Partial discharge analysis in high voltage rotating machines using BPN algorithm.

Sathiyasekar, K. ; Thyagarajah, K. ; Krishnan, A. 等

Introduction

The life of an electrical machine is mainly dependant on the life of its insulation. Therefore the aging process of the machine has to be very closely monitored. Just like the pulse rate, the blood pressure, sugar level etc, of a human are indicative of his health condition, the capacitance, leakage current, dissipation factor, polarization index, surge voltage with standing strength and partial discharge factors are indicative of the insulation condition of an electric rotating machine. The insulation is getting weakened by frequent over load conditions for long periods. This happens due to over load current because of over load, low voltage, single phasing etc., against which the winding can be protected by embedding thermal sensors in the winding incorporating over current sensors and single phasing preventer.

Typically a new machine of 11 kV, 3.75 MW capacity costs about Rs 35, 00,000 /-. By strict periodic preventive maintenance its minimum life expectancy is not less than 10 years. The annual maintenance cost is about Rs 60,000.

Including the cost of periodic monitoring of the condition of the insulation. The installation of the thermal sensors and single phasing preventer costs about Rs 30,000. Thus spending about Rs 60,000 to 70,000 on annual preventive maintenance saves us from a heavy expenditure of about Rs 14,00,000 towards the cost of rewinding, if the winding in burnt out, apart from the loss of production during the lay off period of rewinding the machine.

Stator winding insulation of generators is prone to partial discharge (PD) activity as a result of voids within insulation and air gaps adjacent to insulation under high voltage stress. Partial discharges are "sparks" involving the flow of electrons and ions when a small volume of gas breaks down. The term partial is used since there is a solid insulation, such as epoxy-mica, in series with the void, which prevent a complete breakdown. Depending on the size of the void, the dielectric constant, and the temperature, the stress on the gas within the void may become high enough for breakdown to occur. In most cases, the electric filed will not be uniform and this will tend to lower the breakdown voltage.

Partial discharges are often the result of damages caused by other thermal, mechanical, electromagnetic and chemical stresses acting on the stator winding. These discharges also contribute to the ageing of the machine's dielectric system by eroding away or deteriorating the insulation system. Therefore, partial discharge activity is good indication of insulation deterioration. Partial discharge testing can assess the condition of stator winding insulation and thereafter help to establish a condition based maintenance program. Condition monitoring and predictive maintenance of stator insulation brings users benefits of reliable operation, optimal number of maintenance outages and maximal lifetime of their generators [1].

Electrical Tests on Generator Stator

Carrying out certain nondestructive tests could assess the progressive deterioration of the insulation of the stator. The test parameters selected will give useful information regarding the state and quality of insulation. By monitoring these parameters periodically, trend in the ageing of the insulation could be assessed.

Insulation Resistance (IR) and Polarization Index (PI) Test

The test reflects the surface conditions of the insulation. In HV machines, end windings are likely to be affected by moisture and contamination.

PI is used as an index of dryness and represents cleanliness of the winding. The test parameters identified are IR after one minute voltage application and IR after 10 minutes voltage application at 5000 Volts DC. PI is the ratio of 10 minutes to the one minute value.

Dissipation Factor

The tan [delta] is a valuable quantity that gives integral information about the condition of the insulation. The tan [delta] is considered proportional to the total volume of discharging voids, which increases with the degree of aging. In electrical machines, as in other electrical equipment, the tan [delta] measurement is used as a traditional diagnostic method. The dielectric losses in insulation can be divided into three components, whose sum results in the tan [delta] of the insulation:

Tan [delta] = tan [[delta].sub.c] + tan [[delta].sub.p] + tan [[delta].sub.PD] (1)

Where

tan [[delta].sub.c] = the conducting loss factor caused by free-charge carries, the ions and electrons (conducting losses);

tan [[delta].sub.p] = the polarization loss factor caused by the polarization processes (polarization losses); and

tan [[delta].sub.PD] = the partial discharge loss caused by PD in the insulation (ionization losses).

In practice, in high voltage rotating machines, it is very significant that, in addition to the absolute value of tan [delta] at a certain test voltage, the tan [delta] as a function of the applied test voltage, e.g. from 0.2 to 1.2[U.sub.N], is measured at two designated voltages is named del tan delta ([DELTA]tan [delta]) or tip-up.

In good insulating systems of high voltage rotating machines, the change in tan [delta] is very small with increasing applied test voltage. But, an increase in the number and size of the voids or the information of delimitations in the insulating system during its service life caused by different stresses can lead to a large increase in the value of tan a with applied voltage [6].

