ARCHIVES OF ELECTRICAL ENGINEERING VOL. 66(1), pp.

121-139 (2017)

DOI 10.1515/aee-2017-0009

Comparison of higher harmonic contents in salient pole

synchronous generator with different rotor
construction

KRZYSZTOF LUDWINEK, ROMAN NADOLSKI, JAN STASZAK

Kielce University of Technology, Department of Automatic Control and Computer Science

Tysiąclecia P. P. 7, 25-314 Kielce, Poland
tel. +48 41 342 46 82
e-mail: {k.ludwinek / r.nadolski / j.staszak}@tu.kielce.pl
(Received: 05.02.2016, revised: 13.09.2016)

Abstract: The paper presents a comparison of higher harmonics in induced phase volt-
ages of a stator winding in the no-load state of a three-phase 5.5 kVA salient pole syn-
chronous generator. The comparison is carried out for the synchronous generator with
different salient pole rotor constructions: a non-skewed solid rotor, a non-skewed solid
rotor with radial incisions, and a laminated electrotechnical steel rotor with skewed slots
and damping bars. The calculations of higher harmonics are based on the magnetic field
distributions in the air gap, which are carried out in a 2D model in a FEMM program and
on the induced voltage waveforms in the stator windings registered during experimental
investigations of the 5.5 kVA salient pole synchronous generator in the no-load state.
Key words: salient pole synchronous generator, higher harmonics, solid rotor, rotor with
radial incisions, laminated rotor

1. Introduction

The distribution of the magnetic flux density in the air gap of a salient pole synchronous
generator has an influence on the occurrence of higher harmonics in voltage induced in the
stator windings [1-8]. The shape of the stator and rotor magnetic surfaces has an influence on
the distribution of the magnetic flux density along the air gap circumference, and the magnetic
flux density:
– decreases at the stator and rotor slot opening,
– increases if the length of the air gap decreases.
Knowledge about the effective air gap length allows, for example, an analytically more
precise specification of: self- and mutual inductances, the distribution of flux density in the air
gap, voltages induced in the stator windings etc. [9-14]. The effective air gap length is calcu-
lated from the initial length of the air gap by taking into account the total Carter Factor, which
is the sum of two factors [4, 14, 15]. The first factor is calculated by assuming a smooth rotor

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
122 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

surface and taking into account stator slots. The second factor is calculated by assuming
a smooth stator surface and with rotor slots. In electrical machines the problem of variations in
the distribution of the magnetic flux density due to stator and rotor slot openings was first
solved by F.W. Carter in 1901 [16]. Generally, the variations in the distribution of magnetic
flux density are taken into account in increases of the initial length of the air gap by means of
the Carter factor [4, 14-16]. Variations in the distribution of the magnetic flux density in a one
sided uniform air gap (including the slots on the stator or on the rotor side) can be determined:
– by the quotient of minimum to maximum magnetic flux density [14, 15]
Bmin 2u
= , (1)
Bmax 1 + u 2

– in relation to the quotient of average to maximum magnetic flux density [4]

B0 (u − 1) 2
= , (2)
Bmax 2(1 + u 2 )

where: Bmin is the minimum magnetic flux density, Bmax is the maximum magnetic flux
density, u is the coefficient
u = bs / 2δ 0 + 1 + (bs / 2δ 0 )2
,
δ0 is the initial length of the air gap, bs is the length of the stator slot opening, B0 is the
average magnetic flux density,
As shown in [4] when assuming no eddy currents and the same length of the air gap, if the
relation between the length of the stator slot opening and the stator tooth pitch is bs/τs < 0.625,
then an analytical expression of the flux density variations in an air gap can be described as:
– if 0 < α < 0.8α

⎛ π ⎞
B (α) = ⎜⎜1 − β − β cos α ⎟⎟ Bmax , (3)
⎝ 0 .8 α0 ⎠

– if 0.8α < α < 0.5αd

B (α) = Bmax , (4)

where: β = B0/Bmax, α is the angle of the stator slot opening in relation to the rotor circumfe-
rence [4], α0 = 2bs/d = bs/rs, d is the inner stator diameter, rs is the inner stator radius.
In a salient pole synchronous generator the magnetic rotor asymmetry (in the direct and
quadrature axes) and the shape of a pole shoe with various lengths of an air gap and with the
bars of the damping cage (placed inside the pole shoes) mean that the air gap is more complex
than in cylindrical or asynchronous generators. The air gap length has a significant influence
on the magnetic flux density distribution and on the induced stator voltage waveforms [1-13,
17, 18].
Most frequently, the pole shoes in a salient pole synchronous generator are made of solid
iron or with electrotechnical sheets (up to 2.5 mm in thickness) [14].

