Low-Cost Power Quality System and Automatic Classification Disturbances Using SVM and DWT Analysis

It is a well-known fact that data acquisition in Power Quality (PQ) analysis is based on quality analyzers, the performance of which is standardized, but which have a relatively high purchase price. In this paper we present an equivalent low-cost data acquisition system based on a smart meter, plus the application software, which performs the main functions of a quality analyzer. Based on the acquired signals, we present the main PQ disturbances and their total harmonic distortion (THD) characteristics, but also the characteristics based on the Discrete Wavelet Transform (DWT) analysis, such as: entropy and signal energy. We present the training, testing and validation of an automatic classifier based on support vector machines (SVM) type algorithms that can be used for the automatic classification of PQ disturbances. We present the practical use of the data acquisition system and the automatic classification system by recognizing a sag-type PQ disturbance with harmonics, which is added to the main grid when coupling and uncoupling an uninterruptible power source (UPS).