Identification of parameters in building concentration dispersion model

TL;DR

This paper develops a Bayesian inference-based method to identify parameters in a building pollutant dispersion model, focusing on CO2 transport, validated against known cases and software, with plans for further experimental calibration.

Contribution

It introduces a novel Bayesian approach to parameter identification in building dispersion models, enhancing accuracy and validation methods.

Findings

Numerical results align well with existing software.

Bayesian inference effectively estimates model parameters.

Further experimental calibration is needed for refinement.

Abstract

The aim of this work is to simulate the pollutants transport in buildings. Focusing mainly on the presence of CO2, firstly we resolve the airflow equations for two typical validation cases, the Rao case and the IEA case. These numerical results are compared to the most known software and they are used to evaluate of the evolution of CO2 concentration in the different rooms. In order to obtain the different parameters and filters of the proposed model we use a statistical method based on Bayesian inference. The final comparison of results is coherent but a complementary experimental procedure is necessary to calibrate and refine the model