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 123

The presence of a rotor damping cage during the cooperation of a synchronous generator
with a power grid allows the shortening of many transient processes [19-21] and the presence
of a damping cage is not a subject of discussion. But in the case of the synchronous generator
set (working as a back-up power system), the presence of a damping cage is debatable parti-
cularly in the no-load state. The damping cage can be made of solid iron [14].
Apart from the stator and rotor slotting and the type of stator and rotor windings and
damping cage, the inner eccentricity of the stator and rotor has an influence on the distribution
of the magnetic flux density in the air gap of a salient pole synchronous generator. This
problem is widely discussed in many papers [22-25].
Because of the small initial length of the air gap, the influence of the air gap length on the
eddy currents and power loss in the solid elements in low-power synchronous generators is
more visible than in high-power ones. The depth of eddy current penetration λ is determined
as [14, 15]:

5 ⋅ 10 5
λ= , (5)
f γμ r

where: f is the frequency, γ is the electrical conductivity, μr is the relative magnetic perme-
ability.
Frequently, electrotechnical sheets are used as a method of eddy current reduction in pole
shoes [14]. The other method of reducing the eddy currents shown in this paper is by making
radial incisions across the width of the pole shoe [18]. According to expression (5), the depth
of an incision across the width of the pole shoe depends on the frequency of the magnetic field
and the type of rotor material (its electrical conductivity and magnetic permeability).
This paper presents the influence of the pole shoe radial incision across the width of the
pole shoe:
– on the distribution of the normal magnetic flux density in the air gap of a 5.5 kVA salient
pole synchronous generator,
– on the waveform of voltages induced in the stator windings.
The air gap length δ between the stator and solid pole shoe in relation to the initial length
of the air gap in the longitudinal axis of the examined 5.5 kVA salient pole synchronous
generator is expressed as [18]:
δ0
δ= , (6)
cosα
where: δ is the air gap length between the stator and solid pole shoe, δ0 is the initial length
of the air gap in the longitudinal axis, α is the electrical angle of the pole shoe which is calcu-
lated in relation to the longitudinal axis.
The visualization of air gap length δ between the stator and solid pole shoe and the depth
of incision δi across the width of the pole shoe in relation to the initial length of air gap δ0 and
inner stator radius rs are shown in Fig. 1.

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
124 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

Fig. 2 shows the relative influence of the air gap length δ (6) and of radial incision of
a pole shoe on the effective air gap length in relation to inner stator radius rs and diameter d vs.
the electrical angle of the pole shoe α in relation to the longitudinal axis.

Fig. 1. Visualization of the air gap length δ

δ0 δi and the depth of incision δi across the width
of a pole shoe in relation to the initial length
pole shoe of the air gap δ0 and inner stator radius rs
rs

1.005
stator
1.000
Fig. 2. Relative influence of the air gap
0.995 air gap length acc. to (2)
length δ and the radial incision δi of a pole
0.990
shoe in relation to inner stator radius rs and
d/rs

0.985
0.980
diameter d vs. electrical angle α of the pole
0.975 radial incision shoe
0.970
-60 -40 -20 0 20 40 -60
α (deg)

The introduction of the radial incisions on the pole shoe surface results in even loss of the
steel (Figs. 1 and 2) and has the effect of flattening the resultant magnetic flux density in the
air gap [14, 18]. The effect of flattening the resultant magnetic flux density due to the radial
incisions on the pole shoe surface has a similar effect to the case of the pole shoe saturation
[14, 18]. The advantage of radial incisions on the solid rotor on the pole shoe surface is that
the incisions can be done by utilising a simple lathe. The depth of the radial incisions on the
examined rotor circumference is equal to 2δ0 (Fig. 2).
Due to the complex air gap, more detailed results are obtained using Finite Element Me-
thod (FEM) software for 2D or 3D models [12, 26]. Simulations of the magnetic flux density
distribution and induced voltages with the FEM software are very similar to the results of ex-
perimental investigations due to the fact that the FEM models include many construction
details [11, 12]. But the FEM calculations also have many disadvantages, for example [11, 12,
27, 28]:
– the length of time required for the calculations in FEM programs is longer than in an
analytical way,
– commercial FEM programs are very expensive and require detailed knowledge about the
construction of the electrical machine,
– FEM models require a large number of constructional details,
– for a beginner or even a person with intermediate skills, considering which electromag-
netic parameters have a direct influence on the final results is difficult.

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 125

Therefore, simulations of the magnetic flux density distribution and induced voltages in
electrical machines are often carried out by means of both analytical expressions, circuital
models and/or FEM software [3, 5-12, 27-29].
In this paper the simulation of the magnetic flux density distribution and experimental
investigations of induced voltages are carried out for the salient pole synchronous generator
rated: SN = 5.5 kVA, UN = 400 V (Y), IN = 7.9 A, cosϕN = 0.8, nN = 3000 rpm, the number of
pole pairs pb = 1, the number of stator slots Qs = 24 with a single-layer winding, rotor: with
and without the skew angle αq = 15° (the skew angle is equal to a single stator tooth pitch).
Moreover, magnetic flux density distributions are determined in the air gap of the examined
5.5 kVA salient pole synchronous generator after creating the geometry of the 2D field model
in the FEMM program [9, 11-13, 26]. In calculations of the distribution of the magnetic flux
density the presence of eddy currents is omitted in the 2D field model due to the powering of
the field winding by DC nominal current at no-load If 0 N = const.

