Study On New Windows And Improved FFT-based Harmonic Analysis Methods And Applications

The accurate harmonic analysis of power system is an important task for the research of harmonic pollution, which is also the basis for the identification of excessive harmonic flow, diagnosis of high voltage insulation, measurement of harmonic electric energy, compensation and suppression of power harmonic. The FFT is the most commonly used technique for power system harmonic analysis because of its simplicity and easy implementation in embedded system. However, the direct application of FFT for harmonic analysis may lead to inaccuracies due to the spectral leakage and the picket-fence effects with asynchronous sampling. Hence, it is of great theoretic and practical importance to study and implement novel FFT-based methods for accurate and real-time harmonic analysis.This dissertation proposes new kinds of windows and improved FFT-based harmonic analysis methods, especially on the Triangular self-convolution window (TRISCW), the cosine-combined convolution window(CCCW), the Hanning self-convolution window(HSCW), the Trapezoid self-convolution window(TRASCW), the improved FFT-based harmonic analysis methods with these proposed windows, the application of the proposed methods in harmonic electric energy measurement and dielectric loss angle online measurement.Firstly, the key factors of the FFT-based harmonic analysis method are analyzed. The TRISCW is proposed, which is constructed by self-convolutions of the Triangular window in time domain. With the increase of the self-convolution times, the spectral leakage can be more efficiently suppressed by using the TRISCW. The improved FFT-based harmonic analysis method based on the TRISCW is presented, where the harmonic parameters are calculated by spectral interpolation accurately.Then, the CCCW is proposed, and the cosine-combined cross-convolution window and the cosine-combined self-convolution window are defined and studied. The CCCW has the same major lobe width with the window who is one of the original window and also has a narrow major lobe width. The peak side lobe level and side lobe roll-off rate of the CCCW are the sum of the original windows. The study of the CCCW paves the way for the FFT-based harmonic analysis with the CCCW.This dissertation takes the Hanning window for example, proposes the HSCW and the improved FFT-based harmonic analysis method. The HSCW has a low peak side lobe level, a high side lobe roll-off rate. Hence, leakage errors and harmonic interferences can be considerably reduced. The improved FTT-based harmonic analysis method with the HSCW achieves better performance for applications where the spectral leakage is a main problem, such as the weak harmonic analysis of power system.Subsequently, based on the proposed TRISCW and CCCW, the TRASCW is proposed. The Trapezoid window is optimized by adjusting the window parameters, and the TRASCW with better sidelobe behavior is then obtained by self-convolutions in time domain. The TRASCW has narrow major lobe width, which makes it suitable for applications where the high frequency resolution is required, such as the inter-harmonic analysis.Finally, the proposed FFT-based harmonic analysis methods are implemented to the harmonic electric energy measurement and dielectric loss angle measurement. The harmonic electric energy measurement algorithm based on the HSCW and the dielectric loss angle measurement algorithm based on the TRISCW are proposed and realized in the embedded system, i.e. the digital signal processor. The testing and application results show that the three phase multi-function harmonic power energy meter achieves a precision of grade 0.2S for electrical power measurement, and completely fulfills the requirements for class A power harmonics measurement stated in GB/T14549-1993 Power Quality, Harmonics in Public Power Grid, the terminal can perform high accuracy of the dielectric loss angle online measurement, which can provide a basis for further practical application.