Multi-Domain Simulation and Analysis of
Electromagnetically Actuated Reclosers
Octavian Craciun, Veronica Biagini, Günther Mechler, Gregor Stengel, Christian Reuber
ABB AG, Germany
Abstract— This paper deals with the analysis and simulation
of an electromagnetic actuator for medium voltage reclosers. At
first the actuator structure and its theoretical analytical model is
shortly presented; a detailed finite element (FE) model is then
reported and described. A mid-level simulation model based on
co-energy approach is finally presented, allowing estimating the
effect of different control strategies as well as different design
parameters. Several simulation results provided by the models
are reported and commented referring to different case studies.
The results obtained by a purposely developed prototype are
finally compared with the simulation results in order to verify the
effectiveness of the proposed approaches.
Keywords— medium voltage reclosers, electromagnetic drive
unit, multi-domain modeling, Finite Element Analysis,
optimization
I.
obtained by a purposely developed prototype in order to verify
the effectiveness of the proposed simulation tools.
The paper is organized as follows. After a brief overview
related to the recloser actuation unit technologies available in
the market, a single coil permanent magnet actuator proposed
for MV application is described. Several different simulation
platforms are presented in detail and commented. An accurate
3 dimensional FE model is described in section 2. A midcomplexity model based on the co-energy approach is
addressed in section 3 whereas the simulation results and the
comparison with the experimental data are reported in section
4. The last section presents the findings and the contribution of
the paper.
II.
INTRODUCTION
Medium voltage reclosers now represent an important grid
protection device that connects different grid sources, increase
the network/grid reliability and make possible implementation
of self-healing and auto reconfiguration schemes for overhead
lines. With a high level of renewable energy penetration,
medium voltage networks are becoming bidirectional.
Therefore, the associated switching devices must ensure the
protection of newer types of power systems as well as new
types of loads. The optimal design of medium voltage
reclosers is therefore important in order to enable the required
switching capabilities.
TECHNOLOGY OVERVIEW – MEDIUM VOLTAGE
RECLOSERS
The ABB 3-phase GridShield® recloser is a well know
medium voltage protection device in which single coil
actuators are used main component driving the opening and
closing the device. It has the ability to perform as a recloser,
sectionalized or automated load break switch. The proven
design is rated for 10,000 full load operations.
Even if several different technologies are available and
have been used in order to perform such a function (e.g.
compressed air, minimum-oil and SF6 reclosers), vacuum
interrupters represent nowadays the leading technology in MV
reclosers. They feature in fact several advantages such as
reliability, safety, long life, compactness and high level of
performance.
As a consequence the reclosers performances have been
constantly improved: new modeling and simulation methods
have been developed as well as innovative manufacturing and
testing strategies.
This paper focuses on the modeling and simulation of an
electromagnetic actuator for MV reclosers. Several simulation
models with different order of detail are presented as able to
evaluate both the actuator electromagnetic performances and
the effect of different design parameters. The results obtained
by all these models are compared with the experimental results
978-1-4799-0224-8/13/$31.00 ©2013 IEEE
Fig. 1. ABB 3-Phase GridShield© Recloser
4222
Authorized licensed use limited to: UNIVERSIDADE DE SAO PAULO. Downloaded on September 26,2024 at 10:56:15 UTC from IEEE Xplore. Restrictions apply.
One pole of such a device can be considered as being
composed of two main subsystems: power and actuation. The
first is represented by the power connections and the key
element that ensures the arc extinction - the vacuum
interrupter. The second subsystem can be either mechanical or
an electromagnetic-based actuation unit. The electromagnetic
solution presents several advantages compared to the
mechanical approach, such as fewer components, higher
reliability and less maintenance.
The dynamic characteristics of electromagnetic actuators
are strongly influenced by their shape, material proprieties,
electric and mechanical elements. The magnetic, electric and
mechanical dynamics are actually mutually dependent, with
each affecting the others. Therefore, in order to ensure a fast
and efficient design, it is important to consider, at first, the
Finite Element modeling and simulation that enables virtual
prototyping of electromagnetic actuators. The next step in the
recloser design process is the emulation of its behavior in a
multi-domain modeling approach – including driving
electronics, reduced order electromagnetic model and
mechanical subsystems. Depending on the device complexity,
this solution could present a substantial challenge to identify
the optimal parameters for the system model that will
represent the behavior of the recloser. Also, when it comes to
extremely fine analysis of design parameters, a real prototype
needs to be considered.
attracting magnetic force and the actuator opening operation is
initiated.
