1 Introduction

In space applications, maintenance is either impossible or difficult, and development expenses of mechanical systems are high. The durability and reliability of each component is critical to mission success. Additionally, moving parts have the drawback of being prone to wear and tear while having inherent friction, often necessitating an inevitable need for maintenance. Thus, mechanical parts have a finite lifespan and are prone to failure. Surfaces in direct mechanical contact will inevitably deteriorate. This leads to a hard trade-off between very resource-consuming reparations or forfeiting the lifetime of the components and accepting mission risk.

While this fact is valid for any mechanical system, it is severely amplified for space activities. Furthermore, this fact is unavoidable for any mechanical part in a satellite, but especially in the Attitude and Orbit Control System (AOCS) system. Conventional satellites use in most cases traditional reaction control systems for the basic generation of inertial torque. This category mainly includes the RWs, as well as control moment gyros. Both are based on a rotating disk, which should have a high Moment of Inertia (MoI) I along its main axis of inertia, the rotation axis. These disks are mechanically supported and accelerated or decelerated by electric motors to generate internal torque or store angular momentum in the system. This is indispensable for the attitude alignment of satellites, as well as the compensation of disturbances, especially constantly occurring disturbances such as solar radiation pressure or caused by the non-uniformity of Earth’s gravity field. However, these state-of-the-art mechanical RWs, like any mechanically moving system, have the problem of wear and tear and inherent friction. This mechanical component susceptibility has previously been exposed in several space missions in the past. For instance, the space missions FUSE, Dawn, and Hayabusa, as well as the Kepler space telescope, all have the trait that at least one of their reaction wheel bearings have failed [1]. Other examples include the Neil Gehrels Swift Observatory, which went into safe mode in 2022 due to a malfunction in one of its reaction wheels [2, 3]. Another indication of the seriousness of this issue is the amount of effort spent in the space sector on measuring and classifying the reliability and lifetime of components. An example of this is the European Space Agency (ESA) product assurance data source [4].

Therefore, it is a desirable objective to design systems with a minimum amount of mechanically movable parts. Utilizing ferrofluids, or liquids responsive to magnetic fields, is one such strategy for applications requiring simplicity and reliability, i.e., space. By controlling a ferrofluid with magnetic fields rather than actuating mechanical moving components, it is possible to minimize the number of moving elements [5,6,7,8].

This paper will describe a mechanics-less RW alternative with a high potential for future space application. In this design, ferrofluid cushions reduce the wear and tear of the system.

1.1 The FARGO Project

In 2017, the Space Agency within German Aerospace Agency (DLR) launched the first Überflieger student competition, which enabled students from Germany to submit an experiment proposal. The selected winners were allowed to develop their experiment idea within certain project budget constraints and test it in the form of a micro-payload on the ISS. One of the winners of the first “Überflieger” project was the Small Satellite Student Society–KSat Stuttgart e.V.—with the support of the IRS with their Pump Application using Pulsed Electromagnets for Liquid reLocation (PAPELL) project. The PAPELL experiment showed a successful first proof of concept demonstration of ferrofluid manipulation in the ISS microgravity environment for future space applications. However, the manipulations were limited to simple droplet fluid movements without direct applications as an alternative component [5,6,7,8]. As a direct successor, the Überflieger 2 competition was initiated by the DLR and the Luxembourg Space Agency in late 2020 with Yuri GmbH, SpaceTango, and the DPG (German Physical Society) as co-organizers and technical advisors. For this competition, KSat e.V. submitted the FARGO project idea, which was selected as the winning proposal alongside 3 other teams. The winners will have the opportunity to realize their experiments in the form of a micro-payload of \(20 \times 10 \times {10}\,{\textrm{cm}}^{3}\) with a project budget of \({20\,000}\) €. The experiments will be operated in the microgravity environment of the ISS for at least 30 days in March 2023. The experiments must be largely autonomous and are operated without the intervention of astronauts [9]. The FARGO project developed and tested three different proof of concepts for mechanically-reduced technologies with high potential in application for future spaceflight in the microgravity environment of the ISS [10,11,12,13].

This paper will focus on the Attitude Control System (ACS) experiment, which is called Ferrowheel as it is based on a stator of a brushless DC motor combined with a rotor on a ferrofluid bearing. Hence, the development of this ferrofluid concept from the basic idea and Earth-ground proof of concept to the design and integration of a final working prototype for in-space verification will be described.