2. Distributions of magnetic flux density

The distribution of the magnetic flux density lines is obtained taking into account the non-
linear material properties. Fig. 3 shows two no-load magnetisation characteristics obtained
from the measurement set for the 5.5 kVA salient pole synchronous generator with two rotor
structures. One of the no-load characteristics is for a salient pole synchronous generator with
a factory rotor made of insulated electrotechnical sheets and the second one is for a generator
with a solid rotor with radial incisions.

UphN

solid rotor with incisions

rotor with electrotechnical sheets

If0N

Fig. 3. Comparison of the two no-load magnetisation characteristics of the 5.5 kVA salient pole
synchronous generator with two rotor structures obtained from the measurement set

Figure 3 shows that the presented no-load characteristics are very similar.
Figures 4 and 5 show the distribution of the magnetic flux density lines of the examined
5.5 kVA nonlinear salient pole synchronous generator in the no-load steady state in the Finite
Element Method Magnetics (FEMM) freeware program [26]. The magnetic flux density lines
are obtained from DC current If0N flowing in the field winding. In the FEMM 2D software, the
skew effect in the normal component of magnetic flux density distributions is obtained by
subdividing the active rotor length into 79 skewed axial slides along the axial length (39 radial

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
126 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

incisions and 40 without radial incisions – detailed in Fig. 9b). The normal component of
magnetic flux density Bns with the rotor skewed slots is calculated as [7]:
αq
θ+
2
1
Bns =
αq ∫ B ∂θ ,
αq
n (7)

θ−
2

where: αq is the electrical skew angle, Bns, Bn are the normal component of magnetic flux
density in a non-skewed and skewed rotor, respectively.

(a) (b)
Fig. 3. General view of the distribution of the magnetic flux density lines in the no-load steady state in
the examined 5.5 kVA salient pole synchronous generator: with solid rotor without radial incisions (a),
with solid rotor with radial incisions (b)

The distribution of the magnetic flux density lines of the examined 5.5 kVA nonlinear
salient pole synchronous generator with the rotor made of 0.5 mm insulated electrotechnical
sheets (with an air gap as in the case of the solid rotor without radial incisions) is the same as
shown in Fig. 4a.

UphN

solid rotor with incisions

rotor with electrotechnical sheets

If0N

Fig. 4. Comparison of the two no-load magnetisation characteristics of the 5.5 kVA salient pole synchro-
nous generator with two rotor structures obtained in the measurement set

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 127

Differences in the distribution of the magnetic flux density are more evident in Fig. 5
(detailed view of the air gaps) and in Fig. 6, where the comparison of the distribution of the
normal component of magnetic flux density in the air gap vs. the circumference of the two
rotor types is shown.

(a) (b)

Fig. 5. Detailed view of the distribution of the normal component of magnetic flux density lines in the
no-load steady state in the air gap with rotor: without radial incisions (a), with radial incisions (b)

B
0.5 n - solid without incisions and without skew
B
ns - solid without incisions and with skew
(T)

0
n
B

-0.5

-1
0 60 120 180 240 300 360
ϑ (deg)

(a)
1

B
n - solid with incisions and without skew
0.5
B
ns - solid with incisions and with skew
(T)

0
Bn

-0.5

-1
0 60 120 180 240 300 360
ϑ (deg)
(b)
Fig. 6. Comparison of normal component of magnetic flux density distributions for the examined syn-
chronous generator with: a non-skewed and skewed solid rotor (a), a non-skewed and skewed solid rotor
with the radial incision (b)

Fig. 6 shows a comparison of the distribution of the normal component of magnetic flux
density in the air gap for the synchronous generator with and without the radial incisions and
with and without the skew angle. The distribution of the normal component of magnetic flux

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
128 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

density in the air gap for the 5.5 kVA synchronous generator with a rotor made of insulated
0.5 mm electrotechnical sheets gives the same result as in the case of the solid rotor without
radial incisions.

100
B
nm-solid without incisions and without skew
80
Bnsm-solid without incisions and with skew
(%)

60 B
nim-solid with incisions and without skew
nm

40 Bnism-solid with incisions and skew

B

20

0
1 3 5 7 9 11 13
harmonics

(a)
100
B
nm-solid without incisions and without skew
80
B
nsm-solid without incisions and with skew
(%)

60 B
nim-solid with incisions and without skew
nm

40 B
B

nism-solid with incisions and skew

20

0
15 17 19 21 23 25 27 29 31
harmonics

(b)
100

80 Bnm-solid without incisions and without skew

(%)

Bnsm-solid without incisions and with skew

60
nm

Bnim-solid with incisions and without skew

B

40
Bnism-solid with incisions and skew
20

0
35 37 39 41 43 45 47 49 51 53 55
harmonics

(c)
100

80 Bnm-solid without incisions and without skew

(%)

