International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 12, No. 4, December 2021, pp. 2030~2043

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v12.i4.pp2030-2043  2030

Design and simulation of 5kW BLDC motor with half-buried

permanent magnets using an existing stator body

Budi Azhari, Pudji Irasari, Puji Widianto

Research Center for Electrical Power and Mechatronics, Indonesian Institute of Sciences (LIPI), Bandung-Indonesia

Article Info ABSTRACT

Article history: This paper proposes a design of a 5 kW, 100 volts brushless direct current
(BLDC) motor using an existing stator connected to an inverter and equipped
Received Jun 3, 2021 with Hall sensors. The stator is a radial flux motor-type with 54 slots positioned
Revised Sep 8, 2021 at the outer side of the machine. In this case, the design is focused on the rotor
Accepted Sep 15, 2021 components and winding configuration. However, the inverter aspects are also
taken into account. At the same time, it considers the expected outputs: voltage,
power, speed; and some limitations: maximum current and flux density. Finite
Keywords: element magnetic-based simulation is performed to extract the magnetic flux
distribution, and analytical calculations are then conducted to obtain the output
BLDC motor values and characteristics. The results show the BLDC motor at nominal speed
Cogging torque produces 5.1 kW output power with 122.34 V voltages, 97.09% efficiency, and
Flux distribution torque of 32.82 Nm. The maximum torque and rotation speeds are 51.39 Nm
Half-buried magnets and 4,150 rpm respectively, while the peak-to-peak cogging force is 1.35 Nm.
Torque-speed characteristics It can be concluded that the BLDC motor has a good performance and is
compatible with the connected inverter.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Budi Azhari
Research Center for Electrical Power and Mechatronics
Indonesian Institute of Sciences
154D/21 Sangkuriang st, Coblong, Bandung-West Java 40135, Indonesia
Email: [email protected]

1. INTRODUCTION
Recently, the brushless DC (BLDC) permanent magnet motor has been used in many fields, for
example in industrial machines, medical appliances, robotics, also for currently growing trend: electric
vehicles (EVs) [1]-[4]. It is renowned for its excellent properties: simpler construction, relatively small in
size and weight, high efficiency, high torque density, good dynamic response, low noise, and simple
maintenance [5]-[10]. As an actuator, the BLDC motor has characteristics that basically combine the
advantages of DC brushed motor and alternating current (AC) synchronous motor and at the same time
eliminate their main drawbacks. Because the brushed DC motor suffers mechanical losses from the use of
brushes and commutators, the BLDC motor avoids their presence by using an inverter to convert the input
DC-current into a frequency-controlled AC wave. Furthermore, the inverting process employs electronic
switches whose states’ arrangement can be set to determine the resulted AC wave’s frequency [11].
Therefore, a wider range of speed can be obtained along with a higher torque compared to the AC
synchronous motor or the induction motor [12].
Physically, the BLDC motor has three main parts: the stator, the rotor, and the electronic
components. The stator usually has a similar design as the DC brushed motor’s, with some windings to
generate synchronous flux [13]. Meanwhile, the rotor has several poles of permanent magnet that will be
locked to the current-conducting winding to make a rotation. The number of the windings and the poles

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  2031

generally define the type and the parameter of the motor; also determine the electrical as well as the
mechanical outputs. Because the motor operates with inverted DC voltage, the frequency of the resulted AC
wave will influence its motion. Here, The BLDC motor performs electronic commutation by using solid-state
components, instead of a mechanical commutator like the one in the conventional DC motor.
For designing the BLDC motor, the model can either be approached from a synchronous motor [14]
or be derived from a common brushed DC motor [15]-[17]. Nowadays, the BLDC motor can also be
designed as a slotted machine to optimize the air gap flux density and to achieve higher power density [12],
or as a slotless one, as it can simplify its structure and minimize resulted cogging torque [13]. Nonetheless,
the design process of the BLDC motor needs the desired outputs to be set at the beginning of the process
[18]. From those pieces of information, the settings and dimensions of the BLDC components can be
determined. There will be a challenge for designing a specific BLDC motor with certain output values if a
certain motor part has existed.
In this paper, a 5 kW, 100 volts BLDC motor design is proposed, but by utilizing an existing stator
body. The provided stator was previously used for an asynchronous motor. It has multi slots but no coil yet.
Hence, the rotor and the stator winding will be configured altogether. Similar studies were previously
conducted by [19], [20], but they do not consider the inverter aspects. In this paper, the dimensions of the
rotor components, the winding arrangement, as well as the inverter settings are determined adapting to
properties and dimensions of the available stator. They also consider some output parameters, including
expected terminal voltage, output power, and rotation speed; also aware of several limitations: maximum
current and flux density. Simulations using finite element magnetic-based software are performed to extract
the magnetic distribution, so analytical calculations can be conducted later to obtain the output values, and
characteristics of the motor. From the research, the presented design process that covers the whole
components of the BLDC motor and uses an existing asynchronous motor’s stator can be reviewed or applied
for the next researches.

2. DESIGN OF BLDC MOTOR

In this paper, the design of the BLDC motor consists of three main parts: the existing stator body,
the motion controller to be regulated, and the rotor as well as the winding configuration to be designed. The
first part will be explained, while the later configurations will be discussed subsequently.

