| Autor: |
H. Latała, Natalia Mioduszewska, Krzysztof Nęcka, Marek Wróbel, Anna Karbowniczak |
| Rok vydání: |
2019 |
| Předmět: |
|
| Zdroj: |
Springer Proceedings in Energy ISBN: 9783030138875 |
| DOI: |
10.1007/978-3-030-13888-2_91 |
| Popis: |
The aim of this study was to analyse the influence of different methods of pre-processing of input data such as moving average, subtraction of the mean and smoothing with the 4253H filter on the quality of forecasts of energy yield from a photovoltaic plant developed on the basis of MLP artificial neural networks. Forecasts were conducted at hourly time intervals for three types of cells; mon- and polycrystalline cells, as well as CIGS thin-film cells. The aim of the study was achieved based on the authors’ own research conducted at a PV plant located in Krakow with a total power output of 12.67 kWp. The assessments of the models developed were made based on the total ratio of energy for balancing in the total energy production (ΔESR) and on an analysis of the mean absolute percentage error (MAPE). |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