B
60 nsm-solid without incisions and with skew
nm

B
40 nim-solid with incisions and without skew
B

B
nism-solid with incisions and skew
20

0
57 59 61 63 65 67 69 71 73 75
harmonics

(d)
Fig. 7. Comparison of harmonic contents in the normal component of magnetic flux density distribu-
tions: from 1st to 13th order (a), from 15th to 31st order (b), from 35th to 55th order (c), from 57th to
75th order (d)

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 129

The distribution results presented in Fig. 6 show that in the areas with a longer air gap the
flux density distributions decrease but in the areas with a smaller air gap they increase
(Fig. 5b). The variations in the distribution of the normal component of the magnetic flux
density shown in Fig. 6 are caused by a permeance change between the stator teeth and slots
on the circumference of the stator and rotor (Fig. 5).
Figure 7 shows the comparison of the magnitude of harmonic contents in the magnetic flux
density distributions (Fig. 6) due to the Fourier analysis.
The comparisons shown in Fig. 7 are determined in relation to the amplitude of the funda-
mental component of the magnetic flux density distribution in the air gap for the synchronous
generator with a solid rotor without incisions and without skewed poles. The results in Fig. 7
reveal that the 2δ0 radial incision causes:
– 10% reduction of the magnetic flux density fundamental component (Fig. 7a),
– an increase of the odd harmonics up to the 17th order by a few percent,
– a reduction of the kQs ± 1 higher harmonic amplitudes by a few percent (k is integer
number).
The number of radial incisions has an influence on the increase of the odd harmonics up to
the 17th order (Fig. 7). The radial incisions, which are the percentage loss of steel along the
length of the pole shoe, changed the v-th harmonics of the magnetic flux density from Bnv (for
the non-skewed rotor without incisions) to Bniv (for the non-skewed rotor with incisions) and
from Bnsv (for the skewed rotor without incisions) to Bnisv (for the skewed rotor with incisions).
The effect of flattening the equivalent air gap between the stator and the pole shoe (with
the radial incision) can be reduced by incisions defined according to the relation (6). But such
incision must be made with numerically controlled machine tools instead of a typical lathe.
The normal component of the magnetic flux density distributions for the examined 5.5 kVA
synchronous generator, as presented in Fig. 6, has an impact on the content of higher harmo-
nics in the induced stator phase voltages. As was shown in [9, 18], the wave-shape of the in-
duced stator phase voltages is influenced not only by the magnetic flux density distributions
(Fig. 5) but also by the winding coefficient (which for a single-layer winding depends on the
winding distribution, slot opening and rotor skew coefficients).

3. Voltages induced in the stator windings

In the no-load state of the salient pole synchronous generator (currents ia, ib and ic are equal
to zero) without the power grid and a damping cage, voltages: ua, ub, uc induced in the stator
windings and uf in the field winding, taking into account the electrical angle of the rotor posi-
tion θ, can be derived for the stator windings (in stator coordinates) and for the field winding
in rotor coordinates [11, 12]

dΨ a (θ, i f ) dΨ b (θ , i f ) dΨ c (θ, i f )
= ua , = ub , = uc , (8)
dt dt dt

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
130 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

dΨ f (θ , i f )
+ Rf if = u f , (9)
dt
dθ
=ω, (10)
dt
where: a, b, c and f are the indexes of stator windings and field winding, ua, ub, uc and uf are
the stator phase voltages and field voltage,Ψa, Ψb, Ψc and Ψf are the stator and field linkage
fluxes, θ is the electrical angle of the rotor position, θ = θmp, θm is the mechanical angle of the
rotor position, ω = dθ/dt is the electrical angular velocity, Rf is the resistance of the field
winding, if is the field current, p is the number of pole pairs.
Based on expressions (8)-(10) and the relations between the magnetic fluxes and magnetic
flux densities and taking into account ω = const and if = If0N = const (If0N – nominal field
current at no-load), the induced phase stator voltages ua, ub, uc and uf can be expressed as:

rs l s ∂Ba (θ) dθ ⎫
2N s kw = u a ,⎪
p ∂θ dt ⎪
rs l s ∂Bb (θ) dθ ⎪
2N s kw = ub , ⎬ (11)
p ∂θ dt ⎪
rs l s ∂Bc (θ) dθ ⎪
2N s kw = uc , ⎪
p ∂θ dt ⎭

