Güneş Radyasyon Tahmini İçin Bulanık Zaman Serisi Yöntemleri Ve Fotovoltaik Sulama Sistemi Optimizasyonunda Uygulanması
Güneş Radyasyon Tahmini İçin Bulanık Zaman Serisi Yöntemleri Ve Fotovoltaik Sulama Sistemi Optimizasyonunda Uygulanması
Dosyalar
Tarih
2015-11-10
Yazarlar
Olcan, Ceyda
Süreli Yayın başlığı
Süreli Yayın ISSN
Cilt Başlığı
Yayınevi
Fen Bilimleri Enstitüsü
Institute of Science And Technology
Institute of Science And Technology
Özet
Bir yüzeye ulaşan güneş ışınları, Dünya atmosferinin dinamik özelliklerine göre rastsal olarak değişkenlik göstermektedir. Güneş enerjisi uygulamalarını verimli şekilde planlamak, yönetmek ve sistem tasarımcılarına rehber olabilmek amacıyla, güneş radyasyonunun doğru tahminleri gerekli olmaktadır. Bu kapsamda, tez çalışmasında çeşitli zaman serisi modelleri ve bulanık mantıkla entegre edilmiş yöntemler incelenmiştir. Test verileri kullanılarak deneysel analizler gerçekleştirilmiş ve tahmin yöntemlerinin performansı istatistiksel tahmin hatalarına göre karşılaştırılmıştır. Bulanık zaman serileri, belirsiz veriler için etkin tahmin araçlarıdır ve ekonomi, eğitim gibi alanlarda sıklıkla kullanılmıştır. Bu çalışma, bulanık zaman serilerini düzensiz ve rastsal yapıya sahip radyasyon tahmininde uygulayan ilk çalışmadır. Ek olarak mevcut bulanık modeller 8 farklı hibrid model şeklinde iyileştirilmiş, orijinal modellerin başarılı olan aşamaları birleştirilmiş ve geliştirilmiştir. Radyasyon verisi mevsimsel dalgalanmalar içerdiği için, tahmin hatasını azaltmak amacıyla mevsimsellikten arındırma süreci uygulanmıştır. Sonuçlara göre geliştirilen Deneme-8 modeli, diğer bulanık modeller ve klasik zaman serisi yöntemlerine kıyasla en yüksek performansı göstermiştir. Güneş radyasyonu ve benzer şekilde sıcaklık tahminleri, birçok güneş enerjisi uygulamasında girdi olarak kullanılmaktadır. Tek başına fotovoltaik (PV) su pompalama sistemleri, uzak bölgelerde sulama amacıyla güneş enerjisinden etkin şekilde faydalanmaktadır. Ancak güneş enerjisinin rastsal değişkenliği ve tahmin edilemezliği nedeniyle, PV uygulamaların yayılması zorlaşmakta ve sistem tasarımı karmaşıklaşmaktadır. Bu tür sistemlerin optimal boyutlandırılması başlıca görevlerden biridir. Bu nedenle tez çalışmasında tekno-ekonomik optimizasyon modelleri de önerilmiştir. Böylece su depolama tankı veya pilleri kullanan PV su pompalama sistemi bileşenlerinin kapasitelerini optimal şekilde belirlemek amaçlanmıştır. Önerilen modeller güvenilirlik ve maliyet göstergelerine göre geliştirilmiştir. Güvenilirlik, güç tedarik eksikliği olasılığıyla, maliyet ise yaşam döngüsü maliyetleriyle ifade edilmiştir. Burada önerilen ikinci optimizasyon modeliyle literatüre yenilik getirilmiş ve iki amaç fonksiyonu analitik olarak tanımlanmıştır. Çok amaçlı optimizasyon yöntemlerinden olan ağırlıklandırma tekniği ile uzlaşık çözüm araştırılmıştır. Mevcut boyutlandırma çalışmalarında çok amaçlı boyutlandırma yöntemleri genellikle simülasyonlar ve sezgiseller ile uygulanmıştır. Analitik boyutlandırma yöntemleri ise tek bir amaç fonksiyonuna odaklanmıştır. PV su pompalama sisteminin boyutlandırılması, bitki su ihtiyaçları ve optimal eğim açısı için detaylı analizleri de kapsamaktadır. Uzun dönemli güneş radyasyonu ve sıcaklık zaman serileri gerekliliğinin yanı sıra, sulama ihtiyaçlarının doğru tahminleri belirlenmelidir. PV dizisinin eğimi çoğunlukla enleme eşit bir açıyla kabul edilmektedir. Fakat yıllık, mevsimlik ve hatta aylık sıklıkta optimal eğim açısının hesaplanması ve sistemin bu eğime göre yerleştirilmesi, daha yüksek sistem verimiyle sonuçlanmaktadır. Bu nedenle eğim açısının düzenli olarak değiştirilmesi ve güneş enerjisi çıktısının en büyüklenmesi önerilmiştir. Geliştirilen optimal boyutlandırma modelleri tüm bu iyileştirmeleri içermektedir ve PV su pompalama veya sulama sistemlerinin kapsamlı şekilde optimizasyonunu gerçekleştirmektedir. Eğim açıları yıllık ve mevsimsel olarak farklı senaryolarda değiştirilmiştir. Ayrıca depolama birimi, su tankı ve piller olarak farklılaştırılmış, bu senaryoların performansları karşılaştırılmıştır. Antalya’da yerleştirilen limon bahçesi PV sulama sistemi örnek uygulama olarak varsayılmıştır. Önerilen analitik model (Model-2) sayısal algoritmaya dayanan modelle (Model-1) karşılaştırılmıştır. Boyutlandırma sonuçları 25 yıllık analiz için benzer elde edilmiştir. Böylece analitik modelin sonuçlarının doğruluğu gösterilmiştir. Çeşitli performans parametrelerine göre performans analizi gerçekleştirilmiş, seçilen güneş radyasyonu tahmin yöntemlerinin optimal boyutlandırma sonuçlarına etkisi incelenmiştir.
