An inverse modelling approach to estimate the hygric parameters of clay-based masonry during a Moisture Buffer Value test

Samuel Dubois, Fionn McGregor, Arnaud Evrard, Andrew Heath, Frédéric Lebeau

Research output: Contribution to journalArticle

25 Citations (Scopus)
66 Downloads (Pure)

Abstract

This paper presents an inverse modelling approach for parameter estimation of a model dedicated to the description of moisture mass transfer in porous hygroscopic building materials. The hygric behaviour of unfired clay-based masonry samples is specifically studied here and the Moisture Buffer Value (MBV) protocol is proposed as a data source from which it is possible to estimate several parameters at once. Those include materials properties and experimental parameters. For this purpose, the mass of two clay samples with different compositions is continuously monitored during several consecutive humidity cycles in isothermal conditions. Independently of these dynamic experimental tests, their moisture storage and transport parameters are measured with standard steady-state methods.

A simple moisture transfer model developed in COMSOL Multiphysics is used to predict the moisture uptake/release behaviour during the MBV tests. The set of model parameters values that minimizes the difference between simulated and experimental results is then automatically estimated using an inverse modelling algorithm based on Bayesian techniques. For materials properties, the optimized parameters values are compared to values that were experimentally measured in steady state. And because a precise understanding of parameters is needed to assess the confidence in the inverse modelling results, a sensitivity analysis of the model is also provided.
Original languageEnglish
Pages (from-to)192-203
Number of pages12
JournalBuilding and Environment
Volume81
Early online date28 Jun 2014
DOIs
Publication statusPublished - 30 Nov 2014

Keywords

  • moisture buffer value
  • clay
  • HAM modelling
  • parameter estimation
  • MCMC
  • DREAM

Cite this

An inverse modelling approach to estimate the hygric parameters of clay-based masonry during a Moisture Buffer Value test. / Dubois, Samuel; McGregor, Fionn; Evrard, Arnaud; Heath, Andrew; Lebeau, Frédéric.

In: Building and Environment, Vol. 81, 30.11.2014, p. 192-203.

Research output: Contribution to journalArticle

Dubois, Samuel ; McGregor, Fionn ; Evrard, Arnaud ; Heath, Andrew ; Lebeau, Frédéric. / An inverse modelling approach to estimate the hygric parameters of clay-based masonry during a Moisture Buffer Value test. In: Building and Environment. 2014 ; Vol. 81. pp. 192-203.
@article{8e5243c3a76b4587845a0b0ae372ebda,
title = "An inverse modelling approach to estimate the hygric parameters of clay-based masonry during a Moisture Buffer Value test",
abstract = "This paper presents an inverse modelling approach for parameter estimation of a model dedicated to the description of moisture mass transfer in porous hygroscopic building materials. The hygric behaviour of unfired clay-based masonry samples is specifically studied here and the Moisture Buffer Value (MBV) protocol is proposed as a data source from which it is possible to estimate several parameters at once. Those include materials properties and experimental parameters. For this purpose, the mass of two clay samples with different compositions is continuously monitored during several consecutive humidity cycles in isothermal conditions. Independently of these dynamic experimental tests, their moisture storage and transport parameters are measured with standard steady-state methods.A simple moisture transfer model developed in COMSOL Multiphysics is used to predict the moisture uptake/release behaviour during the MBV tests. The set of model parameters values that minimizes the difference between simulated and experimental results is then automatically estimated using an inverse modelling algorithm based on Bayesian techniques. For materials properties, the optimized parameters values are compared to values that were experimentally measured in steady state. And because a precise understanding of parameters is needed to assess the confidence in the inverse modelling results, a sensitivity analysis of the model is also provided.",
keywords = "moisture buffer value, clay, HAM modelling, parameter estimation, MCMC, DREAM",
author = "Samuel Dubois and Fionn McGregor and Arnaud Evrard and Andrew Heath and Fr{\'e}d{\'e}ric Lebeau",
year = "2014",
month = "11",
day = "30",
doi = "10.1016/j.buildenv.2014.06.018",
language = "English",
volume = "81",
pages = "192--203",
journal = "Building and Environment",
issn = "0360-1323",
publisher = "Elsevier",

}

TY - JOUR

T1 - An inverse modelling approach to estimate the hygric parameters of clay-based masonry during a Moisture Buffer Value test

AU - Dubois, Samuel

AU - McGregor, Fionn

AU - Evrard, Arnaud

AU - Heath, Andrew

AU - Lebeau, Frédéric

PY - 2014/11/30

Y1 - 2014/11/30

N2 - This paper presents an inverse modelling approach for parameter estimation of a model dedicated to the description of moisture mass transfer in porous hygroscopic building materials. The hygric behaviour of unfired clay-based masonry samples is specifically studied here and the Moisture Buffer Value (MBV) protocol is proposed as a data source from which it is possible to estimate several parameters at once. Those include materials properties and experimental parameters. For this purpose, the mass of two clay samples with different compositions is continuously monitored during several consecutive humidity cycles in isothermal conditions. Independently of these dynamic experimental tests, their moisture storage and transport parameters are measured with standard steady-state methods.A simple moisture transfer model developed in COMSOL Multiphysics is used to predict the moisture uptake/release behaviour during the MBV tests. The set of model parameters values that minimizes the difference between simulated and experimental results is then automatically estimated using an inverse modelling algorithm based on Bayesian techniques. For materials properties, the optimized parameters values are compared to values that were experimentally measured in steady state. And because a precise understanding of parameters is needed to assess the confidence in the inverse modelling results, a sensitivity analysis of the model is also provided.

AB - This paper presents an inverse modelling approach for parameter estimation of a model dedicated to the description of moisture mass transfer in porous hygroscopic building materials. The hygric behaviour of unfired clay-based masonry samples is specifically studied here and the Moisture Buffer Value (MBV) protocol is proposed as a data source from which it is possible to estimate several parameters at once. Those include materials properties and experimental parameters. For this purpose, the mass of two clay samples with different compositions is continuously monitored during several consecutive humidity cycles in isothermal conditions. Independently of these dynamic experimental tests, their moisture storage and transport parameters are measured with standard steady-state methods.A simple moisture transfer model developed in COMSOL Multiphysics is used to predict the moisture uptake/release behaviour during the MBV tests. The set of model parameters values that minimizes the difference between simulated and experimental results is then automatically estimated using an inverse modelling algorithm based on Bayesian techniques. For materials properties, the optimized parameters values are compared to values that were experimentally measured in steady state. And because a precise understanding of parameters is needed to assess the confidence in the inverse modelling results, a sensitivity analysis of the model is also provided.

KW - moisture buffer value

KW - clay

KW - HAM modelling

KW - parameter estimation

KW - MCMC

KW - DREAM

U2 - 10.1016/j.buildenv.2014.06.018

DO - 10.1016/j.buildenv.2014.06.018

M3 - Article

VL - 81

SP - 192

EP - 203

JO - Building and Environment

JF - Building and Environment

SN - 0360-1323

ER -