Optimal Scheme For Structural Design of Large Turbogenerator Stator End Winding

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

This article has been accepted for publication in a future issue of this journal, but has not been

fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

Optimal Scheme for Structural Design of Large


Turbogenerator Stator End Winding
Yang Zhao, Bo Yan, Chong Zeng, Song Huang, Changlin Chen, Jianan Deng

 electromagnetic forces. Most of the analytical methods are


Abstract-- A three-dimensional electromagnetic model of the end based on Biot-Savart's law, by which the geometrical details
region of a 600MW turbogenerator is set up. The variations of of the stator end winding cannot be taken into account [3]-[5].
electromagnetic force densities of the stator end winding under rated Recently, the numerical approaches have been employed to
load and three-phase short circuit from full load are obtained by study electromagnetic forces on the stator end winding and its
numerical simulation. Then the electromagnetic forces varying with dynamic responses under the electromagnetic forces.
time and space on the end winding are applied to the three- Scott et al. [6],[7] analyzed the end magnetic field, which is
dimensional fine finite element model established in ABAQUS/CAE assumed to distribute sinusoidally in the circumferential, by
for structural analysis, and the deformation and stress distributions of the finite difference method and obtained the electromagnetic
the stator end winding during rated load and three-phase short circuit forces on the end winding of a turbogenerator in steady state
are numerically analyzed by means of the ABAQUS software.
and transient modes of operation. Khan et al. [8] used a quasi-
Furthermore, the influences of structural and material parameters of
three-dimensional finite element (FE) method to obtain flux
the end winding on its dynamic responses under rated load are
numerically investigated. Based on the numerical results, an optimal
densities and electromagnetic forces on the end winding of a
scheme for the design of the turbogenerator stator end winding is 1000MW two-pole turbogenerator in the case of three-phase
proposed. short circuit from full load. Wen et al. [9] investigated the
transient electromagnetic forces on the end winding in a
Index Terms--Turbogenerator, stator end winding, dynamic 35MW turbogenerator under various conditions by means of a
response, numerical simulation, optimal scheme. transient quasi-three-dimensional method. Liu and Hjärne [10]
set up a three-dimensional FE model to calculate
I. INTRODUCTION electromagnetic forces on the end winding of a 635MVA
turbogenerator under rated load and fault conditions. In the
The stator end winding of a large turbogenerator is subject to numerical model, the rotating excitation fields are equivalently
time-varying electromagnetic forces when the machine is generated by a set of two-phase windings carrying
running. With the increase of the turbogenerator capacity, the sinusoidally alternating currents with space shift of 90 degrees.
electromagnetic forces, which vary with the square of current, However, there is an obvious difference between the
on the winding rise up significantly. Serious vibration of the simplification and the actual situation.
stator end winding during operation may result in excessive Richard et al. [11],[12] set up a three-dimensional
wear, fatigue and even damage [1],[2]. Therefore, it is electromagnetic model of the stator end region of a 600MW
necessary to decrease vibration level of the stator end winding turbogenerator to analyze the electromagnetic forces and a
as low as possible to ensure the operation safety of the three-dimensional mechanical model to analyze deformation
turbogenerator, and the investigation to find an optimal of the end winding under the electromagnetic forces during
scheme of the stator end winding is a focus in the design stage rated steady state and fault conditions. The conductor bars are
of a turbogenerator. simplified as straight line inductors in the electromagnetic
In the past several decades, analytical and numerical model, and the conductor bars and packings are modelled with
methods have been utilized to compute electromagnetic forces beam elements and the radial braces by the shell elements in
and dynamic responses of the stator end winding under the the mechanical model. In references [13]-[15], the
electromagnetic forces on end winding were determined based
Y. Zhao is with the College of Aerospace Engineering, Chongqing on the Biot-Savart’s law to line circuit segments, and the
University, Chongqing 400044, China, and also with the School of Advanced deformations of the stator end windings of different
Manufacturing Engineering, Chongqing University of Posts and turbogenerators were analyzed with FE model, in which three-
Telecommunications, Chongqing 400065, China (e-mail: dimensional solid elements were used to model the conductor
zhaoyang008@cqu.edu.cn).
B. Yan (corresponding author) is with the College of Aerospace bars, spring elements to model bandages and shell elements to
Engineering and the State Key Laboratory of Power Transmission Equipment the other components. Kim et al. [16] analyzed the
& System Security and New Technology, Chongqing University, Chongqing electromagnetic forces on the end winding of a 900kW
400044, China (e-mail: boyan@cqu.edu.cn). induction motor by the three-dimensional FE method, and the
C. Zeng, and S. Huang are with the State Key Laboratory of Power
Transmission Equipment & System Security and New Technology, stress of the end winding parts with only one supporting ring.
Chongqing University, Chongqing 400044, China (e-mail: Lin et al. [17] carried out a three-dimensional electromagnetic
tmshyer@gmail.com; s.huang@cqu.edu.cn). analysis and a three-dimensional mechanical analysis to study
C. Chen, and J. Deng are with Dongfang Electric Machinery Co. Ltd, the vibrations and deformations of an induction machine
Deyang 618000, China (e-mail: chencl@dfem.com.cn; djadfem@163.com).

