Detection And Classification Of Transient Power Quality Disturbance Based On HHT And SVM

With the propulsion of the national smart grid strategy and the rapid economicdevelopment of the power technology and the power economics, the amount of powerelectronic equipment in the power system has been being a sharp increase. theasymmetry, impact resistance and non-linear of the electric load in the system becomessignificant and affect to the power system and the losses of the community caused bypower quality problems is also gradually increased, so how to suppress the disturbingeffect and improve power quality has become an important aspect of the smart gridconstruction and power technology. For this, a method that studying power qualitydisturbances types, especially transient power disturbing types in aspect of type ofeffective detection, analysis and classification is the premise of improving the powerquality problems, meanwhile, is the main integral part of building the perfect powerquality monitoring system.The research status in the feature detection, location has a comparative study in theclassification and recognition to transient power quality disturbances in the beginning ofthis paper. In this, a combining identification method based on HHT method and SVMclassifier method and applied to transient power quality disturbance detection, type ofdisturbances classification and recognition is proposed, making full use of the HHTmethod that having the advantages of the superiority of data processing and analysis ofnonlinear and non-stationary processes and SVM method that having the advantages ofthe statistical validity of the classification in the sample study. With the using the HHTmethod for feature detection to transient power quality disturbance signal data, featureextraction in the detection of characteristic data, selecting the representativecharacteristic and compositing of the eigenvectors used for the identification of the typeof disturbance classification, and generating the sample library based transient powerquality disturbance model, according to principle of crossing order to extract in thesample library to form training sample library and testing sample library, the use ofSVM method in accordance with principles of classification of the two-layer step bystep to achieve the classification and identification of the types of disturbance in thecomposite disturbance signal data. Finally, making the simulation to the above dataanalysis processes in the Matlab environment, and the results of analysis of simulationverifying that the proposed combination method is fast, simple mode, and has a bettersuperiority to the classification accuracy rate, and can provide scientific basis for effective analysis of transient power quality detection and classification. In this paper,the following work has been completed:①Making transient power quality disturbance detection and classificationmethods Research divide into a transient power quality disturbance feature detectionand disturbance type classification of the two parts, and doing a comparative study anda brief summary.②With the HHT method in the detection of transient power quality disturbancefeature detection and parameter extraction method and processes, for example,analyzed by simulation to transient low-frequency oscillation in Matlab platform.③Proposing a combination method that it achieves the classification recognitionand feature extraction of the complex transient power quality disturbance type basedon the HHT method and the SVM classifier.④Doing the numerical simulation analysis to transient power quality disturbancedetection and classification Based on HHT and SVM methods in Matlab environment.