| The existence of numerous harmonic waves will severely endangers the safety operation of electric power system. Harmonic analysis is one of power quality research focuses. Harmonic parameters estimation provides important basis for power quality analysis, which becomes an important branch of harmonic analysis. Fourier analysis is one of most commonly used methods in electric harmonic detection. However, because of the fluctuation of system frequency, the harmonic detection based on Fourier analysis, which under asynchronous sampling, will rise to the phenomenon of frequency spectrum leakage and fence effect, which brings difficulty to accurate measurement of harmonic signals. Window-added and interpolated FFT can effectively reduce impact of spectrum leakage and fence effect, and enhance the accuracy of harmonic detection. Due to cosine window FFT interpolation is easy to structure and correction, it becomes one of the most widely used methods. This thesis analyzed cosine windowed interpolation and proposed an approach for harmonics detection based on power-type sine windowed interpolation FFT. In the approach, the sine window function not only inherits the cosine window function features, but also has more flexible forms, which extends the field of window function.The construction of windowed interpolation FFT method mainly includes two parts: one is windows function design and the other is interpolation function design. This thesis gave a review of many commonly used window function, fully referenced to the constructor methods for cosine windows, presented some thinking to improve Nuttall coefficient design, to the construct sine-class window. The thesis adjusted interpolation by three spectral lines method for cosine windows to sine windows, after grasping the basic idea of interpolation. Moreover, the thesis proposed convolution windows for the performance of windows as well as the accuracy of the harmonic analysis improvements. This thesis discovered the internal relations between cosine-class and sine-class windows by means of time-frequency analysis and mathematical formula, and defined power-type sine windows, promoting that the sine window with exponent one is the most basic unit for power-type sine windows. The definition gave an alternative approach for power-type sine windows construction.Combined with MATLAB software, this thesis made a comprehensive error analysis of factors having an influence on harmonic detection, which included the variety of fundamental frequency, sampling frequency, length of data, DC bias, noise level, quantization bits, harmonic content, and so on. The results indicated that, the power type sine windowed interpolation FFT method has high detecting precision and strong anti- interference capacity. In order to meet different demands of harmonic analyze, the thesis analyzed the time-domain and frequency-domain characteristics of different power type sine windows and their convolution windows with simulated tests.Finally, around the visual design conception, the thesis developed a platform for power type sine windowed interpolation FFT with powerful power type sine windows design and interpolation capability, which can carry out the analysis of integer harmonic efficiently. It can help designers to get rid of repetition of programming and enhance developing efficiency.This thesis showed that, the harmonic detection method based on power type sine windowed interpolation FFT performs good ability in parameter extraction, and has high-detection-rate and robust. Besides, functions are convenient for realization, flexible, and have good portability. It has theoretic and practical merits for offering a much wider choice of harmonic detecting methods for the power system. |
PDF Full Text Request