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

caused by steady state magnetic forces on the end winding. In


the model, the nonlinear magnetization characteristics of the
core are not covered and the support components of end
winding are simplified. Based on an integral formulation for
nonlinear magnetostatic problems, Albanese et al. [18],[19]
obtained the distributions of the magnetic field and forces in
the end winding of a turbogenerator by means of a FE model,
in which beam elements are used to discrete the end winding.
Although numerical methods have been utilized to analyze
electromagnetic forces and dynamic responses of stator end
winding under the electromagnetic forces, simplification of
the electromagnetic models and/or structural models may give
rise to larger errors. The authors of this paper set up a three-
dimensional FE model of the stator end winding of a 600MW
turbogenerator and performed a parameter study on its
dynamic characteristics by means of the ABAQUS software Fig.1. Stator end winding of a 600MW turbogenerator
[20], but no dynamic responses of the stator end winding
under electromagnetic forces were analyzed. TABLE I
In this paper, an electromagnetic model of the stator end PARAMETERS OF THE 600MW TURBOGENERATOR
region of a 600MW turbogenerator, which includes the stator Parameters Value
stepping core, pressure plate, pressure finger and copper shield, Rated output 667MVA/600MW
is set up, and the electromagnetic force densities in the stator Power factor 0.9(lag)
Revolution speed 3000r/min
end winding under rated load and three-phase short circuit
Frequency 50Hz
from full load are determined by the FE method. Then the
Number of phases 3
dynamic responses of the end winding are simulated with the
Number of stator slots 42
FE method, and the deformation and stress distributions of the Number of leads 6
end winding under the electromagnetic forces during rated Number of poles 2
load and three-phase short circuit are obtained. By means of
the numerical method, parameter study on the dynamic Copper
responses of the end winding is carried out and an optimal shield
scheme for the design of stator end winding is proposed. Pressure
plate Stepping
core

II. NUMERICAL SIMULATION OF Pressure


ELECTROMAGNETIC FORCES finger

A. FE Model for Electromagnetic Field Analysis


The stator end winding of a typical 600MW two-pole Z
turbogenerator is shown in Fig.1, and the parameters of this
machine are listed in Table I. Due to the structural symmetry T
R
of the turbogenerator, half of the end region is chosen as the (a)
electromagnetic analysis model, in which the stator stepping
core, involute-shaped conductor bars, pressure finger, pressure
plate, copper shield, rotor shaft and winding are included. To
simplify the model, the stator winding bars which are hollow
copper conductors are replaced by solid ones. r
S
The numerical model of the stator end winding model for
electromagnetic analysis is shown in Fig.2. It is assumed that:
(1) the displacement currents are neglected; (2) the current
densities in the end winding bars are uniformly distributed and
the higher order harmonics are omitted; (3) the stator core is
isotropic and the B-H curve is a single-valued function; (4) the h
influence of the end winding deformation on the magnetic
field is ignored.
(b)
Fig.2. Numerical model of turbogenerator end region for electromagnetic field
analysis. (a) Model without air region. (b) Whole model.

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

B. Boundary Conditions and Currents in End Winding Bars directions. It can be seen that the magnitudes and directions of
The boundary conditions of the electromagnetic model are the electromagnetic force densities at these points differ from
shown in Fig.2 (b). The magnetic field direction is reversed each other, and the dominant frequencies of the forces are
every 180°. Surface S and the surface located at 180° apart 100Hz, which is twice the work frequency of the generator. It
from S satisfy the rotation and antiperiodic boundary is noted that there are some differences, especially the
conditions. Rotational conditions are specified on the rotor amplitudes, of a pair of waves in one work period.
shaft and the rotor winding. In addition, symmetry boundary The electromagnetic force densities on the straight portions
conditions are specified on the surfaces at Z=0, Z=h and R=r. of the end winding in the core are larger, and the force
The parameters of excitation source are determined by two- densities decrease with the distance from the stator core at a
dimensional electromagnetic analysis. The armature currents time. At the points with the same axial position, such as points
and the field currents are 18524.6A and 4356.39A respectively, 2 and 6, the force densities on the upper bar are larger than
and the initial phase angles of the armature current and voltage those on the lower one. Moreover, the radial components of
are -77.5 deg and -51.7 deg respectively under rated load. The the force densities on the straight portion in the core are much
currents in armature and field winding in the first 0.1s after greater than those on the circumferential and axial portions.
three-phase short circuit from full load are shown in Fig.3. It Moreover, the differences among the force components in the
can be seen that the largest currents in armature winding under three directions at the same position out of the core are small.
three-phase short circuit are more than 10 times with those 4
under rated currents; the largest field currents in the field
winding are nearly 8 times to the rated ones in the first period,
and both currents gradually decay with time. It is noted that
3
there is a distortion of the sinusoidal current waveform. Based 5
on the numerical investigation, it is found that if the damping
of the damper winding is large enough, the shape of stator
currents is a standard sinusoidal waveform without distortion. 2
On the other hand, if there is not damper winding or its 6
damping is insufficient, there may be a distortion in the shape
of stator currents. Therefore, the smaller damping of the
Core
incomplete damper winding used in the turbogenerator Core
discussed in this paper may be responsible for the distortion of
the sinusoidal current waveform. 1 7
Fig.4 Typical points on a pair of end winding bars.
5
3x10 3x10
4
5 I
IA
2x10 C 2x10
4
6 6
5 4 3.5x10 2.5x10
1x10 1x10 3.0x10
6 1 6
2.0x10
If /A

