Investigations on Failure of Electromagnetic

Brake units of Motors used in Critical

2019 IEEE 4th International Conference on Condition Assessment Techniques in Electrical Systems (CATCON) 978-1-7281-4331-6/20/$31.00 ©2020 IEEE 10.1109/CATCON47128.2019.CN0033

Applications of Sodium Cooled Fast Breeder

Reactor
Sthitapragyan Pattanayak Sarat kumar Dash S.Sivakumar R.Giribabu
Reactor Facilities Group Fast Reactor Technology Group Reactor Facilities Group Reactor Facilities Group
IGCAR IGCAR IGCAR IGCAR
Kalpakkam, India Kalpakkam, India Kalpakkam, India Kalpakkam, India
sthitapragyan@igcar.gov.in skdash@igcar.gov.in srisiva@igcar.gov.in giribabu@igcar.gov.in

Abstract— In sodium cooled Fast Breeder Reactors; power shaft preventing its movement. The motors and brake units
control is achieved through control rods which are driven by have served their purpose for long years. Recently it is being
Induction motor with Electromagnetic brake. The motor will observed that, some motors are failing to start when start
be energised upon release of brake unit. The brake coil will be command is given. Repeated start attempts, caused
energised to release the brake. The coil is of 48 V DC. If the
irreversible damage to the motor for some cases. Flashover
brake is not released properly due to low flux developed in the
brake coil, inter-turn short, liner wear; then there is a chance across the motor terminal connectors are observed due to
of Induction Motor getting damaged due to high frequency high voltage surge. It was noticed that EM brake is not
Surge generated due to interruption of locked rotor current. releasing even if rated voltage is applied to the EM brake
Hence the performance of brake unit in this application is causing locked rotor condition for the motor. By repeated
highly critical. Voltage build-up time for brake power supply is inching operations in locked rotor condition, high voltages
checked. Magnetic field developed by various EM brake units are being induced across the motor terminal connectors
has been measured and plotted as a function of applied voltage. causing flashover across the connector pins or damage to the
Brake coil has been modelled in FEMM and Electromagnetic windings. Switching-surge of 2 to 3 kV is recorded across
Finite Element Analysis has been carried out for determining
the motor terminal connectors using power quality analyzer
the number of turns and fill factor of the coil. Brake unit has
been dismantled and the effect of brake liner and the friction as shown in Fig.1.
plate on pulling force of coil is observed. The same has been
simulated in FEMM for computing the EM force on movable
armature (disk) at various distances from the brake coil.
Experiments and simulation results were used for diagnosing
the cause of non-release of the EM brake. This paper discuses
various causes of failure to the brake unit and to avoid
recurrence of similar failure event.

Keywords—Electromagnetic analysis of coil, EM brake,

Critical system

I. INTRODUCTION

Fast Breeder Reactors and their experimental test

facilities uses Electrical motors of various ratings. Some of
these motors are used in highly critical applications, which
involve repeated starting, (typically within a second), Fig. 1. Switching surge across the motor terminals
stopping, and speed reversals. For fail safe operation, these
motors are equipped with a pre-compressed spring, which
applies the EM brake in de-energised condition. During II. EM BRAKE
energised condition of EM coil, brake gets released
EM brake consists of Brake coil rated for 48 V DC, a
facilitating the movement of motor shaft. Whenever start
friction plate, brake liners and a spring as shown in Fig.2
command is given, the motor gets 415 Volts 3Φ AC supply
and the EM brake unit is energised with 48 V DC through an A. Brake Assembly
AC to DC converter. Brake power supply is derived from Brake unit of the motor is a spring loaded, energized to
the motor power supply. During de-energised condition, EM release Brake. It consists of a Fixed liner housing (fixed to
brake holds the friction plate which is fixed to the motor the motor body), Coil housing, a movable Armature (soft

1
iron disk) embedded with another liner and a circular
friction disk which is held on a square hub & free to rotate
along with the motor shaft assembled as shown in Fig. 3. A
spring is placed between the coil housing and the movable
armature. The spring force will act in such a way that it will
press the movable armature against the friction plate. The
friction plate is held between two liners, so that the Brake
will always be in action when-ever the coil is de-energized.
The Brake unit has two states;

• Brake Applied Condition: Friction Plate is tightly held

inside the fixed liner and the movable armature, by the
spring force. The motor is not able move as its shaft is
tightly held.
• Brake Released Condition: The Brake coil is energized Fig. 2. Dismantled brake unit
and magnetic force pulls the movable armature against
the spring force, such that the Friction Plate is free to
rotate along with the motor shaft when 3Φ AC supply
is applied to the motor.

