Statistical prediction of the indoor 03 concentration level in eight schools in La Rochelle, France

The ozone (O3) has become an interesting atmospheric pollutant due to the complexity of thetransfer phenomenon to the indoor environment. The study of this phenomenon representedthe main objective of the French national research program PRIMEQUAL PREDIT. Theexperimental work, carried out in eight schools in La Rochelle France, proved that the O3 transfer could be predicted as a function of the outdoor concentration level and the airtightness of the building faade. A neuro-fuzzy model was learned in order to predict the pollutant transfer and finally the indoor O3 concentration level as a function of the two parameters. The prediction consists in 2D mapping of three parameters: I/O ratio, prediction dispersion and credibility. The test procedure of the model underlines its great flexibility. The model can be used in order to predict the indoor ozone level in any kind of building and to any outdoor ozone concentration level.