Development of fast and robust spectral decomposition method for nonlinear industrial loads with current frequency spectrum rich in harmonic and interharmonic content

---

Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Fen Bilimleri Enstitüsü, Fen Bilimleri Enstitüsü, Türkiye

Tezin Onay Tarihi: 2016

Öğrenci: EDA UZ LOĞOĞLU

Danışman: MUAMMER ERMİŞ

Özet:

Nonlinear industrial loads such as AC arc furnaces and induction melting furnaces with resonant converters exhibit highly distorted currents rich in harmonic and interharmonic frequency components. Detection of these harmonics and interharmonics is crucial for applying countermeasures to achieve power quality requirements defined by the regulations of the power system operator. Fast and accurate harmonic and interharmonic detection in a power system experiences some difficulties due to the fact that power system frequency varies in time while the harmonic and interharmonic content is also time-varying. Hence Fourier Analysis based methods are slow and inaccurate to be used with compensation systems such as active filters. The aim of this thesis is to develop methods both fast and robust under time-varying fundamental frequency and spectrum conditions as well as in cases of unbalanced currents drawn by nonlinear industrial loads. Therefore, in this thesis, a novel method which combines the multiple synchronous reference frame (MSRF) analysis with the Kalman filter is proposed in order to take the advantage of the prediction capability of the Kalman filter and the accurate performance of the MSRF analysis. This algorithm, which provides the positive- and negative-sequences of all harmonic and interharmonic components under consideration, is implemented on the NVIDIA Jetson TX1 graphics processing unit (GPU). Results proves that the decomposition of the harmonic and the interharmonic components is achieved accurately with a microsecond latency that allows the proposed algorithm to be used in a real-time operation.