This paper presents a model-based anomaly detection method for finding the faults occurring in the solar photovoltaic (PV) module system. Here, data are collected from the 95kW grid connected solar power plant which is located in Karunya Institute of Technology and Sciences, Coimbatore. For this work, the time frame has taken as 30 days (1 month) and randomly created the different faults and same data are used for prediction of anomalies. Two techniques are compared; one in supervised learning and another one in unsupervised learning. In supervised learning, kNN has taken and in unsupervised learning, Isolation Forest has taken; finally compared two techniques with same data set.
Volume 12 | 04-Special Issue
Pages: 153-162
DOI: 10.5373/JARDCS/V12SP4/20201477