Hazards of insulation life estimation using censored data and high degrees of acceleration

Insulation life studies involve considerable expenditure in terms of time and money. To keep down this expenditure, accelerated life testing is resorted to. The degree of acceleration (value of the accelerating variable) in such testing procedure should be chosen with due caution so that the life data acquired are reliable. Recently, censored data acquisition procedure has come into vogue, in which, the life test at any stress level is terminated either after a certain number of specimens fail (type 1 censoring) or after a specified length of time, subject to the condition that a reasonable number of samples fail. This procedure becomes possible in view of the extremal nature of insulation ageing. Censored testing method enables one to use lower levels of acceleration, thus reducing, considerably, the risk of faulty data entering into the analysis, but the theory and computation procedures are much more involved.