rr lr ∂B f (θ) dθ
2 N f kwf + Rf if = u f , (12)
p ∂θ dt

where: Ba, Bb, Bc and Bf are the stator and field density of magnetic fluxes with and without
the rotor skew (eg. Fig. 6), Ns, Nf are the number of stator and field windings, rs, rr are the
inner stator and outer rotor radius, ls, lr are the length of stator and rotor, kw is the winding
factor: kw = kd kp kb kq [9, 14], kd is the distribution factor, kp is the pitch factor, kbv is the slot
opening factor, kqv is the skew factor, kwf is the field winding factor.
The problem of the influence of the pitch- and winding distribution, and slot-opening and
slot-skewing winding factors (of the 5.5 kVA salient pole synchronous generator) on the re-
duction of the harmonic contents in induced stator voltages in the most frequently used
windings: single-layer, double-layer and triangular, is widely discussed in [9].
Fig. 8 shows the comparison of the lower order harmonics of magnetic flux density (shown
in Fig. 6) and magnetic flux density derivatives for the synchronous generator with a solid
rotor without incisions. Lower order harmonics of magnetic flux density for the generator with
solid rotor with incisions are very similar to ones presented in Fig. 8 and are omitted. The
differences in magnitudes of the harmonics are better visible in Fig. 7. Fig. 8 shows that the
ν-harmonic presence in magnetic flux density is enhanced ν-times in magnetic flux density
derivatives. The ν-harmonic in magnetic flux density enhanced ν-times by derivatives has
significant influence on the increase of the harmonic content in the induced stator voltages-
expression (11).

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
 Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 131

(a) (b)
0.8 0.8

dB /dϑ ( T/rad )
0.4 0.4
B (T)

Bn1 dBn1/dϑ
0 0
Bn3
n

dBn3/dϑ

n
-0.4 Bn5 -0.4 dBn5/dϑ
Bn7 dBn7/dϑ
-0.8 -0.8
0 60 120 180 240 300 360 0 60 120 180 240 300 360
ϑ (deg) ϑ (deg)

Fig. 8. Comparison of the lower order harmonics: of magnetic flux density (a),
magnetic flux density derivatives (b)

Fig. 9 presents winding coefficients for a single layer three-phase winding [9]. The follow-
ing parameters of the 5.5 kVA synchronous generator are used in calculation of the winding
factors of the harmonic ν [9]: the number of phases m = 3, the number of pole pairs pb = 1, the
number of stator slots Qs = 24, the stator slot opening bs = 3 mm, the coil pitch in per unit
y = 1.

(a) (b)
1 1

kp
0.5 kd 0.5
kp , kd

p
k ,k

0 0
b

kq
-0.5 -0.5 kb

-1 -1
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 harmonics
harmonics

0.5
kw

-0.5

-1
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35
harmonics
(c)
Fig. 9. Winding factors for single layer winding of the 5.5 kVA synchronous generator: kp and kd (pitch
and distribution factors) (a), kb and kq (slot opening and skew factors) (b), kw = kp kd kb kq (total factor) (c)

Fig. 10 presents the waveforms of the induced stator voltages of the examined 5.5 kVA
salient pole synchronous generator in no-load state calculated in Matlab from expression (11)

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
132 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

400
ua-solid without incisions
200 ua-solid with incisions

u (V)
0

a
-200

-400
0 0.005 0.01 0.015 0.02
t (s)
(a)
400
uas-solid without incisions and with skew
200 uas-solid with incisions and with skew
ua (V)

-200

-400
0 0.005 0.01 0.015 0.02
t (s)
(b)
Fig. 10. Waveforms of the induced stator voltages under no-load conditions for the examined 5.5 kVA
synchronous generator with the following rotors: solid, skewed and non-skewed (a), solid, skewed and
non-skewed with radial incisions (b)

15
ua-solid without incisions and skew
ua-solid without incisions and with skew
IHDua (%)

10 ua-solid with incisions and without skew

ua-solid with incisions and with skew
5

0
3 5 7 9 11 13 15 17 19 21 23 25
harmonics
(a)
15
ua-solid without incisions and skew
ua-solid without incisions and with skew
IHDu (%)

10
ua-solid with incisions and without skew
a

ua-solid with incisions and with skew

5

0
27 29 31 33 35 37 39 41 43 45 47 49
harmonics
(b)
Fig. 11. Individual harmonic distortion of the induced stator voltages under no-load conditions for the
examined 5.5 kVA synchronous generator: from 3rd to 25th order (a), from 27th to 49th order (b)

and utilising the normal component of magnetic flux density distributions (presented in Fig. 6
for the synchronous generator with the following rotor structures: solid with and without
skew, solid with and without radial incisions). Fig. 10 presents the calculated waveforms of

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 133

the induced stator voltages only in phase a. Voltages ub and uc are very similar to ua (Fig. 10)
and are shifted by the electrical degree of 2π/3. Moreover, Fig. 11 presents the individual har-
monic distortion IHDua in relation to the fundamental component of the induced stator vol-
tages of the synchronous generator with the non-skewed solid rotor.
Comparing the induced stator voltages (Fig. 10), the magnetic flux density derivatives
(Fig. 8) and the winding factors (Fig. 9) it can be concluded that the higher harmonic content
is primarily influenced by the product of the electrical angular velocity ω and derivatives of
magnetic flux density ∂B/∂θ. The product ω(∂B/∂θ) is reduced by the winding factors (Fig. 9).
The greatest reduction of the harmonic content in the induced stator voltages is achieved by
skewing the rotor (Fig. 10c). The rotor skewness has a very small influence on harmonic
reduction of the 3rd and 5th order harmonics (Figs. 9-11).

