The sensitivity of the time-temperature shift process to thermal variations—A note

Abstract

Because of the highly nonlinear relation between time-shift and temperature it is generally not admissible to assume that the appropriate temperature for the shifting process equals the average temperature during the measurement interval. It is shown here that temperature variations around a mean can produce shifts that can be significantly larger than the mean temperature would indicate. Alternatively this paper can aid in providing numerical estimates that indicate for what range of thermal variations acceptable engineering data may be acquired. As a guide one finds that a one degree centigrade variation (±0.5°C) around a mean value limits the shift factor error to less than 4%, which translates into a misrepresentation of two weeks out of a year.