The Kalman filter as a way to estimate the life-model parameters of insulating materials and system

The Kalman filter algorithm is applied to linear life models, valid for insulating materials subjected to electrical and multiple thermal-electrical stresses. Expressions for the state-space equations, and then for the prediction and updating equations, are obtained on the basis of inverse-power and exponential models, thus providing a useful tool for insulating-material characterization. In particular, the extrapolation to stresses lower than the test stresses and the estimation of the endurance coefficient become very sensitive to the test results at the lowest stresses, thus providing evaluations which could be more representative of the service conditions than the normal regression procedure. Examples are reported relevant to the results of life tests performed on XLPE (cross-linked polyethylene) cable models