If you are getting different values for "zone lights total heating energy" when you rotate the building, my guess is that some of your lights are controlled by a daylight sensor. When you rotate the building, it changes how much daylight comes through the windows in the zone. In orientations where more daylight comes through the windows, the daylight sensors dim the electric lights more, leading to a reduction in "zone lights total heating energy."

Yes, I have a daylighting sensor in the zone but the difference in values occurs for the same orientation each time (for example South). when I run the model two or three times for south or north or ..., each time I receive different values for a specific orientation and this difference occurs for all of the orientations. when I run the model with another window construction (clear 6mm), every time I run the model for a specific orientation, I receive the same values and it is ok. The problem happens when I use a glazing material with these properties: Visible transmittance: 0.075 Solar transmittance: 0.065 Visible reflectance (front side): 0.293 Visible reflectance (back side): 0.554 Solar reflectance (front side): 0.243 Solar reflectance (back side): 0.553 emissivity: 0.84 In this case, the Schedule Compact fields has more than 6000 fields! (this schedule is made by radiance and pass to EnergyPlus). But this should not make any difference in the results by running the model each time.