Title:	Real-Time Fault Detection in Massive Multi-Array PV Plants Based on Machine Learning Techniques
Author(s):	C.-C. Hsu, J.-L. Li, Y.-S. Chen
Keywords:	Cluster, Fault Detection, Machine Learning, Fault Diagnosis
Topic:	PV Systems and Storage – Modelling, Design, Operation and Performance
Subtopic:	Operation, Performance and Maintenance of PV Systems
Event:	36th European Photovoltaic Solar Energy Conference and Exhibition
Session:	5BO.6.5
Pages:	1308 - 1311
ISBN:	3-936338-60-4
Paper DOI:	10.4229/EUPVSEC20192019-5BO.6.5
Price:	0,00 EUR
Document(s):	paper

Abstract/Summary:

---

To monitor massive multi-array operational plants in real-time, an inexpensive and effective monitoring system is indispensable. We propose an approach based on machine learning techniques which analyze historical power and irradiance data of PV arrays and do not require additional sensors. The developed system currently monitoring 150 plants including 7028 arrays is deployed on a Spark-cluster distributed platform such that the detection and diagnosis process can be finished within 5 minutes. The system demonstrates good performance. The fault detection accuracy is over 96%.