Capacitance

In absence of cavities or voids within a solid insulation the change of the capacitance with test voltage is inconsiderable. In presence of voids within the insulation the occurrence of a partial discharge within these voids can ionize the gas for several milliseconds. The ionized gas has sufficiently high conductivity that the void is shorted out. This means that the effective thickness of insulation is reduced which leads to an increase of the capacitance.

However, one void shorted out by PD will have no measurable effect on the capacitance of the test object, but if there are many voids, all undergoing PD, then there will be a noticeable increase in the capacitance.

In this case, as long as an increasing number of voids begin to undergo to discharge with rising voltage the value of capacitance will increase continuously. The initial value of capacitance ([C.sub.0]), which is measured at low voltage, is the capacitance at high voltage is the capacitance of the solid insulation alone, because the voids have been shorted out by the PD. By taking the different the capacitance of the voids can be estimated. Therefore, it should be possible to estimate the void content within insulation by the capacitance tip-up measurement.

The electrical stress control coating can also have an influence on the capacitance measuring results of complete machines. The moisture content and winding contamination can also increase the initial value of capacitance ([C.sub.0]). If the end-winding of a stator is polluted with a partly conductive contamination, then the ground potential of the stator core partly extents over the end-winding and this increases the surface area of the capacitor plate [7].

Partial Discharge Measurements

Partial discharges (ionization processes) occur in micro voids and other inhomogeneties present in the body of the insulation. Partial Discharges also occur in gas gaps between stator bar surface and core and in the end winding area. The occurrence of insulation deterioration mechanisms can be determined by measuring partial discharge activity in the winding.

These discharges cause thermal ageing, mechanical erosion and chemical deterioration at localized defect sites leading to failure. In this test, partial discharge magnitudes of the highest partial discharge pulses at various voltages are measured.

Leakage Current Test

The test voltage is applied in steps and the leakage current is measured after maintaining the voltage for 1 minute duration in order to separate the conduction current and the dielectric absorption current. It is generally believed that imminent failure can be predicted from marked rise in the conduction current as the test voltage is raised.

Surge Comparison Test

The Capacitance, Tan delta and Partial Discharge measurements are adequate for testing winding insulation to ground but not the insulation between turns. Surge testing is an accurate method of identifying inters turn faults.

Surge voltage is applied on a winding consisting of a number of coils which in turn consist of many turns. A ringing pattern is seen on the CRT. The fast raising pulse spreads along the coil and creates a voltage gradient along the turns. Since the three phases are wound identically comparing all the phases will show the same single pattern Fig (1). In case of faults in any one of the phases, the wave pattern gets separated indicating a fault.

Experiment on Stator Coils

The measurement of dissipation factor and capacitance tip-up was done by using Schering Bridge at different test voltage at a frequency of 50Hz. The stator coils used for this investigation being to an insulating system with a rated voltage of 11kV.

The insulation is based on the resin rich technology. The coils are finished with semi-conducting anti corona varnish and stress grading tape to prevent partial discharges Fig (2). In practice the electrical stress plays the main role in the development of insulation deterioration and the final breakdown, while other stresses such as thermal, mechanical, thermo-mechanical and environmental stresses are mostly the inception factor for creation of defects in insulating systems.

The electrical stress can cause partial discharges in voids and cavities, which erode insulating materials and may lead to electrical treeing, which is often referred to as the most important degradation mechanism in solid insulation. Therefore, the ability of detection of such deterioration processes is very important and it is necessary to investigate the characteristic parameters, which are able to describe the condition of the insulation [2].

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

The experimental results on the simple epoxy resin specimens confirm that the relative volume of the air-filled voids within insulation can be estimated from the changes of capacitance as a function of the applied voltage [3].

Back Propagation Network

[FIGURE 3 OMITTED]

Architecture

Fig 3 shows a three layer neural network suitable for training with BPN algorithm. The input (first) layer serves only as distribution points; they perform no input summation.

The input signal is simply passed to the weights on their outputs. Each neuron in the subsequent layers produces output signals according to the activation function used. A neuron is associated with the set of weights that connects to its input. This network is considered to consist of three layers. The input or distribution layer is designated as layer 0, the second layer called hidden layer is denoted as layer 1 and the output layer as layer 2.

The activation function used in the BPN is the sigmoidal function. Training is generally commenced with randomly chosen weight values.

Typically, the weights chosen are small (between -1 and +1 or -0.5 and +0.5), since larger weight magnitudes may drive the output of layer 1 neurons to saturation, requiring large amounts of training time to emerge from the saturated state. The learning begins with the feed forward recall phase. After a single pattern vector x is submitted at the input, the layers responses z and y are computed in this phase.

Then, the error signal computation phase follows. The error signal vector must be determined in the output layer first, and then it is propagated toward the network input nodes. Negative gradient descent technique is used for calculation of the error factor and for the calculation of the weight matrix adjustment. First, the weight matrix connects the hidden layer and the output layer is adjusted and then the weight matrix connects the input layer and the hidden layer is adjusted.

The training is stopped if the cumulative error is within the limit or the number of training epoch reaches a maximum set value. The raining algorithm for the BPN network is given later in this chapter [4].

Sigmoidal Activation Function

The sigmoidal function relates the output of a neuron to the weighted input or net input (y) as follows;

f(y) = 1/1 + [e.sup.(-y)] (2)

for binary sigmoidal function.

f(y) = 2/1 + [e.sup.(-y)] - 1 (3)

for bipolar sigmoidal function[5].

An Overview of Training

The objective of training the network is to adjust the weights so that application of a set of inputs produces the desired set of outputs. For reasons of brevity, these input-output sets can be referred to as vectors. Training assumes that each input vector is paired with a target vector representing the desired output; together these are called a training pair. Usually, a network is trained over a number of training pairs. The activation function used for the analysis is bipolar sigmoidal function.

Choice of Learning Rate and Momentum Factor

Weight changes in BPN networks are proportional to the negative gradient of the error; this guideline determines the relative changes that must occur in different weights when a training sample (or a set of samples) is presented, but does not fix the exact magnitudes of the desired weight changes. The magnitude change depends on the appropriate choice of the learning rate [eta]. A large value of [eta] will lead to rapid learning but the weight may then oscillate, while low values imply slow learning. This is typical of all gradient descent methods. The right value of [eta] will depend on the applications. Values between 0.001 and 0.9 have been used in many applications [8].

Back propagation leads the weights in a neural network to a local minimum of the mean squared error. Possibly substantially different from the global minimum that corresponds to the best choice of weights. This problem can be particularly bothersome if the "error surface" is highly uneven or jagged, with a large number of local minima. To avoid this, Rumelhart, Hinton and Williams suggested that the weight changes in the ith iteration of the BPN algorithm depend on immediately preceding weight changes, made in the [(i-1).sup.th] iteration. The implementation of this method is straight forward, and is accomplished by adding a momentum term to the weight update rule,

[[DELTA]w.sub.jk] = [alpha][[delta].sub.k] [z.sub.j](s) + [eta][DELTA][w.sub.jk] (old) (2.3)

Use of momentum term in the weight update equation introduces yet another parameter [alpha], whose optimal value depends on the application and is not easy to determine a prior. A well-chosen [alpha] can significantly reduce the number of epochs for convergence. A value close to 0 implies that the past history does not have much effect on the weight change, while a value closer to 1 suggests that the current error has little effect on the weight change.

Experimental Results And Discussions

This section illustrates the performance of the proposed procedure for the classification of PD measurements. For all the experiments, with the chosen learning rate and momentum factor, the bias for both hidden and output layers are set to 1. The initial weights are randomized between - 0.5 and + 0.5[9].

PD measurements from seven machines are used for training the network. Three machines are of 11kV rating and the rest are of 6.6kV rating. Finally the network is tested with the eighth machine of 11kV rating (Table 1).This problem is tested with conventional BPN algorithm. Three trial sets are used. Each trial weight consists of 10 sets of randomized weight samples. The performance results are shown in Table 2. Epoch Vs Error characteristics for one set of randomized weight samples is shown in Fig 4.

Fig 5-Fig 7 shows that target output and actual output comparison for the test machine of R, Y & B phases respectively.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Conclusion

The condition of insulation of stator winding of the rotating machine is determined by the measurement of leakage current, capacitance, dissipation factor and partial discharge magnitude. We analyzed the above said parameters for seven machines by neural network particularly using BPN algorithm to train the network. Now the above data of the machine under test are fed to the network. From this the condition of the insulation of the 11 kV machine under test can be predicted. This network is found to be suitable for predicting the results close to the actually measured values.

Acknowledgement

We gratefully acknowledge all the testing facilities extended by M/s. CORAL REWINDING INDIA Pvt. Ltd., Erode for test facilities and technical discussion and whole hearted thanks to Er. S.Kanmani, Senior Engineer, Ericssion Ltd., Coimbatore for having given a lot of ideas in soft computing areas.

References

[1] Draft IEEE Guide to the measurement of Partial Discharges in Rotating Machinery, IEEE pp 1434, 1998.

