Title:	Anomaly Detection at Inverter Level via Machine Learning Algorithms under the Absence of O&M Logbooks
Author(s):	A.P. Talayero, N.Y. Yürüsen, A. Llombart Estopinan, J.J. Melero Estela
Keywords:	Evaluation, Performance, Energy Performance, Modelling / Modeling, Inverter
Topic:	PV Systems and Storage – Modelling, Design, Operation and Performance
Subtopic:	Operation, Performance and Maintenance of PV Systems
Event:	37th European Photovoltaic Solar Energy Conference and Exhibition
Session:	5DO.4.2
Pages:	1381 - 1387
ISBN:	3-936338-73-6
Paper DOI:	10.4229/EUPVSEC20202020-5DO.4.2
Price:	0,00 EUR
Document(s):	paper, presentation

Abstract/Summary:

---

There is a need for continuous monitoring and evaluation of the reliability and the performance of Photovoltaic (PV) power plant inverters to secure uninterrupted power generation. The availability of operation and maintenance (O&M) documentation, and the sensor measurements in most PV power plants is very scarce. Therefore, the scope of this study is to evaluate the reliability and the performance of Photovoltaic (PV) inverters even when there are no failure history and no access to data from PV modules. The proposed method aims the diagnosis and prediction of failures using only the available information in industrial PV plants (Power, Radiation, Temperature, etc.). In order to achieve this objective, there is a need to explore appropriate data-driven methods and the correct interpretation of the findings obtained from both individual and merged models in a systematic manner. This approach allows us to cover the entire spectrum of possible variables registered in a PV plant and to perform analysis from different perspectives.