4. Experimental investigations

Experimental investigation of the influence of the rotor incisions on the waveforms of

induced stator voltages of the 5.5 kVA salient pole synchronous generator is carried out by
comparison of the induced stator voltages for the three rotor structures shown in Fig. 12.
Fig. 12 shows [18]:
– a non-skewed solid rotor,
– a non-skewed solid rotor with 39 radial incisions,
– a skewed rotor with insulated electrotechnical sheets and open (not shorted) damping
bars.

(a) (b)

Fig. 12. View of the pole shoe surfaces of

three salient pole rotors: a non-skewed solid
rotor (a), a non-skewed solid rotor with ra-
dial incisions (b), rotor with insulated elec-
trotechnical sheets with skewed slots and
open damping bars (c)
(c)

Experimental investigations of the 5.5 kVA salient pole synchronous generator with the
three rotor structures are carried out using the measurement set shown in Fig. 13.
Fig. 14 presents a simplified block diagram of the measurement set (shown in Fig. 13).

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
134 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

Fig. 13. General view of the measurement set

used for the investigation of the 5.5 kVA
salient pole synchronous generator with the
three rotor structures

MSO 3014

DC if
power
supply uf

ua
ub
DC
uc
motor SG N
P
E

Salient pole synchronous generator

Sn = 5.5 kVA, UN =400 V, In =7.9 A

Fig. 14. Simplified block diagram of the measurement set used for the investigation
of the 5.5 kVA salient pole synchronous generator

Fig. 15 presents the registered waveforms of the induced stator voltages and the field vol-
tage as well as the current of the examined 5.5 kVA salient pole synchronous generator in no-
load state. During the experimental investigations the synchronous generator was running with
the following rotor structures: solid, non-skewed, solid, non-skewed with radial incisions, and
skewed with insulated electrotechnical sheets. Moreover, the field winding of the examined
5.5 kVA salient pole synchronous generator for all cases was powered by a DC voltage source
(Uf = const-shown in Fig. 15c) for the same RMS value of 230 V of the fundamental com-
ponent phase stator voltages. As shown in [11], when powering the field winding of the
5.5 kVA salient pole synchronous generator, in induced ua, ub and uc voltage waveforms (11)
in no-load state, the total harmonic distortion THDu is the lowest. So, during the experimental
investigations the field winding of the examined 5.5 kVA salient pole synchronous generator
is powered using a DC voltage source.
Fig. 16 shows a comparison of the individual harmonic distortion (IHDua) in the induced
stator voltages in phase a shown in Fig. 15. IHDua is calculated due to the Fourier analysis of
voltage ua with respect to the fundamental component.

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 135

(a)

ua ub uc if
100 V/div 100 V/div 100 V/div 500 mA/div

0
4 ms
div

(b)

(c)
Fig. 15. Waveforms of the induced stator voltages under no-load conditions for the examined 5.5 kVA
salient pole synchronous generator with the following rotors: solid, non-skewed (a), solid, non-skewed
with radial incisions (b), skewed, made of electrotechnical sheets (c)

Based on the presented waveforms in the induced voltages in the stator winding in no-load
state of the examined 5.5 kVA salient pole synchronous generator (Fig. 15) and on individual
harmonic distortion (Fig. 16), the total harmonic distortion (THDua, THDua-ri and THDua-e)
counted up to the 75th harmonic is equal to:
– THDua = 16.39% for the generator with a solid, non-skewed rotor without radial in-
cisions (THDua-c = 17.23% calculated for generator with the same rotor construction –
details Fig. 11),
– THDua-ri = 9.58% for the generator with a solid, non-skewed rotor with radial incisions
(THDua-ri-c = 18.34% calculated for generator with the same rotor construction – details
Fig. 11),
– THDua-e-s = 9.30% for the generator with a skewed rotor made of electrotechnical sheets
(THDua-s-c = 10.35% calculated for a generator with the solid, skewed rotor without incisions –
details Fig. 11).

