Introduction

The past two decades have seen a strikingly emerging presence of spacecraft equipped with electric propulsion systems, and specifically Hall thruster-based systems. Various past satellite missions utilizing Hall thruster technologies reported on their on-orbit experience and even attempted to assess propulsion system performance in space. These missions are for Hall thruster technologies ranging in power from less than 100 W [1] up to 4.5 kW [2]. Those missions reported propulsion system cumulative operation duration from several minutes [3] to thousands of hours and ignition cycles [4].

Most space missions, after the spacecraft is injected into initial orbit, perform a preliminary phase that is called In-Orbit Testing (IOT) where all systems are tested for proper operation, after the launch phase. The Electric Propulsion System (EPS) runs cascading housekeeping and health checks with the goal of assessing its health [1,2,3,4,5,6,7]. These include basic PPU electrical checks, pressure and temperature sensors health checks, initial valve activation, and cathode heater startup. IOT is usually performed over the course of several earth revolutions. In the case of the DEIMOS-2 mission the IOT phase led to a re-calibration of the cathode mass flow rate [7].

Throughout the mission there is a need to monitor propellant remaining level in order to track the cumulative impulse used and remaining budget. This is conducted by using two methods – Bookkeeping and Pressure Temperature Volume (PTV) [2, 3, 8]. The bookkeeping method, as its name implies, is based on the integration of the mass flow rate over time to compute the mass used to the point in the mission. As with any integrative method, bookkeeping may accumulate any errors in the intended mass flow rate. Its advantage is that it does not rely on sensor measurements. The PTV method computes the remaining mass using the propellant tank temperature and pressure sensor measurements along with the known internal volume of the tank. This measurement can be performed in real time at any point during the mission and does not require historical mission information. None of the propulsion systems reporting on its mission experience found significant discrepancies between the two methods.

To assess propulsion system performance on orbit the Orbit Determination (OD) method is used. This method uses data of the spacecraft orbit change due to propulsion system activity. All EP systems estimating thrust on orbit used this method. Thorough descriptions of several OD methods are given in refs. [9,10,11,12]. On-ground and on-orbit performance data for a variety of Hall thruster-based missions is presented in Table 1. It can be seen in the table that some Hall thruster-based systems exhibited an improved performance in space while others had better performance on the ground. On average, it can be determined that the performance of these thrusters is approximately the same on orbit as on the ground for the cases the uncertainty was reported making those reports when uncertainty wasn't, less meaningful. It should be noted that even gridded ion engines reported on-orbit thruster performance similar to that measured on the ground [13, 14].

Table 1 On-orbit vs. on-ground thruster performance data for different Hall thruster-based propulsion systems
Full size table

In addition to performance data and regular mission operations some missions reported unexpected events. SMART-1 exhibited several unexpected shutdowns during different phases of the mission [6]. These were resolved by uploading revised software versions which reduced the number of unexpected occurrences. DEIMOS-2 experienced ignition and operation issues from the first few thruster operations. These were tracked back to erroneous cathode flow setting, resulting in a much lower cathode flow rate than required [7]. This issue was resolved by recalibration of the cathode mass flow line. The SPT-100 on-board Inmarsat-4F1 exhibited noisy discharge current during initial operations [5]. These were attributed to thruster outgassing even after propulsion system venting followed by gas purging. Saleh et al. investigated different anomalies and unexpected events with electric propulsion systems on orbit [16]. He concluded that the mean time between minor anomalies, such that do not impede system operation, is on average 116 days between all Hall thruster-based missions. Saleh also concluded that most of these are the so-called infant-minor-anomalies that are associated with the first several operations of the propulsion system.

The VENµS Program

In the recent two decades, Rafael Ltd. has engaged in electric propulsion systems research and development. To date, most efforts have been focused on low power Hall thruster-based propulsion systems [17]. One such propulsion system was developed for the Venµs mission. Vegetation and Environment monitoring on a New Micro Satellite, or VENμS, is a joint program of both the Israeli and French space agencies [18]. Within the frame of the program, two missions have been executed: a scientific mission and a technological mission.

The mission commenced with the launch of the VENμS satellite in August 2017 on-board the Vega launcher, with the In Orbit Test (IOT) phase. The space mission was planned for 4½ years. During this period, the satellite conducted a combination of scientific and technological mission phases.

