675

Incipient Fault Detection and Diagnosis of Induction Motor using Fuzzy Logic
Mini.V.P
Sr. Lecturer Department of Electrical Engineering College of Engineering Trivandrum Kerala, India

Dr.S.Ushakumari
Asst. Professor Department of Electrical Engineering College of Engineering Trivandrum Kerala, India

Abstract: Induction motors are critical components in many industrial processes. Online monitoring of induction motors is becoming increasingly important. The main aim of this paper is to perform the electrical fault analysis of a three phase induction motor using simulation in MATLAB SIMULINK and to design a fuzzy logic algorithm for detecting and analyzing the electrical fault of the motor. Several types of electrical faults such as under voltage fault, unbalance fault, overload fault and earth faults are experienced in three phase induction motor. This paper presents a systematic approach to detect and analyze the electrical faults of an induction motor. The work is explained in three phases. The first phase explains the causes and after effects of different types of electrical faults in an induction motor. The mathematical model and simulation of induction motor with different types of electrical faults are explained in second phase .The design and simulation of fuzzy logic algorithm for identification and analyzing electrical faults are discussed in third part. In this work, fuzzy logic is used to make decisions about the motor condition with high accuracy. Key words: Induction Motor, Electrical Faults, Modeling, Simulation, Diagnosis, Fuzzy logic, Stator current amplitude.

faults such as unbalanced supply voltage, over load, under voltage and earth faults are explained in this paper. The fuzzy logic algorithm for the identification and analysis of electrical faults are also included in this paper. The mathematical model of three phase induction motor used for the fault analysis is developed and simulated using MATLAB SIMULINK toolbox .In fact, fuzzy logic is reminiscent of human thinking process and natural language enabling decisions to be made based on the vague information. In fuzzy logic, the fault condition of the motor are described using linguistic variables. Fuzzy subsets and corresponding membership functions describe the stator current amplitudes, negative sequence components of stator currents and the speed. A knowledge base comprising rules and data- base is built to support the fuzzy inference. The induction motor conditions are diagnosed using a compositional rule of fuzzy inference. The aim of this paper is to present a useful and straight forward method to simulate electrical faults such as under voltage fault, over load fault, unbalanced fault and earth faults. A fuzzy logic algorithm is proposed for identifying and analyzing the faults.
II. INFLUENCE OF ELECTRICAL FAULTS ON MOTOR PERFORMANCE

I. INTRODUCTION Three phase induction motors are commonly used in industrial applications due to their simple construction, high reliability, low cost etc. Early detection of abnormalities will helps avoiding the malfunction of the machine. So, condition monitoring of three phase induction machine can significantly reduce the maintenance cost and its performance is also improved [1-3]. The induction motor performance is affected by three types of faults such as electrical faults, mechanical faults and environmentally related faults. Electrically related faults consist of unbalanced faults, under voltage faults, overload faults and earth faults. The rotor winding failure, stator winding failure and bearing faults are included in mechanically related faults. The external moisture, contamination in the ambient temperature also affects the induction motor performance. The faults due to these external variation and vibration faults are included in the environmentally related faults. Electrical faults in three phase induction motor will produce more heat on both stator winding and rotor winding. Due to the electrical faults, life time of the motor will reduce. The behavior of induction motor under electrically related

Electrically related faults are frequently occurring in three phase induction motors. Because of these electrical faults more heat will produced on stator winding and rotor winding, this will reduce the life time of the machine. In this subsection causes and effects of different types of electrical faults such as unbalanced supply voltage fault, under voltage fault, over voltage fault, over load fault and earth faults are discussed [46].
Causes and effects of electrical faults

Open delta transformers, unbalanced loading, unequal tap settings of the transformer and shunted single phase load etc, are the causes of unbalanced voltage faults. The effects of unbalanced supply voltage faults are, reduction in motor efficiency, increase in stator and rotor copper losses, temperature rise and serious reduction in starting torque. The