2.1. The existing stator design of the BLDC motor

The existing stator body with fixed dimensions is shown in Figure 1. It is a radial flux motor-type
with a cylindrical shape. An axial hollow present at the middle, parallel to the normal of the circular face.
Meanwhile, the detailed dimensions are tabulated in Table 1. Given the available stator structure, it will be
positioned as the outer part of the BLDC motor, while the rotor in the inner part will be designed.
From Figure 1, the existing stator is composed of yoke and slots-teeth regions. The yoke is formed
from sheets of silicon steel material; with a thickness of 0.5 mm each. The magnetization curve of this
material (50H1300-type) is presented in Figure 2. In this matter, the use of ferromagnetic material for the
yoke aims to minimize the magnetic reluctance along the flux passage. Meanwhile, the laminations are
arranged axially (or parallel to z-axis according to Figure 1 to minimize the eddy current loss. To fix the
lamination sheets, they are bolted on six spots at the sheets’ edges. The slots and teeth are in the inner part,
with each amount 54 units. Furthermore, every tooth makes salient at its inner end. This topology is useful to
reduce the teeth' reluctance and to minimize cogging torque [21], [22], hence improve the rotation
smoothness.
Next, the basic model of the BLDC motor is to be stated based on the stator design. Since the
number of the slot is known, it can be considered to determine the number of the winding and the pole pair.
Generally, the higher the number, the torque and the iron loss will be higher, while the lesser number will
make higher speed [23]. Other considerations may also be put into the table, including the manufacturing
difficulties and costs. In this case, the 4 poles-6 windings model of the BLDC motor is chosen. Considering
the existing structure, the 54 stator slots will be divided into 6 groups of slots. They will be allocated for 6
winding sections of 3 phases. Each section is supported by 9 slots, thus the of slots per pole per phase (q) is.
𝑆𝑠
𝑞= (1)
2𝑝𝑚

Meanwhile, the span of the pole is [24]

𝑆
𝑠𝑝𝑎𝑛 = 𝐼𝑁𝑇 [ 𝑠 ] (2)
2𝑝

Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)
2032  ISSN: 2088-8694

Ss is the number of slots, 2p is the pole section number, and m is the number of all phases.

Figure 1. Design of the existing stator

Table 1. Dimension of the existing stator parts

Part, symbol Dimension Unit
Number of slot, Ss 54 units
Inner radius, ri 0.0738 m
Inner diameter, Di 0.1476 m
Outer radius, ro 0.1150 m
Outer diameter, Do 0.2300 m
Axial length, lax 0.1030 m
Yoke thickness, hys 0.0190 m
Slot opening width, Wss 0.00288 m
Salient width, Wtts 0.00571 m
Tooth width, Wts 0.0045 m
Slot width, Wsl 0.0056 m
Slot pitch, s 0.0086 m Figure 2. Magnetization curve of stator’s silicon steel
Slot depth, Dss 0.0200 m
Slot area, as 109.9 mm2

2.2. The BLDC controller topology

To make the BLDC motion controllable, first, three Hall sensors are placed on the stator (Figure 3
(a)), to detect the rotor position. Next, the input DC voltage source supplying the three-phase windings
should be first connected to an inverter. Here, 6 MOSFETs will be employed in a full-wave bridge inverter
circuit (Figure 3 (b)). The MOSFETs will play a role as the switches. The states of the MOSFETs are
regulated based on the Hall sensors’ output to determine the movement of the rotor. When a switch is closed,
DC input current will flow, and vice versa. As the states of ‘closed’ and ‘open’ occur sequentially under a
certain period, square wave current signals will be formed. As presented in the circuit (Figure 3 (b)), the
signals flow to the stator windings, which are connected in a star configuration. The current that flows in a
certain phase winding will generate magnetic flux which flows to the rotor. The magnetic interactions
between the phase winding’s flux and the rotor poles will lock the poles and rotate the rotor. The rotating
direction can be set based on the order of the flowing current to the phase windings, while the rotating speed
can be regulated based on the state’s period or the switching frequency.
In this case, the setting of the Hall sensors, the respective MOSFETs’ states, and the resulted current
flow among the phase windings are presented in Table 2. One Hall sensor changes its state for every 60 o
electrical cycle, so it requires 6 stages to complete one Hall sensor cycle. The flow direction between the
windings depends on the combination of the two conducting MOSFETs. Meanwhile, the generated square
wave signal flows to all phases are shown in Figure 3 (c). Each phase is conducting current for a 120o
electrical cycle, with two phases will deliver current altogether for every conducting period (every 60 o
electrical cycle).

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2030 – 2044
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2033

(a) (b) ©

Figure 3. These figures are; (a) Hall sensors placement among six winding sections, (b) the full-wave bridge
inverter circuit, (c) the resulted square waves in one complete electrical cycle, θ (the below numbers are the
states of the Hall sensors 1,2, and 3 respectively)

Table 2. MOSFETs setting and resulted current flow

Hall sensor MOSFETs’ state Current flow
1 2 3 S1 S2 S3 S4 S5 S6 From phase To phase
0 1 0 1 0 0 1 0 0 A B
1 1 0 1 0 0 0 0 1 A C
1 0 0 0 0 1 0 0 1 B C
1 0 1 0 1 1 0 0 0 B A
0 0 1 0 1 0 0 1 0 C A
0 1 1 0 0 0 1 1 0 C B

2.3. Rotor and winding configurations of the BLDC motor

The 4-poles rotor is configured to produce a sufficient amount of magnetic flux, which is
proportional to the output voltage, power, and maximum torque. Under the operation of 120 o square wave
signals, the flux generated in the no-load condition is.

𝜑𝑠𝑞 = 𝑎𝑖(𝑠𝑞) 𝜏𝑝 𝑙𝑎𝑥 𝐵𝑝𝑚 , (3)

𝜏𝑝𝑚 𝐵𝑎𝑣
𝑎𝑖(𝑠𝑞) = = (4)
𝜏𝑝 𝐵𝑝𝑚

φsq is the resulted flux under the square wave signal (Wb), τp and τpm are circular pole pitch (m) and pole
length (m) respectively, lax is the axial length of the stator (m), Bav and Bpm show the average flux density and
magnet’s flux density, all in tesla. In this design, the value of ai under square wave signal is equal to the value
under sinusoidal mode. The ratio between the two modes is 1/kf, where.

4 𝑎𝑖 𝜋
𝑘𝑓 = 𝑠𝑖𝑛 ( ) (5)
𝜋 2

From (3) and (4), the dimensions of the magnets will determine the resulted flux.
The generated magnetic flux results in magnetic interaction that will lock the poles to certain phase
windings. It will also result in back-electromotive force (back EMF), whose value is affected by the resulted
flux and the winding configuration [25].