[2] Vicki Warren., et. al., (1998) 'Recent Developments in Diagnostic Testing of Stator Windings', IEEE Electrical Insulation Magazine, vol. 14, No. 5.

[3] Binder, E., et. al., (August 2000) 'Development and Verification Tests of Diagnosis Methods from Hydro generators', CIGRE 38 session, paris, France.

[4] James A. Freeman et. al., (2005) 'Neural Networks Algorithms, Application and Programming Techniques', 10th Edition.

[5] Lee, Yang S. 40, 'Modified back-probagation algorithm applied to decision feedback equalization', vision, Image and signal processing, IEEE proceedings, volume 153, Issue:6 pp 805-809.

[6] K. Sathiyasekar, et. al., (June 2007) 'Novel Partial Discharge Analysis in the Stator Winding of 60 MW/11KV Alternator Using Neural Network', Journal of Technical and Vocational Education, volume 24, No: 1, pp 61-66.

[7] K. Sathiyasekar, et. al., (Sep-2007) 'A Novel Partial Discharge Analysis in the Stator Winding Using Neural Network', IEEE-INDICON 2007 & 16th Annual Symposium of IEEE Bangalore Section, at CPRI, Bangalore, pp 49.

[8] K. Sathiyasekar, et. al., (Jan-2008) 'Non-Destructive Testing of Motors by Partial Discharge Method', Resource Utilization and Intelligent Systems--Second International Conference, held on Kongu Engineering College, Perundurai, Tamilnadu, India, vol-II, pp 452.

[9] K. Sathiyasekar, et. al., (Aug-2008) 'Application of BPN algorithm for evaluating insulation behaviour of high voltage rotating machines', International Conference on Digital Factory-ICDF 2008, held on Coimbatore Institute of Technology, Coimbatore, Tamilnadu, India, pp 124.(Won the Best paper award)

[10] K. Sathiyasekar, et. al., (Aug-2008) 'Assessing the condition of stator winding insulation in high voltage rotating machines using BPN algorithm', National Seminar on Reliability & Life Extension Techniques of Electrical Equipments in power system conducted by NPTI, Ministry of Power, Govt of India, Durgapur, India, pp 103-109

K. Sathiyasekar (1), K. Thyagarajah (2), A. Krishnan (3) and A. Ebenezer Jeyakumar (4)

(1) Research scholar, Anna University, Zone-6, Salem.

(2) Principal, PSNA College of Engineering and Technology, Dindigul.

(3) Dean, KSR College of Engineering, Tiruchengode.

(4) Director, Sri Ramakrishna Engineering College, Coimbatore.

(1) [email protected]

Table 1: PD Measurement of 11 kV machine (Used to test
the trained network)

       Classification of PD Measurements BPN

       Input neurons: 3 Functional Inputs: 3
               Hidden neurons: 3
           Output neurons: 1  Bias: 1

      Learning Parameters: Learning Rate = 0.1
             Momentum Factor = 0.85
       Activation Function: Bipolar Sigmoid

     Max Failures: 10 Training Tolerance = 0.59
             Test Tolerance = 0.04

      3         2         1        Trial No.
      5         6         3         Failures
  24343     16943     25763      Minimum Epoch
  30294     30056     29075      Minimum Epoch
  26925     24694     27468        Mean Epoch
1957.50   4945.20   1036.90   Standard Deviation
 272.01    193.65    296.51    Minimum Time(Sec)
 338.51    343.52    334.63    Maximum Time(Sec)
 300.86    282.24    316.13      Mean Time(Sec)
   0.00      0.00      0.00          % Mean
                              Misclassifications

Table 2: Classification of PD Measurements using BPN

                    Applied   Leakage
Phase    Ground     Voltage   Current   Capacitance
        Terminals   in (kV)   in (mA)    in (nF)

  R      Y and B     4.40      135.7      99.73
                     6.60      202.9      99.85
                     8.80      269.5      99.49
                     11.0      339.5      100.3

  Y      B and R     4.40      137.8      101.0
                     6.60      205.8      100.9
                     8.80      275.5      101.3
                     11.0      341.1      100.5

  B      R and Y     4.40      136.1      99.97
                     6.60      203.6      100.1
                     8.80      272.1      100.2
                     11.0      340.7      100.4
                      PD

Phase   Tan Value   Magnitude
         in (%)       (pC)

  R       2.458       1500
          2.586       2200
          2.923       4800
          3.315       6000

  Y       2.421       1200
          2.549       2000
          2.856       4700
          3.372       6500

  B       2.462       1600
          2.592       2200
          2.898       4800
          3.343       6200