The Venµs mission consisted of five consecutive phases named VM1 through VM5 (Venµs Mission X). VM1 consisted of 2.5 years at 720 km altitude, during which Venµs performed super-spectral imaging with a two-day revisit time, and the EPS performed 29 characterization experiments and controlled the orbit. The next phase, VM2, was devoted to a descending orbit transfer, solely performed by the EPS, to 410 km altitude and 1.2° inclination. Then, in VM3, Venµs imaged again with a two-days revisit time, this time with an enhanced resolution due to proximity – while the EPS constantly corrected the orbit in this higher drag environment. The next phase, VM4, consisted of a second orbit transfer, also performed by the EPS, using the remaining Xe propellant to reach 560 km. The next phase, VM5, was solely devoted to imaging in a one-day repeating orbit. Venµs has depleted all of its Xenon propellant – and the orbit is controlled by the remaining chemical propellant. All mission orbits were sun-synchronous and earth repeating.

The scientific mission has been super-spectral imaging, using a camera able to capture 12 narrow spectral bands (bandwidth between 25 to 40 nm). Its goal is to image and monitor agricultural growth, vegetation and water bodies, for environment studies. Since the super-spectral imaging relays mainly on light reflection, the imaging angles must be temporarily constant. Thus, sun-synchronous & earth repeating orbits were chosen for each Venµs phase. Hence the big challenge is controlling the orbit precisely, during all phases and altitudes. This delicate task was assigned to the technological mission and payload – the EPS, which aided by autonomous orbit control algorithms, strictly controlled the required orbit to satisfy the imaging requirements.

Technological Mission

The ‘Technological Mission’ phases consisted of the operation and qualification of the electric propulsion system in space, orbit control and orbit transfer. As its technological payload, VENµS incorporated the Israeli Hall Effect Thruster (IHET), which was developed for this purpose. The thruster was designed to fulfill the requirements of the Venµs small satellite platform of 300 kg, which can produce a relatively low power. IHET’s anode power range is between 300 and 550 W and it consumed a total of 16 kg of Xenon propellant throughout its mission. The IHET is part of Rafael’s 1st generation of Electric Propulsion Systems family, denoted as the R-400EPS.

The two main objectives of the technological mission, space verification and mission validation, were achieved by a series of experiments that tested the EPS performance, followed by five mission phases in which the EPS changed the orbit and maintained it [19].

All Venµs technological mission phase was driven by the Technological Mission Module (TMM), which is a Rafael provided flight software component, residing in the on-board computer (OBC). The TMM was specifically designed for the Venµs technological mission goals. It has three major operation modes: (1) EPS experiment test, (2) Orbit transfer, and (3) Orbit control. The TMM operates autonomously and is fed by commands initiated by the Technological Mission Center (TMC) which is the ground-based facility managing the technological mission. The remarkable autonomous operation of the TMM consists of the ability to estimate the actual orbit on-board the satellite and generate thrust firing EPS commands to correct the orbit to whatever the TMC commanded it [9].

This article describes the operation of the EPS on-board the Venµs satellite mission. First, we present all EPS components and their respective role in the system. We then overview on-orbit mission activities which were divided into five phases. Subsequently, we present propellant consumption calculations using two separate methods, as well as the assessment of propulsion system performance using Orbit Determination (OD). Lastly, we discuss unexpected events such as unplanned thruster shutdowns or several failed ignitions attempts, and the methods used to overcome these while in space.

We emphasize that the focus of this article is on the propulsion system performance and operation on-orbit and on the ground. As such, the article does not discuss the specific orbit determination method used to assess thrust and thrust uncertainty. These are assumed to be given using a representative orbit determination method.

VENµS EPS

System

The Venµs EPS (see Refs. [18, 19]) is designed to support and operate two IHET-300 thrusters, one at a time. Besides the thrusters, the EPS main components are the Propellant Management Assembly (PMA), the Digital Xenon Flow Controller (DXFC), the Power Processing Unit (PPU), and 2 electrical Filter Units (FU). The propellant tank stores highly pressurized Xenon. The system architecture is presented in Fig. 1, and the Venµs propulsion plate is shown in Fig. 2.

Fig. 1
figure 1

Venµs EPS schematic block diagram

Full size image
Fig. 2
figure 2

Venµs propulsion plate

Full size image

One notable feature of the EPS is the ability to throttle the anodic power of its thrusters, according to the instantaneous available power supplied by the bus platform in orbit. This proves to be an essential feature in a LEO satellite, since the power levels produced by the solar arrays are variable and change according to satellite’s true anomaly angle in each revolution. The power level is automatically computed by the TMM and commanded to the PPU which in turn regulates the propellant flow and operates the thrusters at the commanded power level (see Ref. [19]).

