Estimation of Satellite Thermal Network Model Parameters Based on Periodic Solution

Abstract

 Thermal network models are widely used in thermal analysis and design of satellites. A thermal control system is designed to maintain all components within feasible temperature ranges, when a satellite is often considered to be subjected to periodic thermal loads. Therefore, a thermal model should fully represent periodic temperature variations in steady stage. This paper proposes a new approach to determine parameters of a reduced model by using the periodic solution of a detailed model. In order to get the input data for the parameter estimation, the periodic solution of thermal network equations is approximately calculated with low computational costs. A genetic algorithm (GA) and a sequential quadratic programming (SQP) are adopted to enhance both global and local search capabilities in the parameter estimation. Several constraints are introduced to improve the behavior of the reduced model and to reduce the dimension of search space. Thermal analysis and estimation of thermal network parameters of a microsatellite are conducted to verify the methods