Abstract
The next-generation radioisotope thermoelectric generator (NG-RTG) is a high-power RTG under development within the US Radioisotope Power Systems (RPS) Program. The detailed thermoelectric performance of this generator and its extended configurations remains unclear. Using NG-RTG Mod-1 as the baseline, one-eighth symmetry COMSOL models of the standard design and two extended configurations containing 8 and 18 general-purpose heat source (GPHS) modules were developed. To address the long solution times and convergence difficulties caused by the complex RTG structure, a millisecond-scale calculation method was also developed. It combined a one-dimensional thermal-resistance network, a Kirchhoff transformation for non-linear heat conduction problem, and an equivalent representation of the segmented unicouple to calculate load performance from open-circuit temperatures. Each calculation required approximately 20 ms, compared with tens of minutes for the corresponding finite-element model. NG-RTG (Standard) produced 251.5 We power at 28 V, consistent with the latest range reported by NASA. Despite its fewer heat source, the NG-RTG (8 GPHS) maintained a maximum specific power of 4.35 We/kg, whereas NG-RTG (18 GPHS) increased maximum specific power by 8.8%. The maximum relative error in open-circuit temperature was below 1.5%, and the output power error under load was generally within 5%; the maximum power error for NG-RTG (Standard) was only 4 We. The method also remained applicable after fuel decay, with output power errors within 1.8%. These models and the calculation method can support configuration extension, operating-parameter assessment, and rapid engineering iteration for high-power, axially stacked RTGs such as the NG-RTG.

This publication has 14 references indexed in Scilit: