Short term load forecasting using artificial neural networks

Ace p.

Order the writing of a tailor-made Medical studies Term papers

Free quote online

Thesis Format .pdf

Short term load forecasting using artificial neural networks

Download

Read an extract

Reader
Abstract
Contents
Extract

Abstract

Today's electric power industry is undergoing many fundamental changes due to the process of Deregulation. In the new market environment, the power system operation will become more competitive. Therefore the Utilities are required to perform optimal planning in order to operate their system efficiently. Therefore the accuracy of future load forecast becomes crucial. The accuracy of the short-term load forecasts has a significant impact on an electric utility's operations and production costs.

Many conventional statistical methods such as multiple linear regression, time series, general exponential smoothing etc. have been used for forecasting short-term load. Usually, these techniques are effective for the forecasting of short-term load on normal days but fail to yield good results on those days with special events. Further, because of their complexities, enormous computational efforts are required to produce acceptable results. A short-term load-forecasting (STLF) program that uses an integrated Artificial Neural Network (ANN) approach is capable of predicting load for basic generation scheduling functions, assessing power system security, and providing timely dispatcher information. This PAPER PRESENTS the development of an Artificial Neural Network-based short-term load forecasting (STLF).

Keywords: Artificial Neural Network (ANN), Short Term Load Forecasting (STLF), Energy Forecasting

Introduction
Meaning of load forecasting
Importance of STLF
Review of STLF measures
Neutral network and its applications
Program development of STLF
Conclusion and future scope
Bibliography

Get this table of contents for free after login.

Extract

[...] Mehrotra, Mohan and Ranka Artificial Intelligence Neural Network Bart Kosko Modern Power System Analysis Nagrath Kothari Power System Planning R. L. Sulivan McGraw Hill The Art & Science of Protective Relaying Crussell Mason PAPERS Neural Network based Short Term Load Forecasting Model - A. M. Sliaraf, T. T. Lie and H. B. Gooi 1993 IEEE. Short Term Electric Load Forecast Using Artificial Neural Networks Andrew T. Sapeluk, C. Siiheyl Ozveren, Alan P. Birch - 1994 IEEE. Short Term Load Forecasting Using Genetically Optimized [...]

[...] Load Forecasting means prediction of future load on Electrical Power System or part of the system. Load Forecasting is one of the most important power system planning tools. It is very important for the power system to know the load behavior in advance. Load Forecasting, if correct ensures uninterrupted, reliable, secure and economic Electrical energy to the end consumers. Load Forecasting plays a central and dominant role in the economic optimization of electrical power system. The load dispatcher must be in a position to judge the loading on the Electrical Power System on daily basis, weekly basis or yearly basis. [...]

[...] The Paper presented here describes an application of ANN for short term load forecasting in Electrical Power System. Performance & characteristics of a Multi Layer Feed Forward Network (MLFFN) model with various network-training functions is investigated & is used to forecast hourly load of power system. ANN has been trained with hourly load data of the past three weeks & used to predict the load data of the next week for the Month of August 2006. The use of advanced training algorithms like Gradient Descent Algorithm, gradient descent with momentum & high performance algorithm Levenberg-Marquardt makes the training process faster. [...]

[...] One set consists of the hourly load from 1st August to 21stth August 2006 for training the neural network and the second set consists of hourly load data from 22nd August to 28th August 2006 as forecasting period. Training Set 1st August to 21stth August Forecasting Period 22nd August to 28th August 2006 training faster and reduce the chances of getting stuck in local minima. Also normalizing the inputs removes the problem of dependency on initial weights. In particular, scaling the inputs to will work better than scaling them to The function 'premnmx' will preprocess data so that the minimum is & maximum is + PROPOSED STRUCTURE OF ANN It is difficult to formulate a general ANN structure for any given problem. [...]

[...] No complex mathematical equations need to be defined relating the input variables and 4. What is NEURAL NETWORK the load. The models can follow sudden & random effects. Thus, Nonlinear modeling & adoption give edge to these models over other models. The other major advantages of the ANNs are robustness, fault tolerance and ability to perform reasonably well using incomplete databases due to which the ANNs perform better than the other models. Artificial Neural Networks can be trained to solve problems that are difficult for conventional computers or human beings. [...]

pdf