Kesirli bulanık çıkarım sistemi kullanılarak denetleyici tasarımı
Küçük Resim Yok
Tarih
2024
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İnönü Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Bulanık akıl yürütme yöntemleri, belirsizlik içeren problemlerin çözümüne odaklanan yaklaşımlar olarak ifade edilebilir. Günümüzde bulanık mantık yürütme yöntemlerinin daha etkili hale getirilmesine yönelik artan bir ilgi bulunmaktadır. Daha önce yapılan bir çalışmada, fraksiyonel akıl yürütme temel alınarak kesirli indislerin uygun şekilde ayarlanmasıyla bulanık mantık uygulamalarının performansını artırabileceğini ileri süren bir yöntem olan kesirli bulanık çıkarım sistemi (FFIS) önerilmiştir. Bu tezde yapılan çalışmalarda FFIS kullanarak iki farklı sistem için denetleyiciler tasarlanmış, giriş ve çıkış değişkenlerinin evrensel küme aralıkları iki farklı yöntemle ayarlanan Mamdani Bulanık Çıkarım Sistemi (FIS) ile kıyaslamaları yapılmıştır. Birinci çalışmada deneme yöntemi ile ayarlanan Mamdani FIS kullanılarak doğru akım (DC) motorunun pozisyon kontrolü gerçekleştirilmiştir. Daha sonra mevcut Mamdani FIS, farklı kesirli indisler ile oluşturulan FFIS kullanılarak elde edilen sonuçlar ile kıyaslanmıştır. İkinci çalışmada Genetik algoritma (GA) optimizasyon yöntemi ile ayarlanan Mamdani FIS ile klasik yöntemlerle kontrolün zor olduğu düşünülen yüksek dereceli bir sistem için kontrol uygulaması gerçekleştirilmiştir. GA ile ayarlanan Mamdani FIS, uygun kesirli indisler seçilerek oluşturulmuş; FFIS ile değiştirilerek kontrol sonuçları kıyaslanmıştır. FFIS kullanılarak yapılan kontrol uygulamalarında Mamdani FIS tabanlı kontrol uygulamalarına göre daha başarılı sonuçlar elde edilmiştir.
Fuzzy reasoning methods can be expressed as approaches that focus on solving problems involving uncertainty. Nowadays, there is an increasing interest in making fuzzy reasoning methods more effective. In a previous study, Fractional Fuzzy Inference System (FFIS) was proposed as a method based on fractional reasoning, suggesting that it can increase the performance of fuzzy logic applications by appropriately adjusting fractional indices. In the studies carried out in this thesis, controllers were designed for two different systems using FFIS and comparisons were made with the Mamdani Fuzzy Inference System (FIS), where the universal cluster ranges of the input and output variables were adjusted by two different methods. In the first study, position control of the Direct Current (DC) motor was carried out using Mamdani FIS, which was adjusted by trial method. Then, the existing Mamdani FIS was changed to the FFIS obtained using different fractional indices and the control results were compared. In the second study, a control application was carried out with Mamdani FIS adjusted by the Genetic algorithm (GA) optimization method for a high-order system that is thought to be difficult to control with classical methods. Then, the Mamdani FIS adjusted with GA was replaced with the FFIS created by selecting appropriate fractional indices and the control results were compared. As a result, the control applications using the FFIS have achieved more successful results than the Mamdani FIS-based control applications.
Fuzzy reasoning methods can be expressed as approaches that focus on solving problems involving uncertainty. Nowadays, there is an increasing interest in making fuzzy reasoning methods more effective. In a previous study, Fractional Fuzzy Inference System (FFIS) was proposed as a method based on fractional reasoning, suggesting that it can increase the performance of fuzzy logic applications by appropriately adjusting fractional indices. In the studies carried out in this thesis, controllers were designed for two different systems using FFIS and comparisons were made with the Mamdani Fuzzy Inference System (FIS), where the universal cluster ranges of the input and output variables were adjusted by two different methods. In the first study, position control of the Direct Current (DC) motor was carried out using Mamdani FIS, which was adjusted by trial method. Then, the existing Mamdani FIS was changed to the FFIS obtained using different fractional indices and the control results were compared. In the second study, a control application was carried out with Mamdani FIS adjusted by the Genetic algorithm (GA) optimization method for a high-order system that is thought to be difficult to control with classical methods. Then, the Mamdani FIS adjusted with GA was replaced with the FFIS created by selecting appropriate fractional indices and the control results were compared. As a result, the control applications using the FFIS have achieved more successful results than the Mamdani FIS-based control applications.
Açıklama
Anahtar Kelimeler
Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
