Effect of window-to-wall ratio on measured energy consumption in US office buildings

Abstract

Windows can significantly affect building performance, and the window-to-wall ratio (WWR) describes the portion of an exterior wall that consists of windows. Previous research on the effects of WWR and efforts to determine optimal WWR generally rely on iterative simulation or algorithmic optimization. This study seeks to understand the effects of WWR in actual office buildings using survey data reported in the 2012 CBECS (Commercial Building Energy Consumption Survey). Both total annual energy use and four discrete end-uses (heating, cooling, lighting and ventilation) were characterized, and 32 categorical and numerical building characteristic variables were selected for linear regression analysis. Descriptive statistics for energy use intensity (EUI) in ∼1000 office buildings across 6 WWR intervals show increased median total EUI with increasing fenestration, driven by increasing cooling loads. Total EUI generally decreases with year of construction, regardless of WWR. Single-variable regression finds consistent statistical significance for WWR on cooling, lighting, and ventilation energy use, but with a maximum goodness of fit R2=0.04. In contrast, the single variable with the largest explanatory power is cooling degree days (R2=0.22 for cooling energy use). Multi-regression modeling finds a maximum R2 value of 0.34, with WWR appearing as a significant variable in the regression equations for cooling and lighting. In sum, the 2012 CBECS microdata for office buildings suggests that WWR is a significant predictor of energy use for cooling, and to a lesser extent lighting and ventilation, but to a much lower degree than has been found by purely simulation studies.