Indoor air quality (IAQ), as determined by the concentrations of indoor air pollutants, can be predicted using either physically based mechanistic models or statistical models that are driven by measured data. In comparison with mechanistic models mostly used in unoccupied or scenario-based environments, statistical models have great potential to explore IAQ captured in large measurement campaigns or in real occupied environments. The present study carried out the first literature review of the use of statistical models to predict IAQ. The most commonly used statistical modeling methods were reviewed and their strengths and weaknesses discussed. Thirty-seven publications, in which statistical models were applied to predict IAQ, were identified. These studies were all published in the past decade, indicating the emergence of the awareness and application of machine learning and statistical modeling in the field of IAQ. The concentrations of indoor particulate matter (PM2.5 and PM10 ) were the most frequently studied parameters, followed by carbon dioxide and radon. The most popular statistical models applied to IAQ were artificial neural networks, multiple linear regression, partial least squares, and decision trees.

Machine learning and statistical models for predicting indoor air quality.

The effects of air pollution on health have been generating attention for years. A large number of pulmonologists have recently expressed concerns about this in an open letter to Dutch Members of Parliament. Air pollution arises mainly in all kinds of combustion processes; in addition, atmospheric chemical reactions play a role in the formation of ozone and particulate matter. Health effects are both acute (increase in daily mortality and morbidity after days with increased concentrations of air pollution) as well as chronic (shortened life span and increased incidence of respiratory and cardiovascular diseases in areas with elevated concentrations of air pollution). These effects already occur at concentrations that are clearly lower than those currently observed in the Netherlands.

Air pollution and health

The indoor environment directly affects health and comfort as humans spend most of the day indoors. However, improperly controlled ventilation systems can expend unnecessary energy and increase health risks, while improved thermal and air quality can often result in higher energy consumption. One way to approach this dilemma is by understanding the effectiveness of the variables influencing indoor air quality (IAQ) related health and comfort. The objective of this paper is to highlight evidence and variables from empirical and deterministic models, which are combined in analytical models that current machine learning techniques often overlook. This paper reviews the analytical models and identifies the corresponding input variables, discussing their application in models based on artificial neural networks (ANNs) and reinforcement learning (RL). ANN and RL models have accurately described non-linear systems with uncertain dynamics and provided predictive and adaptive control strategies. The first part of this study focuses on the most common thermal comfort models and their variables, mainly related to steady-state and adaptive models. The second part reviews typical models of determining indoor air pollutants and their relationship with ventilation requirements and health effects. Forty-five works closely related to the field are summarized in multiple tables. The last part identifies the factors needed to predict thermal comfort and IAQ.

Measuring the right factors: A review of variables and models for thermal comfort and indoor air quality

A review on interactions between energy performance of the buildings, outdoor air pollution and the indoor air quality

Association between respiratory health and indoor air pollution exposure in Canakkale, Turkey

This study investigated levels of airborne bacteria in 90 indoor sampling sites with no known complaints such as kindergartens, schools, homes and dorms over three seasons. Airborne bacteria sample...

Airborne bacteria levels in indoor urban environments: The influence of season and prevalence of sick building syndrome (SBS)

Volatile Organic Compounds (VOCs) are a concern due to their adverse health effects and extensive usage. Levels of indoor VOCs were measured in six homes located in three different towns in Canakkale, Turkey. Monthly indoor VOC samples were collected by passive sampling throughout a year. The highest levels of total volatile organic compounds (TVOC), benzene, toluene, and xylenes occurred in industrial, rural, and urban sites in a descending order. VOC levels were categorized as average values annually, during the heating period, and non-heating period. Several building/environmental factors together with occupants’ habits were scored to obtain a basic indoor air pollution index (IAPi) for the homes. Bivariate regression analysis was applied to find the associations between the pollutant levels and home scores. IAPi scores were found to be correlated with average indoor VOC levels. In particular, very strong associations were found for occupants’ habits. Furthermore, observed indoor VOC levels were categorized by using self-organizing map (SOM) and two simple scoring approaches, rounded average and maximum value methods, to classify the indoor environments based on their VOC compositions (IAPvoc). Three classes were used for both IAPi and IAPvoc approaches, namely “good”, “moderate”, and “bad”. There is an urgent need for indexing studies to determine the potential sources and/or factors affecting observed VOCs. This study gives a basic but good start for further studies.

Deniz Tasdibi

Papers

Association between respiratory health and indoor air pollution exposure in Canakkale, Turkey

Airborne bacteria levels in indoor urban environments: The influence of season and prevalence of sick building syndrome (SBS)

Estimation of sources and factors affecting indoor VOC levels using basic numerical methods

Assessment of residential exposure to Volatile Organic Compounds (VOCs) and Carbon Dioxide (CO2)