2 Theoretical Background

In this section, a brief background on ferrofluid actuation and its relation to classic and satellite kinematics will be given.

2.1 Ferrofluid Movement

Designing low-wear systems is made possible by removing mechanical connections, which can be replaced by ferrofluid-based systems.

Ferrofluids are a suspension of paramagnetic nanoparticles of a size less than \({10\,\mathrm{\text {n}\text {m}}}\) inside a carrier fluid, in most cases oil. Within this study, EFH-1 from the Ferrotec EFH Series was chosen due to its chemical stability, low viscosity, and high magnetizability [14]. The nanoparticles experience a strong form of paramagnetism. Their very high magnetic susceptibility allows them to be rapidly magnetized by an external magnetic field. The paramagnetic particles are moving freely inside the fluid as long as no external magnetic field is applied. However, paramagnets are always attracted to the strongest point in a magnetic field, so the particles will be uniformly magnetized when an external magnetic field is applied. If a sufficiently strong external magnetic field is applied, these particles align in accordance with the magnetic field lines, thus creating the macroscopic visible magnetic reaction of ferrofluid, as shown in Fig. 1 [11, 15, 16]. A time-varying magnetic field must be applied to the ferrofluid to utilize this motion.

In this work, a stator coil of a Brushless Direct Current (BLDC) motor is used for this purpose. An outrunner BLDC motor consists of an inner stator coil segment and an outer rotating part resembling a ring of permanent magnets. The stator coils generate a time-varying, rotating magnetic field by applying an alternating 3-phase current. Thus, the outer part starts to rotate—the rotor. Such COTS stator coils are used in this work as the basis for moving ferrofluid-bearing systems through an alternating magnetic field.

Fig. 1
figure 1

Movement of a ferrofluid droplet from one electromagnet to another by an applied external magnetic field

Full size image

2.2 Rigid-Body Rotational Motion

Considering the rotational movement of a frictionless rigid body, one sees that the angle \(\phi\) is decisive for the definition of the angular position with respect to the chosen coordinate system, as visualized in Fig. 2. For this description, the center of mass and center of rotation are assumed to be aligned. In such a system, the angular momentum L can be defined as the product of the angular velocity \(\omega\) and the MoI along the rotational axis I, while, respectively, the torque can be defined as the product of the angular acceleration \({{\dot{\omega }}}\) of the rigid body and the moment of inertia I [17, 18]

$$\begin{aligned} L&= I \cdot \omega , \end{aligned}$$
(1)
$$\begin{aligned} T&= I \cdot \alpha = I \cdot {\dot{\omega }} . \end{aligned}$$
(2)

For a mass point m moving with a momentum \(p= mv\) at a distance r around an axis of rotation, the angular momentum L is the cross product of momentum and distance r and is thus perpendicular

$$\begin{aligned} L = p \times r. \end{aligned}$$
(3)

A rotational actuator, which can be assumed as a rigid body, is depicted in Fig. 2. Under the simplifying assumption of frictionless uniformly accelerated rotational motion, the kinematic equations of the rotational motion of a rigid body about a fixed axis can be formulated in dependence on time t

$$\begin{aligned} \phi&= \frac{1}{2} {{\dot{\omega }}} t^2 + \omega _0 t + \phi _0, \end{aligned}$$
(4)
$$\begin{aligned} \omega&= {{\dot{\omega }}} t + \omega _0, \end{aligned}$$
(5)
$$\begin{aligned} {{\dot{\omega }}}&= \alpha = \text {const}. \end{aligned}$$
(6)
Fig. 2
figure 2

Schematic illustration of a rotating rigid body

Full size image

These equations allow the consideration of a single RW. However, if a system of RW and another rotating rigid body is considered, as visualized in Fig. 3, further interrelationships arise. This additional body could, for example, be a satellite in space or a low-friction platform. This can be assumed as a closed system of two bodies. The environment will affect this system with some amount of friction as well as the interaction between the two bodies. While these frictions can be neglected as long as they are very small, they are still noted here in the form of friction losses \(T_\mu\) and \(L_\mu\), respectively

$$\begin{aligned} T_{pl}&= - T_{RW} + T_\mu , \end{aligned}$$
(7)
$$\begin{aligned} I_{pl} \cdot {{\dot{\omega }}}_{pl}&= - I_{RW} \cdot {{\dot{\omega }}}_{RW} + T_\mu , \end{aligned}$$
(8)

and

$$\begin{aligned} L_{pl}&= - L_{RW} + L_\mu , \end{aligned}$$
(9)
$$\begin{aligned} I_{pl} \cdot \omega _{pl}&= - I_{RW} \cdot \omega _{RW} + L_\mu . \end{aligned}$$
(10)
Fig. 3
figure 3

