Building materials VOC emissions - a systematic parametric study

Mathematical models that are used to predict VOC emission of building materials require four basic parameters as input: VOC diffusion coefficient of the material (D-m), VOC partition coefficient (k), material thickness (H) and surface air velocity (v). This paper describes a four-factor simulation design and a statistical procedure for analyzing material VOC time (Y) sensitivities with respect to these four parameter variations and their interaction effects. The parameter effects and their interaction effects are discussed in term of Omega Squared (omega(2)). omega(2) is the ratio of sum of squared deviation of a factor to total sum of squared deviation (SST). The statistical analysis involving the comparison of to 2 of the four parameters and their interactions indicates that two parameters, D-m and H, and their interaction, D-m x H, significantly affect the material VOC time. Other parameters (k, v) and interactions (D-m x k, D-m x v, k x H, k x v, H x v, D-m x k x H, D-m x k x v, D-m x H x v, k x H x v, D-m x k x H x v) have minor effects on the material VOC time. Further statistical analysis of the marginal mean values of the material VOC time for the three significant effects shows that: (1) The material VOC time decreases rapidly as the diffusion coefficient increases. (2) The material VOC time increases significantly as the material thickness increases. (3) There is a clear interaction between the diffusion coefficient and material thickness. As the material thickness increases, the material VOC time increase rate is smaller at high diffusion coefficient than at low diffusion coefficient. (C) 2003 Published by Elsevier Science Ltd.