Reconstructing long-term wind speed data based on measure correlate predict method for micro-grid planning
| Autor: | Asghar Akbari Foroud, Abbas Dideban, Masoud Salehi Borujeni |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Stationary process
General Computer Science Artificial neural network Computer science 020209 energy Computational intelligence 02 engineering and technology Wind speed 020401 chemical engineering Moving average Control theory 0202 electrical engineering electronic engineering information engineering Radial basis function Microgrid Autoregressive integrated moving average 0204 chemical engineering |
| Zdroj: | Journal of Ambient Intelligence and Humanized Computing. 12:10183-10195 |
| ISSN: | 1868-5145 1868-5137 |
| DOI: | 10.1007/s12652-020-02784-4 |
| Popis: | One of the most significant uncertain parameters in microgrid planning is wind speed. Wind speed has a complex dynamic behavior and developing models with high accuracy for this parameter can help decrease microgrid costs and can help improve reliability. Previous methods for wind speed modeling have been mostly statistical methods (e.g., the auto-regressive integrated moving average (ARIMA), the multivariate distribution functions, Copula function). Afore-mentioned methods are suitable for modeling stationary data. However, since the wind speed data are not stationary, the measure correlate predict (MCP) method, capable of modeling the non-stationary data, is used in this paper to generate the wind speed scenarios. Based on radial basis function (RBF) artificial neural network, this work has proposed a novel hybrid computational model to improve MCP method’s performance. The results indicate that the Hybrid-MCP method is more accurate than the conventional statistical methods, which are used to generate the wind speed scenarios (24 h). |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink