LBNL Report Number
This paper presents the results of a study investigating the effects of window U-value changes on residential cooling loads. We used the DOE-2.1D energy analysis simulation program to analyze the hourly, daily, monthly, and annual cooling loads as a function of window U-value. The performance of a prototypical single-story house was examined in three locations: hot and humid, Miami FL; hot and dry, Phoenix AZ; and a heating-dominated location with a mildly hot and humid summer, Madison WI. Our results show that when comparing windows with identical orientation, size, and shading coefficient, higher U-value windows often yield lower annual cooling loads, but lower U-value windows yield lower peak cooling loads. This occurs because the window with the higher U-value conducts more heat from inside the residence to the outside during morning and evening hours when the outside air temperature is often lower than the inside air temperature; and, a lower U-value window conducts less heat from outside to inside during summer afternoon peak cooling hours. The absolute effects are relatively small when compared to total annual cooling which is typically dominated by window solar heat gain effects, latent loads, and internal loads. The U-value effect on cooling is also small when compared to both the effects of U-value and solar heat gain on heating load. Our modeling assumed that U-value and solar heat gain could be independently controlled. In fact, reducing window conductance to the levels used in this study implies adding a second glazing layer which always reduces solar heat gain, thus reducing annual cooling. Thus, when we compare realistic options, e.g., single pane clear to double pane clear, or single pane tinted to double pane tinted, the double pane unit shows lower annual cooling, as well as lower peak loads.