However, it was observed that the Brake unit did not release
even after rated voltage was applied to the Brake coil
drawing rated current. There are only three forces on the
system; 1) Spring Force, 2) Electromagnetic Force and
3)Frictional Force

As the frictional force is observed to be minimal, this can be

Fig. 3. Brake unit in assembled condition
ignored. Hence the only possibility for non-release of Brake
is the Electromagnetic force developed is less than the B. Brake power Supply
spring force. That means, either the spring stiffness has
Power Supply for the Brake unit is 48 V regulated DC
increased or the electromagnetic force has come down or
power supply and it is derived from AC power supply of
both. Hence it has been decided to measure the spring
respective motor. Whenever start command is given for
stiffness and the magnetic field of the coil. It is to be noted
motor, 48 V DC voltage builds up across the output of
that spring force varies in linear relation with distance
regulated DC power source in a finite time. The voltage
whereas electromagnetic force varies as inverse of square of
build up time for all the Brake power supplies was checked
distance. Electromagnetic force is directly proportional to
using power quality analyzer and it was found that for one
the magnetic flux developed. As it is a DC coil, the magnetic
mechanism the delay is more than 100 ms. For other units it
flux ( ) produced is, is varying from 40 ms to 65 ms. This delay causes the Brake

unit to release with a delay. Voltage build up characteristics
= (1 ) for the healthy Brake power supply is shown in Fig. 4.

Reduction in Ampere turns can happen, [2] only if there is

some inter turn short circuit in the Brake coil (this may not
get revealed by winding resistance). Reluctance is a property
of the material and the magnetic structure. Each time a
motor mechanism is operated, the movable armature creates
an impact on the coil base unit, this can cause surface
deformities in the magnetic microstructure of that material
(can’t be seen visually) there by increases the reluctance of
the magnetic path. For investigating these claims, few
experiments and simulations was carried out, these are
described along with the observations and the inferences.

Fig. 4. Voltage build-up characteristics of brake power supply

2
 III. TEST 1

A test setup was made to measure the magnetic field TABLE I. FLUX DEVELOPED BY VARIOUS BRAKE COIL IN
GAUSS VS APPLIED DC VOLTAGE IN VOLTS
inside six identical 48 V DC Brake assemblies (Coil-1 to
Coil-6). A disk made of soft iron material was machined to Voltage Flux Developed by Different Brake Coils in Gauss
the same dimension as that of the original Brake unit Applied
moving armature. The radius of the disk is reduced by 4 in volts Coil-1 Coil-2 Coil-3 Coil-4 Coil-5 Coil-6
mm, so that a Hall Probe can be inserted in the air gap. A
5 129 151 108 121 175 215
nylon (nonmagnetic) guide rod was machined to guide the 10 308 303 226 284 319 393
travel of newly machined disk during the Brake coil 15 490 455 339 445 478 567
energized condition. 20 655 610 449 596 625 745
25 804 746 589 753 777 917
A. Experiment 30 957 863 736 907 926 1102
35 1075 1007 901 1040 1069 1254
As the magnetic field is not perfectly uniform, it is 40 1190 1135 1017 1190 1193 1398
decided to measure magnetic field at four places (say) A, B, 45 1310 1255 1130 1302 1327 1529
C, D in the inter space between the disk and the fixed coil 48 1353 1310 1190 1387 1370 1602
structure. Magnetic field readings were recorded for the six
brake units.

Fig. 5. Locations of measurement of magnetic flux

Average value of flux measured at various points at that

particular voltage has been considered as total flux
developed by that coil and the same has been tabulated vs
applied voltage in table-1.

The Brake coils are energized with a controlled DC voltage

source in volts and the magnetic field readings in gauss were
taken. A curve is plotted between applied voltage and the Fig. 6. Flux developed in Gauss (Y-axis) Vs applied DC voltage in Volts (X-
developed magnetic field. The following curves shown in axis)
Fig. 6 are a comparison between magnetic field build-up
characteristics of different Brake Units.
the rated voltage and has higher slope of Voltage vs
B. Observations Magnetic field characteristics compared to other coils.
• Brake coil-3 is having most inferior magnetic field • Brake coil-6 is a rewound coil. Probably, there may be
development characteristics. some inherent weakness in the aged coils in the Brake
• Brake coil-1, Brake coil-4 and Brake coil-5 have assemblies which are in service for the last 35 years.
almost similar characteristics and the curves almost This can be addressed with rewinding the Brake coils.
converge.
• Brake coil-2 has lower slope compared to the average C. Inference
of the other coils. It develops less magnetic field • There might be some inter-turn short-circuit / failure of
(compared to the average) when applied with rated insulation in some areas of Brake coil causing less flux
voltage. development in some coils. This is later confirmed
• Brake coil-6 has the best electromagnetic using surge comparison test.
performance. It developed highest magnetic field with

3
D. Simulation magnetic pull, so that the Brake can be released
more easily and the response time of the Brake can
Finite Element Electromagnetic Analysis (FEEA) of the also improve.
coil has been carried out to check fill factor of the coil and
determine the number of turns in the coil. Exact dimensions
of the brake coil unit have been taken and the same has IV. TEST 2
been modeled in FEMM. As the Brake unit has cylindrical
symmetry, [1] only the symmetric part of it is needed to be A. Experiment
modeled. The rated current of 0.34 A is circulated through Without friction plate, the brake unit is assembled along
the 0.22 mm dia Brake coil and the number of turns is with the spring. Then the Brake unit is energized with rated
gradually increased till the voltage drop across the Brake voltage of 48 V DC for a few times (inching operation on
coil became 48 V.
Brake unit).

