Supervision and fault detection system for photovoltaic installations based on classification algorithms

Abstract

This   article   presents   a   methodologyfor   the supervision   and   fault   detection   on   photovoltaic   installation, through  the  information  gathered  by  their  SCADA  system. The proposed  methodology  consists  of  the  use  of  a multi-clustering approach  to  analyse  and  classify  the  operating  behaviour  of  the photovoltaic installations, using information of their DCvoltage, generated  current  (per  string),  as  well  as  information relatedto the   climatic   conditions   of   the   park   (i.e.   solar   irradiance, temperature). The   proposed   methodology uses   a   supervised training  algorithm,  based  on a  decision  tree  learning  algorithm, allowing  to  determine  the  appearance  of  anomaly  behaviours  in the    installations    of    photovoltaic    plants,    including    soiling detection,  hot-spots,  tracker  deviations,  electric  connections  and faultsin sensors. The presented methodologyhas been developed in the framework of a CORFO R&D project and validated under real  operating  conditions  in  a  utility-scale  photovoltaic  power plant of one axis mount, located in Chile, with a total power of 40 MWp.