The History Of The Power System Model

Published Date: 02 Nov 2017

Abstract-The protection of transmission line have a great challenging task. A novel approach is used for protection of transmission line is presented in this paper. This paper focuses on fault detection and then classifying the fault. This fault Detection and fault classification is based on artificial neural network. In ANN we used feed-forward back-propagation neural network are designed to detect and classify the fault on a 400 KV, 200 Km transmission line. The features of neural networks, such as their ability to learn, generalize and parallel processing, among others, have made their applications for many systems ideal. The use of neural networks as pattern classifiers is among their most common and powerful applications. The proposed algorithm used the voltage and current signal measured at on end to detect and classify the fault. An improved performance is experienced once the neural network is trained give accurate result. The entire results show that the fault is detected and their classification. Result of performance studies shows that the proposed neural network gives more accurate results.

Key-words:-Transmission line, fault classifier, fault Detector, artificial neural networks, Matlab2009a, simulator.

INTRODUCTION:-

The power system is the most complex machine created by man so far. A Novel approach for protecting the system from failing and faults should be a integral part of such complex systems. Otherwise every time a fault occurs, repairing becomes complex. The greatest threat to the continuity of electricity supply is system faults. Faults on electric power systems are an unavoidable problem. Hence, a well-coordinated protection system must be provided to detect and Classify faults rapidly so that the damage caused to the power system is minimized. It is therefore an everyday fact of life that different types of faults occur on electrical systems, however

ANN operates on the principle of largely inter-connected simple elements operating as a network function. Artificial Neural networks (ANN) are simplified models of biological neuron system. It consists of a massively parallel distributed processing system made of highly interconnected neural computing elements called as "Neurons", which has the ability to learn and thereby acquire knowledge. ANN comprises of number of neurons which forms the basic processing unit. Each neuron is further connected to other neurons by links. Every neuron receives number of inputs which are modified by ‘weights’. The synaptic weights would either strengthen or weaken the signal which is process-ed further. To generate the final output the sum of the weighted output is passed on to a non-linear filter called as ‘activation function or ‘Transfer function’ , plus a threshold value called ‘bias’ which releases the output.

Basically, we can design and train the neural networks for solving particular problems which are difficult to solve by the human beings or the conventional computational algorithms. The computational meaning of the training comes down to the adjustments of certain weights which are the key elements of the ANN. This is one of the key differences of the neural network approach to problem solving than conventional computational algorithms which work step-by-step. This adjustment of the weights takes place when the neural network is presented with the input data records and the corresponding target values.

The goal of this paper is to detect and identify the type of fault in the line and to determine which zone (segment) of the line has become faulty. Back-propagation neural network approach is studied, implemented and modified to perform these tasks. To identify the existence of faults in the system voltage and current signals of a line are observed. These signals are also used to specify the fault type and location. The simulation models of the transmission line system are constructed and the generated information is then channeled using the software MATLAB (Version 7).

II. POWER SYSTEM MODEL

The power system consists of a 400 KV Transmission line having 200 km length. The line are feed from one ends by generators at 13.8 KV as represented in the block diagram. The line models are distributed param-eter lines. The single line diagram of the line is shown in fig 1. Short circuit capacity of the equivalent thevenin sources on each sides of the line is considered to be 100 MVA and X/R ratio is 10. The transmission line is simulated with distributed parameter line model using MATLAB software as shown in fig 2.

This power system was simulated using the SimPower Systems toolbox in Simulink in MATLAB. A snapshot of the model used for obtaining the training and test data sets is shown in Fig 2. In Fig 2. The three phase V-I measurement block is used to measure the voltage and current samples at the terminal A. The transmission line is 200 km long and the three-phase fault simulator is used to simulate various types of faults at varying locations along the transmission line with different fault resistances.

The values of the three-phase voltages and currents are measured and modified accordingly and are ultimately fed into the neural network as inputs. The Sim Power Systems toolbox has been used to generate the entire set of training data for the neural network in both fault and non-fault cases.

Faults can be classified broadly into four different categories namely:

line to ground faults

line to line faults

double-line to ground faults

three-phase faults

Fig 2 Simulation Model in Matlab using simulink

A. Outline of The Proposed Scheme

The main goal of this chapter is to design, develop, test and implement a complete strategy for the fault diagnosis as shown in Fig 3. Initially, the entire data that is collected is subdivided into two sets namely the training and the testing data sets. The first step in the process is fault detection. Once we know that a fault has occurred on the transmission line, the next step is to classify the fault into the different categories based on the phases that are faulted.

B. Overview of the Training Process

Two important steps in the application of neural networks for any purpose are training and testing. The first of the two steps namely training the neural network is discussed in this section. Training is the process by which the neural network learns from the inputs and updates its weights accordingly. In order to train the neural network we need a set of data called the training data set which is a set of input output pairs fed into the neural network. Thereby, we teach the neural network what the output should be, when that particular input is fed into it. The ANN slowly learns the training set and slowly develops an ability to generalize upon this data and will eventually be able to produce an output when a new data is provided to it.

During the training process, the neural network weights are updated with the prime goal of minimizing the performance function. This performance function can be user defined, but usually feedforward networks employ Mean Square Error as the performance function and the same is adopted throughout this work.

C. Overview of The Testing Process

As already mentioned in the previous section, the next

important step to be performed before the application of

neural networks is to test the trained neural network.

Testing the artificial neural network is very important in

order to make sure the trained network can generalize well and produce desired outputs when new data is presented to it. There are several techniques used to test the performance of a trained network, a few of which are discussed in this section. One such technique is to plot the best linear regression fit between the actual neural networks outputs and the desired targets. Analyzing the slope of this line gives us an idea on the training process. Ideally the slope should be 1. Also, the correlation coefficient (r), of the outputs and the targets measures how well the ANN outputs track the desired targets. The closer the value of „r‟ is, to 1, the better the performance of the neural network. Another technique employed to test the neural network is to plot the confusion matrix and look at the actual number of cases that have been classified positively by the neural network. Ideally this percentage is a 100 which means there has been no confusion in the classification process. Hence if the confusion matrix indicates very low positive classification rates, it indicates that the neural network might not perform well. The last and a very obvious means of testing the neural network is to present it with a whole new set of data with known inputs and targets and calculate the percentage error in the neural networks output. If the average percentage error in the ANN output is acceptable, the neural network has passed the test and can be readily applied for future use. The Neural Network toolbox in Simulink by The MathWorks divides the entire set of data provided to it into three different sets namely the training set, validation set and the testing set. The training data set as indicated above is used to train the network by computing the gradient and updating the network weights.

The validation set is provided during to the network

during the training process (just the inputs without the

outputs) and the error in validation data set is monitored

throughout the training process. When the network starts over fitting the data, the validation errors increase and when the number of validation fails increase beyond a particular value, the training process stops to avoid further over fitting the data and the network is returned at the minimum number of validation errors. The test set is not used during the training process but is used to test the performance of the trained network. If the test set reaches the minimum value of MSE at a significantly different iteration than the validation set, then the neural network will not be able to provide satisfactory performance.

III ANN ARCHITECHTURE FOR FAULT DETECTION:-

The selection of the network structure depends upon the

classification problem involving fault patterns, which

ultimately relates to the number of input and output neurons selection. After study, the back-propagation algorithm has been decided as the ideal topology. Even though the basic back-propagation algorithm is relatively slow due to the small learning rates employed, few techniques can significantly enhance the performance of the algorithm. One such strategy is to use the Levenberg-Marquardt optimization technique. The selection of the apt network size is very vital because this not only reduces the training time but also greatly enhance the ability of the neural network to represent the problem in hand. Unfortunately there is no thumb rule that can dictate the number of hidden layers and the number of neurons per hidden layer in a given problem.

In this paper a Multi layer feed forward neural network is selected. Lesser the number of inputs and outputs less is the space required but imposes the problem of insufficient characterization of the patterns to be classified. So after analyzing number of fault

states like: fault types, fault condition, fault resistance, etc number of inputs taken were 6 with 18 hidden layer neurons and 1 output for Fault Classifier and Detector. The transfer functions taken for both are tan-sig for the hidden layer and linear for the output layer.

A. Training The Fault Detection Neural Network

In the first stage which is the fault detection phase, the

network takes in six inputs at a time, which are the voltages and currents for all the three phases (scaled with respect to the pre-fault values) for ten different faults and also nofault case. Hence the training set consisted of about 1100 input output sets (100 for each of the ten faults and 100 for the no fault case) with a set of six inputs and one output in each input-output pair. The output of the neural network is just a yes or a no (1 or 0) depending on whether or not a fault has been detected. After extensive simulations it has been decided that the desired network has two hidden layer

with 18 neurons in the hidden layer.

Fig 3. shows the training process of the neural network with 6-10-8-1 configuration (6 neurons in the input layer, 2 hidden layers with 10 and 8 neurons in them respectively and 1 neuron in the output layer).

From the above training performance plots, it is to be noted that very satisfactory training performance has been achieved by the neural network with the 6-10-8-1

configuration (6 neurons in the input layer, 2 hidden layers with 10 and 8 neurons in them respectively and 1 neuron in the output layer). The overall MSE of the trained neural network is way below the value of 0.0001 and is actually 2.81e-09 by the end of the training process. Hence this has been chosen as the ideal ANN for the purpose of fault detection.

Figure 4. Regression fit of the outputs vs. targets for the Faulty Network (6-10-8-1)

B. Testing The Fault Detection Neural Network

Once the neural network has been trained, its Perfor-mance has been tested by plotting the best linear regression that relates the targets to the outputs as shown in Fig 4.

