Description of TEAM Workshop Problem 28:

An Electrodynamic Levitation Device

Hans Karl, Joachim Fetzer, Stefan Kurz, Gunther Lehner and Wolfgang M. Rucker
Institut fur Theorie der Elektrotechnik, Universitat Stuttgart, Pfaenwaldring 47, 70550 Stuttgart, Germany.

Abstract | This paper presents a new TEAM work- drop due to the motion of the armature.
shop problem taking into account moving bodies. An
electrodynamic levitation device which consists of a II. Description of Model A
conducting plate over two exciting coils shall be ex-
amined. The aim is to determine the dynamic char-
acteristics of the levitating plate. A coupled solution One of the earliest papers on levitation by elds at
of the electromagnetic and the mechanical problem is power frequencies is that due to Belford, Peer and Tonks
necessary for that. [1]. Electrodynamic levitation is based on the induction of
eddy currents in conducting materials. These eddy cur-
rents can be induced by a time varying magnetic eld.
I. Introduction This is the case for the device shown in Fig. 1. A cylin-
drical aluminium plate ( = 3:40 107 1=
m, m = 0:107

The modelling of electromechanical devices, i.e. the so- kg) is located above two cylindrical coils. All three parts
lution of transient coupled electromechanical problems are aligned coaxially. The inner coil has w1 = 960 and
taking into account moving bodies is gaining signicance. the outer coil w2 = 576 turns. The dimensions of the de-
For this reason it is necessary to dene a benchmark which vice are shown in Fig. 2. The levitation height z refers
allows to compare dierent approaches regarding their ad- to the distance between the lower edge of the plate and
vantages and disadvantages for the solution of such prob- the upper edge of the current carrying area (z = 0). For
lems. Typical diculties are the treatment of motion, t 0 the plate rests above the coils at a distance of z =


strategies for remeshing, force calculation, weak versus 3.8 mm due to the thickness of the winding form.
strong electromechanical coupling, ecient time stepping
schemes, etc.
Up to now only TEAM workshop problem 9 and TEAM
workshop problem 17 deal with moving bodies. In prob-
lem 9 a moving body with given constant velocity is con-
sidered. The problem is axisymmetric and innitely ex-
tended in the direction of the velocity. Therefore it is
stationary and lacks the feature of electromechanical cou-
pling. For problem 17 no measured data are available.
Only the description of the underlying experiment is given Fig. 1. TEAM Workshop problem 28: An electrodynamic levitation
and therefore the treatment of the problem is dicult. In device
contrast, the new TEAM workshop problem 28 is a tran- z
sient problem with electromechanical coupling and mea-
sured data being available.
It is intended to split the new problem into dierent
packages with increasing level of diculty. In this pa- 130
per, Model A is presented. Model A is an axisymmetric 3
problem without signicant eddy current reaction to the
exciting coils. In the future, two additional packages will z=0
be provided. Model B will be very similar to Model
A, but the levitating plate will have an eccentric bore
which disturbs the axisymmetry and requires a full 3D 52
modelling. Model C will have a dierent geometry and
operating frequency. The exciting coils will be voltage w1 w2
driven and the reaction due to the motion has to be taken 15 28 41 46,5
into account. This is an important feature because many
actuators are voltage driven and show a typical current Fig. 2. Dimensions (in mm) of the electrodynamic levitation device
 Both coils are connected in series, but with dierent
sense of winding. The device is operated directly be-
tween two outer conductors of the three-phase supply net-
work. With the help of an electronic switch, the instant
of switching on is synchronized onto the driving supply
voltage in a way that there is no electrical transient. For
t 0, sinusoidal currents i(t) ow in the coils in opposite


directions,
i(t) = ^{ sin(2f0 t); ^{ = 20 A; f0 = 50 Hz: (1)
Due to the induced eddy currents a repulsive force is ex-
erted on the plate. After some damped oscillations, the
plate attains a stationary levitation height of z = 11.3
mm. A possible reaction of the induced eddy currents
can be neglected. For this reason, the current in (1) can
be regarded as impressed.
Fig. 4. Measured levitation height of four dierent measurements.
III. Measurement of the levitation height
acceptable. The average values of these measured data
The levitation height is measured by means of laser tri- are given in Table I on the next page and should be used
angulation. The principle is shown in Fig. 3. A laser beam for the comparison with numerical results.
emitted by a laser diode is directed near the center of the It is important to make sure that the plate and the coils
plate. The re ected beam is detected by a position sens- have ambient temperature when the measurement starts.
ing semiconductor detector (PSD). A displacement
z of Due to the ohmic losses the temperature of the device
the plate yields a displacement
s of the re ected beam. raises signicantly during operation. This would increase
With this principle, a high resolution limited only by am- the resistance and cause inaccurate results.
plier noise can be achieved. Another diculty is related to the exact modelling of
In practice, however, there are some undesired eects the coils. For numerical purposes the coils are represented
which complicate the measurements. The rst problem by domains which carry a homogeneous azimuthal current
is that the device can never be perfectly axisymmetric. density. The cross section of these domains is given by the
Any radial displacement of the plate, nonhomogeneous rectangular cross section of the winding space. The coils
winding of the coils and other similar eects disturb the are made of copper wire with 1.2 mm diameter and con-
symmetry. This causes additional oscillations of the plate tain insulating layers. The real current density is neither
around a radial axis. Luckily the experiments showed that strictly homogeneous nor sharply bounded by a rectangle.
these oscillations have only a minor eect on the measured These eects may in uence the plate during its initial lift-
results. Fig. 4 shows the measured levitation height of o phase. It is dicult to estimate the in uence of this
four dierent measurements and gives an idea about the eect.
reproducibility. The dierence of the respective results is
IV. Comparison of Measured and Computed
Results
For the sake of completeness we include a comparison
between the measured data according to Table I and com-
puted results in this description, see Fig. 5. The computed
results have been obtained with the help of the BEM-FEM
code described in [2], [3]. The discrepancy of the maxi-
mum levitation height during the rst half period might
be traced back to the modelling of the coils as explained
in the previous section.

V. Concluding Remarks
This description of TEAM problem 28 superseeds the
preliminary description [4]. The results in Table I
Fig. 3. Measurement of the levitation height by means of laser can be supplied on request by e-mail (hans.karl@ite.uni-
triangulation stuttgart.de). It is hoped eventually to have the entire
 Table I: Measured results of the levitation height
t (s) z(mm) t (s) z(mm) t (s) z(mm) t(s) z(mm) t (s) z(mm) t(s) z(mm)
0.0 3.7 287.4 11.9 574.9 11.6 862.3 11.7 1149.8 11.5 1437.2 11.4
9.9 4.0 297.4 10.4 584.8 11.0 872.3 11.4 1159.7 11.4 1447.2 11.4
19.8 4.9 307.3 9.3 594.7 10.4 882.2 11.2 1169.6 11.3 1457.1 11.4
29.7 6.9 317.2 8.7 604.6 10.0 892.1 11.1 1179.5 11.2 1467.0 11.3
39.6 9.7 327.1 8.7 614.5 9.9 902.0 11.0 1189.4 11.1 1476.9 11.3
49.6 12.8 337.0 9.2 624.5 10.0 911.9 11.0 1199.4 11.1 1486.8 11.3
59.5 15.6 346.9 10.2 634.4 10.3 921.8 11.1 1209.3 11.1 1496.7 11.3
69.4 17.4 356.8 11.4 644.3 10.8 931.7 11.2 1219.2 11.1 1506.6 11.3
79.3 18.0 366.7 12.4 654.2 11.3 941.6 11.4 1229.1 11.2 1516.5 11.3
89.2 18.1 376.7 13.2 664.1 11.7 951.6 11.6 1239.0 11.2 1526.4 11.3
99.1 18.2 386.6 13.6 674.0 12.1 961.5 11.7 1248.9 11.3 1536.4 11.3
109.0 17.8 396.5 13.7 683.9 12.3 971.4 11.8 1258.8 11.4 1546.3 11.4
118.9 16.4 406.4 13.3 693.8 12.3 981.3 11.8 1268.7 11.4 1556.2 11.4
128.9 14.1 416.3 12.7 703.8 12.2 991.2 11.8 1278.6 11.4 1566.1 11.4
138.8 11.5 426.2 11.8 713.7 12.0 1001.1 11.7 1288.6 11.4 1576.0 11.4
148.7 9.0 436.1 10.9 723.6 11.6 1011.0 11.5 1298.5 11.3 1585.9 11.4
158.6 7.2 446.0 10.1 733.5 11.3 1020.9 11.4 1308.4 11.3 1595.8 11.4
168.5 6.7 456.0 9.6 743.4 11.0 1030.8 11.2 1318.3 11.3 1605.7 11.4
178.4 7.3 465.9 9.4 753.3 10.8 1040.8 11.1 1328.2 11.2 1615.7 11.4
188.3 8.8 475.8 9.6 763.2 10.7 1050.7 11.0 1338.1 11.2 1625.6 11.3
198.2 10.7 485.7 10.1 773.1 10.8 1060.6 11.0 1348.0 11.2 1635.5 11.3
208.2 12.6 495.6 10.8 783.0 10.9 1070.5 11.0 1357.9 11.2 1645.4 11.3
218.1 14.3 505.5 11.6 793.0 11.2 1080.4 11.1 1367.9 11.2 1655.3 11.3
228.0 15.6 515.4 12.2 802.9 11.5 1090.3 11.2 1377.8 11.3 1665.2 11.3
237.9 16.2 525.3 12.7 812.8 11.7 1100.2 11.3 1387.7 11.3 1675.1 11.3
247.8 16.3 535.2 13.0 822.7 11.9 1110.1 11.4 1397.6 11.4 1685.0 11.3
257.7 15.8 545.2 12.9 832.6 12.0 1120.1 11.5 1407.5 11.4 1694.9 11.3
267.6 14.8 555.1 12.7 842.5 11.9 1130.0 11.5 1417.4 11.4 1704.9 11.3
277.5 13.5 565.0 12.2 852.4 11.8 1139.9 11.5 1427.3 11.4 1714.8 11.4
 BEM-FEM
Measurement

Fig. 5. Comparison between the measured (Table I) and the com-

puted (BEM-FEM) levitation height
description available on a WWW TEAM page.
References
[1] B.D. Bedford, L.H. Peer, and L. Tonks, \The electromagnetic
levitator," Gen. Elect. Rev., vol. 42, no. 6, pp. 246{247, 1939.
[2] S. Kurz, J. Fetzer, and G. Lehner, \Threedimensional transient
BEM-FEM coupled analysis of electrodynamic levitation prob-
lems," IEEE Transactions on Magnetics, vol. 32, no. 3, pp.
1062{1065, May 1996.
[3] S. Kurz, J. Fetzer, G. Lehner, and W.M. Rucker, \A novel
formulation for 3D eddy current problems with moving bod-
ies using a Lagrangian description and BEM-FEM coupling,"
Submitted to the COMPUMAG 1997.
[4] H. Karl, J. Fetzer, S. Kurz, G. Lehner, and W.M. Rucker, \Pre-
liminary proposal for a new TEAM workshop problem: An elec-
trodynamic levitation device," in Proc. of the TEAM Work-
shop, Graz, Austria, Sept. 1996, pp. 41{42.