Reassessing nanoparticle and lattice modifications in space thermal storage panels: Insights from CFD and a novel reduced-order model

This work examines the transient thermal performance of an aluminium–PCM–aluminium thermal storage panel (TSP) designed for CubeSat-class platforms operating in a 95-minute low Earth orbit. Three-dimensional enthalpy–porosity simulations are coupled with a physics-based resistance–capacitance (RC) network to assess the influence of PCM volume fraction, cavity subdivision, PCM type, nanoparticle additions and internal lattice structures under realistic solar and internal heat loads. It is found that, for a fixed panel footprint, thermal regulation improves as the PCM volume fraction is increased from 0.1 to about 0.4, beyond which the benefit saturates; an optimal fraction of approximately 0.432 makes effective use of latent storage without leaving significant PCM under-utilised during a single orbit. At this loading, distributing the PCM into six cavities reduces the peak base-plate temperature by enhancing PCM–aluminium contact and shortening conduction paths. With this geometry, filling the panel with RT15 lowers the peak temperature by roughly 10 ° C compared with an empty cavity and by about 2 ° C relative to a solid aluminium infill, while Erythritol provides stronger peak suppression and trough elevation by better matching the operating temperature window. A key finding is that nanoparticle enhancement of RT15 and Erythritol and the introduction of internal lattice structures alter local liquid-fraction fields but lead to only nominal panel-level reductions in peak-to-trough temperature difference, typically 1 ° C or less. This weak sensitivity is consistent with a Fourier number based on the cavity thickness that satisfies F o ≫ 1 , so that transverse conduction is already fast and the cycle-integrated latent heat exchange, rather than conduction, controls the response. A compact three-node RC model, constructed from analytical one-dimensional resistances corrected by a single physics-based spreading factor, reproduces the CFD temperature histories with errors below 5% while reducing wall-clock time by nearly two orders of magnitude. Overall, the results suggest that, for CubeSat-class TSPs in orbit, design effort is best directed towards selecting a PCM with an appropriate melting range and moderately tuning the PCM fraction and cavity layout, whereas nanoparticles and aggressive internal latticing do not emerge as efficient design levers in this F o ≫ 1 regime.