EE- Enerji Bilim ve Teknoloji Lisansüstü Programı - Doktora
Bu koleksiyon için kalıcı URI
Gözat
Son Başvurular
1 - 5 / 23
-
ÖgeForecasting of produced output electricity in photovoltaic power plants(Graduate School, 2025-05-16)The global energy market is pivotal to modern economic growth, with renewable energy, particularly solar power, emerging as a key strategy for mitigating climate change and fostering sustainable development. Solar energy, a clean and abundant resource, directly supports UN Sustainable Development Goals (SDGs) such as Goal 7 (Affordable and Clean Energy) and Goal 13 (Climate Action) while indirectly contributing to Goals 8 (Decent Work and Economic Growth), 9 (Industry, Innovation, and Infrastructure), and 11 (Sustainable Cities and Communities). However, the variability of solar energy generation—shaped by factors like solar radiation, temperature, and technological efficiency—poses challenges for grid stability, resource planning, and the broader integration of renewables. Accurate forecasting is critical to addressing these issues, enabling better grid management, informed policymaking, and strategic investments in renewable energy systems. This thesis tackles the challenges associated with solar energy forecasting by utilizing cutting-edge developments in artificial intelligence (AI) and machine learning (ML) to improve prediction accuracy and reliability. It focuses on under-explored areas, including the integration of endogenous and exogenous variables, chaotic modeling, and hybrid model optimization. Through systematic investigation and evaluations, the research proposes innovative techniques validated with real-world data, providing practical insights for managing solar power generation and advancing low-carbon energy systems. By tackling critical gaps in forecasting methodologies, this work supports global sustainability goals, promotes energy security, and accelerates the transition to cleaner and more efficient power networks. This thesis builds upon the extensive literature review to identify and address critical gaps in solar energy forecasting. The research focuses on four key dimensions: (1) integrating both endogenous variables and exogenous variables for a holistic understanding of solar energy generation, (2) employing robust data preprocessing techniques like dimensionality optimization and feature extraction to enhance data quality and model reliability, (3) exploring advanced deep learning architectures and ensemble frameworks to improve predictive performance, and (4) optimizing metaheuristic algorithms for efficient and scalable AI-based forecasting solutions. These four methodological pillars aim to address challenges such as nonlinearity, temporal dependencies, and chaotic behavior in solar power datasets. The study employs five groups of forecasting models—linear, machine learning, deep learning, ensemble learning, and chaotic models—validated using real-world data from the Konya Eregli Solar Power Plant. Among the models, Feedforward Neural Networks (FFNN), Extreme Learning Machines (ELM), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Echo State Networks (ESN) emerged as top performers based on standard evaluation metrics. These seven models were tested under scenarios combining endogenous and exogenous inputs and underwent hyperparameter optimization using the Tree-structured Parzen Estimator (TPE) to maximize accuracy. Together, these advancements contribute to more reliable and precise solar energy forecasting, supporting the integration of renewable energy into power grids. Prior to training the models, the dataset undergoes a thorough cleaning process to ensure the quality and reliability of the data. Outliers are identified and removed using the Local Outlier Factor (LOF) method, followed by an exploratory data analysis to gain deeper insights into the time series characteristics. Once the data is cleaned and understood, feature engineering strategies are applied to extract and transform relevant features. These strategies are specifically tailored to meet the unique requirements of both individual base models and ensemble stacking meta-models. To ensure robust model evaluation and prevent data leakage, the processed dataset is split into training, validation, and test sets, forming a reliable foundation for both base and meta-model development. This study evaluates the performance of all implemented forecasting models using five well-established evaluation metrics: R-squared (R2), Root Mean Square Error (RMSE), Mean Squared Error (MSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). Together, these metrics offer a thorough analysis of the models' accuracy and reliability. Among the fine-tuned base models, the LSTM and GRU models demonstrate the highest levels of accuracy and reliability. These deep learning architectures