Sezgisel optimizasyon algoritmaları ile enerji hub optimizasyonu
Yükleniyor...
Dosyalar
Tarih
2022
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Düzce Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Bu tez çalışmasında, rüzgar ve güneş enerjisi entegreli bir enerji hub (EH) modeli önerilmiştir. Önerilen modelin optimizasyonu için hibrit bir algoritma geliştirilmiştir. Bu kapsamda, tez çalışmasında genel olarak iki konu üzerinde çalışılmıştır. İlk olarak, enerji hub girişindeki enerji kaynaklarının elektrik, rüzgar enerjisi, güneş enerjisi, doğal gaz ve ısı olduğu, çıkışta ise elektrik, ısı, soğutma ve basınçlı hava üretilen bir model önerilmiştir. Bu model temel alınarak 69 EH yapısı oluşturulmuştur. Bu EH yapıları kullanılarak, dört farklı boyutta test sistemi oluşturulmuştur. EH optimizasyonunun ele alındığı çalışmada, toplam enerji hub maliyetinin minimize edilmesi, toplam enerji hub kayıplarının minimize edilmesi ve hem enerji hub maliyetinin hem de kayıplarının minimize edilmesi olarak üç amaç fonksiyonu kullanılmıştır. Literatürde toplam enerji hub maliyetinin minimize edildiği problemlerin aksine, bu çalışmada rüzgar ve güneş enerjisinin maliyet modeli doğrudan maliyet, ceza maliyeti ve rezerv maliyeti olarak hesaplanmıştır. Böylelikle, literatüre yeni bir problem önerilmiştir. Tez çalışmasında ele alınan ikinci konu, meta-sezgisel arama algoritma tasarımıdır. Buna göre, yarı parametre uyarlamalı LSHADE ile hibrit CMA-ES (LSHADE with Semi-Parameter Adaptation Hybrid with CMA-ES, LSHADE-SPACMA) algoritmasının erken yakınsama problemini ortadan kaldırmak amacıyla arama performansının geliştirilmesi amaçlanmıştır. Bu nedenle, algoritmada kullanılan mutasyon stratejisi uygunluk-mesafe dengesi (Fitness Distance Balance, FDB) yöntemi kullanılarak yeniden tasarlanmıştır ve önerilen algoritma FDB-LSHADESPACMA olarak isimlendirilmiştir. FDB-LSHADESPACMA algoritmasının performansını test etmek ve doğrulamak için CEC2017 ve CEC2020 kıyaslama problemleri kullanılarak kapsamlı bir deneysel çalışma gerçekleştirilmiştir. Bu doğrultuda, literatürde sıklıkla kullanılan ve güncel 14 meta-sezgisel arama (MSA) algoritması seçilmiştir. Çalışmada önerilen enerji hub optimizasyon problemleri, FDB-LSHADESPACMA algoritması ile çözülmüştür. Buna göre, EH yapıları kullanılarak oluşturulan dört test sistemi ve üç amaç fonksiyonu kullanılarak on iki durum çalışması oluşturulmuştur. FDB-LSHADESPACMA algoritması ve rakip MSA algoritmalarından elde edilen tüm sonuçlar, Friedman Testi ve Wilcoxon ikili işaret testi kullanılarak değerlendirilmiştir. Elde edilen sonuçlara göre, FDB-LSHADESPACMA algoritmasının rakip algoritmalardan daha iyi sonuçlar elde ettiği ve üstün performans sergilediği görülmüştür.
In this thesis study, an energy hub (EH) model integrated with wind and solar energy was proposed. Moreover, a hibrit optimization algorithm was developed for the optimization of proposed energy hub model. In this context, the thesis study studied on two issues. Firstly, a model was proposed where the energy sources at the input of the energy hub were electricity, wind energy, solar energy, natural gas and heat, while electricity, heat, cooling and compressed air were produced at the output. Based on this model, 69 EH structures were created. By using these EH structures, four different scale test systems were created. In the study in which EH optimization was handled, three objective functions were used as minimizing the total energy hub cost, minimizing the total energy hub losses, and minimizing both the energy hub cost and losses. Contrary to the problems in the literature in which the total energy hub cost was minimized, the cost model of wind and solar energy was calculated as direct cost, penalty cost and reserve cost. Thus, a new problem has been proposed to the literature. The second topic addressed in the thesis study was the design of the meta-heuristic search (MHS) algorithm. Accordingly, it was aimed to improve the search performance of the LSHADE with Semi-Parameter Adaptation Hybrid with CMA-ES (LSHADE-SPACMA) algorithm in order to eliminate the premature convergence problem. For this reason, the mutation strategy used in the algorithm was redesigned using the Fitness-Distance Balance (FDB) method and the proposed algorithm was named FDB-LSHADESPACMA. A comprehensive experimental study was carried out using the CEC2017 and CEC2020 benchmark problems to test and validate the performance of the FDB-LSHADESPACMA algorithm. In this direction, 14 up-to-date meta-heuristic search (MHS) algorithm were selected. Energy hub optimization problems given in this study were solved with the proposed algorithm. Accordingly, twelve case studies were created using four EH test systems and three objective functions. All results obtained from FDB-LSHADESPACMA algorithm and competitor MSA algorithms were evaluated using Friedman Test and Wilcoxon signed-rank test. According to results, it was shown that the FDB-LSHADESPACMA algorithm achieved better results and showed superior performance than the competing algorithms.
In this thesis study, an energy hub (EH) model integrated with wind and solar energy was proposed. Moreover, a hibrit optimization algorithm was developed for the optimization of proposed energy hub model. In this context, the thesis study studied on two issues. Firstly, a model was proposed where the energy sources at the input of the energy hub were electricity, wind energy, solar energy, natural gas and heat, while electricity, heat, cooling and compressed air were produced at the output. Based on this model, 69 EH structures were created. By using these EH structures, four different scale test systems were created. In the study in which EH optimization was handled, three objective functions were used as minimizing the total energy hub cost, minimizing the total energy hub losses, and minimizing both the energy hub cost and losses. Contrary to the problems in the literature in which the total energy hub cost was minimized, the cost model of wind and solar energy was calculated as direct cost, penalty cost and reserve cost. Thus, a new problem has been proposed to the literature. The second topic addressed in the thesis study was the design of the meta-heuristic search (MHS) algorithm. Accordingly, it was aimed to improve the search performance of the LSHADE with Semi-Parameter Adaptation Hybrid with CMA-ES (LSHADE-SPACMA) algorithm in order to eliminate the premature convergence problem. For this reason, the mutation strategy used in the algorithm was redesigned using the Fitness-Distance Balance (FDB) method and the proposed algorithm was named FDB-LSHADESPACMA. A comprehensive experimental study was carried out using the CEC2017 and CEC2020 benchmark problems to test and validate the performance of the FDB-LSHADESPACMA algorithm. In this direction, 14 up-to-date meta-heuristic search (MHS) algorithm were selected. Energy hub optimization problems given in this study were solved with the proposed algorithm. Accordingly, twelve case studies were created using four EH test systems and three objective functions. All results obtained from FDB-LSHADESPACMA algorithm and competitor MSA algorithms were evaluated using Friedman Test and Wilcoxon signed-rank test. According to results, it was shown that the FDB-LSHADESPACMA algorithm achieved better results and showed superior performance than the competing algorithms.
Açıklama
Anahtar Kelimeler
Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