For both the closing and opening processes, the maximum
amplitude of the coil current must be large enough in order to
cause movement over the whole stroke length.
Depending on the recloser rating, different stroke lengths
are included in the actual products. At the same time, the
driving current amplitude and control is adapted accordingly.
Therefore, depending on the application, different variants of
electronic control units are used.
For the opening operation, at the end of the stroke, the off
armature will impact the mechanical components of the stator.
Due to the abrupt stopping of the moving parts, the
components of the actuator are subjected to mechanical stress.
Another solution is to consider the Hardware in the Loop
simulation. This approach enables the coupling of two
subsystems, namely hardware and software, in a controlled
environment. This technique has been initially employed in
the aeronautic industry and nowadays it is a commonly used
technique for studying the behavior of mechatronic devices,
relay systems, distributed generation systems, automotive
applications or humanoid robot applications.
A. Single Pole Recloser Structure
The electromagnetic actuation unit used to drive the recloser is shown in Fig. 2. The main subsystems of this unit are:
stator, the two armatures (corresponding to the on and off
positions), the coil, the permanent magnets, the opening spring
and the stator.
In the closed position, the magnetic flux generated by the
permanent magnets attracts the “on” armature. The open
position is reached when the repelling opening spring is
discharged. The permanent magnets generate magnetic short
circuits at the rear side of the stator.
During the closing process, a coil current generates an
attractive force that overcomes the holding force due to the
short circuits on the rear side of the stator and subsequently the
repelling spring force. At the end of the closing process, the
“on” armature is attracted by the stator pole faces.
For the opening operation, a coil current in the inverse
direction has to reduce the magnetic force of the “on” armature.
Then the repelling spring force becomes greater than the
Fig. 2. Single Pole Recloser Structure
Additionally, once the on armature reaches the final
position relative to the stator, the kinetic energy generates an
impact with the stationary structure. This leads to mechanical
bouncing that generates an over-travel and a back-travel of the
actuator. All these effects can influence the switching
properties of the recloser over its lifetime.
III.
MODELLING AND SIMULATION
The modeling and simulation of a reclosers actuation unit
represents a rather challenging task. It includes in fact the
understanding of several different aspects strongly related to
each other. The electromagnetic actuation unit has to be
deeply analyzed in all its aspects in order to identify a good
design, possible weak points and eventually improve the
overall system performances. The mechanical and electronic
aspects are, on the other hand, significantly affecting the
4223
Authorized licensed use limited to: UNIVERSIDADE DE SAO PAULO. Downloaded on September 26,2024 at 10:56:15 UTC from IEEE Xplore. Restrictions apply.
performances playing a significant role in the overall system.
For such a reason, in order to obtain accurate and realistic
simulations, it is necessary to perform transient multi-domain
analysis able to suitably represent the magnetic and electrical
aspects as well as the mechanical phenomena.
In particular two different models are described hereafter.
The first model is more electromagnetic oriented using FEM
methods with Ansys Multiphysics, whilst the second one is
more system level oriented using a lumped model with
Dymola as simulation tool.
A. Ansys Electrostatic Model
In order to realistically analyze the electromagnetic
behavior of the above described recloser a 3D FE model was
purposely developed using the module available in Ansys
Multiphysics suited for low-frequency electromagnetic
analysis. Aiming to minimize the computational burden only
one quarter of the model was actually developed exploiting all
the available symmetries.
In a first series of 3D magnetostatic analysis neglecting
eddy currents in the actuator conductive parts, coil current and
armature position are varied in a parameter sweep.
Fig. 4 shows the magnetic flux linked to the coil plotted as
a function of current and armature position.
In Fig. 5 the differential coil inductance as a function of
position and current is plotted. In Fig. 6 the magnetic force as
a function of position and current is reported as calculated
using two methods: Maxwell stress tensor and co-energy
variation method. The excellent agreement of the results
provided by the two methods confirms the validity of both
methods, if they are based on the same FEM solution.
The results presented in Fig. 4, Fig. 5 and Fig. 6 are
substantial input for a lumped model system analysis that
includes also control electronics and power electronics of the
actuator coil circuit.
The stator part of the model consists of an inner and two
outer yoke parts made of massive steel.
Between the inner and the outer yokes two permanent
magnets are arranged to supply sufficient flux to achieve an
appropriate holding force in closed position (Fig. 2). In closed
position the gap between the big plate of the armature and the
yokes parts is closed, while the gap between the small disk on
top side is open. Both big and small magnetic steel disks
combined with the non-magnetic shaft form the armature.
