scispace - formally typeset
Search or ask a question

Showing papers by "Mustafa Şekkeli published in 2019"



Journal ArticleDOI
TL;DR: The dynamic performance of the Neuro-Fuzzy controller and the conventional PI controller in different system conditions (such as reference voltage tracking and output load change) of buck type DC-DC converter is shown.
Abstract: Buck type DC-DC converter circuit topology is non-linear due to their switched circuit structure. Conventional control systems are insufficient to control non-linear systems. Neural Networks have important abilities such as learning, optimizing and adaptability. Fuzzy logic and neural networks are used as an adaptive structure based on the fuzzy logic controller. This adaptive structure adjusts the properties of the fuzzy rules and the characteristics of the control system so that the Neuro-Fuzzy controller can be adapted to all different system conditions. In this study, experimental studies were carried out on the dSPACE experiment platform to show the dynamic performance of the Neuro-Fuzzy controller and the conventional PI controller in different system conditions (such as reference voltage tracking and output load change) of buck type DC-DC converter.

6 citations


Journal ArticleDOI
TL;DR: The dynamic performance of the intelligent controller structure in the control of an EPT controlled by NFC is presented, to demonstrate the voltage sag, swell, harmonic, and reactive power compensation performances of the EPT.
Abstract: Along with the rapid advances in power electronics and semiconductor technology, electronic power transformers (EPTs), which are expected to replace conventional power transformers in the future, are being developed and designed. This study presents the dynamic performance of the intelligent controller structure in the control of an EPT. The EPT structure, consisting of the input, isolation, and output stages, is designed for experimental and simulation studies. A three-phase pulse width modulation (PWM) rectifier is used to rectify the grid voltages at the input stage of EPT. The DC-bus voltage of the three-phase PWM rectifier is controlled by a neuro-fuzzy controller (NFC) against the disturbances that may occur in the input stage. In the isolation stage of the EPT, a DC-DC converter circuit is used. Thus, DC voltage obtained from the input part is reduced to the appropriate voltage values. In the output stage, the two-level three-phase inverter circuit is used to provide the necessary voltages for users. The control algorithms for input and output stages of the EPT have been developed using a dSPACE DS1104 controller card, which can work in real time with MATLAB/Simulink. Experimental and simulation studies are realized to demonstrate the voltage sag, swell, harmonic, and reactive power compensation performances of the EPT controlled by NFC.

3 citations


Proceedings ArticleDOI
01 Sep 2019
TL;DR: Type-2 Neuro-Fuzzy Controller which has robust and adaptive structure is proposed for controlling of all stage of SST and the Elliptic Membership Function (MF) is used for T2NFC.
Abstract: As it is known, the most basic and important components in power systems are classical transformers which are passive interface between high voltage and low voltage systems. Classical transformers have undesirable characteristics such as poor voltage regulation, large size/weight and sensitivity to harmonics etc. In parallel with technological developments, solid state transformers (SSTs) which are considered to be one of the most indispensable components for future electrical energy systems, have attracted intense interest. In this paper, SST structure consisting of three stages is modeled in Matlab/Simulink environment. Type-2 Neuro-Fuzzy Controller (T2NFC) which has robust and adaptive structure is proposed for controlling of all stage of SST. In addition, the Elliptic Membership Function (MF) is used for T2NFC. The proposed controller is applied to SST structure and simulation studies are performed to verify the performance of the controller.

3 citations


30 Dec 2019
TL;DR: In this article, Dogrusal devre topolojilerinden farkli olarak, anahtarlamali yapilarindan dolayi DA-DA donusturucu devrelerinin modellenmesi ve denetimi oldukca zordur.
Abstract: Yukselten DA-DA donusturucu, giris gerilim seviyesini giris gerilim seviyesinden daha yuksek bir gerilim seviyesine donusturen bir guc elektronigi devresidir. Dogrusal devre topolojilerinden farkli olarak, anahtarlamali yapilarindan dolayi DA-DA donusturucu devrelerinin modellenmesi ve denetimi oldukca zordur. Durum degiskenleri yontemi ile sistem modelleme denetleyici tasariminda etkin bir sekilde kullanilmaktadir. Durum degiskenleri yontemi ile modellenen sistemin belirli bir calisma noktasinda kucuk sinyal analizi yapilarak sistemin istenilen transfer fonksiyonu elde edilmektedir. Optimal denetim yontemlerinde denetlenen sistemin transfer fonksiyonuna gore tespit edilen denetleyici parametreleri, bilinen bir hata kriteri kullanilarak minimize edilmektedir. LQR denetim literaturde en cok kullanilan optimal denetim yontemlerinden biridir. Bu calismada yukselten DA-DA donusturucunun durum uzayi kucuk sinyal modeli Matlab/Simulink benzetim programinda olusturularak LQR denetleyici ile optimal denetim yapilmistir.

