Master'sOpen Access

Development of deep learning based demand forecasting system for planning electricity generation

Is this your thesis?

This record came from a bulk archive import. If it’s yours, link it to your profile.

2023
0 views
0 downloads

Abstract (EN)

In today's world, where economic and industrial developments continue to increase, electrical energy is an important factor for the development of countries. It is very important to balance the amount of production and consumption when the energy demand is at the highest level and the need is at the lowest level. Production and consumption planning should be done in order to keep the electricity operating system at a certain frequency and to use electricity continuously and effectively. Planning of electricity production has become essential for the uninterrupted and correct operation of the enterprise. In addition, it has been observed that Turkey's electricity infrastructure and energy load balancing is mandatory within the scope of Industry 4.0. In order to minimize the lost energy costs in the system, to plan the production investments correctly, to ensure that the market participants do not suffer economically, and to provide a quality and uninterrupted energy to the system users, it is necessary to estimate the energy demand as closely as possible. Therefore, an energy forecasting plan is needed to keep the electrical energy supply and demand in balance in the system. The electrical energy demand may differ according to regional, demographic and meteorological factors. The use of large data sets related to these factors has positively reflected on the fields of machine learning and deep learning. By using various algorithms, models are made and very high performances are obtained in estimation. In our study, it is aimed to estimate the electricity consumption load by using the Turkish Electricity Consumption Data and meteorological data between the years 2018-2021 on an hourly basis, by adding the official holiday features to them. The prediction performance of the models was tested by using the data in the training of several machine learning models and deep neural network based time series models. The study foresees the elimination of energy supply and demand imbalances.

Author

Muhammet Mustafa Gökçe

How to Cite

Muhammet Mustafa Gökçe (Master Thesis). Development of deep learning based demand forecasting system for planning electricity generation, 2023, Fırat University.

License

Tüm Hakları Saklıdır

This work is shared under the specified license terms.

More theses from Fırat University