# chen2015 .pdf

Nom original:

**chen2015.pdf**Titre:

**untitled**

Ce document au format PDF 1.7 a été généré par 'Certified by IEEE PDFeXpress at 05/28/2015 1:28:55 AM' / Acrobat Distiller 8.1.0 (Windows), et a été envoyé sur fichier-pdf.fr le 05/02/2018 à 01:58, depuis l'adresse IP 41.102.x.x.
La présente page de téléchargement du fichier a été vue 236 fois.

Taille du document: 1.2 Mo (4 pages).

Confidentialité: fichier public

### Aperçu du document

ISBN: 978-1-4799-8903-4 2015 IEEE 11th International Conference on the Properties and Applications of Dielectric Materials (ICPADM)

FEM Simulation and Analysis on Stator Winding

Inter-turn Fault in DFIG

Yu Chen*, Lulu Wang,

Zihao Wang, Attiq Ur Rehman,

Yonghong Cheng

State Key Laboratory of Electrical

Insulation and Power Equipment,

Xi’an Jiaotong University

Xi’an, Shaanxi, China

*E-mail: chenyu@mail.xjtu.edu.cn

Toshikatsu Tanaka

Yong Zhao

IPS Research Center

Waseda University

Fukuoka, Japan

Xi’an TPRI Co., Ltd

Xi’an, Shaanxi, China

Abstract—With the development of renewable energy

technology, Doubly-Fed Induction Generators (DFIG) are widely

used on wind farms. Faults of machines have a direct influence

on the safety and effectiveness of the power grid. In order to

explore failure mechanism, it is of great importance to study the

DFIG fault simulation models and features. In this paper, a

Finite Element Method (FEM) simulation model of stator

winding inter-turn fault in DFIG is presented by the software of

ANSYS Maxwell. Different stator inter-turn short circuit fault

conditions are achieved by changing the stator winding

connection and the values of end leakage-inductance and

resistance. Under normal condition, different branches’ currents

are the same in each phase and different phase currents are

balanced. So the Total Harmonic Distortions (THD) of threephase stator currents are low and phase differences of stator

currents are all about 120°. Additionally, Park’s vector trajectory

is a circle and its eccentricity is approximately zero. However, as

the fault degree enlarges, both the current of faulty-branch and

the eccentricity of Park’s vector trajectory increase. The phase

differences related to faulty-phase and THD of faulty-phase

become larger. Therefore, THD, phase differences and the

eccentricity can be regarded as three fault features to diagnose

fault phase and fault degree.

changing the number of shorted turns and the value of shortcircuit resistance. However, this simulation method cannot

reflect the real condition of fault. Because when an inter-turn

short circuit fault occurs, a new loop emerges between the

faulty turns and the loop current will be much larger than phase

currents. And then temperature rises in the faulty winding,

affecting the insulation materials near the fault location. It may

lead to more turns shorted and even phase to phase or phase to

ground short circuit fault. So the new loop cannot be ignored.

Keywords—DFIG; stator winding inter-turn fault; FEM

simulation; fault features

FEM is a numerical calculation method based on

variational principle. It transforms a variational problem into an

extremum problem of multivariate function by the subdivision,

interpolation and discretization of field.

In this paper, a DFIG model with stator inter-turn short

circuit fault is built in Maxwell 2D. Based on the finite element

method, different shorted turns are simulated by changing the

phase winding connection and the values of end leakageinductance and resistance. The detail procedure is discussed for

inter-turn short circuit fault in the same branch and we analyze

the simulation results to validate this model. The faulty-branch

current, phase differences related to faulty-phase, THD of

faulty-phase and the eccentricity of Park’s vector trajectory of

stator currents are extracted as features under the premise to

ensure that no other damages to the machine.

II.

I.

INTRODUCTION

DFIG is widely used in the wind energy industry. It is an

important system component and is often exposed to complex

environment conditions. According to the survey, 40% of

DFIG failures are related to bearings, while 38% to the stator

and 10% to the rotor [1]. Faults would cause such features as: 1)

unbalanced air-gap voltages and line currents; 2) increased

torque pulsation; 3) excessive heating; 4) disturbances in the

current, voltage or flux waveform [2]. In order to find features

to indicate different kinds of faults in DFIG, researchers have

done a lot of work in modeling, simulating and experimenting.