In the Ansys Design Modeler preprocessing tool we
introduce a position parameter for armature that can be varied
between open and closed position. After changing this
parameter, a new FEM mesh has to be generated, before the
system can be solved.
Fig. 3. Mesh for the Finite Element actuator model
VM4_PAU
1.5
As the differential equation system (Ampere’s law) has to
be solved also in the air space the actuator model region is
surrounded by a suited air box area. A tangential field lines
boundary condition was imposed onto the lateral surfaces.
Flux linked to coil - psi [Wb]
1
In order to calculate the force on the armature applying the
Maxwell stress tensor method, the high permeable volumes of
armature need to be surrounded by at least one layer of
elements with relative permeability of vacuum.
The capability of Ansys Multiphysics for generating user
defined mesh was strongly exploited aiming to obtain a good
mesh distribution. Considering the complexity and the size of
the 3D model in fact it was important to find a right
compromise between complexity and accuracy in order to
achieve accurate results and a reasonable computational time.
A detail of the mesh distribution in the actuator parts is shown
in Fig. 3. The non-linearity of ferromagnetic material was also
considered assuming the material behavior described by its BH curve.
Gap [mm]
0.5
0.2
0
1
2
00
4
6
8
-0.5
10
12
14
max
16
-1
-1.5
-90
-70
-50
Fig. 4. Ansys
3D
Flux linked to the coil
(gap width at big disk)
-30
-10 0
10
Current [A]
30
50
70
magnetostatic
parameter
as a function of current and
90
sweep:
position
4224
Authorized licensed use limited to: UNIVERSIDADE DE SAO PAULO. Downloaded on September 26,2024 at 10:56:15 UTC from IEEE Xplore. Restrictions apply.
Inductance
0.05
gap [mm]
0
0.2
L [H]
0.045
1
2
0.04
4
0.035
6
0.03
8
0.025
10
12
0.02
14
0.015
max
16
0.01
0.005
00
-100
-80
-60
-40
-20
0
0
20
40
60
Current
[A]
80
100
Fig. 5. Ansys
3D
magnetostatic
parameter
Differential coil inductance LD as a function of current and position
sweep:
Magnetic Forces on Armature
Air Gap [mm]
0
0.2
1
2
4
6
8
10
12
14
max
16
Force [N]
10000
8000
6000
4000
Mesh morphing causes an extreme deformation of finite
elements in the air gap, if the gap becomes very small. Using a
morphed mesh as described above, the calculated forces are
systematically calculated low compared to a mesh that is
generated
from
scratch
at
each
position
(Fig. 7). In a transient simulation the calculated force for a
certain time step has to be corrected by a factor that is
determined from curves in Fig. 7. This is even more important
because the start of armature motion is closely related to the
zero-crossing of total force at minimum gap width, (3) prepare
a function call for armature displacement during time iteration,
(4) prepare a function call for force calculation using Maxwell
stress tensor at each time step and (5) prepare the time step
integration loop with the determination and printout of total
coil current including back-emf of eddy currents in the
conductive parts of the actuator, the total force on armature
including the dynamic equation of the armature mass
accelerated by a combination of magnetic and spring forces.
Fm(t) + Fs(t) = M * a(t)
, where Fm is the magnetic force acting on the armature, Fs
the spring force that is oriented in opposition to the magnetic
holding force of the armature, M the mass of the armature and
finally a(t) the armature acceleration.
6000.0000
Magnetic Force [N]
2000
5000.0000
00
4000.0000
-2000
-100
-80
-60
-40
-20
0
0
20
40
60
New mesh for each position
3000.0000
Current [A]
80
100
Fig. 6. Ansys 3D magnetostatic parameter sweep: Force as a function of coil
current and armature position
Morphed mesh
2000.0000
1000.0000
B. Ansys Transient Electromagnetic Model
As the influence of the eddy currents in the electromagnetic
behavior of the actuator is not negligible, a complete transient
3D FE simulation had to be performed. This means to develop
in Ansys a certain number of APDL code snippets in order to
extend the Workbench standard functionalities.
The APDL code snippets have the following functionalities:
(1) change solver settings to enable eddy current calculation.
