IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

CLASSIFICATION OF DIFFERENT FAULTS IN STATOR OF AN ALTERNATOR Manoj U. Bobade1, Prof.Niermal Chhajed 2,Prof. Pooja Ambatkar3 ,Prof.Sneha Palkar 4 [email protected],[email protected],[email protected] , [email protected] 1

Dept. of Electrical Engineering ,2Asst.Prof. Dept.Of Electrical Engineering, Asst. Prof. Dept. Of ENTC, 4Asst. Prof. Dept. Of ENTC, AVBIT, Wardha ___________________________________________________________________________________________ 3

Abstract: Synchronous generator are important elements of power system. Its reliability and proper functioning are crucial in maintaining an uninterrupted power supply to the customers. Their reliability affects the electric energy availability of the supplied area. Hence the alternator protection is critical issue in power system as issue lies in the accurate and rapid discrimination of healthy condition from different faults. It is very difficult to describe the relationship of fault information and terminal parameters by accurate mathematical expression. By applying artificial neural network in alternator, the fault diagnosis can obtain a good result. When there are faulty samples in the training samples of ANN, the severity information of the alternator faults can be directly obtained. This project describes a novel and simple artificial neural networks (ANNs) technique without using rigorous mathematics. In this project various faults were conducted on an laboratory alternator and fault currents were captured using Data Acquisition System. The energies of these current samples were calculated using Discrete Wavelet Transform and were given as input to ANN. The results so obtained are finally compared to classify the faults. Keywords: alternator, current, artificial neural network, data acquisition system, discrete wavelet transform.

Introduction Power system are the largest and most complex human made system, where faults always occurred. Faults can reason workforce and tools safety problem, and can result in substantial financial fatalities. In classify to resolve the problems, faults automatic detection, location and isolation must be employed. Most fault can cause large currents or voltage changing, and they are often detected by traditional protective relay. Whereas, some faults such as high impedance error, grounding error of unsuccessfully earthed distribution system, cause small currents and voltages changing and they are difficult to be detected using traditional protective relay. Synchronous generators are important elements of a power system. Its reliability and proper functioning are crucial in maintaining an uninterrupted power supply to the customers. Some of important fault which may occur on Alternator are Overvoltage, Over speed, Over current, Failure of Prime mover, Unbalanced Loading, Failure of Field and Stator winding faults (which includes Line to earth Faults, stroke To stroke Faults, dual Line To Ground Fault, Three phase Fault and inter turn Fault.)

www.iejrd.in

Page 1

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1 A lot of attention has been focused on generator’s single phase to ground fault which is one of the main causes for shutting down a generator. . Stator’s winding fault must be avoided since the amount of time wasted and the cost for repairing a generator is enormous. Hence, it is necessary to prevent such occurrences by incorporation reliable protection and monitoring schemes.The Literature based on the algorithm which detects and discriminate the faults on the basis of magnitude and direction of the reactive power. More recently two ANN based differential protection scheme have also been introduce to provide protection for generator stator windings. While the first technique uses samples taken from the line-side, neutral-end and field current of the generator, the difference and average of the current entering and leaving the generator windings. This project describes the simple technique based on Artificial Neural Network to discriminate various types of faults in Alternator. Three phase current of Alternator for normal and faulty condition are captured with the help of data acquisition systems. (DSO) with 100MHz bandwidth an adjustable sampling rate of 1GHz is used to capture the current. These currents are then fed as input to Artificial Neural Network which then discriminate healthy and faulty condition.

Alternator: A.Basic of alternator. A.C. generators or alternators operate on the same fundamental principles of electromagnetic induction as D.C. generators.

B. Principle of operation. Alternators generate electricity by the same principle as DC producer. while magnetic field lines cut across a conductor, a current is induced in the performer. In broad, an alternator has a stationary part (stator) and a rotating part (rotor). The stator contains windings of conductors and the rotor contains a moving magnetic pasture. The pasture cuts across the conductors, generating an electrical current, as the mechanical input causes the rotor to turn.The rotating magnetic field induces an AC voltage in the stator winding. Often there are three sets of stator winding, physically offset so that the rotating magnetic field produces a three phase current, displaced by one-third of period with respect to each other. The rotor magnetic field may be produced by induction (in a "brushless" generator), by permanent magnets, or by a rotor winding energized with direct current through slip rings and brushes. Automotive alternators invariably use brushes and slip rings, which allows control of the alternator generated voltage by varying the current in the rotor field winding. Permanent magnet machines avoid the loss due to magnetizing current in the rotor but are restricted in size www.iejrd.in

Page 2

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

owing to the cost of the magnet material. Since the permanent magnet field is constant, the terminal voltage varies directly with the speed of the generator. Brushless AC generators are usually larger machines than those used in automotive applications. An automatic voltage control device controls the field current to keep output current constant. If the output voltage from the stationary armature coils drops due to increase in demand more current is fed into the rotating field coils through the automatic voltage regulator or AVR. This increase magnetic field around the field coils which induces a greater voltage in the armature coils. Thus, the output voltage is brought back up to its original value. Alternator in central power stations use may also control the field current to regulate reactive power and to help stabilize the power system against the effect of momentary faults.

Faults Definition: A fault in a line is any failure which interferes with the normal flow of current in the line. Most of the fault on the power system lead to a short-circuit condition. When such a condition occurs a heavy current(short circuit current) flows through the equipment, causing considerable damage to equipment and interruption of service to the consumer. 3.2 Faults - Types and their Effects It is not practical to design and build electrical equipment or networks so as to completely eliminate the possibility of failure in service. It is therefore an everyday fact of life that different types of faults occur on electrical systems, however infrequently, and at random locations. Faults can be broadly classified into two main areas which have been designated “Active” and “Passive”. Active Faults The “Active” fault is when actual current flows from one phase conductor to another (phase-to-phase) or alternatively from one phase conductor to earth (phase-to-earth). This type of fault can also be further classified into two areas, namely the “solid” fault and the “incipient” fault.The solid fault occurs as a result of an immediate complete breakdown of insulation as would happen if, say, a pick struck an underground cable, bridging conductors etc. or the cable was dug up by a bulldozer. In mining, a rock fall could crush a cable as would a shuttle car. In these circumstances the fault current would be very high, resulting in an electrical explosion. This type of fault must be cleared as quickly as possible, otherwise there will be: 1. Greatly increased damage at the fault location.(Fault energy = 1² x Rf x t where t is time). 2. Danger to operating personnel (Flash products). 3. Danger of igniting combustible gas such as methane in hazardous areas giving rise to a disaster of horrendous proportions. 4. Increased probability of earth faults spreading to other phases. www.iejrd.in

Page 3

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

5. Higher mechanical and thermal stressing of all items of plant carrying the current fault. (Xmers whose windings suffer progressive and cumulative deterioration because of enormous electromechanical forces caused by multiphase faults proportional to the current squared). 6. Sustained voltage dips resulting in motor (and generator) instability leading to extensive shut-down at the plant concerned and possibly other nearby plants.

The “incipient” fault, on the other hand, is a fault that starts from very small beginnings, from say some partial discharge (excessive electronic activity often referred to as Corona) in a void in the insulation, increasing and developing over an extended period, until such time as it burns away adjacent insulation, eventually running away and developing into a “solid” fault. Other causes can typically be a high-resistance joint or contact, alternatively pollution of insulators causing tracking across their surface. Once tracking occurs, any surrounding air will ionize which then behaves like a solid conductor consequently creating a “solid” fault.

Techniques For Classification Of Alternator Faults The techniques for classification of alternator faults include: A) Time domain analysis B) Frequency domain analysis C) Using Artificial Neural Network Algorithm For Fault Discrimination Using Ann. 1.Start and run the alternator at its rated speed. 2.Faults are done on alternator with the help of contactor and take different types of fault reading. 3.The current signal captured with the help of DAS and the energies of these current signals are calculated using DISCRETE WAVELET TRANSFORM and given as input to ANN 4.ANN get trained and classify the faults. FLOWCHART

www.iejrd.in

Page 4

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

EXPERIMENTAL CIRCUIT DAIGRAM Sr.no.

Name of facilities

specification

1

3ph Synchronous Generator

1KVA, 400V, 1.5A, 3000rpm

2

DC Shunt Motor

0.75hp,230V,6.8A,1500rpm

3

Ammeter

0-5 A

4

Rheostat

1750Ω/0.6A

5

Rheostat

750Ω/1.2A

6

Voltmeter

0-600V

7

DSO & CTs

EXPERIMENTAL SETUP

Fig no:1

fig no:2.

Fig no:3 Specifications of ANN:1) Network used: multi-layer feed forward.2) No of Hidden layer(s) : 1. 3) Cross validation : zero. 4) Test percentage : 50%. 5) Transfer function : tanhaxon.6) Learning rule : momentum.7) Maximum Epochs : 1000 8) Step size : 1.00. 9) Momentum : 0.7.Info about DWT:1) Mother Wavelet : db4.2) Decomposition levels : 1 to 5. 3) Window type : rectangular window (i.e. rectwin).4) Number of section for each signal :

Results A. Line-to-ground fault: When one line conductor comes in contact with ground, is said LG fault. It is most severe fault when it is near to generator terminal. www.iejrd.in

Page 5

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

Switching instant of fault current

Fig: graphical representation of current Vs time for LG fault reading Above figure represents the fault current wrt time. On x-axis time is present and on y-axis magnitude of current. fft of of IaLG 2

1.8

1.6

1.4

1.2

x=144.1 y=1.054

X: 144.1 Y: 1.054

1

0.8

0.6

0.4

0.2

0

0

200

400

600

800

1000

1200

Fig: graphical representation of current Vs frequency for LG fault reading

Signa 50 l 0

WAVELET DECOMPOSITION OF SIGNAL

-50

0

a 5 d 1 d 2 d 3 d 4 d 5

500

1000

1500

2000

2500

500

1000

1500

2000

2500

500

1000

1500

2000

2500

500

1000

1500

2000

2500

500

1000

1500

2000

2500

500

1000

1500

2000

2500

500

1000

1500

2000

2500

20 0 -20 0.50 -0.5 -1 0.4 0.2 -0.20 -0.4 0.50 -0.5 1 0 -1 2 0 -2

current Vs frequency for LG fault reading

graphical representation energies at five different levels for 14 differeent sets of readings Above figure shows energies of fault currents at five different levels as in wavelet transform

the energy is decomposed to five different energy levels.

