PSO Training Neural Network MPPT with CUK Converter Topology for Stand-Alone PV Systems Under Varying Load and Climatic Conditions


Yılmaz M., Corapsiz M.

Türk Doğa ve Fen Dergisi, cilt.13, sa.1, ss.88-97, 2024 (Hakemli Dergi) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 13 Sayı: 1
  • Basım Tarihi: 2024
  • Doi Numarası: 10.46810/tdfd.1423852
  • Dergi Adı: Türk Doğa ve Fen Dergisi
  • Derginin Tarandığı İndeksler: TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.88-97
  • Atatürk Üniversitesi Adresli: Evet

Özet

PV panellerden elde edilen güç değeri sıcaklık ve ışınım değerleri gibi çevresel faktörlere bağlı olarak değişmektedir. Bu nedenle değişen iklim koşullarında PV panelden maksimum güç değerinin yüke aktarılması için maksimum güç noktası izleme (MPPT) algoritmaları ve DC-DC dönüştürücü topolojileri kullanılmaktadır. Bu çalışmada, parçacık sürü optimizasyonu (PSO) öğrenen sinir ağı tabanlı MPPT algoritması kullanılarak, yükselten tip dönüştürücü ve CUK dönüştürücü devre topolojilerinin performansları değişken ışınım ve değişken yük koşulları altında incelenmiştir. İlk senaryo olarak panele gelen sıcaklık ve ışınım değerlerinin sabit olduğu varsayılarak analiz edilmiştir. İkinci senaryo olarak değişken yük durumu için akım, gerilim ve güç parametrelerine göre dönüştürücü topolojilerinin performans değerlendirmesi yapılmıştır. Son senaryo olarak ise gün içerisinde güneşin durumuna bağlı olarak panele gelen ışınım değerindeki farklılık incelenmiştir. Çalışmada Canadian Solar CS6P-250P PV panel kullanılmıştır. Anahtarlama frekansı olarak 50 kHz seçilmiştir. Elde edilen sonuçlara göre CUK dönüştürücü devre topolojisinin hem dinamik çevre koşullarında hem de yük değişiminde yükselten tip dönüştürücü devre topolojisine göre maksimum güç noktasına daha hızlı ulaştığı ve bu noktadaki salınımın daha az olduğu görülmüştür. Literatürde yükselten tip dönüştürücü topolojisi ve MPPT kullanan bu yöntemin performansının CUK dönüştürücü topolojisi uygulanarak arttırılması hedeflenmiştir. 


Temperature and irradiance levels are two examples of environmental variables that affect the power value produced by photovoltaic panels. Therefore, in order to transfer the maximum power value from the PV panel to the load under varying climatic conditions, maximum power point tracking (MPPT) algorithms and DC-DC converter topologies are used. In this study, the performances of boost converter and CUK converter circuit topologies are investigated under variable irradiance and variable load conditions by using a neural network-based MPPT algorithm learning particle swarm optimization (PSO). As the first scenario, it is analyzed assuming that the temperature and irradiance values coming to the panel are constant. As the second scenario, the performance evaluation of the converter topologies according to the current, voltage and power parameters is made for the variable load situation. As the last scenario, the difference in the irradiance value coming to the panel depending on the sun's condition during the day has been examined. Canadian Solar CS6P-250P PV panel is used in the study. 50 kHz is selected as the switching frequency. According to the results obtained, it has been observed that the CUK converter circuit topology reaches the maximum power point faster than the boost converter circuit topology both in dynamic environmental conditions and load change, and the oscillation at this point is less. It is aimed to increase the performance of this method, which uses boost converter topology and MPPT in the literature, by applying CUK converter topology.