This includes activation of time step integration for the Ansys
element Solid236, (2) prepare morphing mesh during transient
simulation, where a parameterized field of node coordinates
for all element nodes between the stator parts and the armature
parts is calculated by just solving Laplace equation in that area
while the armature is in an intermediate position. This is
equivalent to the determination of a temperature distribution
between T=1 at the surface of moving parts and T=0 at the
surface of stator parts. If the armature is displaced by dz along
the z-axis direction, the z-position of every node in that area is
shifted by
dz(node) = T(node)*dz(armature)
0.00000
-1000.0000
0 1gap2
minimum
at big disk
3
4
5
6
7
8
9
Position [mm]
10 11 12 13 maximum
14 15 16gap
at big disk
Fig. 7. Magnetic force calculated with new mesh for each position or with
morphed mesh (both as magnetostatic Ansys result)
C. Dymola Model
A 3D finite element (FE) model represents nowadays the
most accurate approach available for electromagnetic
simulation. It allows representing the nonlinearities of
ferromagnetic materials, the eddy currents distribution as well
as the effect of the leakage fluxes. Nevertheless, even when all
the existing symmetries are exploited to reduce the problem
complexity, the computational burden related to a finite
element simulation results too intensive for a system level
analysis aimed to check the behavior of a complete drive. An
intermediate level of accuracy may be obtained with a much
lower computational burden by using the results of the FE
4225
Authorized licensed use limited to: UNIVERSIDADE DE SAO PAULO. Downloaded on September 26,2024 at 10:56:15 UTC from IEEE Xplore. Restrictions apply.
static simulation as an input for a lumped model suitably
developed in Modelica-Dymola simulation environment.
Such model includes an electromagnetic actuator model, a
detailed mechanical system, the supply and regulation
converter and the control system. One of the main advantages
of this software is the possibility to be coupled with external
optimization tools. For the sake of brevity, the mechanical
model as well as the control system will not be addressed in
detail in this paper focusing only on the electromagnetic
actuator model.
The lumped model has been developed using the results of
3-D Ansys electrostatic analysis as shown in Fig. 4, Fig. 5 and
Fig. 6.
This approach is based on the direct evaluation of the force
using the Maxwell stress tensor and the linked flux available
in Ansys post processing.
Alternatively both the force F i, x
and the linked flux
i, x can be determined directly as derivatives of co-energy
C E i, x which is also a result of Ansys 3D transient simulation.
F i, x
dC E i, x
dx
i const
(1)
i, x
dCE i, x
di
x const
(2)
The numerical deviation between both evaluation methods
was found to be less than 1% thus confirming the consistency
of the Ansys results.
Once the main electrical and mechanical quantities (i.e.
linked flux and force) are identified they can be used to
evaluate the lumped parameters and implement the mechanical
and electrical equations.
In particular, focusing on the electromagnetic model, the
main lumped parameters which need to be evaluated are: coil
resistance R, differential inductance LD and motional
coefficient M.
The first parameter is evaluated by considering the coil
geometry, the number of turns and the material characteristics
under the hypothesis of negligible skin effect and temperature
influence. The second and the third are evaluated by
numerically approximating the derivative of the flux respect to
current and position as reported in eq.5.
LD
i, x
i
x const
M
i, x
x i const
(3)
Every parameter is finally introduced in the Dymola model
as look-up tables.
Once determined the above lumped parameters a first
model of the actuator can be implemented under the
hypotheses of negligible eddy currents distribution in the
electromagnetic conducting parts according to the following
equations:
U t
UR t
U t
R it
UE
R it
d
dt
i, x
v(t )
x
i, x di
i
dt
(4)
, where U(t) is the coil voltage, R is the value of coil
resistance, i(t) the coil current and
function of current and position.
i, x
the flux as a
Considering that both the stator and the moving armature
are designed as bulky ferromagnetic material, the effect of the
eddy currents distribution cannot be neglected without
introducing a significant error. The model described by eq. 4
was therefore slightly modified introducing a secondary loop
equation magnetically coupled with the main one meant to
represent the effect of the eddy currents in the conducting
materials. In order to optimize the parameters in the secondary
loop the results provided by Ansys transient 3D can be used.
Defining LD1 , LD2 , M1, M2 respectively the differential
inductances and the motional coefficients for the main coil and
for the secondary eddy current loop the resulting system of
equations works out as:
U1 t
R1 i1 t
M 1 v(t )
LD2
di1
dt
U2 t
R2 i t
M 2 v(t )
L D1
di 2
dt
1
2
i1 , i 2 , x
i2
di 2
dt
i1 , i 2 , x
di1
dt
i1
(5)
In particular the eddy currents will introduce a delay in the
actuator operation.
IV.
ANALYSIS AND VALIDATION
In order to validate the developed modeling and simulation
approaches, a dedicated Hardware in the Loop test rig has
been set-up. This contains appropriate electronic control units
enabling to power the electromagnetic actuator in a controlled
environment, as presented in [6]. The position measurement is
realized by using classical linear position transducers [7]. This
section presents the validation of both Ansys 3D and Dymola
models.