A Thruster Selection Unit (TSU) is responsible for selecting the active thruster, as decided by the TMM and commanded by the PPU.

IHET-300 Thruster

The heart of the EPS is the Hall Effect Thruster (HET), codenamed IHET-300, shown in Fig. 3. It operates on Xenon, which is ionized by electrons emitted from the cathode and accelerated in the form of plasma via a high electric field.

Fig. 3
figure 3

HET-300 Flight Model (FM)

Full size image

The main characteristics of IHET-300 are listed in Table 2 for the End of Life or EOL. This thruster, which employs low power, is specifically designed to be used onboard small and micro satellites. Its useful range of operation is between 250 to 600W. On Venµs platform we limit the operation power to 550W, due to platform power limitations. Thus, it can utilize the instantaneous available power from the satellite as explained above.

Table 2 IHET-300 Main Characteristics @ 300W
Full size table

The IHET-300 mass is about 1.5 kg and its dimensions are 170 × 120 × 90 mm.

Propellant Management Assembly (PMA)

The PMA is responsible for feeding Xenon gas, at controlled pressure and flowrate, to the selected thruster.

It consists mainly of adapted and qualified COTS components. Ref. [5] describes the development and qualification process of most EPS components. A picture of the Digital Xenon Flow Controller and functional schematic are shown in Fig. 4.

Fig. 4
figure 4

Digital Xenon Flow Controller (DXFC): Picture of the Flight Model (left) and block diagram (right)

Full size image

The PMA consists of two (2) pressure regulators connected in series. The pressure regulation works in an open loop mode pre-calibrated on the ground, while the first regulator reduces the tank high pressure feed to nearly the operating pressure and the second regulator gives a more accurate output of 1.8 bara. This setup provides a very accurate pressure at the output of the pressure regulators.

The Digital Xenon Flow Controller (DXFC) (shown in Fig. 4) uses a set of six (6) valves and flow restrictors arranged as a gas manifold to control the total Xenon flow through the anodes; therefore, controlling the discharge power and thrust levels. This flow controller setup is fully redundant in the case some of the valves fail to open. The DXFC controlled by the PPU would select a configuration of valves to maintain the required discharge power within a tolerance of 20W, the control loop checks the discharge power and if required it updates the DXFC valve configuration. Only one controller was installed on board the EPS while the flow was directed to each thruster by opening a series of valves downstream near the thrusters.

PPU

The PPU contains the diverse power supplies and control circuits for operation and management of primarily the thrusters and the whole EPS components. The Sequencer and Control Unit (SCU) is the control module of the PPU. Besides controlling the power supply and the ignition sequence, the PPU also implements communication with the satellite OBC, monitors all signals and prepares the EPS telemetry.

The PPU block diagram is shown in Fig. 5 herein with its main interfaces to the EPS sub-assemblies. The PPU is thermally insulated from the satellite and its radiator on the bottom side is exposed to space for cooling by heat radiation.

Fig. 5
figure 5

Left: PPU block diagram; Right: Venµs PPU Flight Model (FM)

Full size image

Figure 5 also shows the PPU flight model, while Table 3 shows its main characteristics.

Table 3 PPU main characteristics
Full size table

On-Orbit Operation

After the launch of Venµs satellite, a series of In-Orbit Tests (IOT) were performed. These tests checked and assured that the platform and its payloads were ready for the intended mission. The technological mission experiments started as planned in early 2018 and continued until the end of VM1. All other mission objectives were fully accomplished, and all propellant was depleted in 2022.

Overview of electric propulsion system operation

The IOT consists of a large series of trials in which an additional function is checked in turn until a full EPS activation is achieved, including uplink communication and download of telemetry. The success of the IOT is a prerequisite for the next EPS mission phases. Table 4 summarizes the main tests performed.

Table 4 In-Orbit Testing (IOT) set of checks and tests
Full size table

After successful completion of the IOT, the EPS was ready to commence its technological mission execution in each of the five Venµs phases, as shown in Table 5.

Table 5 Summary of the EPS utilization at different mission phases
Full size table

During VM1, the satellite performed each month a technological mission experiment. The goal of these experiments was to estimate the thrust performance of each Thruster Unit (TU). For this purpose, the TMM commanded the EPS a predefined series of firing pulses, that would change the orbit by increasing and decreasing the satellite’s Semi-Major Axis (SMA). The whole maneuver consisted of orbit extrication in three Open Loop (OL) segments and finished with orbit re-insertion in a Closed Loop (CL) segment, as shown in Fig. 6.

