Design and Uncertainty Analysis of a Series of Atomization Experiments in Seven Variables

Abstract

Atomization of liquids is a phenomenon that involves many variables. In order to find correlations for the effects of an extended set of variables, dimensional analysis and experimental design are used, resulting in experimental plans with very reduced number of necessary data points. Means were found to further reduce the cost of the experiments by adjustments of the ranges of the variables and the selection of the parameters and experimental sources of variability. The generation and propagation of uncertainties in the experiments are discussed, and estimates for the true lack of fit of the approximations obtained are found. The selection of a quadratic polynomial approximation seems adequate, given the relative magnitude of the other sources of uncertainty. The effects of some dimensionless parameters with physical meaning are isolated for the first time.