Coupling of a Genetic Algorithm and a Thermoelectric Network Model for Radioisotope Thermoelectric Generator (RTG) Optimization

Kurzfassung

Lunar exploration requires a continuous supply of electrical power, independent of solar irradiation availability, which can be provided by radioisotope thermoelectric generators (RTG). RTG contain a radioisotope heating unit, enclosed in an aeroshell core and coupled with thermoelectric modules (TEM). TEM utilize the temperature difference developed between their hot side, adjacent to the RTG core, and their cold side, connected to radiation fins attached to the outer RTG surface, to convert part of the released heat into electricity through the Seebeck effect. Design and operation of RTG is a complex multi-disciplinary challenge, as they are comprised of multiple components that affect performance. The surfaces of the aeroshell not connected to TEM are covered by thermal insulation, driving most of the source heat through the TEM. Excessive insulation however, could undersibaly increase the system mass which should be minimized. The materials and geometry of the TEM affect the heat-to-electricity conversion efficiency as well as the temperature at its hot side which should not exceed the TEM’s safe operating temperature to prevent device failure. To address these challenges and optimize the RTG for lunar operation, we developed a multi-physics and multi-parametric thermoelectric (TE) network model, which represents the RTG as a series of thermal resistors and TE elements. By solving the heat transfer equation, the temperature difference on all resistors can be calculated, while the TE heat transfer equations on the TEM provide an estimate of the RTG’s performance. We then coupled this model with a Genetic Algorithm for rapid design invesigations to identify the RTG design which maximises the RTG’s specific power, i.e. electrical power output over system mass, while maintaining safe operating temperatures on the TEM. The optimized RTG design, which utilizes a Bi2Te3 TEM, achieves a specific power of 1.42 W/kg, with an electrical power of 13.9 W and a conversion efficiency of 6.9%. This is a significant improvement over the pre-defined specifications of the European Space Agency (ESA), i.e power of 10 W with a 5% conversion efficiency [1]. Our model can further adapt the RTG to incorporate TEM which can operate at higher temperatures, such as Skutterudites or Half-Heusler materials, thus allowing for higher performances.