15 Dec 2019
TL;DR: Turkiye cografi konumu nedeniyle buyuk bir gunes enerjisi potansiyeline sahiptir ve ulkedeki Gunes Enerjisi Santrali (GES) kurulu gucu hizla artmaktadir as mentioned in this paper.
Abstract: Turkiye cografi konumu nedeniyle buyuk bir gunes enerjisi potansiyeline sahiptir ve ulkedeki Gunes Enerjisi Santrali (GES) kurulu gucu hizla artmaktadir. Fakat GES uretimlerindeki degiskenlik bu tesislerin elektrik sebeke sisteminde isletilmesini problem haline getirmektedir. Bu problemin giderilmesi icin GES uretim tahminlerine ihtiyac duyulmaktadir. Bu calismada GES uretimleri icin bir tahmin sistemi onerilmistir. Onerilen tahmin sisteminde Yapay Sinir Aglari (YSA) kullanilmistir. YSA Levenberg-Marquardt ogrenme algoritmasi kullanilarak egitilmistir. YSA egitim, dogrulama ve test sureclerinde Kahramanmaras ilinde yer alan GES gecmis uretim degerleri ve Kuresel Tahmin Siteminden (KTS) alinan bulutluluk tahmin verileri kullanilmistir. YSA yapisi ve giris degerleri degistirilerek analizler yapilmistir. Yapilan analizler sonucunda giris olarak uretim degerleri ile birlikte bulutluluk tahmini kullanan YSA’nin sadece gecmis uretim degerleri kullanilan YSA’ya kiyasla daha basarili oldugu tespit edilmistir.

Journal ArticleDOI
30 Jun 2019
TL;DR: In this article, Matris Laboratuvari (MATLAB) model, Matris Labels (LAB) yazilim gelistirme ortaminda olusturulmustur.
Abstract: Deniz ustu Ruzgâr enerjisi santrali (RES) kurulumunda en onemli konulardan birisi de ruzgâr turbinlerinin iz etkisidir. Bu durum RES'in yillik uretimini dusurmektedir. Bu nedenle RES kurulumunda iz etkisi sonucu meydana gelen uretim kaybi dikkate alinarak RES kurulumu gerceklestirilmelidir. Gunumuzde iz etkisi hesabinda bircok yontem kullanilmaktadir. Bu yontemler arasinda basit ve yuksek derecede dogruluga sahip olan Jensen yontemi daha cok tercih edilmektedir. Bu calismada, iki ruzgâr turbini arasindaki iz etkisi, bir bolgede ayni yukseklikteki degisken ruzgâr hizlari ve frekanslari icin Jensen yontemi kullanilarak modellenmistir. Model, Matris Laboratuvari (MATLAB) yazilim gelistirme ortaminda olusturulmustur. Ayrica olusturulan modeli dogrulamak icin Ruzgâr Atlasi Analiz ve Uygulama Programi (WAsP) yazilimi kullanilmistir. Bu yazilimda iki ruzgâr turbini deniz kiyisinda uzakta bir bolgeye yerlestirilmistir. Bu duruma iliskin turbinlerin brut uretim ve net uretim degerleri hesaplanmistir. MATLAB modeli sonuclari ve ticari olarak kullanilan WAsP modeli sonuclari karsilastirildiginda yaklasik olarak ayni uretim degerleri elde edilmistir. Bu sonuclar gelistirilen modelin deniz ustu RES’lerde yapilacak hesaplamalarda kullanilabilecegini gostermektedir.