Thermal mathematical model correlation through genetic algorithms of an experiment conducted on board the International Space Station

Genetic algorithms have been used for matching temperature values generated using thermal mathematical models against actual temperatures measured in thermal testing of spacecrafts and space instruments. Up to now, results for small models have been very encouraging. This work will examine the correlation of a small-medium size model, whose thermal test results were available, by means of genetic algorithms. The thermal mathematical model reviewed herein corresponds to Tribolab, a materials experiment deployed on board the International Space Station and subjected to preflight thermal testing. This paper will also discuss in great detail the influence of both the number of reference temperatures available and the number of thermal parameters included in the correlation, taking into account the presence of heat sources and the maximum range of temperature mismatch. Conclusions and recommendations for the thermal test design will be provided, as well as some indications for future improvements.