978-1-4244-9477-4/11/$26.00 2011 IEEE

676

efficiency reduction in induction motor due to the unbalanced supply voltage faults results in a higher electricity bill for the same amount of work done. The under voltage fault occurs when a reduced supply voltage with a rated mechanical load on the motor. The increased stator current, excess heating of machine and increase in the stator and rotor copper losses are the after effects of under voltage faults. When any one of the line voltage is greater than 110% of the rated value, the over voltage fault will occur, it will produce harmful effects to the machine insulation. The mechanical load on the motor is to increase beyond the rated value at that time over load fault will occur. Due to the high load torque motor begins to draw more current. The effects of over load faults are increase in stator current and overheating of the machine. Ground faults are more prevalent in motors than other power system devices. The ground fault is identified and analyzed by measuring leakage currents . The effects of earth faults are thermal stress due to fault current and hazards of human safety.
III INDUCTION MOTOR MODEL

Normally, an induction machine is connected to a threephase supply by a three-wire connection. The stator and rotor flux linkages in (4) may be expressed compactly as
ss s L11 q s ss d = L 21 sr r L11 q r sr d L 21 ss L12 sr L11 sr L12 L sr 22 0 L rr 22

L ss 22 sr L12 L sr 22

L sr 21 rr L11 0

i qs s i d r i q r i d

(5)

A. Determination of Inductances

In order to define symmetrical machine inductances, assume that the stator phases as, bs and cs have equal number of winding turns given by Ns, and that the rotor phases ar, br and cr also have equal number of winding turns given by Nr. The stator self-inductances for phases as, bs and cs can be calculated as

2 L asas = L bsbs = L csc s = L ls + L m 3

(6 )

The stator mutual inductances between phases as and bs, bs and cs, and cs and as can be derived as

The model of a symmetrical three- phase induction motor is well known. To derive equations for symmetrical stator winding, and rotor, the following assumptions are made:[1, 711] Each stator phase of the motor has same number of turns, and uniform spatial displacement . Magnetic saturation is not present. With the appropriate subscripts as, bs, cs, ar, br and cr, the voltage equations of the magnetically coupled stator and rotor circuits can be written as :
s s r r v s = rabc i abc + p s , 0 = rabc i abc + p r (1) abc abc abc where p=d/dt. Applying a stationary reference frame transformation to this equation yields the corresponding qd0 equations and (1) becomes

1 L asbs = L bscs = L csas = (Lm ) 3

(7)

The rotor self and mutual inductances can be found by a similar way. As the rotor is also symmetric, the total selfinductances of rotor phases ar, br and cr are equal. Therefore

Larar =Lbrbr =Lcrcr=Llr +

2 Nr Lm =Llr +Lmar 3 Ns2

(8)

where L = 2 N r L . For the same reason, rotor mutual mar m 3 N s2 inductances are also equal to each other and given by
2 1 2 Nr Larbr = Lbrcr = Lcrar = L 2 m 2 3 Ns

s s v s = rqdo i qdo + p s , qdo qdo

(9)

0 1 0 r r rqdo i qdo r 1 0 0 r + p r = 0 ( 2) qdo qdo 0 0 0 In matrix notation, the flux linkages of the stator and rotor windings may be written in terms of the winding inductances and the current as s r s = Lss iabc + Lsr iabc abc abc abc
(3) The stator and rotor qd0 flux linkages are obtained by applying transformation to the stator and rotor abc flux linkages (3), that is s r s = L ss i qdo + L sr i qdo qdo qdo qdo
s r r = Lrs iabc + Lrr iabc abc abc abc

Because of the rotor and stator symmetry the stator to rotor mutual inductances are equal ie Lasar = Lasbr = Lascr,= Lbsar = Lbsbr = Lbscr = Lcsar = Lcsbr = Lcscr. The rotor phase mutual inductances will be 2 Nr L asar = L asbr = L ascr = (10 ) 3 N Lm s where Lmsr = 2 N r Lm . 3 N s2
B. Simulation of the symmetrical Induction Motor

r qdo

= L

rs qdo

s qdo

+ L

rr qdo

r qdo

(4)