Fig. 6
figure 6

Illustration of a typical orbit change set of maneuvers test during the VM1 phase

Full size image

Each OL segment consists of three satellite revolutions, with a one long EPS activation at a constant power between 300 and 550W (meaning constant thrust) for each revolution. In the first segment, EPS activations raised the orbit by aligning the satellite's thruster orientation with the velocity vector. The second and third segments also contained EPS activations for three consecutive revolutions each, but the satellite attitude was commanded by the TMM so that the orbit Semi-Major Axis (SMA) will decrease. At the end of each segment, the EPS was left idle for one revolution, dedicated to orbit measurements and estimation. The overall design goal was to finish the three segments in an orbit lower than the initial one. Subsequently, the CL segment commences. The TMM in CL mode commands the EPS to fire and operate on each consecutive revolution for a duration and altitude instantaneously calculated, until the orbit reached the target value. The orbit data telemetry is then analyzed by the TMC, which estimates what was the force applied that resulted in these orbit changes. Then, all parasitic forces are compensated, resulting in the average thrust actually applied by the TU.

During this VM1 phase the EPS was well characterized in terms of operational parameters (mass flow rates requirements, anode current and anode current noise measurements, system various temperatures, etc.), ignition sequence, downlink and uplink etc., and some parameters calibrations were made through the uplink to the satellite such as mass flow rate to the anode during the ignition. This phase ended on October 15th, 2020, with a total of 29 experiments, 380 ignitions, 86.05 operating hours and a total calculated impulse of 6.7kNs accumulated by the EPS. During this phase of characterization, the EPS was activated across all its discharge power range, which is 300W to 520W. The average firing duration was 20 min for each thruster including the startup ignition sequence.

During VM2, the satellite was commanded to make an autonomous orbit transfer from a 720 km SSO to 410 km SSO. The maneuver started on November 4th, 2020. During this phase, the EPS thrusters were activated in 90% of the revolutions in each day in which TU1 was operated for approx. 25 min and TU2 was operated for approx. 15 min accumulating approximately 40 min of thrusters operation at each revolution. During the remaining revolutions telemetry was downloaded, and additional checks were performed on other satellite subsystems. At each revolution the EPS operated both thrusters to lower the altitude of the satellite. To do so, after one thruster completed its burn, the satellite was rotated about 180° to align the second thruster with the correct velocity vector direction. After several tens of revolutions two significant phenomena were observed, the first is that the maximum continuous discharge power that the EPS delivered was 420W. An investigation was performed on this matter which concluded that the cause of this phenomena was the thermal dissipation on the EPS mounting plate and the mass flow controller, affecting its inlet conditions, therefore reducing the effective mass flow rate. One possible solution to this matter was to use the redundancy of the DXFC to compensate the lower mass flow rate to the anodes. However, the implementation of this solution required some on-orbit testing to revalidate operation, resulting in delaying the scheduled maneuver. At this stage it was decided to limit and leave the maximum discharge power command to 420 W for the VM2 phase in order not to disturb the mission's maneuver schedule.

The second phenomenon was that some beam-out events were reported via telemetry, mostly immediately after thruster ignition. Beam-out events are events in which the discharge spontaneously extinguishes. At this point, it was decided to stop the autonomous maneuver and keep activating the EPS manually using direct scripts which were uplinked to the satellite. To mitigate the beam-out events some tests were performed at Rafael's space propulsion facility resulting in an update of the magnetic field parameters and a change in the ignition routine to start the thrusters with no magnet current, that is also referred to as "Glow mode" ignition. This event proved the benefit of project risk management that was conducted during the development phase, which resulted in two redundant and optional modes of TU ignition and operation. The target orbit of 410 km was reached on September 5th, 2021, with a total of 4,077 ignitions, 1,382.5 operating hours and a total calculated impulse of 110.6 kNs. A total of 23 beam-out events were registered.

The main objective of VM3 in terms of the EPS was to provide autonomous station keeping maneuvers which enable the scientific payload to image with a better resolution than in the previous orbit. This was achieved thanks to the shorter distance to Earth and at the cost of more frequent station keeping maneuvers. By performing this phase, we demonstrated, the feasibility of enabling enhanced imaging resolution without increasing the size and mass of the satellite’s camera. Venµs satellite maintained this orbit for a total duration of one month, while four station keeping maneuvers, each consisting of 9 thruster operations, were required to overcome the higher drag at this altitude. The autonomous module performed a total of 36 thruster ignitions throttling the discharge power between 300 and 480W, accumulating 11.8 thruster operating hours. By the end of this phase of the mission the EPS accumulated a total of 4,113 ignitions, 1394.5 operating hours and a total of calculated impulse of 111.6 kNs.

