Determining the most influential factor in the overall index called - power quality index using Factor Analysis

This paper continues a previous investigation on electricity quality assessment of an electrical substation, part of the EVN Plovdiv distribution network, through a Factor Analysis. The investigated location changed to town Dospat, Bulgaria, and the number of the input variables is increased to seven. The real-time monitoring system is the same - OMNI-QUANT. This instrument measures variety of variables related to the quality of the electricity. However, we used only seven of these variables as follows: line voltages 12, 23, and 31, harmonic distortion for every phase, and three-phase unbalance. The logger measured the data every 10 sec and reported every 600 sec for average, min, and max values for each variable. A Triangular analysis has been used and indexing, after applying a Factor Analysis, thus the input seven variables are reduced to new three factors. The main advantage of the method, which is new in this research field, is the dimensionality reduction of the input variables. The input seven variables could be reduced to three new factors keeping the sample representative. The predominant variables were found to influence the power quality index from the input variables; in our case, those are line voltages variations.