Koehler, MarkMarkKoehlerAoki, YohkoYohkoAokiÖhler, SvenSvenÖhler2025-05-062025-07-152025-05-062024https://publica.fraunhofer.de/handle/publica/487355Due to the increasing proportion of glass surface in modern architecture, as well as their lower sound transmission loss compared to outer walls, environmental noise is transmitted into buildings mostly via windows. Nevertheless, the development of modern insulation glass panes mostly focuses on their energy efficiency and thermal insulation. Especially different kinds of spacers made of plastic, so called "warm edges" are becoming more and more common. However, its effect on the acoustic performance was not studied. The acoustic performance of glass panes is usually estimated by the glass and air layer thickness disregarding the edge bond, and is proven by a test certificate, which refers to the ISO standard measurement of the sound reduction index in a laboratory. Since windows in real construction generally have different dimensions, edge connections and mounting conditions, the acoustic performance of the window in construction often differs from the value specified in the test certificate. This project aims to develop a method to predict the sound transmission loss of glass panes by measuring only certain parameters of the edge bond. A total of 20 glass panes with 5 different formats and 4 pane configurations were tested according to ISO 10140 in third octave and narrow band frequencies. In addition, modulus of elasticity and loss factors of window putty, insulating glass sealant and spacers made of different materials were measured. The measurement data served as input parameters for a finite element simulation model. Finally, the results of the ISO standard measurements and simulations were compared with each other. This paper focuses on the influence of the spacer stiffness on the sound transmission loss and the extent, to which the simulation can accurately predict a standard measurement of the sound reduction index.eninsulation glasswindow panesound transmission losswarm edgeconference paper