Renewable and environmentally friendly green technologies are more preferable and currently show an increasing importance over traditional energy sources such as fossil fuels and nuclear energy. Solar energy is one of the cleanest among renewable resources and has lower maintenance costs compared to other alternatives, e.g. wind, biomass. Prevention of contaminating gases’ emissions and low maintenance costs are the benefits of solar energy whereas high investment costs, dependency on weather conditions and hourly changes in day-time are the main disadvantages. Solar radiation incident on a surface varies randomly due to the dynamic characteristics of Earth’s atmosphere. This random variability and unpredictability of solar radiation makes difficult the penetration of solar energy implementations and complicate the system design. In order to plan, manage solar energy installations efficiently and to guide system designers, accurate solar radiation forecasting is essential. For this purpose, this study analyzes various time series and fuzzy integrated forecasting methods. Solar radiation forecasting methods are generally classified in three types: physical models that depend on meteorological parameters, statistical methods that depend on past radiation values and new approaches including non-parametric statistical methods and stochastic forecasting methods. Besides the literature review of solar radiation forecasting methods, time series models and Box-Jenkins methodology have been described in details. Fuzzy time series (FTS) models have been developed for linguistic and nonlinear data where the assumption of stationarity is no longer needed. FTS are effective forecasting tools with uncertain data and they are widely used in economics, education, etc. This study is the first successful attempt at implementing FTS on radiation which possesses an irregular and random nature. The aim is to find the methods having higher accuracy of forecasting for solar radiation. Therefore experimental tests have been carried out with test data and the performances of forecasting methods are compared according to statistical forecasting errors. These methods are traditional time series, seasonal autoregressive integrated moving average models and fuzzy time series models. Additionally, existing fuzzy models have been improved using 8 different hybrid models which combine and develop aspects of the original FTSs. As the radiation contains seasonal pattern, a deseasonalization procedure has been performed in order to reduce errors. The forecasts were produced with deseasonalized data after seasonal index calculations were performed. To calculate seasonal indices, a “ratio to moving averages” technique was used. The experimental tests have been also realized for original and deseasonalized data. The calculations of fuzzy forecasts have been applied easily with codings developed in MATLAB and time series forecasts using statistical softwares. The results have proved that the proposed Hybrid Model-8 shows higher performance compared to other fuzzy models and traditional time series methods. Furthermore, the deseasonalization procedure applied to fuzzy models has been founded more successful considering the decrease in the forecasting error results. After determining the models which have smaller errors especially with solar radiation test data, the real datasets of both radiation and temperature have been obtained from Turkish State Meteorological Service. Then the missing values have been estimated with interpolation and the forecasting models have been applied with real data of 7 provinces of Turkey for a period 25 years which is considered in this study as the lifetime of solar energy systems. These forecasts of solar radiation and similarly of temperature are utilizable in many solar energy applications. Once solar radiation is properly forecasted, many applications are possible in energy systems that transform solar beams into electrical energy. Designing, planning, optimal sizing and decision making, site selection or optimal allocation of energy systems and also meteorological forecasting need solar radiation data. Accurate prediction of radiation improves the efficiency of these applications and prevents unnecessary expenses with the appropriate sizing of the system’s components. Photovoltaic (PV) systems are developed as stand-alone, grid-connected or hybrid systems. In stand-alone systems, all the electricity demand is supplied by the PV system. Stand-alone PV water pumping systems effectively use solar energy for irrigation purposes in remote areas. However the random variability and unpredictability of solar energy makes difficult the penetration of PV implementations and complicate the system design. An optimal sizing of these systems proves to be essential. The optimal sizing and techno-economic evaluation of PV water pumping systems is basically dependent to the required hydraulic energy and its relation to monthly average daily solar radiation. In general, optimal sizing methods which commonly considered single objective are carried out with simulations and heuristic methods. In the present study, multi-objective optimization is handled and a mathematical model is defined according to reliability and cost parameters. This study also recommends techno-economic optimization models to determine optimally the capacity of the components of PV water pumping system using a water storage tank or batteries. The proposed models are developed regarding the reliability and cost indicators, which are the deficiency of power supply probability and life-cycle costs, respectively. The technical performance of the system is quantified by the ratio of actual power supplied and total power required for pumping water to meet irrigation requirements. It is obvious that to increase the reliability, i.e. minimize the energy deficit, PV system size must be greater in order to cover water supply requirements, while the