A Calibrated Hotbox for Testing Window Systems - Construction, Calibration, and Measurements on Prototype High-Performance Windows
Publication Type
Author
Abstract
Windows have the highest thermal conductance (or U-value) of all the elements in a building envelope. Although the fact that windows also admit (beneficial or detrimental) solar heat gain and (generally beneficial) daylight makes U-value alone a poor predictor of the net seasonal energy performance of a window, reducing the U-value remains an obvious first step in improving window performance.
A small (0.9m X 1.2m [3' X 4'] opening) calibrated hotbox has been built at the Buildings Technology Laboratory of the LBL/DOE Energy-Efficient Windows Program for studying methods of improving window thermal performance. We are particularly interested in comparing generic approaches to window thermal control, in identifying novel methods of improving window performance and in studying the mechanisms by which actual window performance may differ from theoretical. We first describe the construction and instrumentation of the hotbox. Next we discuss the methods by which the box has been calibrated and the limits on its accuracy. Finally we present the results of a study of prototype windows which serves the dual purpose of illustrating our methodology and presenting some interesting options for window improvement.
The thermal hotbox technique is well-known and has been described elsewhere. Since our interest is in research rather than product certification our emphasis in designing the hotbox was on flexibility and economy rather than extreme accuracy. Hence we chose the calibrated rather than the guarded hotbox technique. We have deferred installing fans to produce a moving air film on the cold side of the sample (simulating wind), choosing instead to correct our data from still air to 15 MPH conditions by the known difference in convective coefficient. We record the temperature at a large number of points on and around the window sample; most of these serve as diagnostic tools and are not directly used in measuring the U-value.