A. Double Line-to-ground fault When two line conductors comes in contact with ground is said to be LLG fault. It is unsymmetrical type of fault Switching instant of fault currents

www.iejrd.in

Page 6

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

fft of of IaLL

3. 5

3

2. 5

2

x=141.7 y=0.7152

1. 5

1

X: Y: 141.7 0.7152

0. 5

0

Above figure represents the fault current wrt time.

0

50 0

100 0

150 0

200 0

250 0

graphical representation of current Vs frequency for LL fault reading

On x-axis time is present and on y-axis magnitude of current.

Above figure represents various frequency components present in fault current. On x-axis frequency is present and on y-axis amplitude of current

Triple Line fault: Signa l 1 a 5 d 1 d 2 d 3 d 4 d 5

00 -1 0 0 5 0 5 1 -0 1 1 0 -

1 0. 0 -0. 5 5 0. 0. 0 -0. -0. 4 2 2 4 0. 0 -0. -0. 2 2 4

WAVELET DECOMPOSITION OF SIGNAL 50 0

100 0

150 0

200 0

250 0

50 0

100 0

150 0

200 0

250 0

50 0

100 0

150 0

200 0

250 0

50 0

100 0

150 0

200 0

250 0

50 0

100 0

150 0

200 0

250 0

50 0

100 0

150 0

200 0

250 0

50 0

100 0

150 0

200 0

250 0

current Vs frequency for LL fault reading

Switching instant of fault currents

graphical representation of current Vs time for LLL fault reading

Above figure shows energies of fault currents at five different levels as symmetrical fault..

When all line conductor get short then it is said to be LLL fault. This is most severe fault . This is in wavelet transform the energy is decomposed to five different energy levels.

ANN RESULTS: Using multilayer perceptron for one cycle and testing percentage 50%.

Fig: percentage accuracy Vs PE for one cycle.

Table shows the results of classification of various faults in an alternator by the use of 50% training percentage and multilayer perceptron, it is found that the best results are obtained with 3 processing elements for one cycle. Network description : 1) Number of hidden layer=1 2) Transfer function= tanhexon 3) Processing elements=8 4) Result obtained= 71%

www.iejrd.in

Page 7

IEJRD Journal of Science & Technology E-ISSN: 2349-0721 Volume :1 Issue 1

Best suited network: Time

DFT

DWT

DISCUSSION ON RESULT:1) In artificial neural network multilayer percepton

•

network is used.

2) In MLP no of processing elements used are 1 to 8 and percentage accuracy obtained for

•

LG,LLG, LL, LLL faults are 100% with 4 neurons and 50% testing percentage

.

The above table gives the comparative results of time domain and frequency domain (FFT and DWT ) and best results are obtained in frequency domain and mainly with the help of discrete wavelet transform.

 DISCUSSION ON RESULT: In artificial neural network multilayer perceptron network is used.In MLP no of processing elements used are 1 to 8 and percentage accuracy obtained for LG,LLG, LL, LLL faults are 100% with 4 neurons and 50% testing percentag e.

Conclusion •

Conclusion: This project describes a simple technique based on ANN to classify stator faults.

•

A neural N/W which consists of one hidden layer, one neuron, using multilayer perceptron for training with energy of d-level 1to4 as inputs is found to be the best N/W.

•

Future Scope: This method is an offline method & can be made online by connecting some embedded system consisting of some threshold weights to the experimental set up which is used to capture current waves & input can be given to ANN to classify the faults.

References •

"Protective Relays Applications Guide," The English Electric Company Limited, Relay Division, Stafford, 1975. • C. J. Mozina, IEEE Tutorial on the Protection of Synchronous Generators, IEEE Tutorial Course, IEEE Power Engineering Society Special Publ., no. 95 TP102, 1995. • M. S. Sachdev and D. W. Wind, "Generator differential protection using a hybrid computer," IEEE Trans. Power Apparatus System, PAS-92(1973) 2063-2072. • H. Tao and I. F. Morrison, "Digital winding protection for large generators," J. Electr. Electron. Eng. Aust., 3 (1983), 316-321. “Current Differential Protection of Alternator Stator Winding” N.W.Kinhekar, Sangeeta Daingade, and Ajayshree Kinhekar

www.iejrd.in

Page 8