costs also increase. Therefore a compromise solution of these two objectives is requisite to guide the system designers and investors. The novelty of this study is that the second proposed optimization model is analytically defined for two-objectives and it is able to find a compromise solution using the weighting method, whereas existing sizing methods are carried out with simulations or heuristics and analytical methods are only focused on single objective function. Although solving in analytical way is harder than computational methods, a better approximation to the optimal is effectively reached. The sizing of a stand-alone PV water pumping system comprises a detailed analysis of crop water requirements and optimal tilt angles. Besides the necessity to long solar radiation and temperature time series, the accurate forecasts of irrigation needs have to be determined. The tilt of PV array is commonly set to an angle equal to the latitude. However the calculation of the optimal tilt angle for yearly, seasonally and monthly frequencies, and install the system accordingly results in higher system efficiency. It is, therefore, suggested to change regularly the tilt angle in order to maximize solar energy output. The proposed optimal sizing models incorporate all these improvements and can accomplish a comprehensive optimization of PV water pumping systems. The tilt angles are changed yearly and seasonally in different scenarios. Also the storage units are differed as water tank and batteries in order to compare the performances. A case study is conducted considering the irrigation of citrus trees yard located in Antalya, Turkey. The proposed analytical model (Model-2) is compared with the computational model (Model-1) and it is shown that sizing results are similar for an analysis of 25 years. Thus the accuracy of the results of Model-2 is pointed out. A performance analysis is carried out using several performance parameters and the impact of chosen solar radiation forecasting methods to the optimal sizing results is investigated.
Renewable and environmentally friendly green technologies are more preferable and currently show an increasing importance over traditional energy sources such as fossil fuels and nuclear energy. Solar energy is one of the cleanest among renewable resources and has lower maintenance costs compared to other alternatives, e.g. wind, biomass. Prevention of contaminating gases’ emissions and low maintenance costs are the benefits of solar energy whereas high investment costs, dependency on weather conditions and hourly changes in day-time are the main disadvantages. Solar radiation incident on a surface varies randomly due to the dynamic characteristics of Earth’s atmosphere. This random variability and unpredictability of solar radiation makes difficult the penetration of solar energy implementations and complicate the system design. In order to plan, manage solar energy installations efficiently and to guide system designers, accurate solar radiation forecasting is essential. For this purpose, this study analyzes various time series and fuzzy integrated forecasting methods. Solar radiation forecasting methods are generally classified in three types: physical models that depend on meteorological parameters, statistical methods that depend on past radiation values and new approaches including non-parametric statistical methods and stochastic forecasting methods. Besides the literature review of solar radiation forecasting methods, time series models and Box-Jenkins methodology have been described in details. Fuzzy time series (FTS) models have been developed for linguistic and nonlinear data where the assumption of stationarity is no longer needed. FTS are effective forecasting tools with uncertain data and they are widely used in economics, education, etc. This study is the first successful attempt at implementing FTS on radiation which possesses an irregular and random nature. The aim is to find the methods having higher accuracy of forecasting for solar radiation. Therefore experimental tests have been carried out with test data and the performances of forecasting methods are compared according to statistical forecasting errors. These methods are traditional time series, seasonal autoregressive integrated moving average models and fuzzy time series models. Additionally, existing fuzzy models have been improved using 8 different hybrid models which combine and develop aspects of the original FTSs. As the radiation contains seasonal pattern, a deseasonalization procedure has been performed in order to reduce errors. The forecasts were produced with deseasonalized data after seasonal index calculations were performed. To calculate seasonal indices, a “ratio to moving averages” technique was used. The experimental tests have been also realized for original and deseasonalized data. The calculations of fuzzy forecasts have been applied easily with codings developed in MATLAB and time series forecasts using statistical softwares. The results have proved that the proposed Hybrid Model-8 shows higher performance compared to other fuzzy models and traditional time series methods. Furthermore, the deseasonalization procedure applied to fuzzy models has been founded more successful considering