Different working conditions identification of a PV power plant using hierarchical clustering

This article presents big data analysis, based on the measurements from a stationary power quality (PQ) monitoring system of a 15 kW PV power plant. The study period is one year - 2018. In the article, a unified PQ database was proposed, which consists of PQ parameters and active power level. Then cluster analysis (CA) was performed to indicate the different working conditions of a PV power plant. The Ward algorithm was selected to obtain the automatic classification due to the data features. The importance rate of parameters for each classification was calculated, to indicate which parameters were important in point of different clusters obtained. Additionally, the PQ level comparison between the clusters was realized to assess the indicated working conditions of a PV power plant.