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
136 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

15

u
a-solid without incisions and without skew
IHDu (%)
10 u
a-solid with incisions and without skew
a

u
a-with electrotechnical sheets with skew
5

0
3 5 7 9 11 13
harmonics
(a)
15
u
a-solid without incisions and without skew
u
IHDu (%)

10 a-solid with incisions and without skew

u
a

a-with electrotechnical sheets with skew

0
15 17 19 21 23 25 27 29 31
harmonics

(b)
15
u
a-solid without incisions and without skew

10 u
IHDu (%)

a-solid with incisions and without skew

u
a

a-with electrotechnical sheets with skew

5

0
35 37 39 41 43 45 47 49 51 53 55
harmonics

(c)
15
ua-solid without incisions and without skew
IHDu (%)

10 ua-solid with incisions and without skew

a

ua-with electrotechnical sheets with skew

0
57 59 61 63 65 67 69 71 73 75
harmonics
(d)
Fig. 16. Comparison of individual harmonic distortion in the induced voltages in stator winding in no-
load state of the 5.5 VA salient pole synchronous generator with three rotor constructions: from 3rd to
13th order (a), from 15th to 31st order (b), from 35th to 55th order (c), from 57th to 75th order (d)

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 137

5. Conclusion

The main motivation of this paper is to show the influence of the radial incisions, saliency
and the skewness of the rotor and the stator slotting on the higher harmonic contents in the
magnetic flux density in the air gap (calculated during simulations in the FEMM program) and
in the voltage waveforms induced in the stator winding (registered in a single-layer winding
during experimental investigations) in the no-load state of the salient pole synchronous
generator. The simulations and experimental investigations were carried out for the low power
5.5 kVA salient pole synchronous generator with a solid, non-skewed rotor, with a solid, non-
skewed rotor and radial incisions, and with a skewed rotor made of electrotechnical sheets.
Based on the presented comparison of THDua and THDua for the examined 5.5 kVA synchro-
nous generator, it can be concluded that the introduction of 39 constant depth radial incisions
on the rotor solid surface of the pole shoe equal to 2δ0:
– reduces the amplitude of the magnetic flux density by 10% (Figs. 6 and 7),
– increases by 32% the 3rd harmonic in the component of the normal magnetic flux
density and reduces by 43% its participation in the induced stator voltages,
– in the induced stator voltages gives a similar higher harmonic spectrum to the case of the
generator with the rotor made of electrotechnical sheets.
The higher harmonics that occur in the induced voltages of a single-layer stator winding
are v-times lower than in the normal component of the magnetic flux density. This is due to
the fact that in the single-layer stator winding the harmonic contents are reduced by pitch, slot
opening and skew winding factors.
During the experimental investigation of the three-phase salient pole synchronous
generator with a solid, non-skewed rotor with radial incisions, the total harmonic distortion
was found to be 9.58%. THDua-c obtained from the 2D model in FEMM program (without
eddy current) and expression (11) is 18.34%. It means that there are some differences between
FEMM’s 2D model and the investigated generator that have an influence on the harmonic
content. These differences can be explained by:
– different saturation in the 2D model and in the investigated 5.5 kVA synchronous
generator,
– omission of eddy currents in rotor poles in the 2D model (eddy currents have a shielding
effect on magnetic flux density distribution [30]),
– different shape and nonuniformity of the real air gap along the length of the rotor and
stator than in the 2D model.

References
[1] Dajaku G., Gerling D., Stator slotting effect on the magnetic field distribution of salient pole syn-
chronous permanent-magnet machines, IEEE Transactions on Magnetics, vol. 46, no. 9, pp.
3676-3683 (2010).
[2] Dajaku G., Gerling D., Air-gap flux density characteristics of salient pole synchronous permanent-
magnet machines, IEEE Transactions on Magnetics, vol. 48, no. 7, pp. 2196-2204 (2012).

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
138 K. Ludwinek, R. Nadolski, J. Staszak Arch. Elect. Eng.