𝐸𝑏 = 𝑐𝐸,𝑠𝑞 𝜑𝑠𝑞 𝑛𝑠 , (6)

𝑐𝐸,𝑠𝑞 = 8𝑝𝑁𝑝ℎ 𝑘𝑤1 , (7)

𝑘𝑤1 = 𝑘𝑑 𝑘𝑝 𝑘𝑠 (8)

In (6)-(8), Eb is one phase back EMF (V), cE,sq is a coefficient of configuration, ns is the rated speed of
rotation (rpm), Nph is the number of turns per phase, and kw1 is winding factor in fundamental harmonics.
Furthermore, kd is distribution factor, kp is pole shortening factor, and ks represents pole skewing factor.
The flux flows from the poles to stator teeth and stator yoke through the air gap, and back to rotor
yoke and suitable poles. While the resulted magnetic flux is arranged to meet the expected output, the flux
Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)
2034  ISSN: 2088-8694

density in both yokes and the stator teeth should be kept below 2 T [26] to avoid the saturation phase (based
on the magnetization curve in Figure 2). The equations of the flux density in the respective parts are.
𝜑𝑠𝑞
𝐵𝑌𝑆 = , (9)
2ℎ𝑦𝑠 𝑙𝑎𝑥 𝑓𝑙𝑎𝑚

𝐵𝑝𝑚 𝜏𝑠
𝐵𝑇𝑆 = , (10)
𝑤𝑡𝑠 𝑓𝑙𝑎𝑚

𝜑𝑠𝑞
𝐵𝑌𝑅 = , (11)
2ℎ𝑦𝑟 𝑙𝑎𝑥 𝑓𝑙𝑎𝑚

BYS, BTS, and BYR are flux density in the stator yoke, the stator tooth, and the rotor yoke successively (all in
tesla). Meanwhile, flam is the lamination factor of the respective yoke, hys and hyr are the thickness of stator
and rotor yokes (m), while wts is the stator teeth width (m). Based on (9)-(11), the rotor yoke thickness should
be carefully set, by considering the dimensions of the existing stator’s yoke and teeth. Another to be
determined parameter is winding configuration. Based on (6)-(8), the number of the stator turn is related to
the resulted back-EMF. However, it is also affected by the maximum armature current [27], which depends
on the current density and the inverter signal mode.

3𝑙𝑎
𝐼𝑎 = 𝐽𝑎 𝑎𝑤 ,𝐼𝑎(𝑠𝑞) = √ (12)
2

Ia is armature current (A), Ja is the phase winding current density (A/m2), aw is the conductor front area (m2),
and Ia(sq) is the armature current for operation of 120 o square wave signal. The number of turns in a slot
should not exceed the filling factor (ffs) of the slot. The cumulative winding dimension then also affects the
phase winding resistance (Ω).

𝑁𝑠 𝑎𝑤 = 𝑓𝑓𝑠 𝑎𝑠 (13)
𝜌𝑅 𝑙𝑤
𝑅𝑎 = . (14)
𝑎𝑤

Ns is the number of turns per slot and as is the surface area of the slot (m2). Meanwhile, ρR is the specific
resistance of the wire (Ωm) and lw is the wire length of the phase winding (m).
When the bridge rectifier as in Figure 3 (b) is installed and energized, there will be two conducting
phase windings for every conducting period. According to the formed circuit, the phase-to-phase
instantaneous terminal voltage will be.

𝑣𝑖𝑛 = (𝑒𝜙1 − 𝑒𝜙2 ) + 2𝑅𝑎 𝑖𝑎 + 2𝑋𝑠 𝑖𝑎 . (15)

vin is the input voltage (V) in the motor input terminal. Operation (eφ1 – eφ2) represents the line-to-line back-
EMF voltage (V) generated in the stator. Notation ia is the instantaneous armature current (A), while Xs is the
synchronous winding reactance (H) which the value is obtained from the equations as shown in [28].

𝑋𝑠𝑑 = 𝑋𝑙 + 𝑋𝑚𝑑 ,𝑋𝑠𝑞 = 𝑋𝑙 + 𝑋𝑚𝑞 , (16)

𝐿𝑖 𝑁1 2
𝑙
𝑋𝑙 = 4𝜋𝑓𝜇0 (𝜆𝑠 + 1𝑒 𝜆𝑒 + 𝜆1𝑑 ), (17)
𝑝𝑞 𝐿𝑖
𝑘𝑤1 𝑁1 𝑟𝑝𝑚_𝑜 2 −𝑟𝑝𝑚_𝑖 2
𝑋𝑚𝑑 = 2𝑚𝑓𝜇0 ( ) 𝑘𝑓𝑑 , (18)
𝑝 𝑔′ 𝑑

𝑘𝑤1 𝑁1 𝑟𝑝𝑚_𝑜2 −𝑟𝑝𝑚_𝑖 2

𝑋𝑚𝑞 = 2𝑚𝑓𝜇0 ( ) 𝑘𝑓𝑞 . (19)
𝑝 𝑔′𝑞

X1 is leakage reactance (Ω), while Xmd and Xmq are mutual reactance according to dq-axis respectively. μo is
air permeability (4π × 10−7 H/m), Li is the effective winding length in a radial direction (m), N1 is the number
of series winding in one phase, λs is slot leakage permeance, λe is winding end’s leakage permeance, λ1d is
differential leakage flux permeance, g’d and g’q is equivalent air gaps in dq-axis (m), while rpm_o and rpm_i are
pole’s outer and inner radius (m). Furthermore, kfd and kfq are form factors according to dq-axis.
The electromagnetic power (Pelm in watt) is the output power of the BLDC motor. This and the
torque of the motor are also affected by the magnetic flux (from the poles settings) and the armature current
(based on the winding configuration). They can be calculated by using the equations.