In general, there are two main approaches to model DFIG

faults. From the view of “circuit” analysis, simulation model

can be built in MATLAB/Simulink according to the Multi-loop

Theory [3] and machine mathematical equations [4]. Another

method is from the view of “electromagnetic field” analysis

called FEM. It uses the electro-magnetic software

ANSYS/Maxwell 2D to build the DFIG simulation model.

Models presented in [5-8] for a stator winding inter-turn short

circuit fault are based on Maxwell 2D and achieved by

FEM ANALYSIS OF ELECTRICAL MACHINES

A whole machine is considered as the calculation field for

the aim of analyzing the stator winding inter-turn fault. With

the vector magnetic potential being the solving variable, a twodimensional electromagnetic field model for electrical

machines can be established by the vector magnetic potential

Az. Therefore the expression of electrical machines in

electromagnetic field is shown in equation (1).

⎧∂

⎛

⎪ ⎛⎜ μ ∂AZ ⎞⎟ + ∂ ⎜ μ ∂AZ

⎪ ∂x ⎝ ∂x ⎠ ∂y ⎝ ∂y

⎪

⎨ AZ Γ1 = AZ 0

⎪

⎪ 1 ∂AZ

⎪ μ ∂n = − H t

Γ2

⎩

⎞

∂AZ

∂A

+ vxσ Z

⎟ = −JZ + σ

t

∂

∂x

⎠

（1 ）

In (1), vx is the x component of the rotation speed; JZ is the

z component of the source current; μ is the permeability; σ is

978-1-4799-8903-4/15/$31.00 ©2015 IEEE

244

ISBN: 978-1-4799-8903-4 2015 IEEE 11th International Conference on the Properties and Applications of Dielectric Materials (ICPADM)

TABLE I.

the conductivity; AZ0 is the given value of AZ on the boundary

of Γ1 ; Γ2 is the Neumann Boundary Condition.

III.

Number of poles

Number of slots

Circuit type

Outer Diameter

520/350 mm

Inner Diameter

The number of

turns per phase

Rated output

power

Rated speed

346.4/110 mm

Parameter

MODELING OF DFIG IN ANSYS/MAXWELL

A. Machine Parameters

A 110 kW three-phase doubly-fed induction generator is

used for simulation. The parameters of this machine are given

in TABLE I.

B. Modeling Method

In general, there are three ways to draw Doubly-Fed

Induction Machine geometry model. The simplest and the

most accurate method is to model in ANSYS/RMxprt, then

import it to Maxwell 2D. In order to facilitate the fault set, a

complete model should be imported from RMxprt to Maxwell

2D automatically, seen in Fig.1.

MACHINE OPERATION PARAMETERS

Value

(Stator/Rotor)

4/4

72/60

△/Y

228/100

Parameter

Length

Winding layers

Parallel branches

Conductors per

slot

Coil pitch

Rated power

factor

Value

(Stator/Rotor)

290 mm

2/2

4/1

19/10

15/13

1

110 kW

Rated voltage

380 V

1800 rpm

Frequency

50 Hzh

A zero Vector Potential is added to the outer stator surface.

Three-phase AC current source is given to rotor windings,

while the stator winding excitations are set in external circuit

which has taken stator end leakage-inductances into

consideration.

C. Stator Winding Inter-turn Short Circuit Model

Due to four branches paralleled in per phase of stator

winding, the stator winding inter-turn short circuit fault

consists of two types: 1) inter-turn short circuit in the same

branch; 2) inter-turn short circuit in two different branches.

For the inter-turn short circuit fault in the same branch

(taking phase A for example, phase A consists of four

branches). If the fault is located in the third slot, then this

branch can be divided into one shorted winding (A1_sc) and

two non-shorted windings (A1_1 and A1_3), which can be

seen in Fig.2. The geometry model for stator inter-turn fault in

phase A is seen in Fig.3. The method of field-circuit coupling

is chosen to model the inter-turn short circuit, so the external

circuit should be introduced which is shown in Fig.4.

Additionally, in order to consider the machine end-effect, the

terminal resistance R and end leakage-inductance L for

different shorted turns are calculated and used.

Fig. 2.

The schematic diagram of phase A on stator winding

Fig. 3.

Geometry model for stator inter-turn fault in phase A

IV.

Similar method can be used for modeling inter-turn fault in

different branches. For the inter-turn short circuit in two

different branches, both of them should be divided and the

settings should be made in external circuit.

SIMULATION AND RESULTS ANALYSIS

A. Simulation of Machine under Different Conditions

Machine under healthy condition is simulated in ANSYS

Maxwell at the speed of 1800 rpm. Three phase currents on

stator winding are shown in Fig.5.

