No abstract

Günümüzde bilişim teknolojileri hemen hemen her alanda kullanılmaktadır. Enerji sektörü de bu alanlardan birisidir. Nüfusun gün geçtikçe artmasıyla birlikte bina sayısı ve binaların enerji talebi de artmıştır. Enerji talebini hafifletmenin bir yolu enerji tasarrufu özelliklerine sahip verimli binalar tasarlamaktır. Bu çalışmada sekiz giriş değeri (nispi yoğunluk, yüzey alanı, duvar alanı, çatı alanı, toplam yükseklik, yönlendirme, cam alanı ve cam alanı dağılımı) ve iki çıkış değeri (ısıtma yükü (HL), soğutma yükü (CL)) olan bir veri setinin, makine öğrenmesi algoritmaları kullanarak analizi yapılmıştır. Amaç, konutların ısıtma ve soğutma yükünü tahmin edebilen bir model oluşturmaktır. Bu parametrelerin doğru bir şekilde tahmin edilmesi, enerji tüketiminin daha iyi kontrol edilmesini kolaylaştırmakta ve ayrıca, enerji piyasasında önemli bir sorun olarak görülen enerji ihtiyacına daha iyi uyan enerji tedarikçisinin seçiminde yardımcı olmaktadır. Bu kapsamda, veri seti analiz edilirken makine öğrenmesi algoritmalarından regresyon algoritmaları (Destek Vektör Makinesi (SVM) Regresyonu, Doğrusal Regresyon, Rasgele Orman Regresyonu ve En Yakın Komşu Regresyonu) kullanılmıştır. İki çıkış değeri için sonuçlar deneysel olarak her algoritma için ayrı ayrı hesaplanmış ve elde edilen sonuçlar karşılaştırılmıştır. Çıkan sonuçlara göre analiz yaptığımız veri seti için, tahmin başarımı açısından en yakın sonucu bulan algoritma Rastgele Orman Regresyon algoritması olmuştur.

Bu çalışmada Genetik Algoritma (GA) ve Destek Vektör Makinelerinden (DVM) oluşan melez bir yöntemin CUDA (Compute Unified Device Architecture-Birleşik Hesaplama Aygıt Mimarisi) tabanlı hız optimizasyonu gerçekleştirilmiştir. Makine öğrenmesinde, geliştirilen yöntemlerin yüksek doğruluk değerlerinde başarı vermesi hedeflenir. Ayrıca önerilen algoritmanın sonuçları bulurken hızlı bir şekilde çalışması da yine hedeflenen bir durumdur. Bu çalışmada, özellikle gerçek zamanlı uygulamalarda önemli bir parametre olan hız parametresi dikkate alınmakta ve verilerin hızlı bir şekilde sınıflandırılması için yeni bir GPU (Graphic Processing Unit-Grafik İşlemci Birimi) teknolojisi kullanılmaktadır. Bunun için grafik işlemciler üzerinde programlama yapmamızı sağlayan CUDA programlamadan yararlanılmıştır. Sınıflandırma algoritması olarak genetik algoritmayla optimize edilmiş destek vektör makinesi kullanılmıştır. Deneyler 384 CUDA çekirdeğinden oluşan NVIDIA GeForce 940MX ekran kartına sahip bir bilgisayar üzerinde gerçekleştirilmiştir. Büyük ölçekli veri kümeleri üzerinde yapılan deneylerde, CUDA programlamanın sonuçlar üzerinde pozitif etkilerinin olduğu görülmüştür. Bu şekilde makine öğrenmesi uygulamalarında sınıflandırma aşamasında grafik işlemciler ile gerçek zamanlı uygulamalar için hızlı bir sistemin altyapısı oluşturulabilir.