0 0 6
I/A

2.5x10 6
6 1.5x10
fR/ N/m3

-1x10
5
-1x10
4 2.0x10 7
fR/N/m3

1.5x10
6 2 1.0x10
6
5 4
-2x10 -2x10 1.0x10
6
5 6
5 4 5 5.0x10
-3x10 IB -3x10 5.0x10 3
0.0 0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10 4 5
5 5
t/s t/s -5.0x10 -5.0x10
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
(a) (b) t/s t/s
Fig.3. Currents after three-phase short circuit from full load. (a) Armature 1.0x10
6
5.0x10
5

currents. (b) Field currents. 5


7.5x10 2 2.5x10
5
5
5.0x10
fT/ N/m3
fT/ N/m3

0.0 7
C. Electromagnetic Forces on End Winding under Rated 2.5x10
5
3
Load 0.0 5 5
4 -2.5x10
-2.5x10
5
1
With the model set up above, the distributions of the 5
6
-5.0x10
5 -5.0x10
0.00 0.01 0.02 0.03 0.04 0.05
magnetic flux density and the electromagnetic force density of 0.00 0.01 0.02 0.03 0.04 0.05
t/s
t/s
the end winding from 0s to 0.05s under the rated load are 5
4.0x10
determined by means of the FE method. The time increment 5.0x10
5

for the transient analysis is set to be 6.25×10-4s.


5
3.0x10 5
2.5x10
To analyze the electromagnetic force density characteristics 5 2
fZ/ N/m3

2.0x10
fZ/ N/m3

4 0.0
of the end winding under rated load, Seven points on a pair of 1.0x10
5
7
bars, as shown in Fig.4, are selected. Points 2 and 6 are 0.0 3 -2.5x10
5 5 6
located near the ends of the bar connected with the stator core, 5
1
5
and point 4 is at the nose of the connection between the upper -1.0x10 -5.0x10
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
bar and the lower bar. The time histories of electromagnetic t/s t/s
force densities at these points under rated load are shown in (a) (b)
Fig.5. Time histories of electromagnetic force densities in different directions
Fig.5, in which fR, fT, fZ are respectively the electromagnetic at some points under rated load. (a) At points on upper bar. (b) At points on
force density components in radial, circumferential and axial lower bar.

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

D. Electromagnetic Forces on End Winding under Three- forces determined by the half mode of the end winding in
Phase Short Circuit from Full Load Section II.A can be expanded to the whole structural analysis
The time histories of the electromagnetic force densities in model.
radial, circumferential and axial directions at the typical points L-shaped
Radial
under three-phase short circuit are shown in Fig.6. It is seen bracket
that the electromagnetic force densities in all directions brace
gradually decay with time, which is in accordance with the
change trend of the currents. Compared with the
electromagnetic force densities under the rated load, it is Supporting
known that the maximum values at some points under three- ring Bolt
phase short circuit case are more than hundreds times to those
under the rated load.
Pin
8 7
2.0x10 8.0x10
8
1.6x10 7
7
8
1
6.0x10 External
1.2x10 7
ring
fR/ N/m3

fR/ N/m3

4.0x10
8.0x10
7
2 6
7
7 2.0x10
4.0x10
0.0 3 0.0 Copper
4 5
-4.0x10
7
0.00 0.02 0.04 0.06 0.08 0.10
-2.0x10
7
0.00 0.02 0.04 0.06 0.08 0.10
bracket
(a)
t/s t/s
7 7
2.0x10 2.0x10
7 0.0 5
1.0x10 2
3 -2.0x10
7
0.0
fT/ N/m3

fT/ N/m3

4 -4.0x10
7 7
7
-1.0x10 7
-6.0x10
7 1 6
-2.0x10 -8.0x10
7

7 8
-3.0x10 -1.0x10
0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10
7
t/s t/s
7
3.0x10 5.0x10
7
7 4.0x10
2.0x10 4 7
3.0x10
fZ/ N/m3
fZ/ N/m3

7 3 7
6
1.0x10 2.0x10
7
1.0x10
0.0 5
1 0.0
2 7
7
-1.0x10
7
-1.0x10
0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10
t/s t/s (b) (c)
(a) (b) Fig.7. FE model of stator end winding for structural analysis. (a) Geometric
Fig.6. Time histories of electromagnetic force densities in different directions model. (b) Mesh. (c) Boundary condition.
at some points under three-phase short circuit. (a) At points on upper bar. (b)
At points on lower bar.
Furthermore, Rayleigh damping model is employed in the
analysis. According to [21], the damping ratio of the stator end
III. DYNAMIC RESPONSES OF END WINDING UNDER winding is in the range of 0.01 and 0.03, which is the same
ELECTROMAGNETIC FORCES obtained by the model test carried out in [20]. So the damping
ratio of the stator end winding is set to be 0.02 in the dynamic
A. FE model for Structural Analysis analyses.
The structural analysis model is different from the
electromagnetic model. It is not necessary to take into account B. Dynamic Response under Rated Load
the components such as rotor shaft, rotor winding, pressure With the FE structural analysis model, the dynamic
finger, pressure plate, copper shield and stator core in the response of the stator end winding under electromagnetic
structural analysis model. Due to the larger stiffness of the forces in the case of rated load is simulated by means of the
stator core, the influence of the stator winding on the core is ABAQUS/Standard software. The deformations of the end
ignored. The material properties, connections between winding under rated load at typical moments during the first
adjacent components and boundary conditions are discussed in period are shown in Fig.8. It can be seen that the deformation
detail in [20]. The FE structural analysis model of stator end of the end winding changes with time and the deformed shape
winding is shown in Fig.7. at any time is similar to an ellipse, which is consistent with the
In order to analyze the dynamic responses of end winding distribution of electromagnetic force at a time. Although the
under electromagnetic forces under rated load and three-phase electromagnetic force densities at the nose portion are smaller
short circuit, the electromagnetic force densities obtained in than those at the involute portion, the deformations at the
Section II are applied to the FE model for structural analysis. former portion are larger than those at the later portion, for
Due to the symmetry of the structure, the electromagnetic which the weaker constraint at the nose portion are responsible.