In this subsection, equations are rearranged for the symmetrical induction motor model that has been developed

677

[12-14]. Flux linkages obtained from (2) for a three- wire system is

A. Fuzzy System Input-Output Variables

s q

s d r q r d
Va Vb Vc

(v r i r i ) dt = (v r i r i ) dt = ( r i ) dt = ( + r i ) dt
=
s q s s 11 q s 12 s d s d s s 21 q s 22 s d r r d r r r q r r q r r r d

As stated, the induction motor condition can be deduced by observing the stator current amplitudes and speed are input variables. Interpretation of result is difficult as relationships between the motor condition and the current amplitudes are vague. Therefore, using fuzzy logic, numerical data are represented as linguistic information. In this case, the stator current amplitudes Ia, Ib, and Ic,, negative sequence current I neg and speed N are considered as the input variables to the fuzzy system. The motor condition, MC, is chosen as the output variable. All the system inputs and outputs are defined using fuzzy set theory.
B. Linguistic Variables

(11)

abc to dq

Motor model

dq to abc

Basic tools of fuzzy logic are linguistic variables. Their values are words or sentences in a natural or artificial language, providing a means of systematic manipulation of vague and imprecise concepts.
detection algorithm

Ns Nr

Calc ulatio n of ind. & reis.

For instance, the term set T(MC), interpreting motor condition, MC, as a linguistic variable, could be T(MC )= {NR, AEF, BEF, CEF ABEF ,BCEF, ACEF ,ABCEF, MOL, HOL, MUB, HUB, VHUB, MUVF, HUVF } Similarly, the input variables stator current amplitudes Ia, Ib, and Ic , Ineg and speed N are interpreted as linguistic variables, with T(Q)= {Normal (NR) Slightly high (SLH) medium (M), positive medium (PM), High (H) Very high (VH)}, T(I neg )= {Very very small, very small, small, medium, large, very large}, T(N )= {Very very low, very low, low, medium, positive medium , Normal} where Q= ia, ib, ic respectively
C. Fuzzy Membership Functions Construction

condition

Fig. 1. Block Diagram of fault detection system with induction motor model

The stator current is obtained by inverting (5) and the speed of the machine is obtained from the torque equation as P (T em T mech T damp ) dt r (t ) = (12 ) 2J Tem is the electromagnetic torque impressed on the shaft of the machine and can be expressed as

T em

3 P s s d i q s i ds q 2 2

(13 )

Tmech is the externally applied mechanical torque , Tdamp is the damping torque and J is inertia. Using (11)(13) with resistances and inductances, a motor with symmetrical windings can be simulated. The block diagram representation of the model for the simulation of the fault detection system for symmetrical induction motor is shown in Fig.1.
IV. FUZZY LOGIC BASED DIAGNOSIS APPROACH

Fuzzy rules and membership functions are constructed by observing the data set. For the measurements related to the stator currents and speed, more insight into the data is needed, so membership functions will be generated for all input and output variables. The optimized input membership functions for this problem are shown in Fig.2. The output fuzzy linguistic variables and their membership function value ranges are given in Appendix B for different types of motor condition.
NR SLH M PM H VH

. Fuzzy systems rely on a set of rules. These rules, while superficially similar, allow the input to be fuzzy, i.e. more like the natural way that humans express knowledge. Thus, a power engineer might refer to an electrical machine as somewhat secure or a little overloaded. This linguistic input can be expressed directly by a fuzzy system. Therefore, the natural format greatly eases the interface between the engineer knowledge and the domain expert [15-16].
Fig.2. Input fuzzy sets for ia , ib &ic

678

VI. SIMULATION RESULTS

T o rq u e (N m ),S p e e d (ra d /s e c )