Schematic illustration of a system of two rotating rigid bodies. These resemble a rotational actuator (\(_{RW}\)) and rotating platform (\(_{Pl}\)). The angular accelerations and thus the corresponding torques acting on the bodies in this system are indicated

Full size image

2.3 Satellite Kinematics

This rigid body rotation kinematics can be used to describe the attitude control of a satellite with an applied internal torque. The attitude control with internal moments in a satellite is based on the law of conservation of angular momentum, which states that the angular momentum of a closed system remains constant as long as no external torque sources act on it. Accordingly, an RW as an internal torque source does not change the angular momentum of a satellite but only stores the angular momentum depending on its I and the speed \(\omega\) until saturation at the maximum achievable speed \(\omega _{max}\).

This conservation quality of angular momentum is used in satellites for attitude control. For this purpose, the RW in a spacecraft is accelerated, whereby an angular momentum vector is formed perpendicular to the plane of rotation. Due to the conservation of angular momentum, a second angular momentum vector with the same magnitude but in the opposite direction is formed. This opposite angular momentum is responsible for the rotation of the satellite. Consequently, the satellite rotates in the opposite direction than the RW as visualized in Fig. 3 [19,20,21,22].

In the Ferrowheel experiment, this condition is simulated in a 1-axis rotation system. Instead of a complete spacecraft, a containment is introduced as a rotation platform, in which the rotor of the Ferrowheel is integrated and can perform a rotational movement, which triggers an opposite rotational reaction in the platform itself. The rotation is detected by an optical encoder, which determines the angular position.

3 General System Design Requirements

To test the concept of the Ferrowheel and to compare it with a conventional ACS, technical requirements were defined, based on the requirements for a conventional ACS:

  • The Ferrowheel shall be able to generate a rotating magnetic field.

  • The Ferrowheel shall be able to generate torque up to \({100\,\mathrm{\text {m}\text {N}\text {m}}}\) as the rotor and ferrofluid follow the rotation of the magnetic field of the stator.

  • It shall be possible to measure the torque of the Ferrowheel.

  • The Ferrowheel shall be able to generate a rotational velocity of up to 20,000 RPM [23].

  • The rotation speed of the Ferrowheel shall be incrementally controllable.

  • The Ferrowheel shall be able to move in a clockwise and counterclockwise direction.

  • It shall be possible to measure the temperature of the Ferrowheel.

  • The containment of the Ferrowheel shall be able to compensate the thermal loads.

  • The Ferrowheel shall consume less energy than a conventional ACS.

  • The Ferrowheel shall have a longer lifetime than a conventional ACS.

  • The Ferrowheel shall achieve the same efficiency as a conventional ACS.

A desirable target for rotation speed would be at least \({5000\,\textrm{RPM}}\). A lifetime of up to 5 years and a power consumption less than \({3\,\textrm{W}}\) are also desirable performance goals. These are the key data at which conventional ACS systems of similar size operate [23].

4 Ferrowheel System Design

Fig. 4
figure 4

Design of the Ferrowheel. Above the CAD of the system without cover, below the flight-ready Ferrowheel prototype with viewing windows

Full size image

Figure 4 shows the design of the Ferrowheel. The Ferrowheel consists of a containment structure, a stator (BLDC), a polycarbonate cover with aluminum struts, and a steel rotor. The steel rotor has a diameter of \({70\,\mathrm{\text {m}\text {m}}}\) and is covered with permanent magnets on the outside, in the center, and on the inside. The cut-out dimensions for the permanent magnets on the inside and outside are chosen, so that the magnets protrude slightly beyond the rotor structure geometry. The permanent magnets are arranged with alternating north–south polarity, so that the far magnetic field is practically zero and is focused near the surface.