Geothermal energy is a renewable form of energy, however due to misuse, processing and management issues, it is necessary to use the resource more efficiently. To increase energy efficiency, energy systems engineers carry out careful energy control studies and offer alternative solutions. With this aim, this study was conducted to improve the performance of a real operating air-cooled organic Rankine cycle binary geothermal power plant (GPP) and its components in the aspects of thermodynamic modeling, exergy analysis and optimization processes. In-depth information is obtained about the exergy (maximum work a system can make), exergy losses and destruction at the power plant and its components. Thus the performance of the power plant may be predicted with reasonable accuracy and better understanding is gained for the physical process to be used in improving the performance of the power plant. The results of the exergy analysis show that total exergy production rate and exergy efficiency of the GPP are 21 MW and 14.52%, respectively, after removing parasitic loads. The highest amount of exergy destruction occurs, respectively, in condenser 2, vaporizer HH2, condenser 1, pumps 1 and 2 as components requiring priority performance improvement. To maximize the system exergy efficiency, the artificial bee colony (ABC) is applied to the model that simulates the actual GPP. Under all the optimization conditions, the maximum exergy efficiency for the GPP and its components is obtained. Two of these conditions such as Case 4 related to the turbine and Case 12 related to the condenser have the best performance. As a result, the ABC optimization method provides better quality information than exergy analysis. Based on the guidance of this study, the performance of power plants based on geothermal energy and other energy resources may be improved.

Diabetes is a life-long illness which occurs as a result of lack of insulin hormone or ineffectiveness of insulin hormone. Blood sugar, fructosamine, and hemoglobin A1c (HbA1c) values are widely used for diagnosis of this disease. Although the role of insulin in diagnosing diabetes is great, the HbA1c value is more accurate. This is because HbA1c value gives information about the past two or three months of blood sugar in the treatment of diabetes. This study aims to estimate the HbA1c value with high accuracy. Follow-up data of diabetic patients were used as data. The Orange data mining software is used because it is easy to use in the modeling phase and contains many methods. In this context, the chapter aims to develop an effective prediction model by using a large number of feature selection and classification methods. The results show that the proposed model successfully predicts the HbA1c parameter. In addition, determination of the parameters that are effective in the diagnosis of diabetes has been carried out with the feature selection methods.

In this study, deep-neural-network (DNN)-and artificial-neural-network (ANN)-based models along with regression models have been developed to estimate the pressure, bending and elongation values of ground-brick (GB)-added mortar samples. This study is aimed at utilizing GB as a mineral additive in concrete in the ratios 0.0%, 2.5%, 5.0%, 7.5%, 10.0%, 12.5% and 15.0%. In this study, 756 mortar samples were produced for 84 different series and were cured in tap water (W), 5% sodium sulphate solution (SS5) and 5% ammonium nitrate solution (AN5) for 7 days, 28 days, 90 days and 180 days. The developed DNN models have three inputs and two hidden layers with 20 neurons and one output, whereas the ANN models have three inputs, one output and one hidden layer with 15 neurons. Twenty-five previously obtained experimental sample datasets were used to train these developed models and to generate the regression equation. Fifty-nine non-training-attributed datasets were used to test the models. When these test values were attributed to the trained DNN, ANN and regression models, the brick-dust pressure as well as the bending and elongation values have been observed to be very close to the experimental values. Although only a small fraction (30%) of the experimental data were used for training, both the models performed the estimation process at a level that was in accordance with the opinions of experts. The fact that this success has been achieved using very little training data shows that the models have been appropriately designed. In addition, the DNN models exhibited better performance as compared with that exhibited by the ANN models. The regression model is a model whose performance is worst and unacceptable; further, the prediction error is observed to be considerably high. In conclusion, ANN-and DNN-based models are practical and effective to estimate these values.

Monkeypox disease is caused by a virus that causes lesions on the skin and has been observed on the African continent in the past years. The fatal consequences caused by virus infections after the COVID pandemic have caused fear and panic among the public. As a result of COVID reaching the pandemic dimension, the development and implementation of rapid detection methods have become important. In this context, our study aims to detect monkeypox disease in case of a possible pandemic through skin lesions with deep-learning methods in a fast and safe way. Deep-learning methods were supported with transfer learning tools and hyperparameter optimization was provided. In the CNN structure, a hybrid function learning model was developed by customizing the transfer learning model together with hyperparameters. Implemented on the custom model MobileNetV3-s, EfficientNetV2, ResNET50, Vgg19, DenseNet121, and Xception models. In our study, AUC, accuracy, recall, loss, and F1-score metrics were used for evaluation and comparison. The optimized hybrid MobileNetV3-s model achieved the best score, with an average F1-score of 0.98, AUC of 0.99, accuracy of 0.96, and recall of 0.97. In this study, convolutional neural networks were used in conjunction with optimization of hyperparameters and a customized hybrid function transfer learning model to achieve striking results when a custom CNN model was developed. The custom CNN model design we have proposed is proof of how successfully and quickly the deep learning methods can achieve results in classification and discrimination.

No abstract