Odour-sensitive passenger comfort in small aircraft cabins

Abstract

This work presents a numerical investigation targeting to simulate the slice of a small aircraft cabin as an experimental facility with a controlled environment, to assess passenger comfort when exposed to high volatile organic compound (VOC) concentrations. The mixing and transport of chemical species are evaluated using computational fluid dynamics for 800 s of in-cabin actual flow time and measurements are taken every 10 s from selected computational nodes close to the passengers’ noses. The results are used to create a dataset that trains four different machine learning classifiers, namely, the Random Forest, Support Vector Machine, Logistic Regression and Naive Bayes, and their performance is compared. Moreover, an additional simulation of the cabin with a filtering system utilising high-efficiency particulate air and activated carbon filters is conducted, to evaluate the impact of the molecular weight of the compounds on their residence time, and compare it to the simulation without the filters. Results indicate that the model is insensitive to the inlet air mass flow variation and that the mass of the VOCs measured in the monitored computational nodes remains relatively unaffected, meaning that the impact of the air-conditioning system setting is minor. Additionally, a Boruta feature selection algorithm is used to determine the importance of each measurement of the simulation and to form a dataset that will train the four machine learning classifiers. Furthermore, the comparison of the two simulations, the one with and the one without the filters, indicates that the residence time (RT) of the compounds is independent of their molecular weight, as they all show equivalent percentile reductions, with the naphthalene and styrene showing a 28.5% and 28.3% reduction respectively, compared to the simulation without the filters. Finally, in-cabin flow irregularities are present, disrupting the flow symmetry and suggesting that not all passengers share the same traveling experience.