N. Kumarappan

Day-Ahead Deregulated Electricity Market Price Forecasting Using Recurrent Neural Network

This paper proposes a recurrent neural network model for the day ahead deregulated electricity market price forecasting that could be realized using the Elman network. In a deregulated market, electricity price is influenced by many factors and exhibits a very complicated and irregular fluctuation. Both power producers and consumers need a single compact and robust price forecasting tool for maximizing their profits and utilities. In order to validate the chaotic characteristic of electricity price, an Elman network is modeled. The proposed Elman network is a single compact and robust architecture (without hybridizing the various hard and soft computing models). It has been observed that a nearly state of the art Elman network forecasting accuracy can be achieved with less computation time. The proposed Elman network approach is compared with autoregressive integrated moving average (ARIMA), mixed model, neural network, wavelet ARIMA, weighted nearest neighbors, fuzzy neural network, hybrid intelligent system, adaptive wavelet neural network, neural networks with wavelet transform, wavelet transform and a hybrid of neural networks and fuzzy logic, wavelet-ARIMA radial basis function neural networks, cascaded neuro-evolutionary algorithm, and wavelet transform, particle swarm optimization, and adaptive-network-based fuzzy inference system approaches to forecast the electricity market of mainland Spain. Finally, the accuracy of the price forecasting is also applied to the electricity market of New York in 2010, which shows the effectiveness of the proposed approach.

Hybrid improved binary particle swarm optimization approach for generation maintenance scheduling problem

Abstract This paper presents a model for maintenance scheduling (MS) of generators using hybrid improved binary particle swarm optimization (IBPSO) based coordinated deterministic and stochastic approach. The objective function of this paper is to reduce the loss of load probability (LOLP) and minimizing the annual supply reserve ratio deviation for a power system which are considered as a measure of power system reliability. Genetic algorithm (GA) operators are introduced in the IBPSO to acquire diversified solutions in the search space. Moreover, in this paper, the hybrid IBPSO based economic dispatch (ED) has been decomposed as a sub - problem in the maintenance model that results to a more practical maintenance schedule. A case study for the real power system model in Odisha (India) is considered. Comprehensive studies have also been carried out for the different power system consisting of 5-unit system, 21-unit system and IEEE reliability test system (RTS). It shows that the proposed algorithm can accomplish a significant levelization in the reliability indices over the planning horizon for reliable operation of the power system and demonstrates the usefulness of the proposed approach. The proposed method yields better result by means of improved search performance and better convergence characteristics which are compared to the other optimization methods and conventional method.

Accurate fault location on EHV lines using both RBF based support vector machine and SCALCG based neural network

An appropriate method for fault location on Extra High Voltage (EHV) transmission line using Support Vector Machine (SVM) is proposed in this paper. It relies on the application of SVM and frequency characteristics of the measured single end positive sequence voltage and current measurement of transient signals of the system. This paper is proposing a new hybrid approach for fault location on EHV lines using Radial Basis Function (RBF) basis SVM and Scaled Conjugate Gradient (SCALCG) basis neural network method. Sample inputs are determined by MATLAB. The average error of fault location in 400kV and 150km line is tested and the results prove that the proposed method is effective and reduce the error within a short duration of time using both RBF based SVM and SCALCG based neural network.

Day-Ahead Price Forecasting in Asia's First Liberalized Electricity Market using Artificial Neural Networks

This paper proposes a comparative model for the day-ahead electricity price forecasting that could be realized using multi-layer neural network (MLNN) with levenberg-marquardt (LM) algorithm, generalized regression neural network (GRNN) and cascade-forward neural network (CFNN). In this work applications of various models were applied to national electricity market of Singapore (NEMS), i.e. Asias first liberalized electricity market. The individual price of year 2006 is very volatile with a very wide range. Therefore, accurate forecasting models are required for Singapore electricity market company (EMC) to maximize their profits and for consumers to maximize their utilities. Hence the year 2006 has been taken for forecasting the uniform Singapore electricity price (USEP). The mean absolute percentage error (MAPE) results show that the proposed CFNN model possess better forecasting abilities than the other models and its performance was least affected by the volatility.

Evaluation of composite reliability indices based on non-sequential Monte Carlo simulation and particle swarm optimization

Monte Carlo simulation techniques used in reliability evaluation of power system are based on sequential and non-sequential simulations. This work utilizes non-sequential state transition sampling which can be used to estimate the actual frequency index without requiring an additional enumeration procedure. A state transition sampling technique does not involve sampling of component up and down cycles and storing chronological information on the system state, as the next system state is obtained by allowing a component to undergo transition from its present state. For each sampled contingency state, a minimization load curtailment model is solved using particle swarm optimization algorithm, which gives the status of the sampled state. This approach is applied to Roy Billinton Test System (RBTS) and annualized load point indices and system indices are evaluated. Results obtained are efficient and this approach has been compared with the results of sequential simulation.

Coordination mechanism of maintenance scheduling using modified PSO in a restructured power market

Global electricity market deregulation makes compatible changes and new challenges in power system operation planning problem. Maintenance is required for the generating unit to reduce the risk of capacity outage and to improve availability of units and thereby extending equipment lifetime. Modified particle swarm optimization (MPSO) for the generator maintenance scheduling (MS) generates optimal, feasible solution and overcomes the limitation of the conventional methods such as extensive computational effort which increases exponentially as the size of the problem increases. The objective of this paper is to reduce the loss of load probability (LOLP) and maximize the profit of generating units using levelized risk method (LRM). Market participants submit the MS proposal based on market clearing price (MCP) and they request permission and receive approval for planned maintenance outages from the independent system operator (ISO) in competitive electricity markets. Mainly, we are concerned with a primary framework for ISOs maintenance coordination in order to determine LOLP values in the maintenance time intervals using LRM that uses LOLP convolution algorithm. The ISO will put forward its best endeavor to adjust individual generator maintenance schedules according to the estimated LOLP values. The proposed method is tested on five generating companies model of IEEE reliability test system (RTS) to demonstrate the effectiveness of the proposed method and the applicability of the elucidation scheme for large-scale MS coordination problems.

A Novel Configuration of Unified Power Flow Controller

This paper presents the modeling and design of Unified Power Flow Controller (UPFC) using novel control technique in the inverter. The proposed control technique implements the 150 degree conduction mode of individual IGBTs in the inverter. The UPFC can be operated using PWM controllers to enhance the reactive power compensating and regulating the line voltage and also reduces the harmonic in the transmission line current and voltage. The 150 degree inverter is advantageous and increases the RMS values of output voltages, when compared to 120° mode, and 180° mode. Total required VA rating of the inverters is reduced greatly over wide load conditions. The operating performance of UPFC is demonstrated on Single Machine Infinite Bus (SMIB) system for different case studies. The proposed model considerably improves the system stability by damping the oscillation during the vulnerable conditions.

Double Circuit EHV Transmission Lines Fault Location with RBF Based Support Vector Machine and Reconstructed Input Scaled Conjugate Gradient Based Neural Network

AbstractA new algorithm is developed to enhance the solution for the problems associated with double circuit transmission lines for the mutual coupling between the two circuits under fault conditions and which is highly variable in nature. The algorithm depends on the three-line voltages and the six line currents of double circuit lines at one end. It relies on the application of Support Vector Machine (SVM) and frequency characteristics of the measured single end positive sequence voltage and current measurement of transient signals of the system. Fault resistance, mutual coupling between two circuits and initial prefault conditions are considered. The accuracy of this method has been assessed using a fault simulation software program. In the first state, the accuracy of the method was determined on the basis of SVM reconstructed method. In the second state, this method utilizes voltage and current data acquired at one common end of the two lines. This paper proposes a new hybrid approach for fault locat...

Classification of Day-Ahead Deregulated Electricity Market Prices Using DCT-CFNN

Artificial neural networks (ANNs) are promising methods for the pattern recognition and classification. In this paper applies ANN to day-ahead deregulated energy market prices. The optimal profit is determined by applying a perfect price forecast. A price forecast with a less prediction errors, yields maximum profits for market players. The numerical electricity price forecasting is high in forecasting errors of various approaches. In this paper, discrete cosine transforms (DCT) based cascade-forward neural network (CFNN) approach (DCT-CFNN) is used to classify the electricity markets of mainland Spain and New York is presented. These electricity price classifications are important because all market participants do not to know the exact value of future prices in their decision-making process. In this paper, classifications of electricity market prices with respect to pre-specified electricity price threshold are used. In this proposed approach, all time series (historical price series) are transformed from time domain to frequency domain using DCT. These discriminative spectral co-efficient forms the set of input features and are classified using CFNN. It has been observed that features selected from spectral domain improve the classification accuracy. The proposed model is more effective compared to some of the most recent price classification models.

Fault classification approach for double circuit transmission lines using wavelet transforms

A new approach is developed to enhance the solution of the problems associated with double circuit transmission lines for the mutual coupling (highly variable in nature) between the two circuits under fault conditions. The algorithm depends on the three-line voltage and the six-line currents of double circuit lines at one end. The threshold values involve fault classification and these are done on the basis of the multiresolution analysis (MRA). The main objective is to classify the fault type and it is tested with different values of internal impedance of the generator and the load models of different source resistance, source inductance in double circuit transmission lines using wavelet transform. The result of application of DWT to the double circuit transmission line with mutual coupling is effective and healthy with short duration of time by using Daubechies 4 and 9.