effectively capture the complex temporal dependencies inherent in solar power data. Following LSTM and GRU, other models such as FFNN, XGBoost, RF, ESN, and ELM also perform well, with their efficacy varying based on specific scenarios. These results underscore the strengths of deep learning models while validating the applicability of traditional machine learning and ensemble methods in addressing particular forecasting challenges. To address chaotic behavior in the time series, the False Nearest Neighbors (FNN) method is incorporated as a preprocessing technique for reconstructing the phase space. This step determines an optimized embedding dimension intended to maintain the structural consistency of the original time series. The ESN algorithm is then employed to forecast solar energy production using both the original and embedded data. Results show that optimizing the embedding dimension via the FNN technique significantly improves the model's capability to capture the nonlinear dynamics underlying solar power generation, leading to more accurate forecasts. The enhanced performance of the ESN model on the embedded data underscores the vital role of data preprocessing—particularly embedding dimension selection—in chaotic modeling for solar energy systems. Overall, this approach demonstrates how combining proper data embedding with an optimized model architecture leads to superior predictive capabilities, further contributing to efficient and reliable solar energy forecasting. As the final stage of this thesis, stacking ensemble machine learning techniques are applied to enhance forecasting performance even further. The primary aim is to develop a novel metamodel that surpasses the prediction capabilities of the individual base models. Two categories of stacking ensemble models are investigated: regression-based stacking and neural network-based stacking. Each approach is critically evaluated and compared, leading to the identification of the most accurate model—an essential contribution of this research. In the first category, traditional regression algorithms—Multi-Linear Regression, Ridge Regression, and Lasso Regression—function as the meta-model to combine the predictions of the seven base models. This approach leverages the strengths of diverse base predictors through a linear combination, aiming to minimize overall forecast error. However, it ultimately does not surpass the established accuracy benchmark. In the second category, neural networks (NN) function as the meta-model, exploiting their capacity to capture complex nonlinear relationships within the data. Three distinct neural network-based stacking strategies are developed, including the proposed deep learning-based stacking model. This meta-model is deliberately constructed using only the two best-performing base models, selected based on their individual accuracy metrics. To further enhance predictive performance, advanced techniques and optimized hyperparameters are incorporated. As a result, this final stacking model delivers the highest forecast accuracy among all models investigated. The meta-models' designs undergo several refinements, including the integration of Model Output Statistics (MOS) as a post-processing technique. These adjustments aim to determine whether further tuning can enhance forecasting accuracy. The outcome of these efforts is the Deep Learning-Based Stacking Meta-Model, proposed by this thesis, which predicts solar energy output for the Konya Eregli Solar Power Plant with an impressive precision exceeding 98%. This model not only outperforms the most accurate base model (LSTM) but also achieves the primary objective of this research: setting a new standard for high-accuracy solar energy forecasting. A key strength of the proposed deep learning-based stacking meta-model lies in its simplified architecture, which stands out compared to the other meta-models developed in this thesis. By focusing exclusively on the two most accurate base models, the meta-model balances exceptional forecasting accuracy with reduced computational demands. This minimalistic design speeds up both the training and deployment processes, aligning with the research objective of optimizing metaheuristic model architectures to create more efficient and scalable AI-driven forecasting solutions. This level of predictive precision underscores the transformative potential of advanced ensemble techniques in solar energy forecasting. By achieving improved accuracy, these methods enhance energy management systems, enable smoother grid integration, and support strategic long-term planning in renewable energy sectors. As a result, the proposed deep learning-based stacking meta-model proves to be an essential tool for advancing sustainable energy infrastructures and ensuring a reliable transition to renewable energy systems.
