A statistical analysis of methods using truncated data to shorten thermal evaluation time

On the basis of this study the following conclusions and recommendations are presented. The estimation of thermal-life values for electrical insulation using truncated data has been proven valid, and a substantial savings of aging time can be realized by its use, particularly when applied to motorette tests. The median method offers the greatest degree of accuracy and reliability as compared to the other two methods studied. The probability-plot and fifth-failure methods offer a significant increase in aging time saved but at a sacrifice of about 0.2% average accuracy and 10% reliability. It is recommended that in making a choice between the three methods consideration be given to each individual set of circumstances, weighing such factors as cost limitations, time urgency and reliability needed.