Fig. 5
figure 5

Polarity and direction of the permanent magnets in the rotor. North pole in red and south pole in blue

Full size image

Figure 5 shows the direction of the polarity of the permanent magnets and their arrangement. The poles of the inner and outer permanent magnets point perpendicular to the rotation vector. The poles of the center permanent magnet point parallel to the rotation vector. Due to this arrangement, the rotor is borne in the plane of rotation by the inner and outer permanent magnets and outside the plane of rotation by the center permanent magnets. The inner magnets also generate the torque in combination with the BLDC. The containment consists of two parts and has an outer diameter of \({80\,\mathrm{\text {m}\text {m}}}\) and a height of \({16.3\,\mathrm{\text {m}\text {m}}}\). The main structure of the containment is made of milled PTFE (Polytetrafluoroethylene) and has an outer chamber for the rotor with ferrofluid and secondary fluid, an alcoholic fluid for air displacement, and an inner chamber for the BLDC. The paramagnetic ferrofluid adheres to the permanent magnets of the rotor and thus creates a ferrofluid cushion, which reduces wear and tear due to the fluid properties of the ferrofluid and remains positioned on the magnets due to the paramagnetic properties. The inner diameter of the inner chamber is \({23\,\mathrm{\text {m}\text {m}}}\). The second part consists of a steel plate with an axis interface for the BLDC and is glued to the PTFE. The steel plate increases the moment of inertia and the heat dissipation into the frame structure of the FARGO experiment.

The integration is performed by wetting the permanent magnets of the rotor with ferrofluid and placing the rotor in the outer chamber of the containment. This is then filled with a secondary fluid, closed with the PC (Polycarbonate) lid, and sealed with superglue. The BLDC is glued to the aluminum struts and inserted into the inner chamber via the axis interface in the containment. The aluminum struts are glued to the PC lid and the connectors for the BLDC are attached.

For integration into the FARGO experiment, the Ferrowheel is attached to the ACS shaft with the axis interface running through the center of the containment. This setup is shown in Fig. 6. The shaft is supported by two mechanical bearings, and inserted into the precision frame. The fully assembled precision frame with integrated Ferrowheel will be mounted to the main structural plates at the top and bottom.

The Ferrowheel is controlled by a sensorless field-oriented control algorithm running on a commercial Electronic Speed Controller (ESC) board with firmware for a standard BLDC motor. An optical encoder is used to determine the torque generated. For this purpose, the sensor disk is mounted on the steel plate of the containment as shown in Fig. 6. The sensor head is located on the precision frame. In addition, a camera is mounted on the lower structural plate, which looks into the Ferrowheel, as well as an IR (infrared) sensor opposite the camera, which serves as a backup.

By changing the speed of the rotor, the containment rotates. This rotation is detected by the optical encoder and returned counts with a time stamp as angular position. The first and second derivatives give the angular velocity and the angular acceleration. From these values, the angular momentum and torque can be determined using Eqs. 1 and 2.

Fig. 6
figure 6

CAD of the Ferrowheel integrated into the FARGO experiment. The encoder instrumentation is shown. The ACS wheel is in relative motion to the ACS main containment and is supported by the ferrofluid cushion bearing. The ACS containment is in relative motion to the main frame of the system and supported by the upper and lower mechanical bearings

Full size image

5 Testing

To investigate the functionality of the Ferrowheel, various tests were performed in preparation for the FARGO experiment. In the fundamental test, the basic function was tested. For this, the Ferrowheel was started and the rotor was accelerated to 800 RPM. Afterwards, target speed was changed to 500 RPM, 300 RPM, 1000 RPM and 1200 RPM. This test was used to prove the operation capability of the system. At the same time, all sensors attached to the FARGO experiment for monitoring were tested.

Figure 7 shows the encoder output. From this, the rotation rate torque and angular momentum of the containment were determined. For the given rotor velocity changes, the maximum velocity of the containment was \({30\,\textrm{RPM}}\). The maximum value for the angular momentum was over \({6\,\mathrm{\text {m}\text {N}\text {m}\text {s}}}\) and the generated torque was \({1.8\,\mathrm{\text {m}\text {N}\text {m}}}\).

Fig. 7
figure 7

From top to bottom: data from the encoder and determined angular momentum and torque

Full size image

A second fundamental test was performed to determine an operational range for the Ferrowheel. For stable operation, the minimum rotor rotation rate was 60 RPM on the ground. To determine the maximum rotation rate of the rotor, a maximum power supply current of \({1.4\,\mathrm{\text {A}}}\) was specified for thermal reasons, with which the Ferrowheel can be operated non-destructively for a short time. Under this condition, a maximum rotation rate of 1300 RPM was achieved on the ground. The operating range of 60–1300 RPM can, most likely, be extended in space.

