIR’s (Failure Detection Isolation & Recovery), built into the FPGA functionality to monitor and protect autonomously (i.e. without the need for OBC (On-Board Computer) control or ground intervention). It was previously determined on an earlier coupling test (Part 1C/D) that some of the FDIR was sensitive to electrical noise and the operating environment and were incorrectly tripping the system. It was therefore determined during this test campaign, that a limited number of FDIR’s were selected for test to ensure that in a representative configuration that the FDIR’s functioned as expected.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 Excess Anode Voltage Surveillance is an important parameter and FDIR as dictated by the thruster operating conditions and lifetime, therefore a simulated failure was artificially induced by reducing the main xenon flow into the thruster until the Anode Voltage limit was reached and the Magnet control loop would automatically kick in and reduce the discharge conditions. For this test, the Anode Voltage limit was set to 34.5V (limit set at the PPU interface), and as can be seen in Figure 22, the SOA (Safe Out of Area) flag is detected, and the magnet current automatically started to drop to compensate for the decrease in main flow and increase in Anode Voltage. The PPU is configured with 4 BSM’s (Beam Supply Modules). The SEPS can Figure 22. Excess Anode Voltage FDIR operate with just 3 (one of the BSM’s is for redundancy purposes only). A demonstration of this detection and artificial failure of a beam supply module was configured during this test campaign, during thruster operation. The following test served two purposes, (1) that demonstration of the SEPS operation with just 3 BSM’s is successful and (2) that the BSM temperature Surveillance detects and responds as designed. The temperature threshold for BSM1 FDIR was purposely lowered from 95°C (default) to 32°C. The results in Figure 23 shows that when the temperature of BSM1 reached the threshold, the internal timer started and the temperature continued to rise for 60 s at which point the FDIR was triggered. This is indicated by BM1T.TRIPH going to ‘1’ for a short duration. At the same time BSM1 was turned off and the remaining three BSMs increased their output power to provide continuous beam current for the demanded thrust level. The system state remained in THRUST CONTROL until the operator transitioned the DANS to STANDBY. Therefore, it was concluded that the FDIR Figure 23. BSM Over temperature and Failure FDIR operated correctly and the SEPS operated successfully with only 3 BSMs, simulating a failure case.

V.Conclusion The approach to and results of qualification of the SEPT have been described, as developed through Assembly, Integration and Verification (AIV) activities leading up to the BepiColombo mission. The development of the test strategy along with modelling and with the development of new measurement diagnostics capability has enabled QinetiQ to demonstrate the lifetime applicable to the mission requirements of the SEPT and SEPS. The SEPS Coupling Tests enabled optimisation of the PPU to SEPT interface and demonstrated system performance in a representative flight configuration. In addition it allowed verification of the procedures defined in the SEPS User Manual for flight applications. As well as providing a platform for many requirements verification, which was successfully completed throughout the phases, the test bench also allowed QinetiQ to develop a flight prediction empirical model that allowed the initial in orbit commissioning and first thrust arc to be assessed in terms of expected performance.

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019 Acknowledgments The Author would like to thank the BepiColombo teams within QinetiQ, ESA, Airbus Defence & Space (UK and Germany) for their support on the successful development and qualification of the SEPS and to JPL for the support on the modelling activities undertaken in support of the Endurance Test and lifetime verification. In addition thanks to Airbus DS (Crisa) and Bradford Engineering for their efforts in developing the PPU’s and FCU’s.