the decrease in the forecasting error results. After determining the models which have smaller errors especially with solar radiation test data, the real datasets of both radiation and temperature have been obtained from Turkish State Meteorological Service. Then the missing values have been estimated with interpolation and the forecasting models have been applied with real data of 7 provinces of Turkey for a period 25 years which is considered in this study as the lifetime of solar energy systems. These forecasts of solar radiation and similarly of temperature are utilizable in many solar energy applications. Once solar radiation is properly forecasted, many applications are possible in energy systems that transform solar beams into electrical energy. Designing, planning, optimal sizing and decision making, site selection or optimal allocation of energy systems and also meteorological forecasting need solar radiation data. Accurate prediction of radiation improves the efficiency of these applications and prevents unnecessary expenses with the appropriate sizing of the system’s components. Photovoltaic (PV) systems are developed as stand-alone, grid-connected or hybrid systems. In stand-alone systems, all the electricity demand is supplied by the PV system. Stand-alone PV water pumping systems effectively use solar energy for irrigation purposes in remote areas. However the random variability and unpredictability of solar energy makes difficult the penetration of PV implementations and complicate the system design. An optimal sizing of these systems proves to be essential. The optimal sizing and techno-economic evaluation of PV water pumping systems is basically dependent to the required hydraulic energy and its relation to monthly average daily solar radiation. In general, optimal sizing methods which commonly considered single objective are carried out with simulations and heuristic methods. In the present study, multi-objective optimization is handled and a mathematical model is defined according to reliability and cost parameters. This study also recommends techno-economic optimization models to determine optimally the capacity of the components of PV water pumping system using a water storage tank or batteries. The proposed models are developed regarding the reliability and cost indicators, which are the deficiency of power supply probability and life-cycle costs, respectively. The technical performance of the system is quantified by the ratio of actual power supplied and total power required for pumping water to meet irrigation requirements. It is obvious that to increase the reliability, i.e. minimize the energy deficit, PV system size must be greater in order to cover water supply requirements, while the costs also increase. Therefore a compromise solution of these two objectives is requisite to guide the system designers and investors. The novelty of this study is that the second proposed optimization model is analytically defined for two-objectives and it is able to find a compromise solution using the weighting method, whereas existing sizing methods are carried out with simulations or heuristics and analytical methods are only focused on single objective function. Although solving in analytical way is harder than computational methods, a better approximation to the optimal is effectively reached. The sizing of a stand-alone PV water pumping system comprises a detailed analysis of crop water requirements and optimal tilt angles. Besides the necessity to long solar radiation and temperature time series, the accurate forecasts of irrigation needs have to be determined. The tilt of PV array is commonly set to an angle equal to the latitude. However the calculation of the optimal tilt angle for yearly, seasonally and monthly frequencies, and install the system accordingly results in higher system efficiency. It is, therefore, suggested to change regularly the tilt angle in order to maximize solar energy output. The proposed optimal sizing models incorporate all these improvements and can accomplish a comprehensive optimization of PV water pumping systems. The tilt angles are changed yearly and seasonally in different scenarios. Also the storage units are differed as water tank and batteries in order to compare the performances. A case study is conducted considering the irrigation of citrus trees yard located in Antalya, Turkey. The proposed analytical model (Model-2) is compared with the computational model (Model-1) and it is shown that sizing results are similar for an analysis of 25 years. Thus the accuracy of the results of Model-2 is pointed out. A performance analysis is carried out using several performance parameters and the impact of chosen solar radiation forecasting methods to the optimal sizing results is investigated.
Açıklama
Tez (Doktora) -- İstanbul Teknik Üniversitesi, Fen Bilimleri Enstitüsü, 2015
Thesis (PhD) -- İstanbul Technical University, Institute of Science and Technology, 2015
Thesis (PhD) -- İstanbul Technical University, Institute of Science and Technology, 2015
Anahtar kelimeler
Güneş enerjisi sistemleri; güneş radyasyon tahmini; bulanık zaman serisi; optimal boyutlandırma.,
Solar energy systems; solar radiation forecasting; fuzzy time series; optimal sizing.