BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Australia/Brisbane
BEGIN:STANDARD
DTSTART:19920301T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20180615T043156Z
UID:AF69EF77-7902-441A-AACC-ED31F5EDF5EA
DTSTART;TZID=Australia/Brisbane:20180615T113000
DTEND;TZID=Australia/Brisbane:20180615T123000
DESCRIPTION:While the penetration of renewable energy resources and the com
 plexity of distribution networks are increasing\, highly accurate protecti
 on schemes for Low Voltage (LV) distribution networks are not well develop
 ed yet. The accuracy of the conventional protection frameworks depends gre
 atly on having an adequate number of measurement devices and adequate comm
 unication infrastructure installed. The high cost of monitoring devices mo
 tivates the distribution network operators to employ state estimators and 
 forecasting models instead of increasing measurement and communication dev
 ices. In addition\, the high intermittency in renewable power generation a
 nd unpredictable behaviour of the customer loads have originated from unkn
 own changes in the dynamic behaviour of the distribution systems. This mak
 es the conventional fault detectors with fixed current thresholds less acc
 urate to detect faults\, especially those with low fault currents. Hence\,
  a predictive fault detection framework is required to predict the fault c
 urrent thresholds for each time step.\n\nThis seminar shows how a Kalman f
 ilter as a forecasting-aided state estimator and quantile regression as a 
 distribution free forecasting model can detect dynamic fault current thres
 holds in LV distribution networks. This presentation also shows that how d
 ividing a distribution network into several protection zones with their ow
 n switches can locally isolate faults and keep the rest of the distributio
 n network live. The performance of the Kalman filter and quantile regressi
 on as fault detection frameworks is evaluated by real distribution network
 s and customer data.\n\nCo-sponsored by: Nima Khoshsirat\n\nSpeaker(s): Mr
  Mehdi Shafiei\, \n\nRoom: S620\, Bldg: s blcok\, Room S620\, Level 6\, S 
 block\, Gardens Point Campus\, Queensland University of Technology (QUT)\,
  Brisbane\, Queensland\, Australia\, 4001
LOCATION:Room: S620\, Bldg: s blcok\, Room S620\, Level 6\, S block\, Garde
 ns Point Campus\, Queensland University of Technology (QUT)\, Brisbane\, Q
 ueensland\, Australia\, 4001
ORGANIZER:khoshsirat@ieee.org
SEQUENCE:2
SUMMARY:Smart Fault Detection For Low Voltage Distribution Systems
URL;VALUE=URI:https://events.vtools.ieee.org/m/173603
X-ALT-DESC:Description: <br /><p>While the penetration of renewable energy 
 resources and the complexity of distribution networks are increasing\, hig
 hly accurate protection schemes for Low Voltage (LV) distribution networks
  are not well developed yet. The accuracy of the conventional protection f
 rameworks depends greatly on having an adequate number of measurement devi
 ces and adequate communication infrastructure installed. The high cost of 
 monitoring devices motivates the distribution network operators to employ 
 state estimators and forecasting models instead of increasing measurement 
 and communication devices. In addition\, the high intermittency in renewab
 le power generation and unpredictable behaviour of the customer loads have
  originated from unknown changes in the dynamic behaviour of the distribut
 ion systems. This makes the conventional fault detectors with fixed curren
 t thresholds less accurate to detect faults\, especially those with low fa
 ult currents. Hence\, a predictive fault detection framework is required t
 o predict the fault current thresholds for each time step.</p>\n<p>This se
 minar shows how a Kalman filter as a forecasting-aided state estimator and
  quantile regression as a distribution free forecasting model can detect d
 ynamic fault current thresholds in LV distribution networks. This presenta
 tion also shows that how dividing a distribution network into several prot
 ection zones with their own switches can locally isolate faults and keep t
 he rest of the distribution network live. The performance of the Kalman fi
 lter and quantile regression as fault detection frameworks is evaluated by
  real distribution networks and customer data.</p>
END:VEVENT
END:VCALENDAR