-
ÖgeExperimental and numerical investigation of single and multiple droplet interactions with high-temperature surfaces(Lisansüstü Eğitim Enstitüsü, 2021)The phenomenon of droplet impingement onto solid surface can be seen in several industrial applications such as ink-jet printing, fuel injection process in internal-combustion engines, spray coating, and spray cooling systems. For several decades, the importance of this phenomena has inspired several researchers' investigations in pursuit of a thorough understanding of the interactions between mass, momentum and heat transfer. However, despite many experimental and numerical studies, the effect of droplets on a solid surface is not fully understood. This problem becomes more complicated if the surface is heated during the impact and interacts with other droplets. In the literature, most of the studies have focused on single droplet impact, while those relating multiple droplet impacts are quite more limited. After the single droplet impact onto the solid surface, droplet starts to spread and a circular lamella takes shape. However, after multiple droplet impingement onto a solid surface, the interaction phenomenon will occur if the droplets are too close to each other based on their impact parameters. The hydrodynamic outputs and heat transfer activities of the droplets are very distinct from single droplet cases due to this interaction. This interaction phenomenon leads to uprising sheets which causes lesser spreading area per droplet on solid surface. The challenge of understanding of physical mechanism and modeling the phenomenon of spray cooling comes from the droplets' randomness and untraceable behavior. For that purpose, it is important to examine a simplified method with a known number of droplets. In addition, most of the studies available in the literature have provided limited quantitative data, such as spreading diameter, and lamella height. Recent developments in visualization and imaging technology provide significant advantages in measurement techniques. Spreading phase after droplet impingement is characterized by rapid energy conversions and dissipations within a small area. The major limitation lies in the difficulty of optically reaching the target region for non-invasive physical quantity measurements. The success of particle image velocity (PIV) measurement is especially noteworthy in such circumstances. This technique makes it possible to evaluate complex actions over time in more detail. Therefore, time-dependent radial velocity distribution of the droplets in the spreading phase was measured using the PIV technique. This thesis consists of two main sections: experimental and numerical studies. Experimental studies were carried out in the Visualization Laboratory at Tokyo University, Nuclear Engineering and Management Department. An experimental setup was designed and built to carry out experimental studies. To get simultaneous multiple droplet, a droplet generation and control system has been developed. In the experimental investigations, shadowgraph and PIV methods are used. In the simulation part of this thesis, hydrodynamics behavior and cooling performance of single and multiple droplets were investigated by using Computational Fluid Dynamics (CFD). StarCCM + (version 2019) has been used to perform numerical investigations on a workstation (4 cores 16 GB RAM). New numerical models have been developed for the liquid and gas phases including mass, momentum, and heat transfer equations based on the "Volume of Fluid" (VOF) method. These numerical models were validated with experimental data. 2-D axisymmetric VOF model has been used only for single droplet investigations whereas, 3-D VOF model has been used for multiple droplet interactions. The results obtained in this thesis are summarized as follows: Variation of uprising sheet height with dimensionless time is compared with an existing theoretical model. The theoretical model agrees with experimental results for initial phases. However, the experimental findings do not fully match with the theoretical findings in later phases owing to uprising sheet losing its balance. The contact time of a single droplet in the film boiling has been compared with available correlations in the literature and it has been observed that the experimental results are largely in agreement with available correlations. Also, a new empirical correlation is proposed. It is seen that MAPE value is 3.12% and correlation successfully represents the contact time. It has been observed that the rebound phenomenon for simultaneous and non- simultaneous double droplet cases take place faster comparing with a single droplet. This is more probably due to the less spreading area per droplet owing to droplet interaction. Also, using PIV method, the variation of radial velocity was examined inside the droplets for different temperatures. It was observed that radial velocity is linear over a comparatively wide range of spreading radius, but due to capillary and viscous forces with time, the radial velocity profiles took a non-linear shape in the exterior radial positions. At the high surface temperatures, particularly in later stages, increases the uncertainties in the radial velocity distribution in exterior radial locations because of intense disturbances are formed at the interface by the bubble nucleation. Afterwards, PIV data within the lamella were compared with a theoretical model at ambient temperature. For high We case, the analytical model highly agrees with linear sections of the radial velocity profiles. For moderate We case, during the early spreading phase, the model highly agrees with the linear pieces of radial velocity profiles in the inner radial positions. Partial linearity is still identified in the latter stages, but the theory differs from radial velocity profiles. Furthermore, the spreading velocities within droplet pair are examined at ambient temperature using PIV. Creation of uprising sheet leads to an upward flow which causes this extra stagnation point in the interaction region. Computational models for single droplet were validated by comparing qualitative shadowgraph images, spreading factor as well as radial velocity distributions within the droplet. For multiple droplet case, to validate computational model variation of the dimensionless uprising sheet with time was also compared. Total number of mesh elements are 125 000 and 7.5 × 106 for 2-D Axisymmetric and 3-D model, respectively. When vertical distance is close to each other, uprising sheet created by spreading liquid lamella collision still can be detected. However, when vertical distance is wider the uprising sheet could not be observed. Also, it is observed that decreasing vertical spacing leads to reaching maximum values of total spreading area and heat flow earlier. As the horizontal distance between the droplets gets shorter, the magnitude of interaction increases and the spreading area covered on the surface and heat transfer decreases. A mathematical expression is proposed to predict the dimensionless spreading area per droplet for multiple droplets impacting simultaneously on the solid surface taking into account the variation of spreading factor for a single droplet. Cooling efficiency and performance loss are defined to see the effect of droplet number taking into account the variation of heat flow for a single droplet. It was found that the interaction strongly effects the cooling performance especially in the first stages after the multiple droplet interaction.