A. Ansys 3D transient validation
In order to evaluate the results obtained by using the Ansys
model the outcomes from transient 3D FE simulations were
compared with measured data obtained from a purposely
developed prototype.
The good agreement between experimental and simulation
results may be appreciated in Fig. 8. The measured (red) and
the FE simulation (blue) positions as a function of time are
reported in the same graph highlighting a rather good
agreement – in a range of about 3% if the actuator speed is
considered as reference for the error calculation. The slight
discrepancy between the two curves is due to the simplified
mechanical model implemented, the cumulated effects of
numerical approximations as well due to the tolerances of the
measurement equipment. The simplified linear spring model
4226
Authorized licensed use limited to: UNIVERSIDADE DE SAO PAULO. Downloaded on September 26,2024 at 10:56:15 UTC from IEEE Xplore. Restrictions apply.
adopted together with some neglected second order effects
(e.g. friction, bouncing behavior) introduce, as expected, some
ideal actuator behavior. Anyway in overall terms the results
confirm the validity of the 3D FE Ansys model. Because of
the relatively high accuracy the model can be used in a design
phase for improving the electromagnetic performances of the
actuator before coming to the prototyping phase. Moreover it
can be used to tune the system level lumped Dymola model
especially adjusting parameters related to eddy currents in the
actuator structure.
Fig. 10. Dymola model validation: position [p.u.] as a function of time
V.
CONCLUSION
In this paper the analysis of an electromagnetic actuation
unit for MV reclosers has been investigated referring to two
different modeling and simulation approaches. Firstly, a
complete 3D FE model of the actuator has been developed
using Ansys Multiphysics. This latter was used in order to
create a combined mechanical-electromagnetic reclosers
analysis tool using a lumped parameter approach developed on
Dymola Modelica.
Fig. 8. Ansys 3D transient validation: position [p.u.] as a function of time
This developed modular modeling and simulation method
enables faster and reliable design of medium voltage reclosers.
Different problems can be thus addressed by selecting the
corresponding analysis approach – on the complete system or
only on one subcomponent. The obtained results have been
successfully validated against measurement on prototype.
B. Dymola Model Validation
REFERENCES
[1]
[2]
[3]
Fig. 9. Dymola Model of Controlled Single Phase Recloser Mechanism and
Actuator Sub-Model to be replaced
The main results coming from the system level simulation
realized using Modelica-Dymola are reported in Fig. 10 where
the position as a function of time is presented. The validity of
the approach is again confirmed by the comparison with the
experimental results. The measured (red) and the FE
simulations (blue) positions are again reported in the same
graph highlighting a rather good agreement – within the same
error range as the Ansys 3D validation. Because of the
relatively short simulation time the model can be used for
modifying the control strategy and testing a variety of
operating conditions.
[4]
[5]
[6]
[7]
G. Stengel, G. Mechler, S. Kock, J. Derkx, Multi-domain simulation and
validation of a mechatronic drive system for medium voltage circuit
breakers (in German), Mechatronik 2010 Conference, pp. 133-139,
Technical University of Dresden, Germany.
E. Dullni, H. Finkand C. Reuber, A Vacuum Circuit-Breaker with
Permanent Magnet Actuator and Electronic Control, CIRED Conference
1999.
B. Delinchant, G. Gruosso, F. Wuztr, Two Levels Modeling for the
Optimization of Electromagnetic Actuators, IEEE Transaction on
Magnetics, Vol. 45, No. 3, March 2009.
S. Fang, H. Lin, S. L. Ho, Transient Co-simulation of Low Voltage
Circuit Breaker With Permanent Magnet Actuator, IEEE Transaction on
Magnetics, Vol. 45, No. 3, March 2009.
L. Nowak, Iterative Procedure for 3D Modelling of Electromagnetic
Actuators, IEEE Transactions on Magnetics, Vol. 31, No. 3, May 1995.
O. Craciun, G. Stengel, C. Reuber, Hardware in the Loop Multiobjective Optimization of Electromagnetic Actuators, International
Journal of Distributed Energy Sources, Vol. 9, No. 1, pp. 103-114,
March2013.
http://www.novotechnik.de/uploads/tx_extprodfind/T_TS_e.pdf, August
2013.
4227
Authorized licensed use limited to: UNIVERSIDADE DE SAO PAULO. Downloaded on September 26,2024 at 10:56:15 UTC from IEEE Xplore. Restrictions apply.
Powered by TCPDF (www.tcpdf.org)