Title:
Experimental Measurement and Prediction Modeling of Concrete Thermal Conductivity
Author(s):
Sargam
Publication:
Web Session
Volume:
ws_F23_Sargam.pdf
Issue:
Appears on pages(s):
Keywords:
DOI:
Date:
10/29/2023
Abstract:
The thermal conductivity, k, is a significant property of concrete that affects the heat trans-fer mechanism, design, and energy efficiency of concrete-based structures. This study aimed to determine the effects of contemporary concrete components, such as SCMs, lightweight and recycled aggregates, fibers, and more, on concrete's thermal conductivity through experimentation. The results demonstrated varying degrees of the influence of these parameters, with aggregate mineralogy having a considerable effect while polypropylene fibers having little impact. To address the challenge of measuring k for every concrete structure with a sophisticated test procedure, a machine learning (ML) based prediction model was developed using literature data and experimentally measured data. After assessing a range of parameters and models, a final multilayer perceptron model was adopted that performed well with mean absolute errors of 0.07, 0.14, and 0.10 W/m-K on the training, validation, and independent test sets, respectively. The developed model can provide valuable information for informed decision-making in the design and construction of critical structural elements.