After VM3 was completed, at altitude of 410 km, VM4 phase started, the EPS was commanded to raise the satellite's orbit to a 560 km SSO and ended at VM5 after a successful orbit insertion at this altitude to continue the scientific mission. The VM4 maneuver started on November 9th, 2021, using the same technique as in VM2, but this time it was manually commanded by the Ground Station, on open loop, and ended with the depletion of the Xenon propellant on March 5th, 2022.

Overall, throughout the mission, and until all propellant was consumed, the EPS achieved a total of 6,311 successful ignitions accumulated 2105.1 operating hours, that is 158 kNs and a total throughput of 16 kg of xenon.

Propellant consumption

During the propellant loading campaign in 2017, the propellant tank was pressurized with 16.1 kg high purity Xenon. As of the end of mission, the propellant was entirely depleted. Subsequently and throughout all mission phases, calculations were made to determine the propellant used and remaining. Two conventional methods were used to determine the remaining propellant mass at any time during the mission: (1) PTV method and (2) Bookkeeping method [2, 3, 8]. For the first method, periodic telemetry provides the pressure and temperature data of the Xenon tank. Using this data, the pressure dependent tank volume and basic thermodynamic state tables, the Xenon propellant mass was calculated (Eq. 1).

$${m}_{p}=f\left(T,P,{M}_{w}\right)$$
(1)

where \({m}_{p}\) is the propellant mass, \(T\) and \(P\) are the propellant temperature and pressure, respectively, in the propellant tank, and \({M}_{w}\) is the atomic mass of the propellant. For this mission thermodynamic properties of Xenon provided by the National Institute of Standards and Technology (NIST) were used [20]. This method can be used at any point during the mission and regardless of previous propellant mass estimations used earlier in the mission.

On the other hand, the bookkeeping method is integrative. It integrates the mass flow rate injected by the propulsion system over the operation time to compute the amount of propellant expelled. This amount is then subtracted from the initial propellant mass prior to the respective operation (Eq. 2).

$${m}_{p,i}={m}_{p,i-1}-\dot{\text{m}}\Delta t$$
(2)

where \({m}_{p,i}\) is the propellant mass at the specific maneuver, \({m}_{p,i-1}\) is the propellant mass calculated after the previous maneuver, ṁ is the total thruster and cathode mass flow rate, and \(\Delta t\) is the maneuver time duration. In such a method the error originates from the error in the estimation of mass flow rate and time duration. As such, the error in the bookkeeping method is \({\Delta m}_{p,i}={\sum }_{j=1}^{i-1}{\Delta m}_{p,j}\)

The estimation of the mass flow rate injected into the thruster and cathode were calibrated using mass flow meters before satellite integration. After several measurements of the propellant mass during the qualification phase of the propulsion system, it was estimated that the error in mass flow measurement is 5% of the applied value. This method has the disadvantage of accumulating the computation error from each operation to the next. Additionally, this method relies on the initial calibration of the mass flow controller prior to installation on-board the satellite and may accrue any mass flow calibration errors.

In addition to the two methods, we used a third method based on laboratory experiments on an identical thruster unit. We operated an identical IHET-300 Flight Model (FM) thruster to find the correlation between the thruster discharge power and the mass flow rate. The background pressure during the test was between 1.8 × 10–5 Torr and 2.9 × 10–5 Torr (N2 corrected). We then used the ṁ → Pd relation to assess the mass flow rate used during each thruster operation in space and based on the operating anode power.

Figure 7 shows the propellant consumption with the cumulative impulse of the propulsion system. Propellant consumption is approximately linear, even though the activation time for each thruster varied throughout the mission. This does not come as a surprise as the slope of the trend shown in the figure is proportional to the inverse of the specific impulse. If the specific impulse is approximately constant so will be the slope. It can also be seen that all three methods produce approximately the same result, albeit an increasing deviation from each other as the mission progresses. This increasing deviation is most likely due to a cumulative error due to a slight underestimation of the expelled propellant during all activations. Nevertheless, both the PTV and Bookkeeping methods agree, within their respective uncertainties for all data points, except for at cumulative impulse of ~ 55 kNsec. We can assume that for longer missions, where the total impulse is much higher than in this mission, the accrued error in the Bookkeeping method will overpass the error in using the PTV method.