-
ÖgeExtended exergy accounting (EEA) analysis of Turkish society- determination of environmental remediation costs(Enerji Enstitüsü, 2013)Bu tez çalışmasında, kaynak kullanım verimliliği yönünden incelenmek üzere Turkiye örneği ele alınmış, motod olarak Extended Exergy Accounting (EEA, Genişletilmiş Ekserji Analizi) metodu uygulanış ve ulaşılan sonuçlar sunulmuştur. İlk olarak Dr. Enrico Sciubba tarafından geliştirilerek literatüre katılan ve ekserji bazlı bir kaynak kullanım analizi metodu olan EEA metodu, bugüne kadar literatüre katılmış hiçbir metodolojinin yapısında barındırmadığı bir yenilik sunularak, ele alınan sistemin, enerji yada ağırlık birimleri ile ifade edilebilen girdilerin yanında (enerji akışları ve materyal dışında), sisteme olan ?diğer? girişlerin - kapital, iş gücü ve çevresel etki - ekserji biriminde ifade edilmesi için yeni bir hesaplama metodu sunulmuştur. Metodun arkasındaki zihniyet, sistemin tükettiği kapital, işgücü ve çevresel etkinin giderilmesi için harcanan ekserjinin üretiminde kaynak kullanıldığı ve adı geçen ekserji tüketimlerinin de sistemin toplam kaynak kullanımı içerisinde ele alınması gerektiğidir. İş gücü ve kapitalin ekserji karşılıkları olarak, bunları yaratmak için gerekli olan kaynak tüketiminin ekserji değeri belirlenmektedir. Çevresel etki olarak ise, sistemden çıkan atığın temizlenmesi için gerekli kaynak kullanımının ekserji karşılığı hesaplanır. Sonuç olarak EEA, sistemin hertürlü kaynak tüketimini tek bir birimle (ekserji) ifade ederek birim bütünlüğünün sağlanmasının yanında, bugüne kadar hiç ele alınmamış olan ek akışların da sistem ekserji dengesi içerisine katılması ile ?genişletilmiş ekserji dengesi (extended exergetic balance)? kurulmasını sağlamakta ve adından da anlaşılır şekilde, şu anda literatürde olan en ?gelişmiş? ekserji bazlı kaynak kullanım analizi metodunu sunmaktadır. Özetle, EEA metodu ile yapılan analizlerde, sistemin her safhasında kullanılan malzeme, enerji, kapital, işçilik ve çevresel etki (ele alınan sisteminin atık ve emisyonlarının izin verilen sınırlar dahilinde tutulması için yapılacak işlemler) gibi faktörlerin hepsi analize katılarak ekserji biriminde ifade edilmiş ve sistemin kaynak kullanımı değerlendirmesine katılmıştır.Bu çalışmada, sistem olarak ele alınan Türkiye, EEA metodu ile incelenmiştir. Çalışmanın amacı: eylem yapıcı birimlere, ülke içerisinde kaynak kullanım kalitesinin değerlendirilmesi ve ülkenin daha kararlı ve sürdürülebilir çizgide varlığını devam ettirmesi için en mantıklı ve faydalı müdehale noktalarının bildirilmesidir. Çalışmada yapılan uygulama özetlenecek olursa: EEA ile yapılan ülke analizlerinde mutat olduğu üzere, ele alınan ülke 7 sektörel bölüme ayrılmakta ve birbiri arasındaki ekserji alışverişleri analiz edilmektedir. Bu sektörlerin kendi içindeki ekserji akışlarının yanında çevre ile (Environment, ENV) ve diğer ülkeler (Abroad, A) ile etkileşimi de hesaplamalara