ACTUATOR 2024 ∙ 13. – 14.06.

2024 ∙ Wiesbaden

Optimized high-efficiency MICA actuators for space compressors

Xavier De Lepine, Frank Claeyssen, Francois Barillot, Patrick Meneroud, Aurore Loubet, Timotéo Payre, Nicolas Bour-
geot, Olivia Stadler, David Ferris, Jocelyn Rebufa, Arnaud Barnique, Gérald Aigouy, Hélène Rouxel, Cedrat Technolo-
gies, Meylan, France

Abstract
Moving Iron Controllable Actuators MICA™ are magnetic linear actuators from CEDRAT TECHNOLOGIES (CTEC)
that find interests in applications requiring high-force density and low heating. Because of these attractive performances,
some space applications such as ZBO Gas compressors and Pulse Tube Cryocooler are considering their use in operational
missions. However, this last application requires the demonstration of additional features, notably a high efficiency, a
long lifetime and a good force homogeneity to minimize the generation of exported micro-vibrations. To fulfil these
additional features, CTEC has developed an optimised MICA following the specifications of the Pulse Tube Cryocooler
(PTC) of AIR LIQUIDE. The paper will present the performance progresses obtained on this customized MICA, from
modelling prediction to lab tests of Engineering Models.

1 Introduction 2 Optimization framework

TM
The MICA (Moving Iron Controllable Actuator) is a
CEDRAT TECHNOLOGIES (CTEC) proprietary actua- 2.1 Performance assessment
TM
tor magnetic design concept, based on a polarized variable MICA actuators dedicated to compressor applications
reluctance principle. This concept has been de-scribed in must address several topics that can be contradictory: High
previous CTEC publications [1] and [2] as well as its ad- efficiency is needed to save power in operation, a smoothed
vantages, a high-force density and low heating. behaviour is needed to minimize parasitic forces and extra-
long operation lifetime ([5], [6]) is required to reduce
maintenance to the minimum.
Nevertheless, the very principle of MICATM actuators rely
on the magnetic field concentrated in the radial airgap to
produce axial force. This approach allows significant force
density at the cost of non-linear behaviour along the axis
direction. Consequently, the magnetic designer must find
the best compromise between the homogeneity along axial
Figure 1: CTEC MICATM Moving Iron Magnetic Principle
displacement and the maximum force to achieve. In fact,
Because of these attractive performances, some space ap- the homogeneity of the actuator is strongly related to its
plications such as Gas compressors for Zero-Boil-Off parasitic micro-vibrations, while a large force maximizes
(ZBO) [3] and Pulse Tube Cryocooler (PTC) are consider- efficiency.
ing their use in operational missions such as Lunar Orbital Additionally, an extra-long operation lifetime is easier to
Platform-Gateway (LOP-G) and future satellites of earth achieve with soft guiding blades, while it is recommended
observation. In both cases, the MICA™ is used at reso- to have stiff guiding blades to reject radial vibrations. For
nance in linear oscillation to compress a gas. This has been this purpose, a strategy based on flexible blades [7] has
demonstrated in [3] showing good efficiency. been adopted.
AIR LIQUIDE ADVANCED TECHNOLOGIES (ALAT) Therefore, the optimization approach consists in taking
requested for the PTC application of more demanding ac- into consideration all these parameters to achieve the best
tuator performances [4], in particular a long lifetime, a high solution.
efficiency and low levels of micro vibrations. This new ac-
tuator would provide, in theory, about 6W cooling power 2.2 Modelling framework
in [50K-60K] range as requested. Several tools have been developed to monitor the relevant
At first, MICA™ was assessed by ALAT in Phase 1 parameters of the actuator performance: the input/output
(Gen.1.1) [4]. It was deemed promising but not satisfying. power (magnetic, fluidic & mechanical coupled models),
Therefore, ALAT ordered to CTEC a new MICA™ in Sec- the parasitic force (axial and radial), the flexible bearings
ond Phase, with the objective to cover high power effi- lifetime, etc.
ciency, low exported micro-vibrations, and extra-long life- The Figure 2 presents the electromechanical model of the
time. system that was used to estimate the overall performance
The work presented in this article shows the approach that of the system, including the mechanical parts, the magnetic
has been conducted to achieve this goal. parts, and the fluidic parts.
The mechanical parts are modelled with a linear stiffness
(𝐾𝑏 ) and a viscous damper (𝐶𝑏 ), the latter to express the
structural damping due to fixation.

51
 ACTUATOR 2024 ∙ 13. – 14.06.2024 ∙ Wiesbaden

The magnetic components are considered through the mag-

netic stiffness (𝐾𝑏 ) and the force factor (𝐾𝑚 ), that act on
both mechanical and electrical domains. Besides, eddy cur-
rents that develop either in the statoric part of the actuator
or in the moving part of the actuator are captured with
𝑅𝐸𝐶 (𝜔) and 𝐶𝐸𝐶 (𝜔, 𝑥), respectively.
Similarly, the compression chamber was modelled with a
mechanical stiffness 𝐾𝑔𝑎𝑠 (𝜔, 𝑥) and a viscous damper
𝐶𝑔𝑎𝑠 (𝜔, 𝑥).

Figure 4: A few examples of bearing designs

Similarly, the magnetic circuit must be carefully designed

to achieve several targets: a large actuation force for a
given current at zero displacement (to maximize efficiency
at the resonant frequency), axial force homogeneity to min-
imize axial harmonics (to minimize axial parasitic forces
when two actuators must work in back-to-back setup), low
sensitivity to mechanical tolerances to insure repeatability
Figure 2: Electro-fluid-mechanical model between units, and more importantly, low parasitic radial
forces for a given eccentricity to reduce micro-vibrations.
The input electrical power is called 𝑃𝑒 , while the output
mechanical power dissipated in gas is called 𝑃𝑔𝑎𝑠 . The lat-
ter is the useful power for the purpose of compression.

Figure 5: Estimated exported micro-vibration forces

To find the best compromise, a strategy based on local

magnetic saturations shaped on specific design indicators
was adopted. In Figure 5, an estimation of the exported mi-
cro-vibrations is shown, while Table 1 shows a few inter-
mediate results that were obtained during this iterative pro-
Figure 3: Multiphysics tools used for design and optimization cess. Results are given for 10A 0-p nominal current.

2.3 Trade-off and results of the optimization Phase 1 Gen2.1 Final

(Gen.1.1) (PTC)
process
Stroke ± 5mm ± 5mm ± 5mm
Two key components are central in the optimization pro- Axial homogeneity 58% 86% 76%
cess: the flexible bearings and the magnetic circuit. @±5mm
The flexible bearings must be designed to meet several re-
Force @ 10A ~180 N 130 N ~200N
quirements that can be contradictory, sometimes. In fact,
Rad. mag. Stiff. -540 -190 -420
they are built such that they must withstand extra-long life-
time (about 10 years of continuous working conditions), (N/mm)
while limiting the exported radial parasitic forces due to Outp. mech. Pwr. (W) 130 70 150
magnetic excitation. The first implies low mechanical Sens mech. Tol. (*) High Low Low
stresses in axial direction, while the second implies a quite Axial stiff. Non-lin- Non- <1.8N/m
large radial stiffness so that the moving part of the actuator ear linear m
works as a high pass filter with regard to inner magnetic Mass of Mag. Act. 1.3 kg 1.0 kg 1.4 kg
forces generated by the actuator itself. part
Table 1: Models obtained during optimization process

52
 ACTUATOR 2024 ∙ 13. – 14.06.2024 ∙ Wiesbaden

3 Experimental results
3.1 Setup
Several experimental setups were designed to investigate
the main characteristics of the sub-components: the mag-
netic circuit, the flexible bearings alone, the axial and ra-
dial dynamics, the output power, etc.
In Figure 6 are presented three strategic setups that validate
the magnetic circuit in static and dynamics, the flexible
Figure 7: Output power (Pgas) vs input excitation current
bearings lifetime characteristics and the radial dynamics.
For the validation of the flexible bearings, an approach
similar to [7] was adopted, in which the units were tested 3.2 Performances
at stroke larger than the nominal stroke. More details are In this chapter, the main performances of the actuator are
given in chapter 3.2.1. presented. They include the results of the flexible bearings
in lifetime and the performances in terms of power, effi-
ciency, frequency, etc.

3.2.1 Fatigue performance

The philosophy adopted for the lifetime analysis of the
flexible bearings was to evaluate how far the actual bearing
were from the theoretical Wöhler curve. This approach al-
lows to obtain useful information about the capacity of the
bearing to sustain extra-long working time (more than 10
years of continuous working conditions) during a standard
project.

Figure 6: Experimental setups to validate the performances of

the system

The former setup was designed to demonstrate the capacity

of the actuator to achieve the requirements in terms of: fre-
quency, power, stroke, as defined by the specification. The
challenge is to insure that the stiffness and damping coef-
ficients of the compression chamber (𝐾𝑔𝑎𝑠 and 𝐶𝑔𝑎𝑠 , re-
spectively) would match to the requirement at the right Figure 8: Results of lifetime tests of flexible bearings
stroke (Figure 7 shows the non-linear behaviour of the
chamber). In practice, the flexible bearing is excited such that the
Additionally, a specific setup was needed to collect useful stresses are above the theoretical Wöhler curves provided
information about the performance of the flexible bearings in literature until breakdown. A few tens of components
in terms of lifetime. To achieve this goal, a voice-coil was were tested, and the results were all encouraging: most
used to impose an axial displacement that were much larger pieces broke above the 50% Wöhler curve, which tends to
than the nominal stroke to apply stresses above the Wöhler show that the bearings can work according to the require-
curve. With this approach, it becomes possible to obtain ments. In fact, the dotted line shows the nominal stress seen
useful experimental data regarding theoretical information by the bearings in working conditions.
that exist in literature. In fact, as the objective for the pro-
ject is to achieve more than 10 years in continuous opera- 3.2.2 Exported micro-vibration forces
tion, it cannot be achieved during a standard project, which Tests conducted by the customer showed good correlation
usually last about 1 year or two. More details about this between the estimated forces (Figure 5) and measured
setup and the philosophy behind it are presented in §3.2.1. ones. A typical max level is below 50mN, which is the
The last picture shows a shaker owned by CTEC that was level announced by NG cryocoolers with similar power [8].
used in this project to identify the radial frequency of the
actuator. This result is important as it includes the radial 3.2.3 Dynamic performance
stiffness of the flexible bearings, the magnetic stiffness of
the magnetic circuit, as well as the mobile part of the actu- The performances of the new MICA™ actuator have been
ator (rotor). The result of this test has an immediate impact measured for three units (2xEMs + 1xEQP, they are visible
on the exported forces rejection capacity of the actuator. on Figure 9) for the needs of the project. The outcome re-
sults could then provide a first feedback of the repeatability
of the manufacturing process.

53
 ACTUATOR 2024 ∙ 13. – 14.06.2024 ∙ Wiesbaden

5 Acknowledgement
The authors gratefully thank AIR LIQUIDE ADVANCED
TECHNOLOGIES Sassenage (F) and CNES Toulouse (F)
for supporting these works.

6 Literature
Figure 9: Units produced during the project [1] F. Claeyssen, G. Magnac, O. Sosnicki, "Moving Iron
Controllable actuator", ACTUATOR 2008, 9-11 June
The performance evidenced in the following are : the out- 2008, C3.3, pp. 445 – 448, Messe Bremen
put power 𝑃𝑔𝑎𝑠 (useful for the compression process), the [2] G. Aigouy, X. De Lepine, P. Meneroud, MICA20CS
input electrical power 𝑃𝑒 , the efficiency 𝑃𝑔𝑎𝑠 /𝑃𝑒 , the dis- Proof mass actuator – Conf. ACTUATOR 2022 29-30
placement 𝑥, and the power factor cos 𝜙, with 𝜙 = June 2022, Darmstadt, DE, VDE VERLAG · Berlin
arg⁡(𝑖 ⁄𝑒 ). [3] P. Meneroud, S. Rowe, G. Aigouy, Design of a 2
Figure 10 shows the comparison between the performances stages compressor for mobility applications, using
of three units. The differences between the units lie within compact and efficient Moving Iron Controllable Actu-
the measurement errors, which tends to show that the man- ators, Conf. ISEF 2019, Nancy Fr, 29-31 Aug .2019 –
ufacturing process is robust and repeatable. Proc P2-B-2
[4] P.O. Mine, Air Liquide Pulse Tube Cryocooler Com-
pressor Development (DEMO 0) and Test
Results, 8th European Space Cryogenics Workshop
2023, 5-7 June 2023, Noordwijk, NL
[5] F. NOVÝ, O. BOKŮVKA, L. TRŠKO, M. CHA-
LUPOVÁ, Ultra high cycle fatigue of materials, Dpt
of Materials Engineering, University Of Zilina in
Zilina, Annals of the Faculty of Mechanical Engineer-
ing Hunedoara, Romania - International Journal of En-
gineering, 2012.
[6] W. Illg, C. B. Castle, Fatigue of four stainless steels
and three titanium alloys before and after exposure to
Figure 10: Comparison between the performances of three units
550° f (561° k) up to 8800 hours, NASA, Clearing-
Main parameters are listed in Table 2. house for Federal Scientific and Technical Infor-
mation, Jul. 1965.
Parameters SN06 SN03 SN02 % dev. [7] H. Dang, High Efficiency Pulse Tube Cryocoolers for
𝒇𝒓é𝒔𝒐 59𝐻𝑧 59Hz 59.3Hz 0.5% Aerospace Applications, Shanghai Institute of Tech-
Current 9.9𝐴 9.9A 9.9A 0% nical Physics, Chinese Academy of Sciences, Shang-
Voltage 38.4𝑉 38.9⁡V 39.4𝑉 2.6% hai 200083, China, Advances in Cryogenic Engineer-
𝒀𝟎 4.4𝑚𝑚 4.4mm 4.4mm 0% ing, AIP Conf. Proc. 1573, 378-385, 2014; doi:
𝑷𝒈𝒂𝒛 115𝑊 115W 115W 0%
10.1063/1.4860726
𝑷𝒆 164𝑊 165W 165W 0.6%
𝒄𝒐𝒔𝝓 0.88 0.86 0.87 2% [8] Northrop Grumman High Efficiency Cryocoolers for
Table 2: Performance comparison between units space, 2024, https://cdn.prd.ngc.agencyq.site/-/me-
dia/wp-content/uploads/Northrop-Grumman-Cry-
ocoolers-Datasheet.pdf
4 Conclusions
This works shows that MICATM technology is clearly rele-
vant for compressors in space applications, either in terms
of power efficiency, lifetime and exported micro-vibration
forces. These performances were obtained after a multivar-
iable optimization procedure (magnetic, electrical, me-
chanical, vibrations, etc.), and all indicators were im-
proved, at least in theory, when compared to the initial pro-
posal provided by CTEC in First phase.
Additionally, the theoretical tools were validated with ex-
perimental results in most key parameters: power, guiding
blades lifetime and exported micro-vibrations. Those re-
sults have been reached for three units within a range
smaller that 3% in all parameters, which consolidates the
manufacturing process in terms of quality and repeatabil-
ity.

54