Neural network technologies for forecasting and controlling electricity consumption in energy systems by the genetic method
The paper considers the neural network forecasting technologies in controlling power consumption in energy systems using the genetic method. It is proved that in order to implement the technological management system of a regional grid company, it is possible to use the technical and information pla...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Tomsk Polytechnic University
2025-03-01
|
| Series: | Известия Томского политехнического университета: Промышленная кибернетика |
| Subjects: | |
| Online Access: | https://indcyb.ru/journal/article/view/85/70 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The paper considers the neural network forecasting technologies in controlling power consumption in energy systems using the genetic method. It is proved that in order to implement the technological management system of a regional grid company, it is possible to use the technical and information platform of a hierarchical automated information and measurement system for monitoring and metering electricity. We consider the task of improving the accuracy of short-term forecasting of electricity consumption using deep machine learning methods. The novelty of the work lies in the use of a developed genetic algorithm to select hyperparameters of a neural network that affect the quality of its work, but are not determined in the learning process. The authors have developed the neural network models and carried out the study to find the optimal structure of a neural network, and the influence of specified neural networks hyperparameters on the error in predicting power consumption. The developed management methodology and technologies are applied in the structure of the software modeling system to manage the regional energy system of autonomous consumers. Based on the results of training and testing, the genetic algorithm confirmed the possibility of automating the selection of optimal hyperparameters and obtaining forecasts of greater accuracy and the possibility. |
|---|---|
| ISSN: | 2949-5407 |