Classification of power quality disturbances using wavelet packet energy and multiclass support vector machine

• Ming Zhang ^[1] ; Kaicheng Li and Yisheng Hu ^[1] ; Yisheng Hu
  1. [1] Huazhong University of Science and Technology

     Huazhong University of Science and Technology

     China

     
• Localización: Compel: International journal for computation and mathematics in electrical and electronic engineering, ISSN 0332-1649, Vol. 31, Nº 2 (Special Issue: ARWtr conference Spain), 2012, págs. 424-442
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Purpose – The purpose of this paper is to develop a new method for classification of power quality (PQ) disturbances such as the sag, interruption, swell, harmonic, notch, oscillatory transient and impulsive transient.

    Design/methodology/approach – A PQ disturbances classification system based on wavelet packet energy and multiclass support vector machines (MSVM) is proposed to discriminate seven types of PQ disturbances. The PQ disturbance signals are first decomposed into components in different subbands using discrete wavelet packet transform (DWPT). Statistical features of the decomposed signals are required to characterize the PQ disturbances. A MSVM classifier follows to classify the PQ disturbances.

    Findings – The proposed method could effectively detect information from disturbance waveforms using DWPT and MSVM techniques, which is verified on over 700 samples.

    Research limitations/implications – The classification stage of the proposed method does not differentiate the disturbances occurred simultaneously.

    Practical implications – The proposed method possesses high recognition rate, so it is suitable for the PQ monitoring system for detection and classification of disturbances.

    Originality/value – The paper describes a new and efficient way of classification of PQ disturbances. In this paper, an attempt has been made to extract efficient features of the PQ disturbances using DWPT. It is observed that these features can help correctly classify the PQ disturbances, even under noisy conditions. The MSVM is compared with artificial neural network (ANN) and it is found that the MSVM classifier gives the better result.