[3] Gaussens B., Saint-Michel J., Lécrivain M., Gabsi M., Analytical approach for air-gap modeling of
field-excited flux-switching machine: No-load operation, IEEE Transactions on Magnetics, vol. 10,
no. 10, pp. 1-13 (2012).
[4] Xiaohua B., Na L., Yong F., Fuying L., Novel method of evaluation of Carter factor for closed slot
submersible motor including fringing effect and magnetic saturation, Transactions of China
Electrotechnical Society, vol. 30, no. 12, pp. 220-227 (2015).
[5] Sobczyk T.J., Mathematical model of synchronous generators accounting for saturation due to the
first and the third MMF harmonic, 35th International Symposium on Electrical Machines SME’99,
14-16 June, Kazimierz Dolny, Poland, OWPW, Elektryka, no. 111, pp. 43-51 (1999).
[6] Kutt F., Michna M., Ronkowski M., Chrzan P.J., Non-uniform saturation modelling of synchronous
generator pole shoes, Zeszyty Problemowe Maszyny Elektryczne, no. 103, pp. 121-126 (2014).
[7] Skwarczyński J., Weinreb K., Method of analysis of slot harmonics in the salient-pole synchronous
generators, International Conference on Electrical Machines (ICEM’1990), Aug. 13-15, Boston
MA, USA, pp. 1165-1170 (1990).
[8] Moreira J., Lipo T.A., Modeling of saturated AC machines including air gap flux harmonic com-
ponents, IEEE-IAS Conference Record, 7-12 October, Seattle, Washington, USA, Part 1, pp. 37-44
(1990).
[9] Ludwinek K., An overview of the most important methods of reducing the harmonic content intro-
duced by discretely distributed armature winding, Elektro. Info. (in Polish), no. 7-8, pp. 53-57
(2014).
[10] Kutt F., Michna M., Ronkowski M., Chrzan P.J., Polyharmonic model of synchronous generator
for analysis of autonomous power generation systems, Zeszyty Problemowe Maszyny Elektryczne,
no. 92, pp. 109-114 (2011).
[11] Ludwinek K., Influence of DC voltage and current of field winding on induced stator voltages of a
salient pole synchronous generator, International Review of Electrical Engineering, vol. 9, no. 1,
pp. 62-72 (2014).
[12] Ludwinek K., FEMM utilisation in representation of inductance distributions in a salient pole
synchronous generator circuital model in no-load state, Technical Transactions – Electrical
Engineering, vol. 1-E/2015, pp. 325-341 (2015).
[13] Ludwinek K., Some aspects of inductance distributions modeling in dq0-axes and damping circuits
on the rotor of a salient pole synchronous generator, Technical Transactions – Electrical Engine-
ering, vol. 2-E/2015, pp. 37-52 (2015).
[14] Dąbrowski M., Design of Alternating Current Electrical Machines, WNT (in Polish), Warszawa
(1994).
[15] Pyrhönen J., Jokinen T., Hrabcová V.P., Design of Rotating Electrical Machines, John Wiley &
Sons (2014).
[16] Carter F.W., Note on air gap and interpolar induction, Journal of Institution of Electrical Engine-
ering, vol. 29, no. 146, pp. 923-933 (1926).
[17] Ludwinek K., Influence of representation of the stator to rotor mutual inductances on the induced
phase voltage waveforms in a salient pole synchronous generator, Zeszyty Problemowe Maszyny
Elektryczne, no. 104, pp. 147-154 (2014).
[18] Ludwinek K., Nadolski R., Staszak J., Comparison of higher harmonic contents in salient pole syn-
chronous generator with radial incisions on the solid pole surface, Maszyny Elektryczne – Zeszyty
Problemowe, no. 108, pp. 113-119 (2015).
[19] Vicol L., Banyai A., Viorel I.A., Simond J.J., On the damper cage bars’ currents calculation for
salient pole large synchronous machines, 11th International Conference on Optimization of Elec-
trical and Electronic Equipment, May 22-24, Brasov, Romania, pp. 9-14 (2008).
[20] Nadolski R., Staszak J., Natural hunting pulsations of large power turbogenerators, Archives of
Electrical Engineering, vol. XLV, no. 4, pp. 405-459 (1996).
[21] Bacher J., Maier G., The shielding effect of the built-in damper cage in a synchronous machine,
International Conference on Renewable Energy and Power Quality (ICREPQ'04), 31 March-02
April, Barcelona, Spain, pp. 250-259 (2004).

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM
Vol. 66 (2017) Comparison of higher harmonic contents in salient pole synchronous generator 139

[22] Sahoo S.K., Rodriguez P., Sulowicz, M., Comparative investigation of fault indicators for syn-
chronous machine failures, International Conference on Electrical Machines (ICEM’2014), Sep-
tember 2-5, Berlin, Germany, pp. 1503-1509 (2014).
[23] Rodriguez P., Sahoo S., Pinto C.T., Sulowicz M., Field current signature analysis for fault detec-
tion in synchronous motors, in diagnostics for electrical machines, IEEE 10th International Sym-
posium on Electric machines, Power Electronics and Drives (SDEMPED), 1-4 September, Guarda,
Portugal, pp. 246-252 (2015).
[24] Raziee S.M., Kelk H.M., Alikhani H.R.R., Omati A., Air-gap eccentricity effects on harmonic
contents of field current in synchronous generators, International Review of Electrical Engineering,
vol. 5, no. 1, pp. 83-89 (2010).
[25] Akbari H., An improved analytical model for salient pole synchronous machines under general
eccentricity fault, Progress in Electromagnetic Research, no. 49, pp. 389-409 (2013).
[26] http://www.femm.info/wiki/HomePage, accessed November 2015.
[27] Jedryczka C., Szelag W., Demenko A., Wojciechowski R. M., Description of multiply connected
regions with induced currents using T-T method, Progress in Electromagnetic Research B, no. 43,
pp. 279-294 (2012).
[28] Bastos J.P.A., Sadowski N., Electromagnetic Modeling by Finite Element Methods, Marcel Dekker,
New York (2003).
[29] Gomez E., Roger-Folch J., Gabaldon A., Molina A., Coupling 2D finite element models and circuit
equations using a bottom-up methodology, ACES Journal, vol. 17, no. 3, pp. 225-231 (2002).
[30] Chung Y., Galyda J., Effect of Eddy Current in the Laminations on the Magnet Field, LS Note
No. 200, April (1992).

Brought to you by | Uniwersytetu Technologicznego w Szczecinie - Biblioteka Glówna Zachodniopomorskiego

Authenticated
Download Date | 3/2/17 8:42 AM