For a 7-turn short circuit fault in branch A1, the faultybranch current IA1_1 and the new loop current IA1_sc are starting

to increase from the short-circuit moment. Before t = 0.8 s,

IA1_1 and IA1_sc are equal because their windings are in a series.

At t = 0.8 s, the winding A1_sc is shorted and a large current

is induced in the new loop. The equivalent impedance of

branch A1 decreases and therefore IA1_1 increases. The

waveform of four branches (A1_1 、 A2 、 A3 、 A4) and the

new loop current (A1_sc) under 7-turn short circuit fault is

shown in Fig.6.

Fig. 1.

In order to study the relationship between fault features

and fault degrees, more simulations for different shorted turns

are required. So ten different conditions are simulated and the

simulation results are analyzed below.

Geometry model for normal condition

245

ISBN: 978-1-4799-8903-4 2015 IEEE 11th International Conference on the Properties and Applications of Dielectric Materials (ICPADM)

Fig. 4.

The external circuit for inter-turn short circuit fault

B. Analysis of Simulations Results

1) Analysis of branch currents

In the same slot of branch A1, ten simulations have been

carried out under healthy condition (0-turn short circuit) and

1~9 turns short-circuit fault. Simulation results of branch

current IA1_1 under healthy and fault conditions (3-turn, 6-turn,

9-turn short-circuit fault) are presented in Fig.7. These results

corroborate that the magnitude of faulty-branch becomes

bigger with the increase of fault degree.

Fig. 5.

2) Analysis of three phase currents

In a DFIG online monitoring system, three phase currents

on stator winding are generally collected. And therefore the

analysis of three-phase current is necessary. The phase

differences and THD [9] of stator currents under different

stator inter-turn fault conditions are presented in TABLE II.

For phase A inter-turn short circuit fault, φAB and φCA deviate

120° more and more from 1 to 9 turns shorted, while φBC has

almost no change (about 120°). THD of phase A is bigger

and bigger with the increase of the number of shorted turns,

while THD of phase B and C have little change. In other

words, the phase differences related to faulty-phase deviate

from 120 ° and more deviation means more serious fault.

THD of faulty-phase is bigger than the others. The bigger of

the THD, the more serious the fault is. So it has been

suggested that the fault phase is related to both the phase

difference and THD, which are regarded as two features for

fault degree and fault phase.

Stator three phase currents under healthy condition

Fig. 6.

Branch currents of phase A under 7-turn short circuit fault

condition

According to the principle of magnetic potential balance,

three phase currents can be transformed from three-phase axis

coordinate to two-phase axis. Therefore, the vector I = iα + jiβ

is named as Park’s Vector. Ideally, the Park’s vector trajectory

is a circle centered on the origin. For stator winding inter-turn

fault, it turns out to be an ellipse. The Park’s vector

trajectories of different conditions are shown in Fig.8.

Additionally, the length of semi-major axis, the length of

semi-minor axis and the eccentricity of ellipse for different

conditions are shown in TABLE II. As fault degree enlarges,

the length of semi-major axis lengthens and the length of

semi-minor axis shows in an opposite trend. However, they

are not appropriate to be features of fault. Given eccentricity is

the comprehensive index of them and it increases with the

fault degree, we appreciate it a proper index for fault feature.

Fig. 7.

Current of branch A1 under different stator winding fault

conditions

246

ISBN: 978-1-4799-8903-4 2015 IEEE 11th International Conference on the Properties and Applications of Dielectric Materials (ICPADM)

TABLE II.

The number of

shorted turns

FAULT FEATURES FOR STATOR WINDING FAULT UNDER DIFFERENT CONDITIONS

Phase difference

THD

Parks’ vector trajectory

AB

BC

CA

A

B

C

Length of semimajor axis

Length of semiminor axis

Eccentricity

0

120.011

120.063

119.926

3.39%

3.54%

3.22%

141.835

141.825

0.012

1

122.747

120.071

117.182

3.62%

3.57%

3.22%

143.404

138.809

0.251

2

125.536

120.061

114.404

3.92%

3.54%

3.17%

144.924

135.855

0.348

3

128.357

120.039

111.604

4.39%

3.57%

3.19%

146.517

133.128

0.418

4

131.233

120.017

108.749

4.95%

3.57%

3.21%

148.073

130.495

0.473

5

134.147

119.993

105.860

5.59%

3.55%

3.24%

149.633

127.978

0.518

6

137.102

119.962

102.937

6.29%

3.52%

3.28%

151.194

125.566

0.557

7

140.097

119.925

99.978

7.02%

3.49%

3.33%

152.759

123.253

0.591

8

143.108

119.854

97.038

7.94%

3.48%

3.39%

154.253

121.072

0.620

9

146.127

119.769

94.104

8.75%

3.48%

3.45%

155.810

118.931

0.646

0.646). Therefore, they can be used as fault features to provide

references for fault diagnosis.

