Estimation of natural ventilation parameters by a Bayesian approach

Abstract

The natuml ventilation pammeters are often predicted using the fracer gas method, but accurate prediction by this method is difficult because of fluctuating environmental conditions and the resulting uncertainty Therefore, when estimating these paiameters, the estimates and their uncertainty should be concunently evaluated. In this study, we propose a new estimation method employing a Bayesian approach that yields not only the estimates, but also the probabthty distribution of estimates. We repmduced data fibm a tncer gas experiment by using computational fluid damics (CFD) simulation and estimated the ventilation mite and effective room volume by using two different methods: pmbabthstic estimation based on Bayes' theorem, and deterministic estimation using the quasi-Ne4on method. In the probabilistic estimation, a method of simultaneously estimating the degree of the difference between the measwement and the prediction value required for estimation was used. Additionally, the advantages of the probabthstic approach over the deterministic approach were identified.