To investigate the performance over a longer period, a long-term test was performed at 800 RPM, during which temperature, rotation rate, and phase current were monitored. Figure 8 shows the built-in temperature sensor on the frame of the Ferrowheel.

Fig. 8
figure 8

Position of the internal temperature sensor on the Ferrowheel frame (red box)

Full size image

In addition, a PT100 was placed on the containment of the ACS as shown in Fig. 9.

Fig. 9
figure 9

Position of the PT100 temperature sensor during the long-term test (red box)

Full size image

To create convective conditions as similar as possible to those on the ISS, the cube was sealed as shown in Fig. 10, which reduces direct large-scale convection.

Fig. 10
figure 10

Closed cube to create convective conditions as similar as possible to those on the ISS

Full size image

The duration of the experiment was \({160\,\mathrm{\text {min}}}\). Due to the thermal stability of PTFE, the temperature limit of the PT100 is set to \({80\,\mathrm{{{}^{\circ }\text {C}}}}\). Figure 11 shows the results of the temperature measurement.

Fig. 11
figure 11

Temperature curve of the internal temperature sensor and the PT100 during the long-term test at 800 RPM

Full size image

The temperature of the PT100 increases sharply at the beginning than the temperature sensor on the frame. As can be seen in Fig. 9, the PT100 is located directly on the steel plate of the containment, which is also directly connected to the BLDC, i.e., the heat source. The temperature sensor on the frame is distanced from the BLDC. For this reason, the temperature at the sensor on the frame is lower. The temperature drop during the \({90\,\mathrm{\text {min}}}\) operation time results from the opening of the cube during the experiment for measurements via a thermal camera. It can be seen in Fig. 11 that the temperatures converge toward a constant value over time, i.e., thermal equilibrium. For a constant rotation rate of 800 RPM, the temperature of the PT100 converges toward \({55\,\mathrm{{{}^{\circ }\text {C}}}}\) and the temperature of the sensor on the frame toward \({40\,\mathrm{{{}^{\circ }\text {C}}}}\). These temperatures are far from the limit of \({80\,\mathrm{{{}^{\circ }\text {C}}}}\). When the Ferrowheel is switched off after \({90\,\mathrm{\text {min}}}\), the temperature drops quickly within \({30\,\mathrm{\text {min}}}\). The heat is dissipated via the steel axis into the frame and the cube. Figure 12 shows the thermal image at the start, after \({90\,\mathrm{\text {min}}}\) and after \({160\,\mathrm{\text {min}}}\).

Fig. 12
figure 12

Thermal image of the Ferrowheel during the long-term test. From top to bottom: before the test, after \({90\,\mathrm{\text {min}}}\), and after \({160\,\mathrm{\text {min}}}\)

Full size image

In the beginning, the frame and the Ferrowheel are still at ambient temperature and only the electronics are heating up. After \({90\,\mathrm{\text {min}}}\), the heating of the Ferrowheel and the frame is visible on the thermal image. After \({160\,\mathrm{\text {min}}}\), the Ferrowheel has cooled significantly, and the heat has largely dissipated into the frame. For the experiment itself, the temperature sensor on the frame can be used to monitor the Ferrowheel.

To observe changes in performance over time, the rotation rate and phase current were measured and plotted in Fig. 13.

Fig. 13
figure 13

Phase current and rotation rate curve during the long-term test of the Ferrowheel at 800 RPM

Full size image

As can be seen in Fig. 13, the rotation rate is constant over \({90\,\mathrm{\text {min}}}\) and is stably controlled by the system around 800 RPM. The phase current changes more during operation, depending on the current required by the BLDC in combination with the ESC to keep the rotation rate of the rotor constant. Thus, the phase current alternates strongly. It is noticeable that the phase current drops approximately \({1\,\mathrm{\text {A}}}\) after \({10\,\mathrm{\text {min}}}\) at a constant rotation rate. A reason for this observation may be the heating of the stator. The hot stator also heats the rotor and the ferrofluid, which reduces the viscosity of the ferrofluid. As a result, less power is needed to turn the rotor, since warm ferrofluid is less viscous. For this reason, the current consumption drops to \({81\,\mathrm{\%}}\) of the initial current consumption when the ferrofluid heats up. The initial power of the system is \({6.48\,\textrm{W}}\). Similar to the current, the power drops by \({19\,\mathrm{\%}}\) to \({5.28\,\textrm{W}}\) after \({10\,\mathrm{\text {min}}}\).