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

The time histories of displacements at points 3 and 4, as positions are assumed to be fixed in the structural analysis
shown in Fig.4, of the end winding under rated load are shown model. Because there is no supporting structure on the
in Fig.9, in which U, UR, UT, UZ, are the displacement, the connections between leads and bars, it is reasonable that the
radial, circumferential and axial components of the maximum stress at some times appears there. Moreover, the
displacement respectively. It is seen the vibration arrives at binding constraints may induce the maximum stress at some
stable state after about 0.1s, and the amplitudes of a pair of times appearing at the contact zones between bars and external
vibration in a work period are different because the rings.
corresponding electromagnetic force amplitudes are different It is known that the maximum Von Mises stresses of the
as shown in Fig.5. It is evident that the radial displacement end winding in the first five periods are less than 10MPa,
components are larger than those in circumferential and axial
which is much smaller than the ultimate strength of the
directions, and the winding bars mainly reflect radial
material. Therefore, the strength of the stator end winding is
deformation in the form of bending. In addition, the vibration
enough under the rated load. However, fatigue damage of the
amplitude of point 3 is smaller than that of point 4, and the
maximum radial displacement of point 4 at the nose portion is components may take place and it is worth to be investigated
74μm, close to the test result 75μm of the same level in the further.
turbogenerator presented in [22]. 80 80

60 60

U/m

U/m
40 40

20 20

0 0
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
t/s t/s
80 80
60 60
40 40
UR/m

UR/m
20 20
0 0
(a) (b) -20 -20
-40 -40
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
t/s t/s
80 80
60 60
40 40
UTm
UTm

20 20
0 0
-20 -20
-40 -40
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
t/s t/s
80 80
60 60
(c) (d) 40 40
UZ/m

UZ/m

Fig. 8 Deformations of stator end winding at typical moments under rated load 20 20
(deformations are multiplied by 2500). (a) t=0.01s. (b) t=0.0125s. (c) t=0.015s. 0 0
(d) t=0.0175s. -20 -20
-40 -40
Because the electromagnetic force densities at positions 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
t/s t/s
close to the stator core are large and the winding is subject to (a) (b)
bending deformation, the maximum stress of a single winding Fig. 9 Time histories of displacements at typical points of an end winding bar
bar under electromagnetic force appears at the knuckle of the under rated load. (a) Point 3. (b) Point 4.
winding. However, the maximum stress of the whole end
winding occurs at another location different from that on the C. Dynamic Response under Three-Phase Short Circuit
single bar because of the binding constraint of the structure. It The deformations of the stator end winding under three-
is noted that the stress distribution of the end winding varies phase short circuit at typical moments in the first period are
shown in Fig.10. It can be seen that the deformation of the end
with time and the maximum stress occurs at different position
winding changes with time and its deformed shape at any time
with the change of time although it mostly appears at the ends
is closed to an ellipse too. Moreover, the deformations of the
of the bars connecting with the core, the connections between stator end winding in this load case are much larger than those
leads and bars and the contact zones between bars and external in the case of rated load.
rings. The maximum stress at some times appears at the ends The time histories of displacements at point 4, as shown in
of the bars connecting with the stator core because these Fig. 4, of the stator end winding under three-phase short

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

circuit are shown in Fig.11. Compared with the results under


rated load as shown in Fig.9, it is evident that the largest
displacement at the nose portion under three-phase short IV. PARAMETRIC STUDY ON DYNAMIC RESPONSE OF
circuit is over one hundred times more than that under rated STATOR END WINDING
load. The influence of the structural parameters and material
It is noted that the maximum Von Mises stress of the end
parameters on the dynamic response of the stator end winding
winding even exceeds 800MPa in the case of three-phase short
under rated load, based on which an optimal design scheme
circuit, which is larger than the ultimate strength of the
may be selected, are performed by means of the FE method in
material.
Based on analysis of the deformation and stress, it is seen this section.
that the end winding is dangerous in the case of three-phase The parameters and their variation ranges, which may be
short circuit, which may even give rise to abrupt collapse of available in the design of the stator end winding, are
the structure. summarized as follows:
(1) Stiffness of the rings including the supporting ring at the
nose portion and the external rings: 0.5Er, Er, 1.5Er, 2.0Er,
where Er is the Young’s modulus of the original rings.
(2) Stiffness of the radial braces: 0.5Eb, Eb, 1.5Eb, 2.0Eb,
where Eb is the Young’s modulus of the original radial braces.
(3) Number of the spacer blocks and binding forms. In the
original design of the investigated turbogenerator, there are 7
spacer blocks between any pair of adjacent upper bars and 6
spacer blocks between any pair of adjacent lower ones. In
addition, every 9 upper bars and every 3 or 9 lower bars are
(a) (b) bound together respectively. In Case 2, the number of spacer
blocks between adjacent upper bars increases to 8 from 7. In
Case 3, the number of spacer blocks between adjacent lower
bars increases to 8 from 6, and every 9 lower bars are bound
together. In Case 4, the number of spacer blocks between
adjacent upper and lower bars increases to 8 from 7 and 6
respectively, and every 9 lower bars are bound together.
(4) Connections between the L-shaped brackets and the
bolts and the connections between the pins and the copper
brackets: sliding and bonding.
(5) Number of pins bonded to the radial braces: 0, 6, 18, 30,
36 and 42.
(c) (d) In the parameter study, point 3 at involute portion of the
Fig.10 Deformations of stator end winding at typical moments under three- upper bar, point 4 at nose portion and point 5 at the involute
phase short circuit (deformations are multiplied by 20). (a) t=0.0102s. (b) portion of the lower bar are selected as evaluation points to
t=0.0118s. (c) t=0.0137s. (d) t=0.0156s. investigate the influences of the structural parameters and
6000
6000 material parameters on the dynamic response of the stator end
4000
4000
winding under rated load.
UR/m