The mathematical model of induction motor is developed and simulated by MATLAB SIMULINK tool box. The simulation results are given below for a 2 hp motor [1] with parameters given in Appendix A. The results are taken during acceleration from standstill to full speed. Fig.3 shows the speed and torque variation of symmetrical machine under normal conditions. In this condition, the starting torque is about 80Nm and the speed is about 155 rad/sec and the motor will maintain the constant speed within 0.5seconds. Fig .4, shows the amplitude of stator current and negative sequence current of the motor under normal working condition.
200 Speed Torque T o rq u e (N m ), S p e e d (ra d /s e c ) 150

Fig.5 shows the torque and speed response of the machine under unbalanced fault condition(Va = 230V, Vb= 190V and Vc = 160V). Due to unbalance in supply voltage the motor will take more time (about 3.5sec) to attain the steady state speed and the starting torque gets reduced and also the torque response consists ripples.
160 140 120 100 80 60 40 20 0 Speed Torque

100

-20

0.5

1.5

2.5 Time(Sec)

3.5

4.5

50

Fig. 5. Speed and Torque under unbalanced fault condition

-50

0.5

1.5

2.5 Time(Sec)

3.5

4.5

Fig. 3. Speed and Torque Variation under normal condition

Fig.6. shows the stator phase currents ia, ib and ic under unbalanced fault condition. Due to this unbalanced supply voltage fault the currents are in unbalanced condition, Because of these unbalanced full load current and low starting torque, the premature failure of motor winding and nuisance of over load tripping will occur.

30 ic ib ia

25

18 16 14 12 C u rren t (A ) 10 8 6 4 2 Ia Ib Ic

20 C urrent (A )

15

10

0.5

1.5

2.5 Time(Sec)

3.5

4.5

0.03

0
0.025

0.5

1.5

2.5 Time(Sec)

3.5

4.5

0.02 Ineg(A )

Fig. 6. Stator currents i a , ib and i c under unbalanced fault condition

0.015

0.01

0.005

0.5

1.5

2.5 Time(Sec)

3.5

4.5

Fig.4. Stator and negative sequence currents under normal condition

Fig.7.and Fig.8 show the torque speed characteristics and stator current waveforms of the induction motor in under voltage fault condition. The supply maximum voltage is reduced to 60% of rated voltage on all three phases for simulating the under voltage fault. During under voltage fault the stator current become highly increasing, due to these high current more heat is produced and it will take very large time to attain the steady state speed.

679

160 140 T o rq u e (N m ),S p e e d (ra d /s e c ) 120 Speed Torque

30 ia ib ic

25
100 80 60 40 20 0 -20

20 C u rre n t (A )
0 0.5 1 1.5 2 2.5 Time(Sec) 3 3.5 4 4.5 5

15

10

Fig. 7. Speed and Torque variation on under voltage fault condition.

0.5

1.5

2.5 Time(Sec)

3.5

4.5

Fig. 10. currents ia , ib and i c on single phase to ground fault condition

0.06 ineg 0.05

0.04 Ineg(A)
0 0.5 1 1.5 2 2.5 Time(Sec) 3 3.5 4 4.5 5

6 4

0.03

0.02
2 0

0.01

Fig. 8. currents ia , ib and i c on under voltage fault condition

0.5

1.5

2.5 Time(Sec)

3.5

4.5

The speed and torque characteristics of three phase induction motor under single phase to ground fault is shown fig.9.
200 Speed Torque

Fig. 11. currents I neg for single phase to ground fault condition

150 T orque(N m ), S peed (rad/s ec )

100

50

The over load fault is created by increasing the mechanical torque on the motor by changing the load torque value in the simulation environment. Fig.12 shows the amplitudes of phase currents under the overload fault condition. Due to the overload fault, the phase current is increased to more than 7 times of the rated current. Due to the high current, the motor will over heat.

