Thermal analysis of multilayered composite walls: experimental investigation and comparison with tolerance modeling and FEM
Year: 2026
Type: article
Abstract: This paper presents a comparative study of experimental, analytical, and numerical results on steady-state temperature distribution in multilayer, two-component structures. The analytical approach uses the local homogenization model (LHM) from tolerance modeling (TM) and the one-dimensional resistance method, with numerical simulations conducted in COMSOL Multiphysics and Mathematica. The study focuses on a periodic structure consisting of six alternating layers of oriented strand board (OSB, 18 mm) and expanded polystyrene (EPS, 20 mm), analyzed within a (0, 228 mm) $$\times $$ × (0, 240 mm) $$\times $$ × (0, 240 mm) domain. Analytical and numerical models show strong convergence for the 1D problem. The LHM predictions reproduce the overall trends but exhibit a systematic shift when compared with the FEM results (the difference is less than $$0.5 ^{\circ }$$ 0 . 5 ∘ C for the analyzed structures). This discrepancy does not affect the overall heat flux assessment and can be easily corrected by applying a shape function shifted by $$\eta /2)$$ η / 2 ) . Experimental measurements conducted in a climate chamber agree well with 3D simulations, with an average deviation of 4.2%. A second structure made of autoclaved aerated concrete (AAC, 50 mm) and EPS (20 mm) exhibited an average deviation of 6.1% due to imperfect thermal contact between layers. The paper also examines a two-component Functionally Graded Material (FGM) structure designed using the TM method to achieve a target macro-temperature profile.
Authors: Olga Szlachetka, Ivan Giorgio
Institutions: Warsaw University of Life Sciences, University of L'Aquila
Language: English
Topic: Composite Material Mechanics
Subfield: Mechanics of Materials
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,590
Funder: Narodowe Centrum Nauki
Award: G6554613254
OpenAlex ID: W7162114412
