TU Darmstadt / ULB / TUbiblio

SoundLab AI-Machine learning for sound insulation value predictions of various glass assemblies

Drass, Michael ; Kraus, Michael Anton ; Riedel, Henrik ; Stelzer, Ingo (2022)
SoundLab AI-Machine learning for sound insulation value predictions of various glass assemblies.
In: Glass Structures & Engineering
doi: 10.1007/s40940-022-00167-z
Artikel, Bibliographie

Kurzbeschreibung (Abstract)

Modern architecture promotes a high demand for transparent building envelopes and especially glass facades. Commonly, facades are designed to fulfill a multitude of objectives such as superior aesthetic appearance, a higher degree of weathering reliability, quick installation, high transparency as well as economic and ecologic efficiency. For such glazing applications, often an assessment of acoustic properties and especially sound insulation abilities are required. Because of the complexity of such an experimental or computational investigation given the framing systems and glass unit compositions, a reliable and fairly accurate estimation of sound insulation properties of such systems becomes time-consuming and demanding. This paper provides a Machine Learning (ML) based estimation tool of acoustic properties (weighted sound insulation value Rw, STC and OITC) of different glazing set-ups. A sufficiently rich database was used to train several machine learning algorithms. The acoustic properties are determined by comparing the third-octave or octave band spectrum of the sound reduction index with a reference curve (typical curve for solid construction elements) specified in the standard DIN EN ISO 717-1. Sound insulation values can currently only be determined by complex and expensive experimental investigations or numerical simulations for certain glass set-ups. Hence, there is no efficient tool for convenient and reliable estimation of the sound insulation performance of glazing systems available at the moment. To this end, the engineering team led by the authors conducted extensive studies on various glazings consisting of different glass assemblies with varying glass, cavity and interlayer thicknesses and different types of interlayer and gas fillings. Based on our research outcomes, a comprehensive web-based prediction program, the so-called AI Tool, has been developed recently. This program can provide a quick analysis and accurate prediction of arbitrary glazing set-ups, interlayers and glazing infills. A series of laboratory tests were conducted to validate the predictions by the AI Tool. The goal of this program is to provide designers, engineers, and architects an effective and economically efficient tool to facilitate the design w.r.t. acoustical properties.

Typ des Eintrags: Artikel
Erschienen: 2022
Autor(en): Drass, Michael ; Kraus, Michael Anton ; Riedel, Henrik ; Stelzer, Ingo
Art des Eintrags: Bibliographie
Titel: SoundLab AI-Machine learning for sound insulation value predictions of various glass assemblies
Sprache: Englisch
Publikationsjahr: Februar 2022
Titel der Zeitschrift, Zeitung oder Schriftenreihe: Glass Structures & Engineering
DOI: 10.1007/s40940-022-00167-z
URL / URN: https://doi.org/10.1007/s40940-022-00167-z
Kurzbeschreibung (Abstract):

Modern architecture promotes a high demand for transparent building envelopes and especially glass facades. Commonly, facades are designed to fulfill a multitude of objectives such as superior aesthetic appearance, a higher degree of weathering reliability, quick installation, high transparency as well as economic and ecologic efficiency. For such glazing applications, often an assessment of acoustic properties and especially sound insulation abilities are required. Because of the complexity of such an experimental or computational investigation given the framing systems and glass unit compositions, a reliable and fairly accurate estimation of sound insulation properties of such systems becomes time-consuming and demanding. This paper provides a Machine Learning (ML) based estimation tool of acoustic properties (weighted sound insulation value Rw, STC and OITC) of different glazing set-ups. A sufficiently rich database was used to train several machine learning algorithms. The acoustic properties are determined by comparing the third-octave or octave band spectrum of the sound reduction index with a reference curve (typical curve for solid construction elements) specified in the standard DIN EN ISO 717-1. Sound insulation values can currently only be determined by complex and expensive experimental investigations or numerical simulations for certain glass set-ups. Hence, there is no efficient tool for convenient and reliable estimation of the sound insulation performance of glazing systems available at the moment. To this end, the engineering team led by the authors conducted extensive studies on various glazings consisting of different glass assemblies with varying glass, cavity and interlayer thicknesses and different types of interlayer and gas fillings. Based on our research outcomes, a comprehensive web-based prediction program, the so-called AI Tool, has been developed recently. This program can provide a quick analysis and accurate prediction of arbitrary glazing set-ups, interlayers and glazing infills. A series of laboratory tests were conducted to validate the predictions by the AI Tool. The goal of this program is to provide designers, engineers, and architects an effective and economically efficient tool to facilitate the design w.r.t. acoustical properties.

Fachbereich(e)/-gebiet(e): 13 Fachbereich Bau- und Umweltingenieurwissenschaften
13 Fachbereich Bau- und Umweltingenieurwissenschaften > Institut für Statik und Konstruktion
13 Fachbereich Bau- und Umweltingenieurwissenschaften > Institut für Statik und Konstruktion > Fachgebiet Statik
Hinterlegungsdatum: 03 Feb 2022 06:43
Letzte Änderung: 03 Feb 2022 10:56
PPN:
Export:
Suche nach Titel in: TUfind oder in Google
Frage zum Eintrag Frage zum Eintrag

Optionen (nur für Redakteure)
Redaktionelle Details anzeigen Redaktionelle Details anzeigen