Modeling of and experiments with heat storage in regenerators packed with phase-change material (PCM) are discussed, as well as the second-law thermodynamic efficiency for the ideal phase-change regenerator (PCR). An algorithm to solve the coupled partial differential equations for heat transfer and storage in the PCR on the bed scale and on the PCM scale is presented. The bed is discretized via the tanks-in-series approximation. The PCM scale is solved by orthogonal collocation applied to the equations, transformed to immobilize the melt/solid interface and eliminate the effect of spherical geometry. Parametric studies show the effects of specific dimensionless groups. A novel PCM consisting of n-octadecane retained by capillary forces in a porous silica support is used in a lab-scale PCR to verify the model. It visually changes from opaque to semi-transparent when the wax melts, thereby allowing the melt front within the bed to be tracked. Experiments with heated or cooled CO2 passing through the PCR are described. The measured outlet temperature compares qualitatively with model predictions. The model quantitatively predicts melt front movement in the first 60% of the bed. Discrepancies between the model and experiments are linked to significant heat losses.