(a) Healthy condition

However, this paper only discusses and simulates the interturn short circuit in the same slot of one branch. Although

different numbers of shorted turns are simulated, it cannot

include all the conditions of inter-turn fault. In addition, other

kinds of fault may also induce thus features, so inter-turn short

circuit in different slots and rotor winding inter-turn fault

should be simulated and analyzed in the near feature.

(b) 3-turn short-circuit

ACKNOWLEDGMENT

This work is supported by the Headquarters Science and

Technology Projects of China Huaneng Group, under project

number “HNKJ13-H20-05”.

REFERENCES

(c) 6-turn short-circuit

[1]

(d) 9-turn short-circuit

Fig. 8.

Park’s vector trajectory under different stator winding fault

conditions

[2]

V. CONCLUSIONS

This paper discusses how to build a FEM model for DFIG

with stator winding inter-turn short circuit fault in detail. The

simulation model is implemented in ANSYS Maxwell. Ten

simulations of different conditions have been carried out by

changing the phase winding connection and the value of end

leakage-inductance and resistance in external circuit. Branch

currents of faulty-phase and three phase currents on stator

winding are analyzed. Some features are calculated, i.e.,

branch currents, phase currents, phase differences, THD of

three stator phase currents and the eccentricity of Park’s vector

trajectory of stator currents.

[3]

[4]

[5]

[6]

[7]

The simulation results show that the increase of fault

degree from healthy condition to 9 turns shorted causes the

increasingly growth of the faulty-branch current, the phase

differences related to faulty-phase (φAB: from 120.011 to

146.127, φCA: from 119.926 to 94.104), the THD of faultyphase (THDA: from 3.39% to 8.75%), and the eccentricity of

Park’s vector trajectory of stator currents (from 0.012 to

[8]

[9]

247

N. Goudarzi, D. Zhu W, “A review on the development of wind turbine

generators across the world”, International Journal of Dynamics and

Control, vol. 1, pp. 192-202, 2013.

L.M. Popa, B. Jensen, E. Ritchie, “Condition monitoring of wind

generators”, Industry Applications Conference, Ias Annual Meeting,

Conference Record of the. IEEE, vol.3, pp.1839-1846, 2003.

J.Q. Li, Y.Z Ren, D. Wang, “Multi-loop Mathematical Model and

Calculation of Inductances in Doubly Fed Induction Generation”, Large

Electric Machine and Hydraulic Turbine, vol. 6, pp. 5-10, 2013.

L.L. Wang, Y. Zhao, W. Jia, B. Han, Y. Chen, “Fault diagnosis based on

current signature analysis for stator winding of Doubly Fed Induction

Generator in wind turbine”, Electrical Insulating Materials (ISEIM),

Proceedings of 2014 International Conference on, pp. 233 – 236. 2014.

S.Y. Wei, Y. Fu, H.Z. Ma. “Stator winding inter-turn short-circuit

diagnosis and experimental research on doubly-fed induction generator”,

Power System Protection and Control, vol. 38, pp. 25-28, 2010.

J.Q. Li, M. Li, D.Y. Wang. “Influence of stator turn-to-turn short-circuit

on magnetic field of DFIG”, Electrical Machines and Systems (ICEMS),

2011 International Conference on, pp.1 – 5, IEEE, 2011.

S. Seman, Transient performance analysis of wind-power induction

generators. Helsinki University of Technology, 2006.

M. Hacene, K.E. Hemsas, “Design and simulation analysis of outer

stator inner rotor DFIG by 2D and 3D finite element methods”,

International Journal of Electrical Engineering & Technology, vol. 2,

2012.

L. Fan, S. Yuvarajan, R. Kavasseri, “Harmonic analysis of a dfig for a

wind energy conversion system”, IEEE Transactions on Energy

Conversion, vol. 25(1), pp. 181 – 190, 2010.