Application of the Kalman filter for electrical endurance characterization of insulating materials and systems

The Kalman filter algorithm is applied to the results of accelerated life tests performed on insulating materials and systems carried out for electrical endurance characterization purposes. It consists of a recursive procedure that applies to the failure-time percentiles obtained from life tests, on the basis of the inverse power or the exponential models, and provides estimates of the endurance indices which give particular weight to the results of the life tests performed at the lowest test stresses. This allows more meaningful endurance characterization and accurate life prediction. This paper provides the algorithms for Kalman filter applications, giving emphasis to error estimation, as well as the method for calculation of confidence limits of life-line parameters and predicted failure times. The proposed Kalman filter procedure is applied to the experimental results of accelerated life tests relevant to different insulating materials and systems. © 1996 IEEE.