CONDITION-BASED MAINTENANCE OF ELECTRICAL MACHINES
Jeffery L. Kohler, Ph.D., Senior Member
[National Instite for Gocupationl Safety and Health
Piusburgh Research Laborstary
P.O, Box 18070
Picsburgh, PA 15236
Joseph Sottile, Ph.D., Member
‘University of Kentucky
234A Mining and Mineal Resources Bldg.
Lexington, KY 405060107
Frederick C. Trutt, Ph.D.., Fellow
‘University of Kentucky
453 Anderson Hall
Lexington, KY 40806-0046
Absract-Twentyfive years ago, the former US. Bureau of
‘Mines funded s research project akmed at developing the
aabling technology for incipient fllure prediction In eect
Dower system components at means of reducing the inure
find fatale that somedmes occur when equipment
‘malfuntions, Over the ening years, interest in ths has waxed
realved using an online monitoring technique. The detection
cf turntoturn defects is expelallyIroportant because they are
Deleved to represent the begining stage of most moter wining
‘ars
IntRopucTION
‘The detection of incipient failures and the replacement of
the components just priot to failure would eliminate the
consequences of unexpected equipment failures. The
capability to do this has long been desired by industy,
however, it has been an illusive goal, Failure prediction of
‘mechanical components, pericularly bearings isan old art
form, with a history of $0 or more years. The desire to spply
predictive techniques to electrical equipment hasbeen slower
fo develop. The measuring and. subsequent wacking of
insulation resisance was an early form of predictive
‘maintenance, providing reasonably good indications of
eterioration of ground insulation systems. One of the fest
instances ofa continuous on-line monitoring system to delet
electrical deterioration was reported in 1]. That system and
approach was also notable because it estiblished a link
between improved pecsonnel safety andthe application of on-
line monitoring for incipient flues
software technologies advanced, it became more practical to