A strategy to locate partial discharges in power transformers using acoustic emission

Abstract

 This  paper  presents  a  solution  for  the  problem  related  to  Partial  Discharges  (PD)  source  localization  in  oil-isolated  transformers,  using  Genetic  Algorithms.  As  known  in  literature,  the  simple  PD  detection  is  not  enough  to  take  a  decision  about  intervening,  so  the  localization  is  necessary  to  assess  the  risk  and  to  plan  corrective  actions.  The  parameters  extracted  from  the  acoustic  signals  collected  by  piezoelectric  sensors installed outside of the transformer, provide the correct position  of  PD.  From  parameters  extracted  of  acoustic  signals  collected   from   piezoelectric   sensors   installed   outside   the   transformer,   the   proposed   algorithm   provides   the   position   where    the    PD    is    occurring.    This    work    compares    the    performance of the method using Genetic Algorithms and those using  iterative  method,  which  is  normally  used  to  solve  this  problem.  The  conclusion  describes    the  advantages  of  the  proposed solution.