0
30 ia ib ic

-50

25

-100

0.5

1.5

2.5 Time(Sec)

3.5

4.5

5
20 Current(A )

Fig. 9. Speed and Torque when single phase to ground fault occurs

15

Corresponding currents in three phases such as a, b and c and leakage current Ineg are shown in fig.10and fig.11 respectively. During normal running condition, at 1sec phase A is earthed and cleared at 2sec for simulating the single phase to ground fault. Due to single phase to ground fault more pulsation is occurred in speed and torque this will reduce the life time of the motor. Due to single phase to ground fault the unbalance is obtained in the stator currents, because of these unbalance in stator current turn to turn insulation failure and motor insulation failure will also occurred.

10

0.5

1.5 Time(Sec)

2.5

Fig. 12. Currents i a , ib and ic on over load fault condition

The stator current rms amplitudes, speed and leakage current amplitudes have been applied to the fuzzy logic fault detection algorithm and the corresponding fuzzy rule viewer

680

for different fault conditions of the motor is shown in Figs. 13-17. Fig.13. shows the result of fuzzy detection with certain output membership function value (0.875) for the healthy condition of the motor when all the stator currents and speed are in normal position and the negative sequence current is almost zero. Figs. 14, 15, and 16 show the fuzzy rule viewer for double phase to ground fault in an induction motor , unbalanced fault and under voltage fault of the motor respectively.

Fig. 15. Fuzzy inference diagram for unbalance fault.

Fig.13. Fuzzy inference diagram for healthy condition

The motor condition will be decided according to the output membership function value considered in the design of fuzzy logic detection algorithm. The output membership function value ranges are given in Appendix B for different types of motor fault condition.

Fig.16. Fuzzy inference diagram for under voltage fault.

Fig.17.shows the result of fuzzy detection with certain output membership function value (24) for the over load condition of the motor when all the stator currents are very high, speed is in very very low position and the negative sequence current is very small.

Fig.14. Fuzzy inference diagram for double phase to ground fault.

For example fig.14 shows the result of the fuzzy rule viewer with certain output member function value (16.4) for double phase to ground fault (ACEF). Fig.15. Shows the fuzzy rule viewer for certain output membership value (27.3) then the motor is in unbalanced fault condition (MUB). In fig. 17 the output is 32.9 then the motor is in under voltage fault condition (VHUVF).

Fig. 17. Fuzzy inference diagram for over load fault. VII. CONCLUSION AND FUTURE SCOPE

The symmetrical induction motor model has been adopted for analyzing different types of electrical faults. The model is based on general machine parameters so that it is not

681

necessary to know detailed geometry or physical layout of the windings. The electrical faults such as earth fault, unbalanced fault, under voltage faults and overload faults can be easily simulated by using the developed model. Fault severity can be analyzed by varying the supply voltage for under voltage and unbalanced fault conditions, The results from the simulation such as stator current rms amplitudes, leakage current and speed amplitudes have been used as inputs of the fuzzy fault detection algorithm to detect the type of fault. The earth fault also can be done easily with the model. Simulation results of stator current have been used as inputs to the fault detection algorithms to detect the condition of the motor. A method of using fuzzy logic to interpret stator current signal, speed and leakage current signals of induction motor for its electrical fault condition monitoring was presented. The fuzzy decision system achieved with high diagnosis accuracy. The work done can be extended to identify and analyze the rotor fault of an induction motor by fuzzy logic algorithm or by any other diagnostic techniques.
Appendix A: Motor Parameters Line Voltage Horse power Rated speed No. of Poles Total no. of turns per phase Stator winding resistance Stator leakage inductance Rotor leakage inductance Rotor resistance Magnetizing inductance = 400 V = 2 hp = 1500 rpm at 50 Hz =4 = 252 = 4.05 ohm = 13.97 mH = 13.97 mH = 2.6 ohm = 538.68 mH

[2]

W. T. Thomson, A review of on-line condition monitoring techniques for three-phase squirrel cage induction motors- past, present and future, Proceedings of the IEEE International Symposium on Diagnostics for Electrical Machines, Power Electronics and Drives, IEEE SDEMPED99, 1999, pp. 3-18. M. Arkan, Stator Fault Diagnosis in Induction Motors, University of Sussex, Ph.D. Thesis, 2000. M,Sudha & P.Anbalagan A Protection Scheme for three phase induction motor from incipient faults using embedded controller Asian Jounal of Scintific research 2(1):28-50,2009 ISSN 1992 -1454 Asian network for scientific information.