Fig. 7
figure 7

Propellant consumption as a function of the cumulative impulse throughout the mission using three methods – Bookkeeping, PTV and Bookkeeping based on lab test with an identical thruster

Full size image

It is interesting to note that the error associated with the “Bookkeeping Method” increases with EPS activation as discussed above. This is due to the integrative nature of this method, that relies on the previous assessment of the remaining mass when computing the propellant mass at a given time. Since the temperature and pressure of the DXFC has changed throughout the mission, it was necessary to correct the calculation of the mass flow controller according to the temperature and input pressure. The DXFC was calibrated on the ground at pressure and temperatures of 1.8 bara and 20 °C. On orbit, the input pressure was measured by the low-pressure transducer located upstream of the DXFC. The value of the pressure varied during firing due to temperature changes. The temperature was measured using a thermistor located on the body of the DXFC. Since the DXFC is located close to the PPU it was subjected to heat flux that raised its temperature. Figure 8 shows DXFC temperature during a five-day long EPS activations cycle (tens of thruster activations) at VM2 (second mission phase). It can be seen from the figure that the DXFC temperature increases from ~ 22 °C to 28 °C during each activation cycle. The corresponding correction was made to assess the mass flow rate deviation from the calibrated values. In any case, the PPU operated the EPS in a closed-loop manner, while adjusting the mass flow rate to match the available/commanded power.

Fig. 8
figure 8

Temperature of the DXFC during VM2 (phase 1) EPS operation. Three operation cycles are captured

Full size image

The errors associated with the “PTV Method” are different in nature since this method relies on real-time pressure and temperature measurements. The uncertainty in using the PTV method originates from the uncertainty in the measurement of the pressure gauge and thermistor. To compute the uncertainty in propellant mass remaining in the propellant tank the mass was computed for the high and low margins of the measured temperature and pressure. The challenge with the “PTV Method” for Xenon is that Xenon density is very sensitive to changes in pressure in the 60–80 bar range, at temperatures around 25 °C. This can be observed in Fig. 9 where propellant mass is plotted against tank pressure at the measured tank temperatures (~ 25 °C). Small variations in pressure may lead to large variations in measured propellant mass. For this reason, propellant mass uncertainty can go up to over 0.5 kg during the intermediate phases of the mission when the propellant mass is approximately half of the initial loaded mass. The “PTV Method” errors presented in Fig. 7 are based on measuring device error of 0.6 bar for the high-pressure gauge and 0.2 °C for the temperature measurement by the propellant tank thermistor.

Fig. 9
figure 9

Propellant mass as a function of tank pressure at the measured tank temperatures (~ 25 °C). The horizontal errorbars represent the 2% pressure gauge reading uncertainty

Full size image

On-orbit performance estimation

Propulsion system performance was estimated on-orbit using Orbit Determination (OD). To use this method the orbit of the satellite was determined using GPS location before a satellite maneuver was performed. Subsequently, the propulsion system was operated, and the new orbit was determined after the maneuver was completed. This, and similar OD methods, were used in previous Hall thruster-based missions and are thoroughly described in refs. [9,10,11,12]. It should be noted that the authors of this paper are no professionals in orbit determination and are unfamiliar with the details of the specific method used to determine the propulsion system thrust from the orbit changes performed during the Venμs mission. The propulsion system thrust assessment was conducted by the project Orbital Mechanics group, and is not the focus of this paper. The thrust, and thrust uncertainty, values reported in this mission overview paper were obtained using an OD method that is a good representative of OD methods as listed in Table 1.

In addition, an identical Hall thruster FM was operated in a vacuum chamber at Rafael to measure its performance and compare to the on-orbit data. The identical FM was at its beginning of life and only operated during its acceptance test previously. To adequately compare thruster operation the lab test operation was conducted for 20 min, approximately the same duration as its on-orbit counterparts. Background pressure during lab operation was between 7.2 × 10–6 Torr and 1.16 × 10–5 Torr (Xe corrected). Past laboratory acceptance tests with the units operating in space were deemed unreliable due to exceedingly high chamber pressure.

