LBNL Report Number
The validity of a method to reduce the total number of computer simulations which must be run to determine the U-values of a window product line with multiple glazing options is examined. The accuracy and limits of this method, which uses the WINDOW4 and FRAME simulation programs, is evaluated by comparing the edge, frame, and total window U-values calculated on the basis of single point FRAME simulations to those U-values as calculated on the basis of four point FRAME simulations combined with linear interpolation of frame and edge U-values by WINDOW4. The accuracy of this procedure is examined for two frame types, a low thermal conductivity wood-framed casement and a high thermal conductivity aluminum-framed casement, using both aluminum spacers and insulating spacers over a wide range of glazing types. The effect of center-of-glass U-value, overall glazing thickness and spacer type on frame and edge-of-glass U-values is discussed. It is shown that the agreement between total window U-values as calculated by the single point and four point simulation methods is better than 1% for double and triple-glazed windows with aluminum spacers, better than 1% for double-glazed windows with insulating spacers, and better than 2% for triple-glazed windows with insulating spacers.