[3] [4]

[5] Kersting.W.H. Causes and effects of unbalanced voltage supplying on an induction motor I EEE Transactions on Industrial applications 2001. 137.PP 165 170 [6} Colak.I, H.Celik and S. Demirbas On line protection systems for Induction motors IEEE Transactions on Energy Conversions manage 2005 vol.46 . PP2773-2786 [7] M. Arkan, D. Kostic-Perovic, P. J. Unsworth, Online stator fault diagnosis in induction motors, IEE Proceedings on Electric Power Applications 148 (6) (2001) 537-547. M. Arkan, D. Kostic-Perovic, P.J . Unsworth, Modelling and simulation of induction motors with inter-turn faults for diagnostics,Electric Power System Research, Science Direct, May 2005, pp. 57-66. Prof.Vilas N. Ghate et.al. Generalized model of Three phase induction motor for fault analysis in IEEE REGION 8 SIBIRCON 2008

[8]

[9]

[10] Rangarajan .M. Tallam, et.al. Transient Model for Induction Machines With Stator Winding Turn Faults IEEE TRANSACTIONS ON INDUSTRY
APPLICATIONS, VOL 38, NO 3,pp 632- 637,

Appendix B: Fuzzy output membership function value ranges and corresponding motor condition
Output membership value range
-2 to 2 2.2 to 4.6 4.8 to 7.2 7.4 to 9.8 10 to 12.4 12.6 to 15.0 15.2o 17.6 17.8 20.2

[11] A. Consoli,T. A. Lipo,Orthogonal axis models for asymmetrically Connected induction machines, IEEE Transactions on PowerApparatus System PAS- 101(12) (1982) 12] P. Krause, O . Wasynczuk, S. D. Sudhoff, Analysis of Electric Machinery, IEEE Transactions on Industrial Electronics , 1995, ISBN 07803-1101-9. [13] G. M. Joksimovic, J. Penman, The Detection of Inter-Turn Short Circuits in the Stator Windings of Operating Motors, IEEE Transactions on Industrial Electronics, vol.45, no.5, October 2000 [14] C. M. Ong, Dynamic Simulation of Electric Hall PTR, 1998, ISBN013-723785-5 Machinery, Prentice

Fault condition
Normal AEF(Earth fault at A Phase) BEF(Earth fault at B Phase) CEF(Earth fault at C Phase) ABEF(Earth fault at A B Phase) BCEF(Earth fault at BC Phase) ACEF(Earth fault at AC Phase) ABCEF(Earth fault at all Phase)

Output membership value range

Fault condition

MOL(Medium over load) HOL(Heavy over load) MUB(Medium unbalance) HUB(Highly unbalance) VHUB(Very high unbalance) MUVF(Medium under voltage fault) VHUVF(Very high under voltage fault)

[15] H. J. Zimmermann, Fuzzy Set Theory and Its Applications. Kluwer Academic Publishers, 1991. [16] R. SaravanaKumar, K.Vinoth Kumar, and Dr. K.K.Ray, Fuzzy Logic Based fault detection in induction machines using Lab view IJCSNS International Journal of Computer Science

and Network Security, VOL.9 No.9, PP 226-243, September 2009

REFERENCES
[1] V.P.Mini,Sivakottiah, and Dr. S.Ushakumari,Fault Detection and Diagnosis an Induction Motor using Fuzzy Logic Proceedings of the IEEE Region 8 International Conference on Computational Technologies in Elect, and Electronics Engineering (SIBIRCON), July 2010, pp. 459- 464.