6 Fulfillment of the Requirements

The Ferrowheel generates a rotating magnetic field and the rotor and ferrofluid follow the field. The Ferrowheel also generates a torque up to \({3\,\mathrm{\text {m}\text {N}\text {m}}}\) that can be measured. The rotor of the Ferrowheel can be controlled incrementally in a range between \({60\,\textrm{RPM}}\) and \({1300\,\textrm{RPM}}\) clockwise and counterclockwise. The desirable target of \({5000\,\textrm{RPM}}\) was not achieved and is a target for future development. It is possible to measure the temperature of the Ferrowheel and set a limit for thermal destruction. In addition, the containment can compensate for thermal loads over a long period for lower rotational velocities and over a short period for higher rotational velocities. The required power of the Ferrowheel of over \({5\,\textrm{W}}\) is much higher than the required power of less than \({3\,\textrm{W}}\) of a conventional ACS. A goal of future developments would be to reduce the required power by optimizing the gap between the rotor and the stator. Due to the power gap, the Ferrowheel also does not reach the efficiency of a conventional ACS. A lifetime test has not been performed yet.

Not all requirements were accomplished yet, especially the performance targets compared to a conventional ACS. However, the Ferrowheel fulfilled all basic requirements and reached a high level in the short development time of a single year.

7 Conclusion

In the research discussed here, a working prototype of a novel RW based on a ferrofluid bearing rather than a mechanical bearing was presented. The Ferrowheel was explained from its concept study, through development to final design and integration. In addition, preliminary verifying tests were detailed. The design-critical points of the system were highlighted and the general functionality on the ground was shown. Furthermore, this research shows that ferrofluid-based RWs have the potential to replace conventional RWs in the future.

The COTS BLDC has turned out to be a critical point of the development, as both the control is not trivial and the part is fragile when manually produced and therefore tends to fail during integration, which has usually led to a complete replacement of the BLDC. In addition, it became apparent that such a rotating system requires precise manufacturing on the order of hundreds \(\upmu {\textrm{m}}\), especially in the area of the rotor, to avoid residual friction or grinding and therefore wear and tear. The temperature development of the BLDC, especially at high speeds, turns out to be a main design driver, as temperature is a very critical point to avoid components overheating or melting. Therefore, further design for heat mitigation and cooling systems should be considered. As a further point, the potential for reducing the system mass compared to conventional RWs should be investigated, as such a mass reduction could be possible with this concept and would be highly beneficial for space missions.

In conclusion, the system discussed here cannot yet keep up with conventional RW systems but is on a promising path toward competitiveness. If one compares the decades of heritage of conventional systems with the development time that has elapsed so far for ferrofluid systems, it becomes clear that these systems have already reached a considerable level in a very short time, within a student project and that future potential is apparent.

Both short-term high rotation rates of over 1000 RPM and long-term operation of almost \({2\,\mathrm{\text {h}}}\) at lower rotation rate ranges were already possible with the system. Even longer operating periods are also possible but are subject to future testing. Thus, this paper represents the entire development of the system up to the integration into the FARGO cube for implementation on the ISS. The Ferrowheel was successfully put into operation on the ISS in March 2023. The basic functionality of the system was thus demonstrated both on the ground and in space. The data obtained from this operation will provide an important basis for future research on this system [24].

8 Outlook

The work on ferrofluid as a competitive technology for mechanical RWs is very promising and offers great potential for the future of space activities. Therefore, building on this work, research on the Ferrowheel will continue in the future in the cooperation between KSat e.V. and IRS. Further research and development work may lead to the Ferrowheel replacing a conventional ACS in spacecraft.

As already mentioned, current issues include optimizing the precision of the production of all components, in particular the rotor, and implementing a cooling system to avoid high temperatures. In addition, the electric control of the BLDC should be optimized for usability, range and reliability.

The next big step for the Ferrowheel experiment after this development was the operation on board of the ISS for 30 days in March 2023. This provided in-space verification of the system, and the resulting data will be of significant value for the further development of the system [24].