Thermal sensor calibration using non-integer system identification

Abstract

A widely used approach to measuring heat flux in harsh environments such as high enthalpy plasma flows, fusion plasma and rocket motor combustion chambers is based on solving the inverse heat conduction problem in a semi-infinite environment. This approach strongly depends on model parameters and geometrical aspects of the sensor design. In the present paper the surface heat flux is determined by solving the inverse heat conduction problem using an a priori identified system as a direct model within the estimation process. Calibration measurements using modern laser technique are used to identify the physical thermal system, where it turns out that a non-integer model appears most suitable to this particular problem. It is concluded that with the new method the heat flux sensor is significantly improved. Furthermore, its applicability is extended to very short measurement times.