The second step in the testing process is to create a separate set of data called the test set to analyze the performance of the trained neural network.

After the test set has been fed into the neural network and the results obtained, it was noted that the efficiency of the neural network in terms of its ability to detect the occurrence of a fault is a 100 percent. Hence the neural network can, with utmost accuracy, differentiate a normal situation from a fault condition on a transmission line.

Figure 5. presents a snapshot of the trained ANN with the 6 – 10 –8 – 1 configuration and it is to be noted that the number of iterations required for the training process were 55. It can be seen that the mean square error in fault detection achieved by the end of the training process was 2.81e-09 and that the number of validation check 6 by the end of the training process.

The structure of the chosen neural network for fault detection is shown in Fig 6. with the input layer, hidden layers and the output layer labeled. It is to be noted that there are 6 neurons in the input layer, 2 hidden layers with 10 and 8 neurons in them respectively and one neuron in the output layer

IV. ANN ARCHITECHTURE FOR FAULT CLASSIFICATION:-

For fault classification a Multi layer feed forward neural network is selected. So after analyzing number of fault states like: fault types, fault location, fault inception angle, fault resistance, etc number of inputs taken were 45 with 50 hidden layer neurons and 1 output for Fault Classifier. The transfer functions taken for both are tan-sig for the hidden layer and linear for the output layer.

Simulations were carried out for different fault scenarios for getting various fault patterns. The system model as well as ANN training was done using MATLAB 7.0. The simulations done were for different fault types like LG, LL, LLL, LLLG faults. The fault resistances taken in to consideration were 1Ω and 20Ω for faults occurring at various locations. The instantaneous phasor magnitudes of voltage and current are extracted from each phase by CTs and PTs placed in the system. A sampling rate of 3 KHz, which corresponds to 50 samples per cycle for a 60 Hz cycle, was selected. Preprocessing is a useful method which can significantly reduce the size of the neural networks based classifiers and improve the performance and speed of training process. The voltage and current signals were processed using an anti-aliasing filter for removing the unwanted frequencies from the sampled signal.

V. PERFORMANCE RESULTS

The proposed performances were evaluated by applying the pre-processed data to it. The performance of Fault Classifier are shown in figure 8. It can be seen that Fault Classifier is able to perform well in classifying the fault in 60 epochs.

The performance results are shown in table 2 for few fault types classified at individual and overall checking phase of the ANN. The performance of the fault classifier shows that LL fault has least performance compared to others, whereas LLL and LG has got far better performances. Overall Performance for all types of faults results to 0.00064644. In case of LL fault

the reduction of the fault impedance results in to difficulty in classifying the fault to greater accuracy.

CONCLUSION

It has been shown through the implementation done on a Power System model comprising EHV transmission line, an ANN based Fault Classifier and Fault Detector gives satisfactory results. The training of the ANN is done by numerous simulations done on MATLAB 7.0 software package. The various fault scenarios have been taken into consideration which results into variety of patterns presented to the NN. The inputs given to Fault Classifier contain instantaneous voltage and current magnitudes at a reference end. The performance results shows that the ANN based Fault Classifier and Fault Detector provides accurate results.

REFRENCES

[1] , A. J. Mazon, I. Zamora, J. Gracia, K. Sagastabeitia, P.Eguia, F. Jurado, J. R. Saenz, Fault Location System on Double Circuit Two-Terminal Transmission Lines Based on ANN’S

[2] Bretas A.S. and Phadke A.G. (2003): Artificial neural networks in power system restoration.—IEEE Trans. Power Delivery, Vol. 18, No. 4, pp. 1181–1186.

[3] Purushothama G.K, Narendranath A.U., Thukaram D.and Parthasarathy K. (2001): ANN applications in fault locators — Electrical Power & Energy System.

[4] T. Dalstein and B. Kulicke, Neural network Approach to Fault Classification for High Speed Protective Relaying," IEEE Trans. on Power Delivery

[5]. S. Haykin, Neural Networks, IEEE Press, New York,1994.

[6]. M. Kezunoic, "A Survey of Neural Net Application to Protective Relaying and Fault Analysis", Eng. Int. Sys., Vol. 5, No. 4, Dec. 1997, pp. 185-192.

[7] S. Saha, M. Aldeen, C.P.Tan, "Tan Fault detection in transmission networks of power systems," Scince Direct Electrical Power and Energy Systems 33, pp 887–900, 2011.

[8] Cichoki A, Unbehauen R, "Neural networks for optimization and signal processing", John Wiley & Sons, Inc., 1993, New York.

[9] Haykin S, "Neural Networks. A comprehensive foundation", Macmillan Collage Publishing Company, Inc., 1994, New York.

[10] El-Sharkawi M, Niebur D, "A tutorial course on artificial neural networks with applications to Power systems", IEEE Publ. No. 96TP 112-0, 1996.