A novel method for analysing indoor radon concentration measurements

Abstract

Radon is a radioactive gas which, when it accumulates in a room, can have a negative effect on the persons in it. The mathematical model presented in the study, based on statistics and log-normal distribution, allows recommendations to be developed on optimal statistical parameters for radon measurements in buildings. This paper presents a novel method for analysing indoor air quality based on indoor radon concentration measurements using machine learning methods. The study used a k-means algorithm to isolate three periods with similar radon concentration parameters. An assessment of the variability of the radon measurements depending on the height of the location of the detectors in the room was carried out, from which it was concluded that similar distributions are obtained at the height of the breathing zone or higher. The results of the study indicate that short-term active measurements taken during the winter period underestimated the median of long-term measurements by only 5 %. Weekly measurement data from the winter period was sufficient to estimate the expected annual average for the building. The conclusions obtained in the article lead to the initiation of a discussion on past requirement passive long-term radon measurements. The ability to reproduce the algorithm under different conditions of building location and use will allow a global evaluation of short-term radon measurements to be evaluated in the context of long-term measurements.