dahil edilmektedir. Söz konusu 7 sektörel bölüm ve kapsadığı faliyetler şunlardır:EX (Madencilik Sektörü): Hammadde çıkarma ve işleme (Petrol ve doğal gaz çıkarma ve rafineri işlemleri dahil)CO (Dönüşüm Sektörü): Enerji üretim tesisleri (rafineriler, ısı ve elektrik üretimi)AG (Tarım Sektörü): Tarım ve hayvancılık faliyetleriIN (Endüstri Sektörü): Endüstriyel faliyet kolları (rafineriler hariç)TR (Ulaştırma Sektörü): Ulaştırma faliyetleriTE (Servis sektörü): Servis faliyetleri (otel, eğitim, danışmanlık vs. hizmetleri)DO (Hanehalkı): Ev içi kullanım ve üretime dayalı faliyetlerYukarıda özetlenen EEA metodolojisinin Türkiye uygulamasının tez içinde sunulmasının yanısıra, bugüne kadar literatürde ilk defa görülür şekilde, sektörel katı, sıvı ve gaz atıkların çevresel etki maliyetleri, EEA metodu içerinde sunulan orjinal tanım ve teori doğrultusunda hesaplanmıştır. Diğer bir değişle, bugüne kadar literatürde uygulanan: atık temizleme faliyetlerinin gerektirdiği parasal yatırımın ekserji karşılığını ?çevresel etki maliyeti? olarak kabul eden pratik fakat sentetik ve metodun doğasını yansıtmayan yaklaşımın dışına çıkarak, çevresel etki maliyetleri, gerçek sistemler ele alınarak, EEA içerisinde sunulan orjinal tanımına uygun olarak hesaplanmıştır.Çevresel etkinin ekserjetik maliyetinin hesaplanmasında ele alınan sistemlerin ticari olarak aktif, teknik olarak bilinen ve yaygınlıkla kullanılan sistemler olmasına dikkat edilmiştir. Bu amaçla,1) günümüzde atık su ve katı atık islahı için sıklıkla kullanılan ve atıktan, yaklaşık 98% saflıkta metan oranına sahip olan -bir nevi doğal gaz alternatifi- bir tür yakıt (biyogaz) ürtilmesini sağlayan anaerobik çürütme (anaerobic digestion) prosesi2) dönüştürülebilir atıklar için geridönüşümtabanlı sistem seçimleri yapılmış ve bu çalışma dahilinde analiz edilmiştir.Katı atık söz konusu olduğunda, atık türlerinin atık kompozisyonu içindeki oranları değişmekle beraber, DO, IN ve TE Sektörlerin katı atık bileşiminin ayni maddelerden oluştuğu göz önüne alınarak aynı proses zinciri içinde atık giderimi incelenmiştir. Özetle: atığın organik kısmı anaerobik çürütme prosesine tabi tutularak elde edilen biyogaz bir kojenerasyon tesisinde yakıt olarak kullanılmış ve elektrik ve ısı üretilmiştir. Inorganik kısım ise olabilecek maksimum oranda geridönüşüme uğradıktan sonra, geridönüşümsüz kısım yakma tesinde yakılarak ısı ve elektrik üretilmiştir. Geridönüsüm işlemleri sırasında oluşan artık kısım, düzenli depolama yapılmıştır. EX Sektör atığı, doğadan gelip tekrar depolama yolu ile doğaya terk edildiğinden incelenmemiştir. CO Sektör atığı içerisinde de yukarıda sayılan sektörlerin atık bileşiminde bulunan maddeler olduğundan yukarıda özetlenen atık giderimi sistemlerine ek olarak, rafineri atıkları için IGCC (integrated gasification combined cycle, entegre gazlaştırma kombine çevrim) sistemi ile enerji üretimi yapılmıştır. AG Sectör katı atığı