The thrust stand used to record thruster performance is a swinging arm type pendulum thrust stand with an uncertainty of 3% of the measured thruster. Thrust stand displacement is monitored by optical means and reduced to zero using a magnetic force actuator. Each thrust measurement is recorded by recording the thrust reading while the thruster is firing, turning off the thruster, and recording the zero-thrust point while the thrust is off and gas flow discontinued. The thrust stand is calibrated by using a force actuator emulating a set of known forces onto the stand at the location of the thruster.The on-orbit and lab thrust data are presented in Fig. 10. It can be observed in the figure that the thrust increases approximately linearly with discharge power for on-orbit and lab operations. It can also be observed that the thruster operating in lab environment measured higher thrust values than both on-orbit thrusters by approximately up to 5% on average, although within the measurement uncertainties. We can attribute this to the difference in background pressure, although the background pressure during the lab test is considered sufficient for the estimation of performance in space [ref. [21]]. Also, ‘Thruster 2’ exhibited higher performance than ‘Thruster 1’ by about 5% yet within thrust assessment error bars.

Fig. 10
figure 10

On-orbit thrust estimation using the Orbit Determination (OD) method for maneuvers performed by each thruster and measured thrust in the lab using an identical thruster at the beginning of life after operation of 20 min

Full size image

Another set of laboratory experiments were conducted to understand thruster operation differences at the Beginning of Life (BOL) and End of Life (EOL), as well when it is considered operating cold (after 20 min of operation) and when it is considered hot (> 2 h operation). To do this another laboratory thruster was used that completed the lifetime qualification test of 130 kNsec. The qualification thruster is considered at its EOL. Experiment results are presented in Fig. 11. It can be seen in the figure that there is little discernable difference between “hot” and “cold” operations of the thrusters in lab, whether the thrusters are at BOL or EOL. It can also be seen that the EOL thruster performance is slightly lower than that of the BOL thruster. We believe that this behavior is attributed to the increased inner thruster volume due to ceramic channel erosion. The greater volume at the thruster EOL causes a reduced density within the discharge channel, that in turn, may lead to lower ionization, thus lower mass utilization efficiency, as previously observed with low power Hall thrusters [22].

Fig. 11
figure 11

Thrust vs. discharge power as produced by two IHET-300 Hall thrusters in lab, at the Beginning of Life (BOL) and End of Life (EOL) after 20 min and after 120 min of operation

Full size image

Note that EOL thrust data on-orbit is unavailable since orbit determination was performed only at the beginning of the mission during the first phase (VM1).

Mission challenges and solutions (On-Orbit Troubleshooting)

As previously mentioned, some difficulties were encountered with EPS operation during In-Orbit Testing (IOT) and VM2 where the satellite descended from 720 to 410 km. For each issue that arose the team in charge of the EPS worked together with the mission operation team to troubleshoot the issue in a prompt manner. Some issues only required tracking of the EPS parameters, and some required a more extensive investigation to reach an optimal solution. A short review of the main issues is presented in this section.

Ignition difficulties during IOT

Before the system was delivered to the satellite a long End-to-End firing campaign was performed in which the whole propulsion plate was installed on a dedicated mounting fixture installed in the vacuum test facility at Rafael. During this test all the EPS functions and sequences were tested. In particular, the anode mass flow rate parameters for ignition were determined.

During IOT the same test was repeated in space. It was found that the discharge current was lower than expected after anode breakdown resulting sometimes in spontaneous beam-out after a few seconds of operation. Some tests were performed on orbit to determine the source of this issue. Thanks to the flexibility of the EPS and particularly the PPU, operational parameters were uploaded in each test to evaluate their effects on the EPS ignition ability. Some examples for parameters testing were magnet current during ignition, cathode mass flow rate, anode mass flow rate, valve timing sequences etc. The first solution was to initiate anode breakdown with no magnetic field. Although this solution succeeded in solving the ignition problem, it was discarded due to the high currents observed on the discharge current circuit which may cause a PPU shutdown trigger. After several trials the most acceptable solution was to increase the mass flow rate through the anode during ignition and reduce it to nominal values in the subsequent seconds.

Limited Maximum Discharge Power

During AIT and VM1 the thrusters were operated only for a short number of revolutions, and the maximum discharge power was 520 W.

When VM2 started the EPS was operated intensively to execute the orbit change. Each thruster was operated on each revolution and for a longer duration per operation. It was observed that the maximum continuous discharge power that the EPS delivered was 420 W when the command was 480 W. When telemetry parameters were investigated it was found that the mass flow controller (DXFC) reached its maximum nominal output, meaning that the discharge power control loop was functioning as intended and the issue was most probably hardware related. The respective investigation concluded that the cause of this phenomenon is the thermal dissipation on the EPS mounting plate and the mass flow controller, affecting its thermal inlet conditions, therefore reducing the effective mass flow rate through the DXFC mass flow restrictors (see Fig. 8). A possible solution was to open the sixth DXFC valve (redundant valve) to compensate for the lower mass flow rate to the anodes. However, the implementation of this solution required some on orbit testing, resulting in delaying the scheduled maneuver at this stage. In coordination with the mission operation team, it was decided to limit the maximum discharge power command to 420 W for this phase in order not to disturb the mission's maneuver.

