Stochastic spacecraft thermal design optimization with low computational cost

This paper presents a strategy for a quick determination of the optimum configuration for radiators and solar absorbers in a spacecraft thermal design, to minimize heater power consumption and maximize temperature margins. It is particularly useful when applied to multimission platforms in which the thermal design is adapted for different orbits and operational modes. A two-step approach is adopted wherein a simplified thermal model is developed to search for the optimum radiator/solar absorber areas, and then the results are implemented in a. detailed thermal model to verify the temperature distribution, thereby reducing computational time, a common drawback in complex engineering optimization problems. If necessary, small adjustments are then made in the radiator/solar absorber configuration. The search for the optimum design is accomplished using a recently proposed global search metaheuristic, called generalized extremal optimization. Based on a model of natural evolution, it is easy to implement and has only one free parameter to adjust, making no use of derivatives. This paper presents the strategy as applied to the thermal design of the Brazilian Multimission Platform now under development.