AN APPROACH FOR DATA MINING OF POWER QUALITY INDICES BASED ON FAST-ICA ALGORITHM

Payman Moallem, Abolfazl Zargari, and Arash Kiyoumarsi

Keywords

Power quality, data mining, ICA algorithm, pattern classification

Abstract

Power quality monitoring of many sites of an electric power system produces an enormous amount of unstructured data covering the various types of power quality indices. The collected data are not in a suitable form to give insights to the general power quality conditions of a particular site or a particular area within the network. This paper proposes a global power quality index (PQI) which is based on data mining and pattern classification approaches. Firstly, the continuous and discrete PQIs are annually analyzed, normalized and merged. Then, the analyzed indices are classified according to their cost coefficient, and the power quality levels for all distribution sites are determined using the Fast-independent component analysis (ICA) data mining algorithm. Finally, an application example for future explanations is presented.

Important Links:



Go Back