Akademik Veri Yönetim Sistemi
Tezin Türü: Yüksek Lisans
Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü, Türkiye
Tezin Onay Tarihi: 2014
Öğrenci: MEHMET BARIŞ ÖZKAN
Danışman: PINAR KARAGÖZ
Özet:As a result of rapid depletion of non-renewable energy resources, the importance of the e efficient utilization of renewable energy sources has increased all over the world and in our country in recent years. Wind has an important role in renewable energy sources with its high potential. However, compared to other renewable energy sources, wind has a spatial and temporal discontinuity characteristic so there is a need for estimating and planning of wind power generation. Wind Power Plants (WPPs) inform their wind power production forecasts for the day-ahead to an energy market and they get pro t according to correctness of their declared forecasts. So, the accuracy of estimates of power generation is very important from the economic point of view for WPP owners. In addition, forecasts must be as accurate as possible for efficient and effective administration of energy by electric transmission and distribution operators. Transmission System Operators (TSOs) regulate the energy grid of all country according to energy forecasts. Because of these factors, a reliable wind power forecast system is crucial for both WPP owners and TSOs. The accuracy of the wind power estimations is directly proportional to effective use of Numerical Weather Prediction (NWP) data. NWP data have many parameters such as wind speed, wind direction, temperature, pressure, humidity. Data mining methods and models play an important role in order to use these parameters for wind power generation forecasts effectively. The main forecast models in the literature are grouped as physical, statistical and hybrid models. Statistical models are based on constructing a mathematical modelling between past real power data and past NWP data. In this thesis, a new statistical short term (up to 48h) wind power forecast model, namely Statistical Hybrid Wind Power Forecast Technique (SHWIP), which is based on the data mining methodologies, is presented. The main aim of the model is clustering the weather events according to most important NWP parameters for improving the accuracy of the wind power forecasts. It also combines the power forecasts obtained by from three different NWP sources and produces a hybridized final forecast. The model has been verified at Wind Power Monitoring and Forecast System for Turkey (RİTM) since June 2012 and the results of the new model are compared with well-known statistical models and physical models in the literature.