Then, the Friction Plate has been put back inside the
Brake unit and assembled along with the spring. Again the
Brake unit is energized a few times.

B. Observation
• When the Friction Plate is not there, Brake unit
was drawing the rated current but brake did not get
released. After pressing the movable armature
externally, the brake releases. But the guiding path
of the movable armature was free and certainly no
friction was observed there; certainly the spring
only is holding the Brake.
• When the disk is inserted back and the Brake unit
is re-assembled, the Brake releases every time the
coil is energized. It is also observed that the
Fig. 7. Brake unit magnetic field pattern movement of the disk is only 0.6 mm as the Brake
coil energizes.
E. Observations:
The following points were observed from the
simulation; C. Inferences
Total current = 0.34 A, Voltage Drop = 47.7534 V When the Friction Plate is removed from the Brake
Voltage/Current = 140.451 Ω, Power = 16.2362 W unit, the movable armature stays at 1.7 mm additional
Number of turns = 2300, Fill Factor = 50% distance away from the coil unit. The spring
compression is less (linearly) but the magnetic force
F. Inferences: reduces exponentially. Hence the magnetic force may
• As the Voltage drops across the coil approaches 48 not be sufficient to pull the movable part. Again
Electromagnetic FEM simulation is required to verify
V DC, the resistance and power dissipation in the
the above statement.
coil is approaching the name plate values, hence the
number of turns in the Brake coil must be around D. Simulation
2300 turns. The liner dimensions are accurately noted and coil is
• The fill factor of coil is 50% only. Hence, if modeled in FEMM along with the liner and its groove.
carefully designed; a few additional turns can be Magnetic force on the movable armature is calculated in
incorporated in the same space (without changing three different cases; i.e 1) movable armature is sitting
on the coil base, 2) it is at 0.6 mm distance from the coil
the dimensions and saturating the core) during the base and 3) it is at 2.3 mm distance from the coil base.
rewinding of Brake coil. This will increase the The simulation results are as follows,

4
 Fig. 8. Movable armature is sitting on the coil base Fig. 10. Movable armature is at 2.3 mm distance from the coil base
(i.e 0.6 mm + 1.7 mm)

E. Observations:
• When the movable armature is sitting on the coil
base, the magnetic force is 680 units
• When the movable armature is at 0.6 mm
distance from the coil base, the magnetic force is
293 units
• When the movable armature is at 2.3 mm ( i.e 0.6
mm + 1.7 mm) distance from the coil base, the
force is 49 units only

F. Inferences:
As the movable armature has moved just 1.7 mm
extra, the magnetic force has reduced from 293 units to
49 units. This is equivalent of a linear wear of 1.7 mm.
An SS plate of 1.1 mm is put in place of friction plate
for simulating a liner wear of 0.5 mm. It was observed
that the brake did not release even if rated voltage of 48
V is applied. Brake released sluggishly when the applied
voltage is 52 V DC. It concludes that even a small
increase in air-gap between movable armature from the
coil base can reduce the magnetic force drastically.
Then, the resulting magnetic pull is not sufficient to pull
Fig. 9. Armature is at 0.6 mm away from the coil base the movable armature against the spring tension.

5
 V. Conclusion

• Non-release of brake is due to the combined effect

of liner wear and the inter-turn short present in the
aged brake coils.
• Inter-turn short was identified in aged brake coils,
which is the main cause for generating less
magnetic flux when applied with rated voltage.
• Reduction in liner thickness affects the position of
movable armature adversely; which results in
sharp reduction in magnetic pull acting on it upon
energizing the brake. Hence liner thickness plays a
crucial role in proper release of the brake. After
use for long years change in position of the
movable armature would have happened causing
the magnetic forces to be insufficient against the
spring force
• Recurrence of such event can be avoided by
rewinding the brake coils and replacement of brake
liners. As a part of periodic condition monitoring,
the magnetic field produced by the brake coils and
brake pick up and drop off time need to be
checked at regular intervals for detecting any fault
in the brake unit in the incipient stage itself.

REFERENCE

[1] Finite Element Method Magnetics Version 4.2

User’s Manual January 30, 2018

[2] Electromagnetic Devices, by H. C. Roters,

Wiley, 15-Jan-1941

[3] Analysis of DSRDM Electromagnet at Operating

Temperature, SK Dash, STD/99118/DN/3004/R-A