olarak ele alınan hayvan ve bitki artıkları, anaerobik çürütme prosesinden geçirilmiş, oluşan biyogaz enerji üretiminde kullanılmıştır. TR Sektör atığı, tamamen farklı bir bileşime sahip olduğundan, sektöre özel bir yaklaşımla, taşıtların parçalanmasından sonra geri dönüşüm prosesi yapılmış, atık lastikler ise yakılarak ısı ve elektrik üretiminde değerlendirilmiştir. Geri dönüşüm işlemi artıkları ve yanmadan arta kalan kül, düzenli depolama ile yok edilmiştir. TR Sektör atığı olarak, sadece kara yolu atıkları incelemeye alınmıştır. Türkiyedeki ulaştırma sisteminin ne derece kara yoluna dayandığı dikkate alınırsa atığın büyük kısmının kara yolu taşıtlarından üretilmesini beklemek mantıklıdır. Ayrıca diğer ulaştırma motlarının ürettiği atık üzerine veri yoktur.Gaz emisyonlar için, güvenli ve düzenli bir veri analizinin ulaşılabilir olduğu CO2, CH4 ve N2O gazları ve bunların giderilmesi ele alınmıştır. Zaten kendisi bir yakıt olarak kullanılabilir olan CH4 enerji üretiminde değerlendirilerek, bu sistemin EEA analizi sunulmuştur. CO2 giderimi için CO2'nun Ca ile reaksiyonu sonucu CaCO3 üretimine dayanan bir sistemden faydalanılmıştır. N2O için ise N2O'nun yüksek sıcaklıkta dekompozisyonuna dayanan bir sistem incelenmiştir.Sıvı atıklar için ise, Türkiye'nin DO Sektörü tarafından üretilen evsel sıvı atık ele alınmıştır. Türkiye'ye özgü datalar incilendiğinde, atığın bir kısmının hiç işlem görmediği, bir kısmının ise çeşitli kademelerde arıtma proseslerine uğradıktan sonra ?arıtma çamuru? oluştuğu ve bu çamurun düzenli depolama ile gömüldüğü bilinmektedir. Bu çalışmada hem hiç proses görmemiş atık suyun hem de üretilen çamurun anaerobik çürütülmesi yolu ile bertarafının çevresel etki maliyetleri bulunmuştur. Diğer sektörel atıklar için de, çevresel etki maliyetinin evsel sıvı atık ile aynı olduğu kabul edilerek diğer sektörler için işlem yapılmıştır. Bu yaklaşımın gerekliliği, her sektörlerün atık su bileşimlerine ait bir veri kaynağının Türkiye için olmaması ve bu derece ayrıntılı bir analizin zaman ve hacim olarak sınırlı böylesi bir tez çalışması içinde mümkün olmadığı göz önüne alınarak açıklanabilir.Diğer bir çevresel etki ekserji maliyeti araştırması, sektörlerden atmosfere deşarj edilen ısının giderimi için yapılmıştır. Söz konusu ısı, en büyük oranda ve en yüksek sıcaklıkda baca gazları yolu ile atmosfere verildiği için baca gazları ele alınmış ve ortalama baca gazı bileşimlerinden yola çıkarak, atık gazların çevre ile aynı sıcaklığa getirilmesi için kullanılan ORC (Organic Rankine Cycle) sisteminden elektrik üretilminin EEA analizi sunulmuştur.Yukarıda anlatılan çevresel etki ekserji maliyetleri ve sektörel verimler sonuç bölümünde özetlenerek sunulmuştur. Bulunan sonuçların ayrıntılı incelemesi de sonuç bölümünde görülmektedir. Sonuçlara göre EX, CO, AG, IN, TR, TE ve DO Sektörlerin EEA analizi verimleri 91%, 43%, 0,13%, 57%, 48%, 87% ve 99% olarak belirlenmiştir.