Beam-out after ignition

After VM2 started some beam-out events were reported via telemetry, mostly immediately after thruster ignition and mainly on Thruster Unit 1 (TU1). At this point, it was decided to stop the autonomous maneuver and keep activating the EPS manually using direct scripts which were uploaded to the satellite. To mitigate the beam-outs events some tests were performed at Rafael's space propulsion facility. An identical Flight Model (FM) IHET-300 thruster was fired to accumulate the total operating time and the total accumulated ignitions performed already with TU1 in space, to simulate as close as possible the thruster operation history.

The ground testing at this step did not reveal any issues and the beam-out phenomena was not recreated during the lab test. At this point other tests were performed with the goal of treating the symptoms to facilitate uninterrupted thruster ignition and thruster operation. One possible option was to reduce the intensity of the magnetic field so to reduce discharge resistivity and to facilitate a spontaneous reignition in the case of a quick beam-out, while taking advantage of the fact that the keeper-emitter circuit could be kept on during thruster operation. A test of magnet current sensitivity was performed and thrust, and anode current oscillations were measured. Figure 12 presents thruster discharge current oscillations, main frequency and thrust as measured in the lab. The data shows that regardless of the magnetic field current, instabilities are below 15% of the discharge current and with frequencies in the ~ 10 kHz range. These frequencies are well above the cutoff frequency of the PPU filter. The thrust results show the decrease with magnetic coil current. This decrease can surpass 10% when the magnet current is below 1.5 A. This investigation affirmed that the magnet current can be decreased by 25% of its designed value with little consequences to thruster performance and current oscillations.

Fig. 12
figure 12

Relative discharge current instability STDV, instability frequency and thrust as a function of electromagnet current as measured in a lab test. The nominal magnet current is 2.5 A

Full size image

Thanks to the investigation, updated magnet current parameters were uploaded to the satellite. In addition, it was decided to increase the cathode mass flow rate during TU ignition and TU steady state operation. Lastly, the ignition sequence was updated to ignite the thruster in "Glow mode", that is with no magnet current during anode circuit breakdown. To do so, and to reduce the risk of the high inrush currents on the discharge current circuit that may cause a PPU shutdown trigger, the four anodes power supply modules were turned on consequently to reduce the total output capacitance at the anode output line during circuit breakdown.

Conclusions

Venµs is a sophisticated mission where the propulsion system was responsible for a variety of complex tasks, from orbit maintenance to orbit transfer (up and down) and thrusters’ characterization. The technological mission operated the EPS to demonstrate LEO mission enhancement capability and validate the EPS in space.

The EPS consisted of two throttleable Hall thruster, PPU, Propellant Management Assembly (PMA), and a 9-L propellant tank carrying 16 kg of Xenon. The system was capable of throttling operation in the 300–600 W discharge power range.

The EPS successfully passed its In-Orbit Testing (IOT) phase and proved to operate properly according to specifications and expectations. The subsequent mission phases included orbit maintenance demonstration, 720 km → 410 km orbit transfer, orbit maintenance at high drag environment, and 410 → 560 km orbit transfer. Throughout all mission phases the EPS operated within the designated power range, yet most of the time under discharge power of 450 W.

Two methods were used to compute the remaining propellant mass – “Bookkeeping” and “PTV” methods. Both methods proved successful, agree to within their respective uncertainties, and corresponding measurement errors were computed for each method.

Thruster performance was measured on-orbit, using the “Orbit Determination” method and compared to laboratory experiments conducted on the ground with identical thrusters, one at its beginning of life, while the other after it past the lifetime test of 130 kN-sec. The measured performance on-orbit was found to be on average approximately the same as the performance measured on the ground within the measurement uncertainties. Additional tests were conducted to assess performance variation between BOL and EOL thrusters, as well as a cold (< 20 min) and hot (> 120 min) operation.

Lastly, we presented several repeating events in which the propulsion system suffered from sudden beam-outs or ignition difficulties. The methods used to assess a possible solution and implement it were presented.

Although designed as a fully redundant EPS, until the end of the technological mission there was no need to switch any redundant component or sub-system. All initial assigned components and branches – completed the mission successfully.