Machine Learning Photovoltaic String Analyzer
Photovoltaic (PV) system energy production is non-linear because it is influenced by the random nature of weather conditions. The use of machine learning techniques to model the PV system energy production is recommended since there is no known way to deal well with non-linear data. In order to dete...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2020-02-01
|
Series: | Entropy |
Subjects: | |
Online Access: | https://www.mdpi.com/1099-4300/22/2/205 |
id |
doaj-f86d018085524fcfa34c75648f221641 |
---|---|
record_format |
Article |
spelling |
doaj-f86d018085524fcfa34c75648f2216412020-11-25T02:03:34ZengMDPI AGEntropy1099-43002020-02-0122220510.3390/e22020205e22020205Machine Learning Photovoltaic String AnalyzerSandy Rodrigues0Gerhard Mütter1Helena Geirinhas Ramos2F. Morgado-Dias3Instituto de Telecomunicacoes of the Instituto Superior Tecnico of the University of Lisbon, 1049-001 Lisbon, PortugalALTESO GmbH, 1010 Vienna, AustriaInstituto de Telecomunicacoes of the Instituto Superior Tecnico of the University of Lisbon, 1049-001 Lisbon, PortugalLaboratory for Robotics and Systems in Engineering (LARSyS), Madeira Interactive Technologies (M-ITI) and Institute and Interactive Technologies Institute (ITI), 9020-105 Funchal, PortugalPhotovoltaic (PV) system energy production is non-linear because it is influenced by the random nature of weather conditions. The use of machine learning techniques to model the PV system energy production is recommended since there is no known way to deal well with non-linear data. In order to detect PV system faults, the machine learning models should provide accurate outputs. The aim of this work is to accurately predict the DC energy of six PV strings of a utility-scale PV system and to accurately detect PV string faults by benchmarking the results of four machine learning methodologies known to improve the accuracy of the machine learning models, such as the data mining methodology, machine learning technique benchmarking methodology, hybrid methodology, and the ensemble methodology. A new hybrid methodology is proposed in this work which combines the use of a fuzzy system and the use of a machine learning system containing five different trained machine learning models, such as the regression tree, artificial neural networks, multi-gene genetic programming, Gaussian process, and support vector machines for regression. The results showed that the hybrid methodology provided the most accurate machine learning predictions of the PV string DC energy, and consequently the PV string fault detection is successful.https://www.mdpi.com/1099-4300/22/2/205machine learning prediction modelspv stringpv faulthybrid methodologyensemble methodology |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Sandy Rodrigues Gerhard Mütter Helena Geirinhas Ramos F. Morgado-Dias |
spellingShingle |
Sandy Rodrigues Gerhard Mütter Helena Geirinhas Ramos F. Morgado-Dias Machine Learning Photovoltaic String Analyzer Entropy machine learning prediction models pv string pv fault hybrid methodology ensemble methodology |
author_facet |
Sandy Rodrigues Gerhard Mütter Helena Geirinhas Ramos F. Morgado-Dias |
author_sort |
Sandy Rodrigues |
title |
Machine Learning Photovoltaic String Analyzer |
title_short |
Machine Learning Photovoltaic String Analyzer |
title_full |
Machine Learning Photovoltaic String Analyzer |
title_fullStr |
Machine Learning Photovoltaic String Analyzer |
title_full_unstemmed |
Machine Learning Photovoltaic String Analyzer |
title_sort |
machine learning photovoltaic string analyzer |
publisher |
MDPI AG |
series |
Entropy |
issn |
1099-4300 |
publishDate |
2020-02-01 |
description |
Photovoltaic (PV) system energy production is non-linear because it is influenced by the random nature of weather conditions. The use of machine learning techniques to model the PV system energy production is recommended since there is no known way to deal well with non-linear data. In order to detect PV system faults, the machine learning models should provide accurate outputs. The aim of this work is to accurately predict the DC energy of six PV strings of a utility-scale PV system and to accurately detect PV string faults by benchmarking the results of four machine learning methodologies known to improve the accuracy of the machine learning models, such as the data mining methodology, machine learning technique benchmarking methodology, hybrid methodology, and the ensemble methodology. A new hybrid methodology is proposed in this work which combines the use of a fuzzy system and the use of a machine learning system containing five different trained machine learning models, such as the regression tree, artificial neural networks, multi-gene genetic programming, Gaussian process, and support vector machines for regression. The results showed that the hybrid methodology provided the most accurate machine learning predictions of the PV string DC energy, and consequently the PV string fault detection is successful. |
topic |
machine learning prediction models pv string pv fault hybrid methodology ensemble methodology |
url |
https://www.mdpi.com/1099-4300/22/2/205 |
work_keys_str_mv |
AT sandyrodrigues machinelearningphotovoltaicstringanalyzer AT gerhardmutter machinelearningphotovoltaicstringanalyzer AT helenageirinhasramos machinelearningphotovoltaicstringanalyzer AT fmorgadodias machinelearningphotovoltaicstringanalyzer |
_version_ |
1724947376744955904 |