References

1Rawlin, V.K. et al, “An Ion Propulsion System for Nasa’s Deep Space Missions”, AIAA-99-4612 2Koppel, C.R., Marchandise, F., and Prioul, M., “The SMART-1 Electric Propulsion Subsystem around the Moon: In Flight Experience”, AIAA 2005-3671. 3Kuninaka, H., Shimizu, Y., Yamada, T., Funaki, I., and Nishiyama, K., “Flight Report during Two Years on HAYABUSA Explorer Propelled by Microwave Discharge Ion Engines”, AIAA 2005-3673. 4Brophy, J.R. et al, “Status of the Dawn Ion Propulsion System”, AIAA-2004-3433. 5Brophy, J.R. et al, “The Ion Propulsion System for Dawn”, AIAA-2003-4542. 6Nishiyama, K., Satoshi, H., Kazuma, U., Kuninaka, H., “The Ion Engine System for Hayabusa 2”, IEPC-2011-309. 7 H. Gray, J. Bolter, K. Kempkens, P. Randall and N. Wallace “BepiColombo – The Mercury Transfer Module” IEPC-2019- 606, 36th International Electric Propulsion Conference, Vienna, September 15-20, 2019 8 R. Lewis, J. Perez-Luna, N. Coombs and F. Guarducci, “Qualification of the T6 Thruster for BepiColombo”, IEPC-2015- 132, 34th International Electric Propulsion Conference, Hyogo-Kobe, July 4-10, 2015 9J. Snyder, D. Goebel, R. Hofer, J. Polk, N. Wallace and H. Simpson, “Performance evaluation of the T6 ion engine”, AIAA 2010-7114, 46th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Nashville, July 25-28, 2010. 10John T. Yim, George C. Soulas, Rohit Shastry, Maria Choi, Jonathan A. Mackey, and Timothy R. Sarver-Verhey6, “Update of the NEXT ion thruster service life assessment with post-test correlation to the long duration test “IEPC-2017-061, 35th International Electric Propulsion Conference Georgia Institute of Technology • Atlanta, Georgia • USA October 8 – 12, 2017 11N. Wallace and M. Corbett, “Optimisation and assessment of the total impulse capability of the T6 ion thruster”, IEPC-2007- 231, 30th International Electric Propulsion Conference, Florence, September 17-20, 2007. 12I. Rudwan, N. Wallace, M. Colletti and S. Gabriel, "Emitter depletion measurement and modelling in the T5&T6 Kaufman- type ion thrusters", IEPC-2007-256, 30th International Electric Propulsion Conference, Florence, September 17-20, 2007. 13I. Rudwan, N. Wallace and M. Kelly, “Dispenser Temperature Profile Measurement in the T5 and T6 Kaufman-type Ion Thrusters”, IEPC-2007-256, 30th International Electric Propulsion Conference, Florence, September 17-20, 2007 14H. Simpson, N. Wallace, D. Fearn and M. Kelly, “A Summary of the QinetiQ Hollow Cathode Development Programme in Support of European High Power Hall Effect and Gridded Thrusters”, IEPC-2003-0214, 28th International Electric Propulsion Conference, Toulouse, March 17-21, 2003 15P. Randall, R. Lewis, S. Clark, K. Chan, H. Gray, F. Streidter and C. Steiger, “BepiColombo – MEPS commissioning activities and T6 ion thruster performance during early mission operations”, IEPC-2019-615, 36th International Electric Propulsion Conference, Vienna, September 15-20, 2019 16R. Wirz, I. Katz, D. Goebel and J. Anderson, “Electron Backstreaming Determination for Ion Thrusters”, Journal of Propulsion and Power, Vol. 27, No. 1, January-February 2011 17 J. Anderson, I. Katz and D. Goebel, “Numerical Simulation of Two-Grid Ion Optics Using a 3D Code”, AIAA 2004-3782, 40th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Fort Lauderdale, July 11-14, 2004 18V. Chaplin, J. Polk, I. Katz, J. Anderson, G. Williams, G. Soulas and J. Yim, “3D Simulations of Ion Thruster Accelerator Grid Erosion Accounting for Charge Exchange Ion Space Charge”, AIAA 2018-4812, 54th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Cincinnati, July 9-11, 2018 19N. Wallace, I. Rudwan, I. Mikellides, D. Goebel, "QinetiQ T6 ion thruster neutraliser lifetime assessment in support of the ESA BepiColombo mission", SP2014_2980886, Space Propulsion Conference, Cologne, May 19-22, 2014. 20K. Kempkens, F. Striedter, H. Gray, S. Clark, K. Chan and N. Wallace “BepiColombo – MTM and MEPS Integration and Verification” IEPC-2019-494, 36th International Electric Propulsion Conference, Vienna, September 15-20, 2019 21S. Clark, M. Hutchins, I. Rudwan, N. Wallace, J. Palencia “BepiColombo Electric Propulsion Thruster and High Power Electronics Coupling Test Performances” IEPC-2013-133, 33rd International Electric Propulsion Conference, The George Washington University, Washington, D.C. USA

The 36th International Electric Propulsion Conference, University of Vienna, Austria September 15-20, 2019