2000
U/m

0
2000 A. Stiffness of Rings
-2000
0
The supporting ring and the four external rings contribute
0.00 0.02 0.04 0.06 0.08 0.10 -4000
t/s
0.00 0.02 0.04 0.06 0.08 0.10 for the global structure stiffness of the end winding, which
t/s may affect its dynamic response under electromagnetic force.
(a) (b) The numerically determined maximum displacements at
6000 6000 typical points of the stator end winding as its rings with
4000 4000 different Young's modulus are listed in Table II. It can be seen
UT/m

2000 2000 that the maximum displacements at the typical points decrease
UZm

0 0 obviously with the increase of the Young's modulus of the


-2000 -2000 rings.
-4000 -4000
0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10
t/s t/s
(c) (d)
Fig. 11 Time histories of displacements at point 4 of an end winding bar under
three-phase short circuit. (a) Displacement. (b) Radial displacement
component. (c) Circumferential displacement component. (d) Axial
displacement component.

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

TABLE II connections may lead to larger thermal stress in the


EFFECT OF STIFFNESS OF RINGS ON DEFORMATION
components.
OF END WINDING
Maximum displacements (μm)
Points TABLE V
0.5Er Er 1.5Er 2.0Er EFFECT OF CONNECTIONS BETWEEN ADJACENT
3 53.02 52.73 47.77 45.79 COMPONENTS ON DEFORMATION OF END WINDING
4 79.58 73.57 61.77 56.66 Maximum displacements (μm)
Points
5 40.03 38.00 35.94 34.47 Sliding connection Bonding connection
3 52.73 54.07
B. Stiffness of Radial Braces 4 73.57 62.47
Similar to the rings, variation of the Young's modulus of the 5 38.00 25.15
radial braces may change the structural stiffness of the end
winding. The maximum displacements at typical points of the E. Number of Pins Bonded to Radial Braces
stator end winding as the radial braces with different Young's With the increase of the number of pins bonded to the radial
modulus are shown in Table III. It is seen that the stiffness of braces, the structural stiffness of the stator end winding
radial braces obviously affect the vibration amplitude of end enhances. The dynamic responses of the stator end winding in
winding. the cases that respectively 0, 6, 18, 30, 36 and 42 pins are
bonded to the radial braces are numerically investigated. In
TABLE III each case, the pins bonded to the braces are symmetrically
EFFECT OF STIFFNESS OF RADIAL BRACES ON arranged along the circumference of the end winding. The
DEFORMATION OF END WINDING
numerically obtained results are given out in Table VI. It can
Maximum displacements (μm)
Points be seen that as the increase of the number of pins bonded to
0.5Eb Eb 1.5Eb 2.0Eb
radial braces, the maximum displacements at typical points
3 54.84 52.73 50.25 47.44
decrease. However, as the number is greater than 18, its
4 72.69 73.57 65.35 64.15
influence on the dynamic responses seems limited. Therefore,
5 55.92 38.00 32.59 32.73
it may be an efficient way to bond some of the pins to the
radial braces to reduce the vibration of the end winding.
C. Number of Spacer Blocks and Binding Forms
Four cases of different number of spacer blocks and binding TABLE VI
forms introduced above are numerically investigated. The EFFECT OF NUMBER OF PINS BONDED TO
effect of number of spacer blocks and binding forms on the RADIAL BRACES ON DEFORMATION OF END WINDING
deformation of the stator end winding is listed in Table IV, Maximum displacements (μm)
from which it is implicated that the change of the number of Points Number of pins bonded to radial braces
spacer blocks between adjacent bars and the binding forms 0 6 18 30 36 42
affects the deformation of the end ending very limitedly. 3 52.73 49.73 45.41 43.68 43.59 43.08
4 73.57 67.11 63.90 63.74 63.86 64.30
TABLE IV 5 38.00 34.49 33.02 33.14 33.29 33.31
EFFECT OF NUMBER OF SPACER BLOCKS AND BINDING
FORMS ON DEFORMATION OF END WINDING
Maximum displacements (μm) F. Optimal Scheme of End Winding Design
Points Original In the design of the stator end winding of a turbogenerator,
Case 2 Case 3 Case4
(Case 1) its natural frequencies corresponding to specific mode shape,
3 52.73 51.63 50.52 49.46 such as ellipse for a two-pole generator, have to avoid the
4 73.57 72.44 70.36 69.11 resonance range of 94Hz-115Hz. On the other hand, the
5 38.00 38.37 34.24 34.43 deformations of the end winding under different load cases
have to be sufficiently small. To find an optimal scheme, the
D. Connections Between Adjacent Components two conditions, prevention of resonance and least deformation,
To avoid high level stress in the components of the stator have to be satisfied. It is noted that the stress in the end
end winding induced by the temperature variation during winding is not taken into account because the maximum Von
operation of the turbogenerator, the bolts are designed to Mises stresses of the end winding in all the parametric study
freely slide in the holes of the L-shaped brackets, and the pins cases are far less than the ultimate strength of the material
to slide in the chutes of the copper brackets, as shown in Fig. under rated load.
7(a), in the original design. In the FE structural analysis model, The authors of this paper performed the parameter study on
these connections are modelled by “translator” in ABAQUS the dynamic characteristics of the stator end winding [20], and
software [20]. It is assumed that the L-shaped brackets and the the influence law of the parameters on the natural frequencies
bolts as well as the pins and the copper brackets are bonded corresponding to elliptic mode are shown in Fig.12. The
together, and the dynamic responses of the end winding under influences of some structural and material parameters on the
electromagnetic force are numerically simulated. The dynamic response of the stator end winding are numerically
maximum displacements at typical points in the cases of investigated above, and the influence laws of these parameters
sliding and bonding connections are shown in Table V. It is are shown in Fig. 13.
obvious that the variation law of the maximum displacements From Fig.12, it is known that the natural frequencies of
at different points are different. However, constraint of the elliptic mode are usually smaller than the lower bound of the

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

resonance frequency 94Hz. In view of resonance prevention, elliptic frequencies and deformations under dynamic
the further the natural frequency leaves from the lower bound, electromagnetic forces, as shown in Figs.12(c) and 13(c). That
the safer the structure is. So a design scheme with the lowest means the effects of changing the number of spacer blocks and
frequency and the smallest deformation may be the best choice. binding forms on dynamic characteristics and dynamic
However, there is a contradiction between the two aspects. response of the stator end winding are very limited.
From an overall perspective of Figs.12 and 13, the further the

Maxmium displacement/m

Maxmium displacement/m
natural frequency leaves from the lower bound, the larger the 80 80
maximum displacement of end winding is. So decreasing the
60 60
deformation of end winding under electromagnetic forces with
the premise that natural frequencies of the elliptic modes avoid 40 40
the resonance frequencies should be considered in the
optimization design. 20 20
From Figs. 12(a) and (b), it is seen that the natural
frequencies of elliptic mode of the stator end winding, in the 0 0
0.5 1.0 1.5 2.0 Er 0.5 1.0 1.5 2.0 E
variation range of the Young's modulus of the rings or radial Young's modulus of rings Young's modulus of radial braces b
braces, are smaller than the lower bound of the resonance (a) (b)
frequency. On the other hand, the deformation of the end

Maxmium displacement/m
Maxmium displacement/m
80 80
winding is smallest as the Young's modulus of the rings or the
radial braces is largest, i.e. two times of the original material, 60 60
as shown in Figs. 13(a) and (b). Therefore, to increase the
stiffness of the rings and/or the radial braces is a possible 40 40
optimal scheme for the design of the end winding.
20
20
100 100
0
1 2 3 4 0
Natural frequency/Hz
Natural frequency/Hz

90 90 Case Sliding fixed


(c) (d)
80 80
Maxmium displacement/m

70 70 80
Elliptic mode 1 Elliptic mode 1
60 Elliptic mode 2 60 Elliptic mode 2 60
Elliptic mode 3 Elliptic mode 3
50 50
0.5 1.0 1.5 2.0 Er 0.5 1.0 1.5 2.0 E 40
Young's modulus of rings Young's modulus of radial braces b
(a) (b) 20

110 120 0
Elliptic mode 1 Elliptic mode 1 0 6 12 18 24 30 36 42
100 110
Natural frequency/Hz

Natural frequency/Hz

Elliptic mode 2 Elliptic mode 2 Number of pins bonded to radial braces


90 Elliptic mode 3 100 Elliptic mode 3 (e)
90 Fig. 13 Influences of different parameters on the maximum displacements of
80 end winding. (a) Stiffness of rings. (b) Stiffness of radial braces. (c) Number
80 of spacer blocks and binding forms. (d) Connections between adjacent
70
70 components. (e) Number of pins bonded to radial braces.
60
60
50 If the connections between the L-shaped brackets and the
1 2 3 4 50 Sliding
Case Fixed bolts as well as the pins and the copper brackets are bonded,
(c) (d) the natural frequencies of elliptic modes are larger than the
upper bound of resonance frequency 115Hz and the maximum
100
deformation of end winding under electromagnetic force
Natural frequency/Hz

