Big difference between glazing conduction by HB and in steady-state

I am trying to compare the Energy Balance results from HB with my own steady-state calculation. The glazing conduction from ‘Honeybee_construct energy balance’ is quite different from what I calculated with the equation of ‘U_value * window area * delta T’. I attached with the excel of comparison. Does anyone have an idea about the big difference?

forum-glazing comparison between HB and SS.xlsx (146.8 KB)

@Tiantian ,

I haven’t gotten the chance to dig into your file yet but I have a strong hunch about why you would get different results between E+ and a steady state calculation.

Firstly, I imagine that your steady state calculation is only looking at the difference in indoor and outdoor air temperature across the glass and does not account for solar radiation that gets absorbed by the glass when the sun is shining on it and then conducts to the interior of the zone. Because E+ models the heat flow across the window with several nodes at different depths into the construction, E+ can account for this absorbed solar that then conducts to the interior and this energy will show up under the glazing conduction term out of the energy balance. This could explain why you have higher peaks in the E+ conduction than in your steady state calculation.

As for the reason why you might have more heat loss through the window at night than your steady state calculation, this might be the result of the fact that E+ dynamically calculates the resistance of air films on either side of the window construction and these air films can contribute a lot to the U-values (especially when the glass isn’t well insulated). To compute the resistance of the indoor air film, E+ uses using things like the orientation of the surface in E+ and the temperature difference between the interior window surface and the room air at a given point in time. To compute the resistance of the outdoor air film, it uses things like the outdoor wind speed at a given hour in the epw. Since it seems like your steady state calculation doesn’t have a sophisticated means of dynamically calculating the air films at each timestep, it seems likely that it could be overestimating the benefit of the interior air films compared to the actual building and climate that you are simulating.

Agreed, and another contributing factor to underestimating conduction loss here may be the radiation loss from the window to the clear sky. That is, clear (as opposed to overcast) skies have a high absorptance to emittance ratio and thus your windows will experience net radiation loss to the sky; as a function of your emittance-weighted view factor to your surroundings.

@Tiantian, I think the easiest way to account for these surface convection, and radiation exchanges without tracking all energy interactions through your glass, is simply to substitute the surface temperatures of your glazing assembly in your Fourier’s Law calculation, in place of the interior room and exterior wetbulb temperatures.

I attempted a similar experiment a while ago here: Eplus Conduction Rate and Uvalue of the wall check, albeit for opaque constructions during night time conditions. Since glazing should have negligible thermal mass, this same strategy should work fine for glass during the daytime.


1 Like