Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/C0369497-AB33-48C6-BBE6-2856AB280CB8 | ||
| Year | 2024 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Ioannis Ioannou, "Analysis of wind data from the island of Crete using time series methods", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2024 https://doi.org/10.26233/heallink.tuc.101021 | ||
| Appears in Collections |
The growth of the Earth's population has resulted in accelerating energy demand as well as environmental problems. Wind energy can compensate for some of these problems, which is why its exploitation is an industry that is growing very rapidly. Quantitative analysis and prediction of wind speed is of major importance for energy production and for decision-making. It is therefore important to find reliable mathematical models with the smallest possible estimation errors for wind speed prediction. The aim of this thesis is the prediction of wind speed with time series models on the island of Crete and more specifically in the regions of Sitia and Kissamos. Historical data are used which were extracted from the Soda (SOlar radiation Data) page which provides reanalysis data. The study processes hourly step data for the year 2021, providing a statistical analysis of wind speed for the two regions and determining the optimal probability distributions for the data. Next, the ARIMA (Autoregressive Integrated Moving Average) time series model is applied which produces hourly forecasts by season for the year 2021. The forecasts are for the last day of each season. The SARIMA (Seasonal Autoregressive Integrated Moving Average) model is also applied from 2011 to 2020 with a monthly step, taking the average values for each month of the analysis period. The objective is to investigate the seasonality of the wind speed in order to predict it for the year 2020 (based on previous years). The optimal models were selected based on the AIC -- Akaike Information Criterion and BIC -- Bayesian Information Criterion but based on the behaviour of the time series residuals. Then, using as statistical measures the root mean square error (RMSE) and the mean absolute error (MAE), the accuracy of the forecasts of different models is evaluated quantitatively.
Copyright © DIAS 2013
