BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Amsterdam BEGIN:DAYLIGHT DTSTART:20230326T030000 TZOFFSETFROM:+0100 TZOFFSETTO:+0200 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD DTSTART:20231029T020000 TZOFFSETFROM:+0200 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20230728T130454Z UID:8AE6FF04-EAB3-4EF6-AB5A-28F4EAF8B510 DTSTART;TZID=Europe/Amsterdam:20230511T140000 DTEND;TZID=Europe/Amsterdam:20230511T150000 DESCRIPTION:As the continuous consumption of fossil fuels has caused seriou s diseases\, environmental pollution\, and distributing the ecological bal ance\, renewable energy sources (RESs) such as solar\, wind\, hydroelectri c\, and geothermal energy have started to attract great attention all over the world. The use of renewable and low-carbon energy sources plays a sig nificant role in supplying electrical energy demands for sustainable and e nvironmentally friendly energy production. Photovoltaic (PV) power generat ion is one of the remarkable energy types to provide clean and sustainable energy. However\, losses of electricity production are generally caused b y the presence of various anomalies influencing the operation systems in P V plants. Therefore\, rapid fault detection and classification of PV modul es can help to increase the reliability of the PV systems and reduce opera ting costs. In this study\, an efficient PV fault detection method is prop osed to classify different types of PV module anomalies using thermographi c images. The proposed method is designed as a multi-scale convolutional n eural network (CNN) with three branches based on the transfer learning str ategy. The convolutional branches include multi-scale kernels with levels of visual perception and utilize pre-trained knowledge of the transferred network to improve the representation capability of the network. To overco me the imbalanced class distribution of the raw dataset\, the oversampling technique is performed with the offline augmentation method\, and the net work performance is increased. In the experiments\, eleven types of PV mod ule faults such as cracking\, diode\, hot spot\, offline module\, and othe r classes are utilized. The experimental results show that the proposed me thod gives higher classification accuracy and robustness in PV panel fault s and outperforms the pre-trained deep learning methods and existing studi es.\n\nSpeaker(s): Deniz\, \n\nBldg: K1\, The western Norway University o f Applied Science\, Bergen\, Vestfold\, Norway\, Virtual: https://events.v tools.ieee.org/m/359414 LOCATION:Bldg: K1\, The western Norway University of Applied Science\, Berg en\, Vestfold\, Norway\, Virtual: https://events.vtools.ieee.org/m/359414 ORGANIZER:Myou@hvl.no SEQUENCE:3 SUMMARY:An Efficient Fault Classification Method in Solar Photovoltaic Modu les using Convolutional Neural Network” URL;VALUE=URI:https://events.vtools.ieee.org/m/359414 X-ALT-DESC:Description: <br /><p>&nbsp\;<span class="fontstyle0">As the con tinuous consumption of fossil fuels has caused serious diseases\, environm ental pollution\, and distributing the ecological balance\, renewable ener gy sources (RESs) such as solar\, wind\, hydroelectric\, and geothermal en ergy have started to attract great attention all over the world. The use o f renewable and low-carbon energy sources plays a significant role in supp lying electrical energy demands for sustainable and environmentally friend ly energy production. Photovoltaic (PV) power generation is one of the rem arkable energy types to provide clean and sustainable energy. However\, lo sses of electricity production are generally caused by the presence of var ious anomalies influencing the operation systems in PV plants. Therefore\, rapid fault detection and classification of PV modules can help to increa se the reliability of the PV systems and reduce operating costs. In this s tudy\, an efficient PV fault detection method is proposed to classify diff erent types of PV module anomalies using thermographic images. The propose d method is designed as a multi-scale convolutional neural network (CNN) w ith three branches based on the transfer learning strategy. The convolutio nal branches include multi-scale kernels with levels of visual perception and utilize pre-trained knowledge of the transferred network to improve th e representation capability of the network. To overcome the imbalanced cla ss distribution of the raw dataset\, the oversampling technique is perform ed with the offline augmentation method\, and the network performance is i ncreased. In the experiments\, eleven types of PV module faults such as cr acking\, diode\, hot spot\, offline module\, and other classes are utilize d. The experimental results show that the proposed method gives higher cla ssification accuracy and robustness in PV panel faults and outperforms the pre-trained deep learning methods and existing studies.</span>&nbsp\;</p> END:VEVENT END:VCALENDAR