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2030 – 2044
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2035

𝑃𝑒𝑙𝑚 = 𝑒𝑙𝑙 𝐼𝑎(𝑠𝑞) , 𝑒𝑙𝑙 = 𝑒𝜙1 − 𝑒𝜙2 , (20)

𝑐𝐸,𝑠𝑞 𝜑𝑠𝑞
𝑇𝐼_𝑛 = 𝑘 𝑇 𝐼𝑎(𝑠𝑞) , 𝑘 𝑇 = . (21)
2𝜋

TI_n is the torque under nominal current (Nm). Next, the torque characteristics give information about the
torque-speed relation, especially the limit values that should be obeyed to produce optimum performance. It
is closely related to the rotation speed, as described by the equations as shown in,

𝐼 𝑇
𝑛 = 𝑛𝑜 [1 − ] = 𝑛𝑜 [1 − ]. (22)
𝐼𝑜 𝑇𝑜

To is stall torque (Nm), or the maximum torque that stops the rotation of the motor. no and Io are no-load
rotation speed (rpm) and no-load current (A). They can be calculated by using the equations as shown in,
𝑣𝑖𝑛 𝑣𝑖𝑛
𝑛𝑜 = , 𝐼𝑜 = . (23)
𝑐𝐸,𝑠𝑞 𝜑𝑠𝑞 𝑅𝑎

When the motor is energized, some losses will be dissipated, both electrically and mechanically. The
copper loss (Pcu) and the core loss (Pc) are two dominant electrical losses, all are stated in watt. The core loss
itself is the sum of hysteresis and eddy current loss. These electrical losses are affected by the armature
current, the winding configuration, and the resulted maximum flux density. The values of all those losses can
be calculated using the equations as shown in,
2
𝑃𝑐𝑢 = 𝑚𝑎 (𝐼𝑎(𝑠𝑞) ) 𝑅𝑎 , (24)

𝑃𝑐 = 𝐶ℎ 𝑓𝐵𝑝 𝑎+𝑏𝐵𝑝 + 𝐶𝑒 𝑓 2 𝐵𝑝 2 . (25)

Notation ma is the number of active phases during a conducting period, Ch is hysteresis coefficient = 0.0025
while Ce is eddy current coefficient = 7.94x10 -5. Notations of a and b are constants that depend on the electric
steel material that is used. In this case, the value of a = 1.8317 and b = 0.0035. Furthermore, Bp is the resulted
maximum flux density in the respective body (T).
Next, the mechanical loss (Pm, in watt) also appears during the motor’s work. When the rotor moves,
there will be losses due to friction in bearing (Pb, in watt) and air gap (Pgap, in watt). The losses’ values can
be calculated using the following equations,

𝑃𝑚 = 𝑃𝑏 + 𝑃𝑔𝑎𝑝 , (26)

𝑃𝑏 = 𝑘𝑓𝑏 𝑚𝑟 𝑛𝑠 × 10−3 , (27)

𝑃𝑔𝑎𝑝 = 2𝐷𝑟 3 𝑙𝑎𝑥 𝑛𝑠 3 × 10−6. (28)

kfb is bearing friction coefficient whose value ranges between 1-3, mr is the rotor total mass (kg), and Dr is the
rotor outer diameter (m). From (26)-(28), the losses’ values are also affected by the rotor dimension.
Moreover, additional losses (Padd in watt) including loss from leakage flux in slots and leakage flux in
windings also exist. In this case, the value of the remaining losses is approximated 20% of the core loss.
Finally, the total loss (Pl_tot in watt) and the efficiency (η in %) of the motor is.

𝑃𝑙_𝑡𝑜𝑡 = 𝑃𝑐𝑢 + 𝑃𝑐 + 𝑃𝑏 + 𝑃𝑔𝑎𝑝 + 𝑃𝑎𝑑𝑑 , (29)

𝑃𝑒𝑙𝑚
𝜂= × 100. (30)
𝑃𝑒𝑙𝑚 +𝑃𝑙_𝑡𝑜𝑡

Another obstacle when rotating the rotor is the cogging torque phenomena. It is a fluctuating torque
resulted from changing attractive forces between the rotor poles and the ferromagnetic stator body. The
phenomenon will cause ripple during the rotation and if the force is quite strong, it will cause vibrations and
noises [29]. Next, the starting torque needs to increase. The cogging torque is stated by the formula as shown
in,

1 𝑑ℜ
𝑇𝑐 = − 𝜑 2 . (31)
2 𝑑𝜃𝑎

Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)
2036  ISSN: 2088-8694

Ɍ is the air gap reluctance (At/Wb) and θa is the rotor position during rotation (rad).

3. DESIGN AND ANALYSIS METHOD

The flowchart for the design and analysis process of the BLDC motor with an existing stator body is
shown in Figure 4. After elaborating on the stator design (Figure 1 and Table 1), the rotor design and the
winding configuration are determined. However, prior to that step, some parameters need to be stated as the
design references. They are including the expected output values, the desired settings’ values, and the limit
values for some parameters. The references’ values are listed in Table 3.
The design process is then conducted considering the references’ values and using the analytical
calculations of (1)-(31). The selected inverter configuration is taken into consideration. In (3)-(7), (9)-(12),
(20), (21), and (23), (24), the generated square wave signals from the inverter affect the resulted dimensions
and output values. The electronic switching configuration is also considered, as the resulted current and
polarity among the winding phases at a certain time will determine the magnetic flux flow.

Table 3. References values for design

Part, symbol Dimension Unit
Inverter voltage, Vin 100 V
Maximum inverter current, Imax 120 A
Number of the pole, 2p 4 -
Nominal frequency, f 50 Hz
Nominal rotation speed, ns 1,500 rpm
Number of phases, m 3 -
Number of active phases during
2 -
conduction, ma
Conduction period, Tc 0.02 s
Inverter voltage, Vin 100 V

Figure 4. Flowchart of the design, simulation, and analysis

After the design is complete, the magnetic flux distribution is extracted through finite element
magnetic-based simulation using FEMM software under no current and nominal current conditions. The
magneto-static principle is applied, in which the magnetic fields are treated as a time-invariant function. In
this case, the flowing current that produces certain magnetic fields is fixed at a specific state, so that the
magnetic parameters at that state can be extracted. To change the design condition (i.e., rotating the rotor) or
the phases’ current values, the state is updated, and the new parameters’ values can be obtained again. Hence,
to model time-variance progress, several different states are captured sequentially. In magneto-static mode,
the relation between the magnetic field and the current density satisfies the following equations.

