BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20240310T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20231105T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20240126T022415Z UID:01610285-89F5-4B22-AD17-EE3C21DAE38F DTSTART;TZID=America/New_York:20240124T103000 DTEND;TZID=America/New_York:20240124T120000 DESCRIPTION:Traditional signal detection methods require an explicit prescr iption of partial discharge features that characterize the state of the ca ble system. When the sample size is sufficiently large\, deep learning mod els allow complex interrelations of auto-generated features. A distinct ch allenge is the characterization of the waveform signal\, which depends on cable length.\n\nIn this webinar\, an overview of the developed deep learn ing models for an extensive partial discharge dataset to automate the dete ction of underground cable defects will be provided. The developed deep le arning models outperform predictions with traditional methods. In addition to classifying partial discharge signals\, the models identify source loc ations of the defects within a cable system through recurrent Neural Netwo rks. Additional assessments include advanced data augmentation strategies and interpretability to verify the potential use of the model for predicti ve maintenance.\n\nCo-sponsored by: Power and Energy Systems Research Labo ratory\n\nSpeaker(s): Steffen Ziegler\n\nVirtual: https://events.vtools.ie ee.org/m/387548 LOCATION:Virtual: https://events.vtools.ieee.org/m/387548 ORGANIZER:ieeeunh@newhaven.edu SEQUENCE:14 SUMMARY:Artificial Intelligence-Based Fault Detection and Localization for Underground Cables URL;VALUE=URI:https://events.vtools.ieee.org/m/387548 X-ALT-DESC:Description: <br /><p>Traditional signal detection methods requi re an explicit prescription of partial discharge features that characteriz e the state of the cable system. When the sample size is sufficiently larg e\, deep learning models allow complex interrelations of auto-generated fe atures. A distinct challenge is the characterization of the waveform signa l\, which depends on cable length.<br /><br />In this webinar\, an overvie w of the developed deep learning models for an extensive partial discharge dataset to automate the detection of underground cable defects will be pr ovided. The developed deep learning models outperform predictions with tra ditional methods. In addition to classifying partial discharge signals\, t he models identify source locations of the defects within a cable system t hrough recurrent Neural Networks. Additional assessments include advanced data augmentation strategies and interpretability to verify the potential use of the model for predictive maintenance.</p> END:VEVENT END:VCALENDAR