A sequence of two optimum experiments involving a three-layer set-up to retrieve simultaneously the directional thermal conductivities of fiber-reinforced (FR) composites is simulated through finite element (FE) analysis. The experimental set-up consists of a thin electrical heater sandwiched between two larger, identical plates of composite material. The heater heat capacity effect on the identifiability of the out-of- and in-plane thermal conductivities (namely, k x and k y ) of the composite is assessed by applying the approximation error model (AEM). This is a technique useful to compensate the modeling errors affecting the accuracy of a reduced model, avoiding the complexity and computational costs of a complete model when computing an inverse solution. In this paper the FE model (acts as the complete model) considers the heater as a lumped capacitance body, whereas the forward problem (reduced model) completely neglects this disturbance by modeling the sample as an orthotropic rectangular plate partially heated at the front boundary and exposed to free convection at the backside (unknown heat transfer coefficient h). The combined application of FEM and AEM makes the heater heat capacity effect clearly recognizable. As it was not comprehensively investigated for the analyzed set-up, this is the novelty of the paper. Temperature measurements at the unheated surface of the sample are simulated through the FE model and canonical Monte Carlo (MC) method. This simulated data is used in the Gauss-Newton algorithm, under both the reduced model and AEM approach, to estimate simultaneously k x , k y and h . Results of the MC simulations show that the AEM approach ensures the identifiability of all unknowns, while under the reduced model only k y can be retrieved with high accuracy.

Assessment of the identifiability of directional thermal conductivities of FR composites using an optimized three-layer experimental set-up and the approximation error model

D'Alessandro, Giampaolo
;
2026-01-01

Abstract

A sequence of two optimum experiments involving a three-layer set-up to retrieve simultaneously the directional thermal conductivities of fiber-reinforced (FR) composites is simulated through finite element (FE) analysis. The experimental set-up consists of a thin electrical heater sandwiched between two larger, identical plates of composite material. The heater heat capacity effect on the identifiability of the out-of- and in-plane thermal conductivities (namely, k x and k y ) of the composite is assessed by applying the approximation error model (AEM). This is a technique useful to compensate the modeling errors affecting the accuracy of a reduced model, avoiding the complexity and computational costs of a complete model when computing an inverse solution. In this paper the FE model (acts as the complete model) considers the heater as a lumped capacitance body, whereas the forward problem (reduced model) completely neglects this disturbance by modeling the sample as an orthotropic rectangular plate partially heated at the front boundary and exposed to free convection at the backside (unknown heat transfer coefficient h). The combined application of FEM and AEM makes the heater heat capacity effect clearly recognizable. As it was not comprehensively investigated for the analyzed set-up, this is the novelty of the paper. Temperature measurements at the unheated surface of the sample are simulated through the FE model and canonical Monte Carlo (MC) method. This simulated data is used in the Gauss-Newton algorithm, under both the reduced model and AEM approach, to estimate simultaneously k x , k y and h . Results of the MC simulations show that the AEM approach ensures the identifiability of all unknowns, while under the reduced model only k y can be retrieved with high accuracy.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11697/291501
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact