Energy Systems: Technology and Management

A New Developed Model for Cascaded Optimal Neural Networks for Short Term Electricity Demand Forecasting

Document Type : Original Article

Authors

Faezeh Amirteimoury 1
Hossein Jalali 2

1 Electricity Distribution Company, South of Kerman

2 Technical Expert at Ramin Power Generation Company, Khuzestan, Ahvaz.

https://doi.org/10.48306/estm.2025.232464
Abstract
Abstract: Short-term load forecasting ( STLF ) is precisely required for planning, operation, and control of the power system. This forecasting method is used by electricity installations, system operators, generators, and energy marketers. In this paper, by applying the feature selection ( FS ) technique on the basis of the Mutual Information ( MI ) criterion, the most effective of these data are selected. The proposed method in this paper is to use the Composite neural network ( CNN ), which consisting of three feedforward neural networks of the type multi-layer perceptron that are connected in series. Also, the Particle swarm optimization method ( PSO ) is used to achieve optimal values of the impact parameters in the forecasting process, which are included in the feature selection section and neural network. this method has been applied to load data of year 97 cities of Kerman and its results emphasize the efficiency of this method in the forecasting of electric load.

Keywords

Short-term load forecasting
feature selection
feed-forward neural network
composite neural network
particle swarm optimization
Power systems

Home

Guide for Authors

Submit Manuscript

Reviewers

Contact Us