-
ÖgePlanning And Stochastic Evaluation Of Combined Cooling Heat And Power Systems Under Uncertainty(Energy Institute, 2019-01-29)CCHP (Combined Cooling Heat and Power) systems are the most well-known technologies for efficient energy usage and it usually refers to simultaneous production of cooling, heating and power from a single energy source. CCHP plants are built as decentralized systems, and they are operated close to where it is needed. Thus, CCHP systems are considered as more efficient, profitable, reliable and environmentally friendly systems compared with conventional generating plants. Nonetheless, CCHP systems or any other energy conversion systems should be designed and operated effectively to gain the expected advantages. It is not an easy decision for a SME (small and medium size enterprises) to invest in CCHP systems. Decisions for investments are generally taken by the conventional method, which relies on the result of an economic analysis with the assumption that variables will remain stable over the time the analysis is made. However, this kind of systems is dynamic and all the parameters are subject to change until the day the CCHP system expires economically. Thereby, CCHP systems work under uncertainty conditions during their economic life. The technical and financial performance of the system is affected by various parameters which include the fluctuation of energy loads, working hours, energy prices, exchange rates and interest rates. Accordingly, evaluating only the scenario where the current values of variables are taken into account may not help investors in decision making owing to uncertainties, the probability of occurrence of uncertainties and their outcomes. The analysis held in this study has been based on real and current operational data of an existing industrial facility located in Istanbul. Beside that, This study has two stages to assess the uncertainties in CCHP systems. The main purpose of the first part of the study is to specify a model and a methodology to select the best CCHP scheme in the presence of uncertainties. Differing from previous studies, this study examined the uncertainties in CCHP systems and evaluated the impacts of these uncertainties on the operational decision-making process as well as the stochastic impacts on the decision making process of the given investment. In the first stage, the system has been evaluated as a sole CHP system in the light of the updated value of the variables, then the system has been designed as a CCHP system by adding the absorption chiller with the intention of covering the cooling demand partly or fully. Setting the correct load capacity and scheduling is important while deciding on whether the most profitable system should be CHP or CCHP for a given plant. For this propose, macro program in Microsoft Excel has been run in order to determine the most proper capacity for the absorption chiller that will maximize the total annual saving. After determining the most proper cooling load of the absorption chiller, the system has been re-evaluated in the economical aspect. In another subsection, sensitivity analysis has been applied with the purpose of seeing the impact of the variables on the result. Following this, a new formula has been created to analyse and calculate the effects of variables on the result of the objective function on a percentage basis. Genetic algorithm is used to see the best case scenario in given constrains of uncertain variables. The result of this forms a reference for the comparison of the actual situation with the best case scenario. As a last step, possible results of the total annual saving have been re-calculated by using probabilistic models under non-parametric stochastic method. The analyses conducted in first stage have specifically addressed the variables that affect the economic feasibility of the investment and the uncertainties that may affect the investment any time in the systems economic life span. The main objective is to analyse all the possibilities and changes of the uncertain parameters during the life of the system to help investors see the possibility of the occurrence of the best and worst case scenario before making investment decision. Moreover, it is shown that some certain criteria should be satisfied in order for CCHP power plants to be more feasible. The results concluded from this stage are mentioned more in details in the last section of the manuscript. This stage of study has revealed that an evaluation made solely by considering the current values of the variables of the system is not sufficient to analyze the profitability of the investment. Apart from the conventional evaluation, the random changes of the independent variables at any time should be evaluated in order to see how they affect the profitability. Accordingly, it has been concluded that the deterministic evaluation is not sufficient to assess CCHP systems by its own and the stochastic evaluation gives a broader point of view in terms of overseeing all possible risks. The first stage of the thesis presented a very wide range of possibilities to assess the profitability of the system. At second stage, a re-evaluation was carried out in order to make a clearer analysis considering the historical data of the independent variables that affect the applicability of the system and the correlations between the variables, if any, and the probability density functions of the variables. Second stage of the study has been aimed to estimate how the profitability of a CCHP system, which is considered investable based on current values, will change throughout its economic life by adopting stochastic methods. Accordingly, the system has been analysed under four different simulation methods, namely parametric method, historical trend method, Monte Carlo method and scenario-based method, and their results have been compared. Among all the studied methods, the Monte Carlo and the historical trend methods directly take historical data as a reference. The parametric method, on the other hand, uses only the parameters of the mean and standard deviation from the historical data as a reference and thereafter assumes that all parameters will follow the normal distribution. Differing from these methods, the scenario-based method tries to determine where the objective function will be concentrated by considering all probable scenarios. In this regard, the parametric method gives results across the widest range, offering an unclear prediction about future results. The Monte Carlo method gives the highest mean value, while the historical trend method gives probabilities in a narrower range. The scenario-based method, meanwhile, offers a broader prediction than the historical trend method and also predicts a lower mean value for tas. Second stage of the study has showed that Investments in energy systems, including CCHP systems, face uncertainty. To answer whether an investment will remain profitable in the midst of these uncertainties, different methods can be applied either using past data or considering all possible scenarios. Although each method used in this study has certain advantages and disadvantages, all four methods can be used to evaluate CCHP systems at the investment stage. Since prices in almost all countries, particularly in the energy market, may not move in line with the historical trend, this study has shown that the scenario-based method is most appropriate to adopt given the comparisons and contrasts it provides.