90 decreases, as shown in Figs. 12(d) and 13(d). However,


80
thermal stress may increase due to the fixed connections.
Therefore, it may be a possible optimal scheme if the
70
Elliptic mode 1
maximum thermal stress is less than the allowable value.
60 Elliptic mode 2 As the number of the pins bonded to the radial braces
Elliptic mode 3 increases, the natural frequencies increase and the deformation
50 of the end winding under electromagnetic force decreases, as
0 6 12 18 24 30 36 42
Number of pins boned to radical braces shown in Figs. 12(e) and 13(e). However, as the bonded
(e) number is greater than 18, some of the natural frequencies of
Fig. 12 Influences of different parameters on natural frequencies of elliptic elliptic modes may locate in the resonance range and the
modes. (a) Stiffness of rings. (b) Stiffness of radial braces. (c) Number of
spacer blocks and binding forms. (d) Connections between adjacent
deformation of the end winding change very limited. So
components. (e) Number of pins bonded to radial braces. bonding 18 pairs of pins radial braces may be another optimal
scheme.
In the four cases of changing number of spacer blocks and It can be seen that increasing the stiffness of rings may most
binding forms, there are a few differences of the natural greatly decrease the maximum displacements of the end

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

winding, and the natural frequencies of elliptic modes can [2] S. Wan, C. Zhan, X. Yao, Y. Deng and G. Zhou, "Investigation on
electromagnetic forces and experimental vibration characteristics of
avoid resonance at the same time. So it is the optimal scheme stator end windings in generator," Transactions of the Canadian Society
for design of the stator end winding. for Mechanical Engineering, vol. 38, no. 3, pp. 331-345, 2014.
It is known that various types of end winding support [3] Calvert J F, "Forces in turbine generator stator windings," Trans. of the
system, such as binding form and clamping form, are used in American Institute of Electrical Engineers, vol. 50, no. 1, pp. 178-194,
1931.
two-pole or four-pole large turbogenerators. Usually, the [4] D. Harrington, "Forces in machine end windings," Electrical
support structures are composed of cones, rings and radial Engineering, vol. 72, no. 2, pp. 153-153, 1953.
brackets. For a large turbogenerator, a support structure with [5] M. Ohtaguro, K. Yagiuchi, and H. Yamaguchi, "Mechanical behavior
enough strength and stiffness is necessary to prohibit high of stator endwindings," IEEE Trans. Power Apparatus and Systems, vol.
Pas-99, no. 3, pp. 1181-1185, 1980.
level vibration of the end winding. From the above discussion [6] D. J. Scott, S. J. salon, and G. L. Kusik, "Electromagnetic forces on the
on the specific turbogenerator, it is known that increasing the armature end windings of large turbine generators I-steady state
stiffness of support structure and changing the connection conditions," IEEE Trans. Power Apparatus and Systems, vol. Pas-100,
stiffness among the axial elastic components properly will no. 11, pp. 4597-4603, 1981.
[7] S. J. Salon, D. J. Scott, and G. L. Kusik, "Electromagnetic forces on the
improve the performance of the end winding. end windings of large turbine generators II-transient conditions," IEEE
Trans. Power Apparatus and Systems, vol. Pas-102, no. 1, pp. 14-19,
1983.
V. CONCLUSIONS [8] G. K. M. Khan, G. W. Buckley and N. Brooks, "Calculation of forces
and stresses on generator end-windings.-part I: forces," IEEE Trans.
In this paper, a three-dimensional electromagnetic model of Energy Conversion, vol. 4, no. 4, pp. 661-670, 1989.
the end winding of a 600MW turogenerator is set up, and the [9] X. Wen, R. Yao and J. A. Tegopoulos,"Calculation of forces on the
electromagnetic force densities of the stator end winding stator end windings of turbogenerator by the tranisent quasi-3D
method," IEEE Trans. Magnetic, vol. 32, no. 3, pp. 1669-1672, 1996.
under rated load and three-phase short circuit from full load [10] Y. Liu and S. Hjärne, "Analysis of forces on coils ends of formed stator
are numerically investigated. The dynamic responses of the windings," Proceeding of international conference on electrical
stator end winding under electromagnetic forces are carried machines and systems, Seoul, Korea, Oct. 2007, pp. 1019-1024.
out by means of the ABAQUS software. Finally, the [11] N. Richard, F. Duffeau, A. C. Léger and N. Szylowicz, "Computation of
forces and stresses on generator end windings using a 3D finite element
influences of structural and material parameters on the method," IEEE Trans. Magnetics, vol. 32, no. 3, pp. 1689-1692, 1996.
dynamic response are studied and an optimal scheme is [12] N. Richard, "Calculation of electromagnetic forces on large generator
proposed based on the dynamic characteristics and end-windings under fault conditions using a three-dimensional finite
deformation of the end winding under rated load. It is element method," Mathematics and Computers in Simulation, vol. 46,
pp. 257-263, 1998.
concluded as follows. [13] K. Senske, S. Kulig, J. Kauhoff and D. Wünsch, "Vibrational behaviour
(1) The distribution and magnitude of electromagnetic force of the turbogenerator stator end winding in case of electrical failures,"
density of the stator end winding change with time, and the Conférence Internationale des Grands Rèseaux Electriques, ,
magnitude of the electromagnetic force density decreases with Yokohama, Oct. 1997, pp. 1-12.
[14] O. Drubel, S. Kulig and K. Senske, "End winding deformations in
the distance from the stator core at a moment. different turbo generators during 3-phase short circuit and full load
(2) Both of the displacements and stresses of the stator end operation," Electrical Engineering, no. 82, pp. 145-152, 2000.
winding under the rated load are small, and the maximum [15] A. Grüning and S. Kulig, "Electromagnetic forces and mechanical
stresses of all component are much less than the ultimate oscillations of the stator end winding of turbogenerators," Recent
Developments of Electrical Drives, Springer Netherlands, 2006, pp.
strength of the material. But in the case of three-phase short 115-126.
circuit from full load, both the displacements and stresses are [16] K. C. Kim, H. W. Lee, Y. D. Chun and J. Lee, "Analysis of
at least dozens of times those in the case of rated load. electromagnetic force distribution on end winding for motor reliance,"
(3) Increase of stiffness of rings and radial braces will IEEE Trans. Magnetic, vol. 41, no. 10, pp. 4072-4074, 2005.
[17] R. Lin, A. N. Laiho, A. Haavisto and A. Arkkio, "End-winding
obviously decrease the deformation of the end winding. vibrations caused by steady-state magnetic forces in an induction
Variation of the number of spacer blocks and binding form machine," IEEE Trans. Magnetics, vol. 46, no. 7, pp. 2665-2674, 2010.
slightly affect the dynamic response of stator end winding. [18] R. Albanese, F. Calvano, G. DalMut and F. Ferraioli,
Bonding connections between the L-shaped brackets and the "Electromechanical analysis of end windings in turbogenerators,"
COMPEL: The Interactional Journal for Computation and
bolts as well as the pins and the copper brackets will Mathematics in Electrical and Electronic Engineering, vol. 30, no. 6,
apparently decrease the deformation of the end winding. pp. 1885-1898, 2011.
Bonding 18 pairs of pins radial braces may be another optimal [19] R. Albanese, F. Calvano, G. DalMut, F. Ferraioli, et al. "Coupled three
scheme. dimensional numerical calculation of forces and stresses on the end
winding of large turbo generators via integral formulation," IEEE Trans.
(4) Increasing the stiffness of rings can be chosen as an Magnetics, vol. 48, no. 2, pp. 875-878, 2012.
optimal scheme for design of the stator end winding, although [20] Y. Zhao, B. Yan, C. Chen, J. Deng and Q. Zhou, "Parametric study on
there may be other choices. dynamic characteristics of turbogenerator stator end winding," IEEE
(5) Generally, increasing the stiffness of support structure Trans. Energy Conversion, vol. 29, no. 1, pp. 129-137, 2014.
[21] F. Li, S. Jiang, Z. Wang and C. Ji, "Experimental analysis of dynamic
and changing the connection stiffness among the axial elastic characteristics of the "rigi-flex" type stator endwinding of a large
components properly will improve the performance of the end turbogenerator," Proceedings of the CSEE, vol. 12, no. 5, pp. 16-25,
winding of a large turbogenerator. 1992.
[22] X. Yao, "Characteristic analysis of the forces acted on stator end
windings of turbo-generator," North China Electric Power University,
VI. REFERENCES 2013.
[1] S. Exnowski, "Excitability of different modes of vibration of stator end
windings," IECON 2012-38th Annual Conference on IEEE Industrial
Electronics Society, Montreal, Canada, Oct. 2012, pp. 1781-1785.

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TEC.2016.2597151, IEEE
Transactions on Energy Conversion

10

Yang Zhao was born in Hebei, China, on


December 16, 1988. He received the B. Eng. and
Ph.D. in mechanics from Chongqing University,
China in 2010 and 2015, respectively. He is
currently a lecturer with the Chongqing University
of Posts and Telecommunications. His research
interests include engineering mechanics and
vibration control of large turbogenerators.

Bo Yan was born in Chongqing, China, on May 2,


1965. He received the B. S., M. Eng. and Ph.D. in
solid mechanics from Chongqing University, China,
in 1985, 1988 and 1999, respectively. He joined the
Department of Engineering Mechanics, Chongqing
University in 1988 and is a professor now. His
research interests include solid mechanics and
engineering mechanics. He is focusing on the
research work on dynamic response of transmission
line under wind action and after ice-shedding and
galloping of overhead transmission lines currently.

Chong Zeng was born in Chongqing, China, on


November 2, 1990. He received the B.Eng. and M.
Eng. in electrical engineering from Chongqing
University, China in 2013 and 2015, respectively.
He is currently a graduate student for Ph.D. degree.
His research interests include analysis of
electromagnetic field in electrical machines and fault
diagnosis of electrical motors.

Song huang was born in Sichuan, China, on July,


1972. He received the B.Eng., M.Eng. and Ph.D. in
electrical engineering from Chongqing University,
Chongqing, China in 1994, 2000 and 2005,
respectively. He joined the Department of Electrical
Engineering, Chongqing University in 1994 and is
currently an associate professor. His research
interests include calculation and analysis of
electromagnetic field in electrical machines,
electromagnetic design of electric machines and
motor control.

Changlin Chen was born in Chongqing, China, on


March 21, 1964. He received the B. Eng. in
mechanics from Lanzhou University, China in 1985.
He joined the Research & Testing Center of
Dongfang Electric Machinery Co. Ltd. and is a
professor of engineering now. His research interests
include mechanical strength, vibration theory,
design and development of large electrical machine.

Jianan Deng was born in Chongqing, China, in


May, 1965. He received the B. Eng. in mechanics
from Lanzhou University, China in 1987. He joined
the Research & Testing Center of Dongfang Electric
Machinery Co. Ltd. and is a professor of
engineering now. His research interests include
stiffness, strength, dynamic characteristics and fault
diagnosis of large electrical machine.

0885-8969 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy