Вы находитесь на странице: 1из 3

#

Current Transformer Modelling.


Yann Le Floch(1)(2), Christophe Gurin(1), Dominique Boudaud(3), Grard Meunier(2), Xavier Brunotte(1)
(1)
Cedrat Technologies, Meylan, France ; (2) Laboratoire d'Electrotechnique de Grenoble, UMR 5529 INPG/UJF - CNRS, ENSIEG, France;
(3)
Schneider Electric, Grenoble, France.

Keywords - Air gap, Electric circuit, With this assumption, the relation
Transient, Nonlinear material, Nodal between current and voltage is [3]:
element, Shell element, Time
stepping, Reduced magnetic scalar
potential.
Abstract - This paper presents the To compute t 0k , we have two
modelling of a current transformer by solutions.
various methods with the FLUX3D The first solution is to use edge
software. The technique used is elements, which is natural in order
based on the Finite Element Method
to take into account the surface
coupled with electric circuits. A
Figure 2: Description of the current condition t0k x n = 0 on .
magnetic scalar potential reduced
transformer. The other one is to compute nodal
versus T0 formulation (T0 -) taking
into account the electric circuits with t0k. For this purpose, we compute
an air-gap is used for this purpose. t0k in the air (0) such as :
The air-gap is described either by a II. Description of the current t0k = h0k - grad(k)
thin volume region or by a surface transformer
region. The transformer is constituted by a Where :
magnetic core surrounded by two - h0k is the magnetic field due to a
I. Introduction secondary coils connected in series. unit current in the inductor k,
The study deals with a current The finite element modelling (in time calculated with Biot and Savart's
transformer used in a low voltage stepping and circuit equations) formula (nodal value) in the air (0).
circuit breaker made by Schneider represents 1/8th of the device (see - k is the reduced-total increment
Electric (see figure 1). FLUX3D figure 2). The simulated curves [4] [5] calculated with a unit current
software allows to take into account correspond to a primary sinusoidal in the inductor k such as :
nonlinear transient magnetic excitation (I0 = 11.137 A and f = 50 grad(k) x n = h0k x n on = t 0.
problems coupled with electric Hz) and a purely resistive load. The
circuits. This software enables to total simulation time (40ms) Thus, on , we respect the
model in an effective way the corresponds to the transient mode conditions: tok x n = 0
current transformers by introducing of the sensor. because t0k = h0k - grad(k) and we
a thin volume air-gap. This solution compute k as follow :
can be used when modelling simple III. Formulation: T0 - h0k x n = grad(k) x n
devices such as the current The present formulation (T0 - ) [1]
transformer presented in this paper. [2] to treat coupling between Now, we will see which solution we
When modelling more complex electric circuits and magnetic choose to model our current
devices, difficulties due to the devices is : transformer.
geometrical description and the In magnetic circuit (t) :
meshing of the thin volume air-gaps )
H = -grad( B = H IV. Modelling air-gaps
can occur. We would like then to One of the difficulties of the current
model the thin volume air-gap in. In air and in air-gap (0) : transformer modelling is to take
another way by using shell into account thin air-gaps. In our
B = o H
elements which are surface case, for a 40 mm long device the
elements with a thickness. Thus, a air-gap thickness is 50 m. This
new version which allows to take With : m the number of inductors. scale difference makes the device
into account electric circuits and t0k is calculated in the 0 region with difficult to geometrically describe it
surface air-gaps has been a unit current in the inductor k, such and to mesh it (see figure 4).
developped. We will describe the as: t0k x n = 0 on = t 0
improvements obtained thanks to Thus, we would like to model thin
the introduction of a surface air-gap volume air-gaps by surface air-gaps
with the electric circuits. with a tickness. For this purpose,
we have to use surface elements
with potential jump (shell element).
Our experience in magneto-statics
leads to use shell elements with a
nodal approximation [6].
The solution is then to use the
formulation presented above with
the nodal t 0k which enables to
describe the air gap with shell
Figure 1: Photo of the current Figure 3: Formulation T 0 - elements. Firstly, we will present in
transformer used for the modelling. configuration.
(continued on page 6)

Number 38 - January 2002 - CEDRAT - CEDRAT TECHNOLOGIES - MAGSOFT


$
Current Transformer Modelling.
Yann Le Floch , Christophe Gurin , Dominique Boudaud , Grard Meunier , Xavier Brunotte(1)
(1)(2) (1) (3) (2)
(continue)
(1)
Cedrat Technologies, Meylan, France ; (2) Laboratoire d'Electrotechnique de Grenoble, UMR 5529 INPG/UJF - CNRS, ENSIEG, France;
(3)
Schneider Electric, Grenoble, France.