𝛻 × 𝐻̄ = 𝐽 ;𝛻 • 𝐵̄ = 0 (32)

H and B represent the vectors of the magnetic field (A/m) and flux density (T). Relation between both
parameters can be expressed in vector potential A,

𝐵̄ = 𝜇𝑟 𝜇0 𝐻̄ = 𝛻 × 𝐴̄ (33)

𝜑𝑡 = 𝐵𝑡 𝑎𝑡 (34)

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2030 – 2044
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2037

ϕt is the flux magnitude in a salient (Wb), Bt is the salient flux density (T), and at is the tooth surface area
(m2). At this stage, it needs to be ensured that the flux densities in the stator and rotor do not exceed the
maximum allowable value, otherwise the rotor and the winding need redesigning. To find the back-EMF, the
flux linkage is gained from the circuit properties extracted through FEMM simulation. The value can be
approached from the sum of the flux magnitudes in teeth which are covered by the respective phase coil.

𝜆 ≈ 𝑁𝑠 ∑𝑡𝑒𝑒𝑡ℎ_
𝑖
𝑐𝑜𝑣 𝑒𝑟𝑒𝑑
𝑘𝑖 𝐵𝑛𝑖 𝑎𝑡 (35)

λ is the flux linkage of a coil phase (Wb), k is a coefficient. The back-EMF is then

𝑑𝜆 𝑑𝜃𝑎
𝐸𝑏 = 𝑑𝜃 (36)
𝑎 𝑑𝑡

θa is the angular position of the rotor (rad), and dθa/dt represents the nominal angular speed (rad/s). From
those values, several output parameters can then be calculated, including the line-to-line voltage, terminal
voltage, electromagnetic power, dissipated losses, efficiency, as well as stall torque. The characteristics of
torque and speed can also be plotted based on the calculations. Finally, the cogging force will be measured
through the simulation for checking the smoothness during the rotor rotation.

4. RESULTS AND DISCUSSION

In this section, the resulted design of the rotor, as well as the winding configuration, will be
described first. Then the output of the BLDC motor will be explained subsequently.

4.1. Rotor design and winding configuration

The resulted design of the rotor is shown in Figure 5, while the dimensions of its all parts are
tabulated in Table 4. In the rotor, each pole section consists of five segments of the magnet; all are made of
NdFeB material with N45H-type which has a remanence of 1.32 T and a maximum operating temperature of
120 oC. Every magnet is separated one another by the rotor yoke column. The magnetic polarities of the two
opposite pole sections are the same. Hence, two opposite pole sections have the “North" poles in the outer
part and the inner part for the other opposite ones. Furthermore, the magnets are “half-buried”. Their outer
surfaces still face the stator and are in contact with the air gap. However, the main bodies are fixed inside the
rotor yoke, while the edges of the outer magnets’ surface are held and locked by parts of the rotor yoke.
The use of several magnet segments is to divide the magnetic flux flow to the rotor yoke and to increase the
overall surface area of the magnet. Therefore, the maximum flux density can be lowered. Then the hysteresis
and eddy current losses can also be reduced. Meanwhile, the total flux flow to the stator can still be kept,
since the additional air gap from the height of the clippers at the edges of the permanent magnets is relatively
small. Furthermore, the division of each pole into several magnet segments will more distribute the attractive
force (with the stator body) along the longer pole arc. The use of the half-buried permanent magnet
configuration with edges’ clippers is to more easily fix the magnets’ position into the rotor.
On the other hand, the winding employs a spiral-type configuration as shown in Figure 6 (a). Each
phase’s windings fill certain stator slot numbers Figure 6 (b) in a certain way. Every slot is divided into two
layers: the inner and outer layers. Each layer is filled with different phase windings Figure 6 (c). In Figure 6
(c), the positive sign means the wire is coming out and the negative means it is coming into. As only eight
slots in each section are filled, there will be six empty slots in total, out of 54 slots. The configuration is
chosen by evenly distributing the six empty slots to reduce cogging torque and to avoid unbalanced magnetic
pulls. However, such configuration causes the winding to shorten, shortening four slots to give kw = 0.853.
From these arrangements, the value of kd, kp, and ks can also be obtained. They are recorded in Table 4.
The resulted rotor movement based on its configuration and the inverter input is shown in Figure 7.
The arrows in the poles show their magnetic orientations. Meanwhile, the sequence of the flowing phase
current from Figure 7 (a) to Figure 7 (f) is based on Figure 3 (c) and Table 2. For instance, in Figure 7 (a) the
electrical cycle is from 0 to 60 degrees. The current is flowing from phase A to phase B (Table 2). Thus, the
positive current flows in phase winding A, whereas the negative one flows in phase winding B. According to
the configuration shown in Figure 6 (c), the current in the phases of A and B will generate certain magnetic
fields in certain directions. In Figure 7 (a), the arrows in the stator show the direction of the fields, and the
letters show which phase produces them. The same analysis can be carried out for Figure 7 (b)-(f). Based on
the 4 poles-6 windings configuration and the order of the conducting phases, one pole pair needs one
complete electrical cycle to make a complete rotation. Thus, every pole section needs two complete electrical
cycles to make a full mechanical rotation.
Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)
2038  ISSN: 2088-8694

Figure 5. Resulted rotor design, with the arrows in the poles’ magnets (left) show their magnetic orientations

Table 4. Resulted rotor dimensions and winding parameters

Part, symbol Dimension Unit
Radius of the rotor axis, rax_r 24.15 mm
Pole outer radius, rpm_o 72.5 mm
Pole inner radius, rpm_i 61.5 mm
Pole pitch, τp 20.2 mm
PM length, τpm 14 mm
PM thickness, hpm 5 mm
Height of the PM edges’ clipper, hcl 1.2 mm
Width of the PM edges’ clipper, Wcl 2 mm
Wire diameter, Dw 3.7 mm
Turn per phase, Nph 64 -
Coil per slot, Ns 2 -
Wire per coil, Nc 2 -
Current density, J 3.2 A/mm2
Distribution factor, kd 0.964 -
Pole shortening factor, kp 0.885 -
Pole skewing factor, ks 1 -
Winding factor, kw1 0.853 -

(a) (b)

(c)

