Research On A Combination Of Time-domain & Frequency-domain Analysis Method To Power Quality

With the rapid development of industrial technology and national economics, power quality has become an important index that every department of electrical companies pays much attention to. With the forming and boosting of electricity market reformation, the utility must meet the increasing requirement of not only power quantity but also power quality from customer.To effectively improve the power quality, we need to research on the disturbance phenomena. The identification and classification of disturbance is the premise to depress the wicked influences produced by them. This paper proposes an appropriate combination of time-domain and frequency-domain analysis method to research on power quality events. In time-domain, the analysis method on characters of instantaneous load is taken. In frequency-domain, the wavelet transform is used in this thesis. The result of the simulation indicates the usefulness and veracity of this new method.Fourier Transform, the most classical signal processing analysis method, is ideal for the calculation of magnitude of sinusoidal signals in their steady state. Based on this, this thesis brings out a exertion of Fast Fourier Transform (FFT) to detect the wave distortion disturbance. By the way, it also extracts of the magnitude features of short-term disturbance and long-term disturbance.The characteristic features of short-term disturbance and long-term disturbance, such as voltage sag and voltage swell, are concentrated on magnitude and time duration. Therefore, it is suitable to be analyzed in time-domain area. This paper innovates the characters of instantaneousload to analyze these disturbance events. After calculating the deviation factor by accounting the imitation resistance and imitation inductance, we can point the onset and the end time of the disturbance by rule and line. Then the idiographic type of the disturbance can be given out.In power system, the transient disturbance signal always represents as the singular signal. The characteristic features of them are mainly distributed in frequency domain area. Wavelet transform, an efficient mathematical tool for singularity detection, can offer plentiful time-frequency information on the signal processing. In this paper, a method using multiresolution analysis and modulus maximum based on wavelet transform is presented, bywhich the singularity and the main frequency containing can be accurately achieved.The practicality of the proposed algorithm in this thesis is validated by four digital simulations, which are voltage fluctuation, voltage sag, voltage swell, and impulsive disturbance. The results of the simulation show that the new analysis method based on the combination of time-domain and frequency-domain, in which the entropy is used as feature extraction, possesses efficiency and feasibility.