a short way the shell For the thin volume air-gap and the
elements and surface air-gap, the currents
its limitation obtained are not sinusoidal due to
and, in a second the saturation of the magnetic
part, the t 0k material (see figure7). The shapes of
computation. the resulting waves for both
simulations are the same (see
A. Shell elements figure 7) and are accurate in
comparison with measurements
As mentioned before, we (less than 5% of variation on the
can model air-gaps with shell Figure 4: Surface mesh of the air- whole simulation period). The more
elements. Indeed, the magnetic gap and the magnetic circuit. accurate is the provided B(H) curve
field is mainly normal to the air- of the magnetic material, especially
gap surface, so there is a jump of We use now these shell elements at the saturation bend, the smaller
the magnetic scalar potential in the with the T0 - formulation with a is the variation between simulation
thickness direction. Therefore, the nodal t0k presented below. and measurements.
new element will be a surface
element in the plane of the air gap B. t0k Computation with shell elements The contribution of the surface air-
and will have double nodes (see gap leads to strong improvements
figure 5). Each couple of double When we compute k for the in term of computation time which
nodes will have the same inductor k, we impose: kib - kit = is divided by 4 (see table I) without
coordinates and the shell element constant = 1 on shell elements modifying the results (see figures 7
(Notation on figure 5). This and 8).
constant is the current in
the inductor k (1A) On figure 8, the isovalues of the flux
because of the Ampre's density in the air are almost
law [5]. identical, made smoother with the
surface air-gap. This difference is
This reduced-total increment due to the t0k calculated with edge
enables to make the potential elements used with the volume air-
jump between the two sides of gap and with nodal elements used
Figure 5: Prismatic element (a), the air-gap surface. with the surface air-gap.
Shell element with potential jump (b).
V. The results VI. Conclusion
will be considered as a conventional We have performed two FLUX3D software is therefore a
prismatic element [6]. However, simulations, one with a thin volume powerful tool for modelling and
shell elements have tickness air-gap and an edge t 0k , and analyzing low voltage current
limitations. The ratio between the another with a surface air-gap and transformers. The difficulties of the
air-gap tickness and the device a nodal t0k. We compare these two current transformer modelling is to
length has to be smaller than 1/10 computations with measurements take into account thin air-gaps.
and higher than 1/105. given by Schneider Electric.
(continued on page 8)

Figure 7 :
Induced current
in the
secondar y
circuit (B2).

Table 1:
Figure 6: Reduced-total increment Computation
( B2 ) calculated with a unit current time for the
in the inductor B2 and the surface various
mesh of the magnetic circuit. methods (for
80 time steps) with Pentium II 450 MHz, 512Mo of RAM.

Number 38 - January 2002 - CEDRAT - CEDRAT TECHNOLOGIES - MAGSOFT


&
New Axial Field Electric Machine for Energy Storage.
(continue) Olivier Gergaud, Bernard Multon, Hamid Ben Ahmed, LSiR - Antenne de Bretagne de l'ENS de Cachan.

Conclusion - An effort of the radial moment of - An axial stiffness of 5 N/mm for


The 3D finite element computation the order of magnitude of milli- reluctant efforts.
allowed us to evaluate the Newton-meter for Laplace efforts,
distribution of the magnetic flux
density in the air-gap, due to the
inductor coil. The interaction
between the inductor magnetic field
and the induced currents, allowed
an appropriate evaluation of the
parasitic efforts that are exerted on
the magnetic suspension that
operates perfectly centered
(vertical, axial and angular), but
also in the case of various non -
alignment. Finally, a 2D finite Figure 6: Normal component of the Figure 7: Normal component of the
element computation allowed us to magnetic flux density under the disk of magnetic flux density under the disk.
evaluate the efforts of the reluctant average radius, at different heights.
type.
When applied to the validation
mock up (0.1 Nm at 10,000 rpm),
the computations showed that at
rating operating conditions of the
machine, the magnetic bearings
should be dimensioned in order to
support:
- A radial and axial stiffness of the
order of magnitude of Newton per
millimeter,

Figure 8: Orthoradial component of the Figure 9: Radial component of the


magnetic flux density in the air-gap. magnetic flux density in the air-gap.

Current Transformer Modelling.


Yann Le Floch , Christophe Gurin , Dominique Boudaud , Grard Meunier , Xavier Brunotte(1)
(1)(2) (1) (3) (2)
(continue)
(1)
Cedrat Technologies, Meylan, France ; (2) Laboratoire d'Electrotechnique de Grenoble, UMR 5529 INPG/UJF - CNRS, ENSIEG, France;
(3)
Schneider Electric, Grenoble, France.

To avoid the problems linked to air-


[3] F. Piriou and A. Razek, A Non-linear
gap geometrical descriptions and
Coupled 3D Model for Magnetic Field and
meshing, a new computation of t0k Electric Circuit Equations, IEEE Trans. Magn.,
is introduced which allows to take vol. 28 n2 (1992),
into account both circuit equations (a)
[4] J. Simkin and C.W. Trowbridge, On the
and surface air-gaps with thickness. used of a total scalar potential in the
This contribution strongly improves numerical solution of field problems in
problem description (geometry and electromagnetics, Int. J. Num. Meth. Eng.,
mesh of thin volume regions), Vol. 14 (1979),

computation times (4 times faster) [5] H.T. Luong, Y. Marchal, P. Labie, C. Guerin
as well as the smoothness of the and G. Meunier , Formulation of magnetostatic
isovalue results. problems in terms of source, reduced and
total scalar potentials, Proccedings of 3rd
References International Worshop on Electric And Magnetic (b)
[1] O. Biro, K. Preis, W. Renhart, G. Vrisk, Field, Liege (Belgium), 6-9 May 1996,
K.R. Richter, Computation of 3D Current
Driven Skin Effect Problem Using a Current [6] C. Guerin, G. Tanneau, G. Meunier, X.
Vector Potential , IEEE Trans. Magn., vol. 29 Brunotte, J.B. Albertini, Three dimensinal
n2 (1993), magnetostatic finite elements for gaps and
iron shells using magnetic scalar potentials,
[2] G. Meunier, H.T. Luong, Y. Marchal, IEEE Trans. Magn., vol. 30 n5 (1994). Figure 8: Flux density (Tesla) at time
Computation of Coupled Problem of 3D Eddy t=0.033s with volume air-gap (a) and
Current and Electrical Circuit by using T0 - T - with surface air-gap (b).
Formulation, IEEE Trans. Magn., vol. 34
n5 (1998),

Number 38 - January 2002 - CEDRAT - CEDRAT TECHNOLOGIES - MAGSOFT