-
ÖgeIncipient Fault Detection in Wind Turbines(Energy Institute, 2019-06-21)The global goal of increasing the share of renewable energy supplies in the overall energy consumption has resulted in a rising focus on technological developments in this field. Wind energy is one of the promising options amongst renewable energy sources with a growing number of investments and rising installation number and capacities. Due to the increasing demands from wind energy industry, the requirement of more effective wind farm operations has emerged. Wind turbine maintenance systems are essential parts towards achieving this requirement. Today, maintenance of wind turbines is mostly based on preventive and corrective actions. However, these approaches are inadequate to meet current demands from wind energy industry. With the developments in computational capabilities and data collection systems, a high potential of using advanced data-driven techniques has appeared for the maintenance of wind turbines. This thesis proposes a predictive maintenance approach using data which were collected from a wind turbine Supervisory Control and Data Acquisition System (SCADA). SCADA is the primary interface between the wind farm operators and wind turbines which allows remote and local control and monitoring. Various kinds of data are collected by SCADA systems such as wind parameters, temperature values, operational and status data. It is a built-in part in most medium and large-scale modern wind turbines. Therefore, a major advantage of using SCADA data for fault detection purposes is that additional hardware costs are not required. However, there are imperfections in the data such as low sampling frequency and high ratio of missing values. To handle these disadvantages, a suitable approach is required which was provided by Artificial Neural Networks (ANN) in this thesis. Moreover, wind turbines are highly non-linear systems with complex control parts and ANN models are also powerful on handling such complex systems. By this way, this thesis aims to design a cost-effective maintenance system for the overall wind turbine. Firstly, a sensor validation technique to detect faults of temperature sensors was designed. The method solely uses sensor measurements to detect calibration drifts by analyzing a set of sensors located on components with similar temperature characteristics. Auto-Associative and Multi-Input-Single-Output ANN structures were employed. The concurrent use of them provided the best outputs on the detection of the simulated calibration drift. The results prove that, validation of sensors can be realized by continuously monitoring sensor readings. It is advantageous as there is no need of dismantling sensors to test their calibration. Also, this method is a cost-effective solution in terms of not requiring redundant sensor use. After the sensor validation part, a 3-level fault classification system to detect, isolate and predict wind turbine faults was realized. The types of faults attempted in this part are frequent and non-fatal wind turbine faults. Distinguishing these kind faults is a challenging task because they do not show as strong indications as fatal faults do. However, as they are observed frequently in all wind turbines and decrease turbine performance, detection of them is a significant research topic. The core part of algorithms employed in this part is ANN models, in addition to them assistive methods were also designed to increase the fault classification performance. For the initial step of this part, feature construction and selection techniques were employed to find out an effective subset of inputs to be used as inputs of ANN models. These pre-processing tasks are important to design fast and accurate models as performance of algorithms strongly depend on the feature representation of input data in artificial intelligence applications. Raw data collected by the SCADA system were used to generate new features that possibly give more information about the hidden relations indicating fault occurences comparing to the raw features. In the feature selection step, both raw and constructed features were analyzed to identify a subset of relevant features to reduce computational burden and increase accuracy of models. Two different feature selection methods were used in a hybrid way, which are filter and wrapper-based methods. The results show that, the feature construction and selection algorithms designed are useful especially in terms of reducing false fault alarms which is an important issue in fault detection systems built using SCADA data. Finally, a 3-level classification scheme for wind turbine faults was designed using ANN models. By this way, a complete system was formed that provides required information by wind farm operators to take actions or measures in case of a current or an upcoming fault. In the detection level, the status of the turbine was analyzed to find out if the turbine is in a normal or a faulty mode. In the fault isolation level, the specific subsystem subjected to fault was attempted to be found. Therefore, this level includes distinguishing detected faults from each other. Finally, in the fault prediction level it was aimed to predict faults in advance to inform operators for possible prevention or repairing actions. We have obtained comprehensive results proving that the proposed methods are effective in all levels of fault classification. Our findings support the idea that despite the shortcomings of SCADA data, ANN models used with assistive methods are powerful on the classification of wind turbine faults. As a result, this thesis contributes to efforts of designing a cost-effective predictive maintenance approach for wind turbines.
- 1 (current)
- 2
- 3
- 4
- 5