| Title |
Forecast of energy produced by solar power plants / |
| Translation of Title |
Saulės elektrinių pagamintos energijos prognozė. |
| Authors |
Narkutė, Vaiva |
| Full Text |
|
| Pages |
39 |
| Keywords [eng] |
Forecasting, Time Series, Solar Power Plants, Vanilla LSTM, Stacked LSTM, Bidirectional LSTM, deep learning, neural networks |
| Abstract [eng] |
In this paper, based on materials from international scientific articles, forecasts of energy produced by solar power plants are analyzed. The main goal of this paper is to analyze the theoretical and practical part of forecasting energy produced by solar power from lighting and weather prognosis. 3 data sets of solar power plants in different locations in Lithuania are used for analysis. General LSTM model is compared with statistical Linear Regression and, after it, Vanilla LSTM, Stacked LSTM, and Bidirectional LSTM are compared with each other for forecasting energy produced by solar power plants. The best-fitting parameters of Stacked LSTM are adapted to forecasting the energy production of each solar power plant. Furthermore, minimum and maximum forecasts of produced energy are prepared for having intervals between the minimum and maximum possible value of the average energy production forecast. |
| Dissertation Institution |
Vilniaus universitetas. |
| Type |
Master thesis |
| Language |
English |
| Publication date |
2022 |
