Pradipta K. Dash

Frequency estimation of distorted power system signals using extended complex Kalman filter

The paper proposes an extended complex Kalman filter and employs it for the estimation of power system frequency in the presence of random noise and distortions. From the discrete values of the 3-phase voltage signals of a power system, a complex voltage vector is formed using the well known /spl alpha//spl beta/-transform. A nonlinear state space formulation is then obtained for this complex signal and an extended Kalman filtering approach is used to compute the true state of the model iteratively with significant noise and harmonic distortions. As the frequency is modeled as a state, the estimation of the state vector yields the unknown power system frequency. Several computer simulations test results are presented in the paper to highlight the usefulness of this approach in estimating near nominal and off-nominal power system frequencies.

Mining for similarities in time series data using wavelet-based feature vectors and neural networks

This paper presents a comparison between different wavelet feature vectors for data mining of nonstationary time series that occurs in an electricity supply network. Three different wavelet algorithms are simulated and applied on nine classes of power signal time series, which primarily belongs to an important problem area called electric power quality. In contrast to the wavelet analysis, the paper presents a new approach called S-transform-based time frequency analysis in processing power quality disturbance data. Certain pertinent feature vectors are extracted using the well-known wavelet methods and the new approach using S-transform. Neural networks are then used to compute the classification accuracy of the feature vectors. Certain characteristics of the wavelet feature vectors are apparent from the results. Further in large data sets partitioning is done and similarities of pattern vectors present in different sections are determined. The approach is a general one and can be applied to pattern classification, similarity determination, and knowledge discovery in time varying data patterns occurring in many practical sciences and engineering problems.

Stock market prediction of S&P 500 and DJIA using Bacterial Foraging Optimization Technique

The present paper introduces the bacterial foraging optimization (BFO) technique to develop an efficient forecasting model for prediction of various stock indices. The connecting weights of the adaptive linear combiner based model are optimized by the BFO so that its mean square error(MSE) is minimized. The short and long term prediction performance of the model is evaluated with test data and the results obtained are compared with those obtained from the multilayer perceptron (MLP) based model. It is in general observed that the proposed model is computationally more efficient, prediction wise more accurate and takes less training time compared to the standard MLP based model.

Local linear wavelet neural network for breast cancer recognition

Breast cancer is the major cause of cancer deaths in women today and it is the most common type of cancer in women. Many sophisticated algorithm have been proposed for classifying breast cancer data. This paper presents some experiments for classifying breast cancer tumor and proposes the use local linear wavelet neural network for breast cancer recognition by training its parameters using Recursive least square (RLS) approach to improve its performance. The difference of the local linear wavelet network with conventional wavelet neural network (WNN) is that the connection weights between hidden layer and output layer of conventional WNN are replaced by a local linear model. The result quality has been estimated and compared with other experiments. Results on extracted breast cancer data from University of Wisconsin Hospital Madison show that the proposed approach is very robust, effective and gives better classification.

Particle swarm optimized multiple regression linear model for data classification

This paper presents a new data classification method based on particle swarm optimization (PSO) techniques. The paper discusses the building of a classifier model based on multiple regression linear approach. The coefficients of multiple regression linear models (MRLMs) are estimated using least square estimation technique and PSO techniques for percentage of correct classification performance comparisons. The mathematical models are developed for many real world datasets collected from UCI machine repository. The mathematical models give the user an insight into how the attributes are interrelated to predict the class membership. The proposed approach is illustrated on many real data sets for classification purposes. The comparison results on the illustrative examples show that the PSO based approach is superior to traditional least square approach in classifying multi-class data sets.

Pattern Classification using Polynomial Neural Network

In this paper we present polynomial neural network (PNN) model using the group method of data handling to generate a nonlinear time series for classification of patterns. The proposed method considers nonlinear characteristics of the datasets and tries to evolve a polynomial using polynomial neural network that will approximate it to arbitrary token values representing the different classes in the dataset. The approach suggested finds the coefficients of PNN model by means of least square estimation technique. The PNN evolves its layers and number of neurons in each layer after evaluating the fitness function till it attains satisfactory result. Empirical result shows that PNN designed classifier performs better than many other classifier models on selected data sets using less number of features

Particle Swarm Optimized Polynomials for Data Classification

Data classification is an important area of data mining. Several well known techniques such as decision tree, neural network, etc. are available for this task. In this paper we propose a particle swarm optimized polynomial equation for classification of several well known data sets. Our proposed method is derived from some of the findings of the valuable information like number of terms, number and combination of features in each term, degree of the polynomial equation etc. of our earlier work on data classification using polynomial neural network. The PSO optimizes these polynomial equations. The polynomial equation that gives the best performance is considered as the model for classification. Our simulation result shows that the proposed approach is able to give competitive classification accuracy compared to PNN in many datasets

Fault Classification and Ground detection using Support Vector Machine

This paper presents a new approach for the fault classification and ground detection in transmission line in large power system networks using support vector machine (SVM). The proposed method uses post fault current and voltage samples for 1/4th cycle (5 samples) from the inception of the fault as inputs to the SVM. SVM-1 is trained with current and voltage samples to provide faulty phase involved and SVM-2 is trained with peak of the ground current to provide the involvement of the ground in the fault process. The SVMs are trained with Gaussian kernel with different parameter values to get the most optimized classifier. The proposed method converges very fast and thus provides fast and accurate protection scheme for distance relaying

Reduced polynomial neural swarm net for classification task in data mining

In this paper, we proposed a reduced polynomial neural swarm net (RPNSN) for the task of classification. Classification task is one of the most studied tasks of data mining. In solving classification task of data mining, the classical algorithm such as polynomial neural network (PNN) takes large computation time because the network grows over the training period (i.e. the partial descriptions (PDs) in each layer grows in successive generations). Unlike PNN our proposed network needs to generate the partial description for a single layer. Particle swarm optimization (PSO) technique is used to select a relevant set of PDs as well as features, which are then fed to the output layer of our proposed net which contain only one neuron. The selection mechanism used here is a kind of wrapper approach. Performance of this model is compared with the results obtained from PNN. Simulation result shows that the performance of RPNSN is encouraging for harnessing its power in data mining area and also better in terms of processing time than the PNN model.

Distance protection of compensated transmission line using computational intelligence

A new approach for protection of transmission line including TCSC is presented in this paper. The proposed method includes application of Fuzzy Neural Network for distance relaying of a transmission line operating with a thyristor controlled series capacitor (TCSC) protected by MOVs. Here the fuzzy neural network (FNN) is used for calculating fault location on the TCSC line. The FNN structure is seen as a neural network for training and the fuzzy viewpoint is utilized to gain insight into the system and to simplify the model. The number of rules is determined by the data itself and therefore, a smaller number of rules are produced. The network parameters are updated by Extended Kalman Filter (EKF) algorithm. with a pruning strategy to eliminate the redundant rules and fuzzification neurons resulting in a compact network structure . The input to the FNN are fundamental currents and voltages at the relay end, sequence components of current, system frequency and the firing angle with different operating conditions and the corresponding output is the location of the fault from the relaying point The location tasks of the relay are accomplished using different FNNs for different types of fault (L-G,LL-G,LL, LLL).