Forecasting wind turbine failures and associated costsinvestigating failure causes, effects and criticalities,modeling reliability and predicting time-to-failure, time-to-repair and cost of failures for wind turbines using reliability methods and machine learning techniques
Başlık çevirisi mevcut değil.
- Tez No: 622184
- Danışmanlar: PROF. DR. DANIŞMAN YOK
- Tez Türü: Doktora
- Konular: Enerji, Makine Mühendisliği, Energy, Mechanical Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2019
- Dil: İngilizce
- Üniversite: Colombia University
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 207
Özet
Electricity demand is rapidly increasing with growth of population, development of technologies and electrically intensive industries. Also, emerging climate change concerns compel governments to seek environmentally friendly ways to produce electricity such as wind energy systems. In 2018, the wind energy reached 600 GW total capacity globally. However, this corresponds to only about 6% of global electricity demand and there is a need to increase wind energy penetration in electricity grids. One way to enhance the competitiveness of wind energy is to improve its reliability and availability and reduce associated maintenance costs. This study utilizes a database entitled“Wind Monitor and Evaluation Program (WMEP)”to investigate, model and improve wind turbine reliability and availability. The WMEP database consists of maintenance data of 575 wind turbines in Germany during 1989-2008. It is unique as it includes details of turbine model and size, affected subsystem and component, cause of failure, date and time of maintenance, location, and energy production from the wind turbines. Additional parameters such as climatic regions, geography number of previous failures and mean annual wind speed are added to the database in this study. In this research, two metrics are considered and developed such as time-to-failure or failure rate and time-to-repair or downtime for reliability and availability, respectively. This study investigated failure causes, effects and criticalities of wind turbine subsystems and components, assessed the risk factors impacting wind turbine reliability, modeled the reliability of wind turbines based on assessed risk factors, and predicted the cost of wind turbine failures under various operational and environmental conditions. A well-established reliability assessment technique - Failure Modes, Effects and Criticality Analysis is applied on the WMEP maintenance data from 109 wind turbines and three different climatic regions to understand the impacts of climate and wind turbine design type on wind turbine reliability and availability. First, climatic region impacts on identical wind turbine failures are investigated, then impacts of wind turbine design type are examined for the same climatic region. Furthermore, we compared the results of this investigation with results from previous FMECA studies which neglected impacts of climatic region and turbine design type in section 5.4. Two-step cluster and survival analyses are used to determine risk factors that affect wind turbine reliability. Six operational and environmental factors are considered for this approach, namely capacity factor (CF), wind turbine design type, number of previous failures (NOPF), geographical location, climatic region and mean annual wind speed (MAWS). Data are classified as frequent (time-to-failure80 days) failures and we identified 615 operations listing all these factor and energy production from 21 wind turbines in the WMEP data base. These factors are examined for their impact on wind turbine reliability and results are compared. In addition, wind turbine reliability is modeled by machine learning methods, namely logistic regression (LR) and artificial neural network (ANN), using the considered
Özet (Çeviri)
a
Benzer Tezler
- Rüzgar çiftliklerinin meteorolojik parametreler üzerindeki etkisi : Model perspektifinden inceleme
The impact of wind farms on the meteorological parameters: The investigation from the model perspective
TARIK KAYTANCI
Doktora
Türkçe
2024
Enerjiİstanbul Teknik ÜniversitesiMeteoroloji Mühendisliği Ana Bilim Dalı
PROF. DR. ŞÜKRAN SİBEL MENTEŞ
- Türkiye'de rüzgar enerjisi üreten türbinlerin enerji tahmini: Markov süreci yaklaşımı
Forecasting wind energy produced from wind turbine in Turkey: Markov process approach
YASİN ATCI
- Wind speed prediction using linear prediction methods
Lineer öngörü metodları ile rüzgar hızı öngörüsü
ZAFER CANAL
Yüksek Lisans
İngilizce
2015
Enerjiİstanbul Teknik ÜniversitesiEnerji Bilim ve Teknoloji Ana Bilim Dalı
YRD. DOÇ. DR. BURAK BARUTÇU
- Forecasting wind power generation with ensembling techniques
Topluluk yaklaşimlariyla rüzgar enerjisi üretimi tahmini
YUNUS EMRE ÖZERTAŞ
Yüksek Lisans
İngilizce
2024
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolOrta Doğu Teknik Üniversitesiİstatistik Ana Bilim Dalı
DOÇ. DR. FULYA GÖKALP YAVUZ
- Atmospheric effects on short term wind power forecasting
Kısa vadeli rüzgar enerjisi üretim tahminlerinde atmosferik etkiler
YÜKSEL KALAY
Yüksek Lisans
İngilizce
2021
Enerjiİzmir Yüksek Teknoloji EnstitüsüEnerji Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ FERHAT BİNGÖL