Why do I frequently get energy imbalance when I use shell conduction, even if my calculation looks well converged (stable monitors and low residuals)?
Heat flux at the sides of the shell walls are not taken into account inside the heat flux reports in Ansys Fluent software (> Report > Fluxes > Total Heat Transfer rate).
This means that at these locations, heat flux is accounted for correctly in Fluent software, but cannot be post-processed. If the heat flux crossing the side faces is significant, this could lead to an imbalance in the reports between the wall and wall-shadow (more than 30% imbalance reported in some cases).
To make sure this imbalance is not due to convergence difficulties, a good practice is performing a quick sensitivity analysis by reducing either the shell conduction thickness or the conductivity of the solid material, and then checking changes in the heat flux report accordingly.