Research Of Power Quality Monitoring Method Based On Improved All Phase FFT And TT-Transform

In recent years, people put forward higher requirement to power quality. But the structure of electrical load has changed, so the power quality has been seriously polluted. Therefore, governance and improvement of the power quality is imperative. Though, power quality monitoring as the basis of power quality control has been extensively studied, there are still many issues to be studied further.The definition and classification of power quality problems is discussed. Then the harmonic of steady-state and voltage sag, voltage swell, voltage interruption, transient pulse, transient oscillation of transient disturbances are mainly studied.In the harmonic detection, windowed interpolation and correction FFT is analysed and the results of harmonic detection based on four common window functions are compared. Because the accuracy of phase detection based on windows interpolated and correction FFT is not high, harmonic phase detection method based on double blackman all-phase FFT is referenced for the first time. The experimental results show that the double blackman all-phase FFT has huge superiority.In the disturbance detection, the transient power quality disturbance detection based on S-transform is analysed. As to the imprecise positioning of S-transform, disturbance detection method based on TT-module matrix's line maximum sequence is put forward. The experimental results show that this method has a better positioning effect for not only single disturbances, but also complex disturbances.In the feature extraction of power quality disturbances, the disturbance signals are decomposesd with db4 wavelet packet and energy information of ten designated node are extracted. In order to make feature vectors more comprehensive, the feature extraction method based on TT-transform is analysed. Eventually, a total of twelve feature vectors are applied to classifier.In the disturbance identification, because the common classification methods such as neural networks and support vector machine have drawbacks, so relevance vector machine is adopted in this thesis. Then power quality disturbances are learned and classified based on binary tree relevance vector machine. The disturbance identification results show that binary tree relevance vector machine has advantages of real-time and accuracy.