Figure 6. These figures are, (a) winding configuration, (b) slot numbering, (c) winding arrangement based on
the numbering

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2030 – 2044
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2039

(a) (b) (c)

(d) (e) (f)

Figure 7. Rotor position sequence during the rotation on the electrical cycle of, (a) 60o, (b) 120o, (c) 180o,
(d) 240o, (e) 300o, (f) 360o

4.2. Output of the BLDC motor

The flux distributions of the resulted motor under no current and nominal current conditions are
shown in Figure 8. In no current condition (Figure 8 (left)), the flux density is concentrated around the
magnets as the magnetic flux sources. However, when nominal current is injected, the windings will generate
reverse magnetic flux that opposes the poles’ magnetic flux. In that situation, the flux density in the stator
yoke will increase, especially in the stator yoke area which is passed by magnetic flux flowing between two
opposite polarity poles (Figure 8 (right)). Regarding the value, the maximum flux density reaches slightly
above 2 T, but is still in the tolerable saturation range of the stator’s 50H1300 material [26].
The air gap flux density in a stator tooth fluctuates as the rotor rotates. Its value depends on its
relative position with respect to the rotor. Figure 9 (left) shows the graph of the flux density in a tooth when it
is passed by the rotor. The resulted five-wave peaks show the amount of magnet segment in each pole
section. The flux and flux density reach the maximum values as the tooth directly faces a magnet segment.
When nominal phase current is injected, there are increases in negative values of the flux densities. In this
case, the positive flux direction is coming out of the rotor’s poles. The increases are the result of the stator
phase windings generating opposite magnetic flux in opposite directions with respect to the poles’ magnetic
flux. The negative flux will also reduce the resultant flux magnitude. From Figure 9 (left), the average flux
density in no current condition is 0.4713 T, while under nominal current it decreases to 0.2810 T.
The resulted flux magnitude will determine the back EMF as well as the output torque. The first
term is proportional to the required input voltage to produce certain amounts of torque and power. In Figure 9
(right), the back-EMFs for all phases are presented. For each phase, the graph forms a trapezoidal shape as in
the usual BLDC motor [30], with the width of its peak is 120 o electrical cycle, and transition from zero
voltage to the peak is about 60o.
Table 5 then tabulates the outputs of the BLDC motor at nominal speed. The output power is 5.1
kW, which slightly exceeds the desired output of 5 kW. The copper loss accounts for the largest loss
proportion suffered by the BLDC motor, while the estimated additional loss has the smallest value. From the
resulted losses, the efficiency of the motor at the nominal speed is obtained at 97.09%. Meanwhile, the
maximum torque or the stall torque is 51.39 Nm and the rated torque is 32.82 Nm.
The output torque-speed characteristic is further shown in Figure 10 (a). The torque is linearly
opposite to the rotation speed. In this design, the maximum allowable speed is 4,150 rpm, in which the
motor’s torque is down to zero. In Figure 10 (b), the relation between the output power and efficiency
towards the rotation speed is shown. The maximum power of the motor, 5,584.25 W occurs at 2,100 rpm but
is followed by the drop of efficiency to 96.63%. After that point, the motor’s output power and efficiency
Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)
2040  ISSN: 2088-8694

continue to decrease as both core loss and mechanical loss continue to increase at high rotation. Meanwhile,
the cogging torque graph is presented in Figure 11. The peak-to-peak range of the torque is 1.35 Nm (0.43
Nm to -0.92 Nm), which is relatively small (4.1% of the rated torque). Thus, the resulted torques are
expected to not making significant noise and ripple during rotor movement.
From the results, all aspects that contribute to the work of the BLDC motor have been presented and
analyzed. The winding and rotor configurations are suitable for the inverter setting. A similar study has been
partially conducted [19]. This research completes the analysis by involving the inverter components. The
resulted flux density is still within the range of tolerance. Meanwhile, the output voltage and output power
have reached the expected values. The torque and rotation speed show sufficient values and suit the targeted
results. Furthermore, the peak-to-peak cogging torque is relatively insignificant to disturb the rotation of the
rotor. Hence the designed BLDC motor in overall shows a relatively good performance. This design process
of BLDC motor from an existing stator can then be reviewed and applied in similar cases in other researches.
Further improvement for the next research can be performed by selecting optimum materials with smaller
reluctance. Hence the output voltage and power can be maximized.

Figure 8. Flux distribution of the LPMG design under no current (left) and nominal current (right)

Figure 9. Flux densities in a tooth under no current and nominal current (left) and back-EMF of all phases
(right)

Table 5. Output values of the BLDC motor

Part, symbol Dimension Unit
Line-to-line back EMF, ell 122.34 V
The input voltage, vin 198.25 V
Armature current under square wave signal, Ia(sq) 42.14 A
Electromagnetic power, Pelm 5,155.29 W
Copper loss, Pcu 74.77 W
Core loss, Pc 33.75 W
Mechanical loss, Pm 39.51 W
Additional loss, Padd 6.75 W
Total loss, Pl_tot 154.78 W
Rated torque, T 32.82 Nm
Stall torque, To 51.39 Nm
No-load current, Io 65.98 A
No-load rotation speed, no 69.19 rps
Efficiency, η 97.09 %

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2030 – 2044
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2041

(a) (b)

Figure 10. These figures are, (a) torque-speed characteristics, (b) graph of output power-efficiency towards
rotation speed

Figure 11. Cogging torque of the BLDC motor

5. CONCLUSION
A BLDC motor, connected to an inverter circuit and equipped with Hall sensors, has been designed
based on an existing stator body. The motion converter and the rotor with half-buried permanent magnets are
designed to adapt to the dimension and topology of the stator. Some desired outputs are stated, including
expected voltage, output power, and rotation speed. Meanwhile, several limitations, including maximum current
and flux density, are also considered. The flux distribution is extracted by using finite element magnetic-based
software of FEMM, and it is found that the maximum flux density is still in the tolerable range. Analytical
calculations show that: At nominal speed, the BLDC motor can produce acceptable results, which are 5.1 kW of
output power with 122.34 V of voltage, 97.09% of efficiency, and 32.82 Nm of rated torque; the maximum
torque and rotation speed obtained from the torque-speed characteristics are 51.39 Nm and 4,150 rpm,
respectively; the cogging torque curve shows a relatively small peak-to-peak value of 1.35 Nm. The results
conclude that the BLDC motor has a good performance and is compatible with the connected inverter. Next
research can be performed by optimizing the materials used to minimize the reluctance.

