Numerical Study on the Dynamic Performance of Timber–Steel–Concrete Composite Floor Unit

Sammanfattning

This study investigates the serviceability behaviour of a prefabricated timber-steel-concretecomposite floor unit through a script-based finite element model in ABAQUS/CAE. The workalso includes a python-based digital twin that consist of approximately 13 000 rows of code,which was not included in the appendices. The model, combining glulam timber beams, thin-walled metal webs (Posi-struts), and a concrete slab, was validated against full-scale experimentaleigenfrequencies and acceleration-response envelopes, reproducing the first four vertical bendingmodes within a maximum deviation of 5.58%. A parametric assessment was carried outcovering eight posi-strut layout configurations, four glulam stiffness grades (GL22h -GL32h),four concrete stiffness grades (C20/25 -C35/45), and nine damping levels. The serviceabilityresponse was evaluated in terms of fundamental natural frequency, static deflection under a 1kN point load, and load-normalised RMS accelerance in the 1.5 -8.0 Hz walking band. The results indicate that strut arrangement governs the global stiffness distribution more stronglythan strut count alone, with Layout 1 producing the highest natural frequency (7.51 Hz) andlowest static deflection (6.85 mm). Glulam stiffness had the greatest influence on frequency anddeflection, while concrete stiffness variation within practical grade ranges had a limited effect.Damping was identified as the parameter most directly controlling the peak RMS accelerationamplitude, with a 10% increase reducing the peak accelerance by 9.09%. All configurationssatisfied the Eurocode 5 minimum frequency criterion and were classified within Level III forstiffness and acceleration. Under the JRC-ECCS framework, the baseline floor corresponded tocomfort class E at 1% damping, acceptable for hotel use.