I'm trying to replicate the surface temperature calculation in EnergyPlus for a research project on cool roofing. I created a run from one of the PNNL buildings that uses the TARP algorithm as documented in Section 3.5.5.2 in the EnergyPlus Engineering Reference. I can match the roof convection coefficient with these equations quite well, except during an hour or two in the mid-morning or early-evening when the roof temperature is rapidly changing. During those times, the answer is 10-50% off. I suspect there is some undocumented curve smoothing happening but I'm not sure. Has anyone run into this before?
Question-and-Answer Resource for the Building Energy Modeling Community