ACKNOWLEDGEMENTS
The authors gratefully thank the Deputy of Engineering Science, Indonesian Institute of Sciences for
the financial support under Electrical Vehicle project, number 26/A/DT/2021.

REFERENCES
[1] D. Potnuru, K. Alice Mary, and Ch. Sai Babu, “Experimental implementation of Flower Pollination Algorithm for
speed controller of a BLDC motor,” Ain Shams Engineering Journal, vol. 10, no. 2, pp. 287-295, June 2019, doi:
10.1016/j.asej.2018.07.005.
[2] Z. U. Bayrak and M. T. Gencoglu, “Application areas of fuel cells,” 2013 International Conference on Renewable
Energy Research and Applications (ICRERA), 2013, pp. 452-457, doi: 10.1109/ICRERA.2013.6749798.
[3] H. Seol, D. Kang, H. Jun, J. Lim, and J. Lee, “Design of Winding Changeable BLDC Motor Considering
Demagnetization in Winding Change Section,” IEEE Transactions on Magnetics, vol. 53, no. 11, pp. 1-5, Nov.
2017, doi: 10.1109/TMAG.2017.2695890.
[4] J. Feng, K. Liu, and Q. Wang, “Scheme based on buck‐converter with three‐phase H‐bridge combinations for high‐
speed BLDC motors in aerospace applications,” IET Electric Power Applications, vol. 12, no. 3, pp. 405-414,
March 2018, doi: 10.1049/iet-epa.2017.0615.
Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)
2042  ISSN: 2088-8694

[5] M. Ridwan, D. C. Riawan, and H. Suryoatmojo, “Particle swarm optimization-based BLDC motor speed controller
with response speed consideration,” 2017 International Seminar on Intelligent Technology and Its Applications
(ISITIA), 2017, pp. 193-198, doi: 10.1109/ISITIA.2017.8124079.
[6] M. Kumar and K. Chaursiya, “Position control of brushless DC motor using harmony search algorithm
optimization technique,” 2017 Int. conf. of Elect., Comm. and Aerospace Tech. (ICECA), 2017, pp. 754-757.
[7] B. A. kumar, C. Kamal, D. Amudhavalli, and T. Thyagarajan, “Reformed Stator Design of BLDC Motor for
Cogging Torque Minimization Using Finite Element Analysis,” 2018 4th International Conference on Electrical
Energy Systems (ICEES), 2018, pp. 481-484, doi: 10.1109/ICEES.2018.8442389.
[8] U. Nagalingam, B. Mahadevan, K. Vijayarajan, and A. P. Loganathan, “Design optimization for cogging torque
mitigation in brushless DC motor using multi-objective particle swarm optimization algorithm,” COMPEL: The
International Journal for Computation and Mathematics in Electrical and Electronic Engineering, vol. 34, no. 4,
pp. 1302-1318, July 2015, doi: 10.1108/COMPEL-07-2014-0162.
[9] B. Alsayid, W. A. Salah, and Y. Alawneh, “Modelling of sensored speed control of BLDC motor using
MATLAB/SIMULINK,” International Journal of Electrical and Computer Engineering (IJECE), vol. 9, no. 5, p.
3333, October 2019, doi: 10.11591/ijece.v9i5.pp3333-3343.
[10] M. S. Raja and B. Geethalakshmi, “Modified Rotor Material for Minimization of Torque Ripple for Interior
Permanent Magnet BLDC motor,” Materials Today: Proceedings, vol. 5, no. 2, pp. 3639-3647, 2018, doi:
10.1016/j.matpr.2017.11.614.
[11] A. A. Obed, A. L. Saleh, and A. K. Kadhim, “Speed performance evaluation of BLDC motor based ondynamic
wavelet neural network and PSO algorithm,” International Journal of Power Electronics and Drive Systems
(IJPEDS), vol. 10, no. 4, pp. 1742-1750, December 2018, doi: 10.11591/ijpeds.v10.i4.pp1742-1750.
[12] M. Khelifa, M. Mordjaoui, and A. Medoued, “An inverse problem methodology for design and optimization of an
interior permanent magnetic BLDC motor,” International Journal of Hydrogen Energy, vol. 42, no. 28, pp. 17733-
17740, July 2017, doi: 10.1016/j.ijhydene.2017.02.017.
[13] A. Shabanian, A. A. P. Tousiwas, M. Pourmandi, A. Khormali, and A. Ataei, “Optimization of brushless direct
current motor design using an intelligent technique,” ISA Transactions, vol. 57, pp. 311-321, July 2015, doi:
10.1016/j.isatra.2015.03.005.
[14] I. Anshory, I. Robandi, and Wirawan, “Monitoring and optimization of speed settings for Brushless Direct Current
(BLDC) using Particle Swarm Optimization (PSO),” 2016 IEEE Region 10 Symposium (TENSYMP), 2016, pp.
243-248, doi: 10.1109/TENCONSpring.2016.7519412.
[15] D. Kumpanya, S. Thaiparnat, and D. Puangdownreong, “Parameter Identification of BLDC Motor Model Via
Metaheuristic Optimization Techniques,” Procedia Manufacturing, vol. 4, pp. 322-327, 2015, doi:
10.1016/j.promfg.2015.11.047.
[16] V. K. S. Patel and A. K. Pandey, “Modeling and Performance Analysis of PID Controlled BLDC Motor and
Different Schemes of PWM Controlled BLDC Motor,” International Journal of Scientific and Research
Publications, vol. 3, no. 4, pp. 1-14, 2013.
[17] R. Palanisamy, R. Sahasrabuddhe, M. K. Hiteshkumar, and J. A. Puranik, “A new energy regeneration system for A
BLDC motor driven electric vehicle,” International Journal of Electrical and Computer Engineering (IJECE), vol.
11, no. 4, pp. 2986-2993, August 2021, doi: 10.11591/ijece.v11i4.pp2986-2993.
[18] Y. B. A. Apatya, A. Subiantoro, and F. Yusivar, “Design and prototyping of 3-phase BLDC motor,” 2017 15th
International Conference on Quality in Research (QiR): International Symposium on Electrical and Computer
Engineering, 2017, pp. 209-214, doi: 10.1109/QIR.2017.8168483.
[19] M. Y. Marusina, A. A. Silaev, and D. A. Nevmerzhitsky, “Method of Converting an Induction Motor to a
Permanent Magnet Synchronous Motor (Ls-Pmsm) with Linear Starting Using FEA for Mechatronic Systems,”
2019 International Conference "Quality Management, Transport and Information Security, Information
Technologies" (IT&QM&IS), 2019, pp. 356-358, doi: 10.1109/ITQMIS.2019.8928320.
[20] Y. K. Chin and J. Soulard, “A Theoretical Study on Permanent Magnetsynchronous Motors for Electric Vehicles,”
in 6th International Power Engineering Conference (IPEC) 2003, Singapore, Nov. 2003, pp. 435-440.
[21] F. D. Wijaya, B. Azhari, and H. Stephanus, “Optimum Permanent Magnets Configuration in Flat-Quasi Linear
Permanent Magnet Generators,” International Journal of Electrical and Computer Engineering (IJECE), vol. 6, no.
6, pp. 2589-2602, December 2016, doi: 10.11591/ijece.v6i6.11966.
[22] H. Arof, A. M. Eid, and K. M. Nor, “Cogging force reduction using special magnet design for tubular permanent
magnet linear generators,” 39th Int. Uni. Power Eng. Conf., 2004. UPEC 2004, vol. 1, pp. 523-527, 2004.
[23] J. S. Park and K.-D. Lee, “Design and Implementation of BLDC Motor with Integrated Drive Circuit,”
International Journal of Power Electronics and Drive Systems (IJPEDS), vol. 8, no. 3, pp. 1109-1116, September
2017, doi: 10.11591/ijpeds.v8.i3.pp1109-1116.
[24] I. Boldea and S. A. Nasar, “The Induction Machine Handbook,” CRC Press, p. 968, 2010.
[25] J. F. Gieras, “Permanent magnet motor technology: design and applications,” 3rd ed. Boca Raton: CRC Press, 2010.
[26] P. Irasari, H. S. Alam, and M. Kasim, “Magnetic Simulation and Analysis of Radial Flux Permanent Magnet
Generator using Finite Element Method,” Journal of Mechatronics, Electrical Power, and Vehicular Technology,
vol. 3, no. 1, pp. 23-30, July 2012, doi: 10.14203/j.mev.2012.v3.23-30.
[27] M. Yazdani-Asrami, M. Alipour, and S. A. Gholamian, “Optimal ECO-Design of Permanent Magnet Brushless DC
Motor Using Modified Tabu Search Optimizer and Finite Element Analysis,” Journal of Magnetics, vol. 20, no. 2,
pp. 161-165, June 2015, doi: 10.4283/JMAG.2015.20.2.161.

Int J Pow Elec & Dri Syst, Vol. 12, No. 4, December 2021 : 2030 – 2044
Int J Pow Elec & Dri Syst ISSN: 2088-8694  2043

[28] P. Irasari, P. Widiyanto, and M. F. Hikmawan, “Desain dan Simulasi GMP Fluks Aksial Berbasis Dimensi Magnet
Permanen Komersil,” ELKOMIKA: Jurnal Teknik Energi Elektrik, Teknik Telekomunikasi, & Teknik Elektronika,
vol. 8, no. 3, p. 602, August 2020, doi: 10.26760/elkomika.v8i3.602.
[29] C. Kamal, T. Thyagarajan, M. Selvakumari and D. Kalpana, “Cogging torque minimization in brushless DC motor
using PSO and GA based optimization,” 2017 Trends in Industrial Measurement and Automation (TIMA), 2017,
pp. 1-5, doi: 10.1109/TIMA.2017.8064815.
[30] M. A. Ibrahim, A. Kh. Mahmood, and N. S. Sultan, “Optimal PID controller of a brushless DC motor using genetic
algorithm,” International Journal of Power Electronics and Drive Systems (IJPEDS), vol. 10, no. 2, pp. 822-830,
June 2019, doi: 10.11591/ijpeds.v10.i2.pp822-830.

BIOGRAPHIES OF AUTHORS

Budi Azhari received the B. Eng. and M. Eng degrees in Electrical Engineering both from the
Department of Electrical Engineering and Information Technology, Universitas Gadjah Mada in
2016 and 2017 respectively. His research topics are focused on the design and analysis of the
electric machine, mainly for renewable energy conversion. He is currently working as a
researcher on the Electrical Vehicle Project at Research Center for Electrical Power and
Mechatronics, Indonesian Institute of Sciences. (Scopus: 57221973939, ORCID: 0000-0003-
0094-8902)

Pudji Irasari completed B.Eng. degree in Universitas Brawijaya in 1994 and M.Sc. degree at
Oldenburg University, Germany in 2003. Her research interests including analysis of
renewable energy conversion, electric machine design, and electrical power system
assessment. She is currently conducting a project on water turbines for electricity generation
at Research Center for Electrical Power and Mechatronics, Indonesian Institute of Sciences,
collaborate with Institut Teknologi Bandung, Indonesia. (Scopus: 56087989100)

Puji Widiyanto received his B. Eng. and M. Eng. degrees both in Mechanical Engineering,
first in STT Mandala and later in Institute Technology Bandung. Having focused his research
on the structure and mechanical design of electrical machines, he is now working as a
researcher at the Research Center for Electrical Power and Mechatronics, the Indonesian
Institute of Sciences. Currently, he is conducting a project on water turbines for electricity
generation, collaborates with Institut Teknologi Bandung, Indonesia (Scopus: 56966777600)

Design and simulation of 5kW BLDC motor with half-buried permanent magnets using an … (Budi Azhari)