Análise de sensibilidade e propagação de incertezas em evaporação de gotas

Abstract

The droplet evaporation process, as well as the phenomenon of combustion, according to historical data is one of the oldest mankind processes, therefore both are widely used in everyday and industrial applications. Nevertheless, the antiquity and use of the process do not correspond to a complete and effective understanding of it, since the purely deterministic knowledge of these is usually loaded with errors and uncertainties (random and epistemic). In this context, the present work, seeking to increase the knowledge and mastery of the evaporation process of fuel droplets, performed a global sensitivity analysis and uncertainty propagation on the evaporation process of a n-heptane droplet in an inert nitrogen medium. Being the referred system initially evaluated by a deterministic perspective, through which it was possible to implement a set of equations related to thermal and mass changes associated with the Nusselt and Sherwood dimensionless numbers, the NuSh-nat equations. In a later step, the computational code of droplet evaporation was associated with the stochastic libraries Chaospy, Uncertainpy, and SALib so that the global sensitivity analysis was performed under three different systems of droplet evaporation, extracting from these the most relevant terms to the evaporation process so that later, in the stage of uncertainty propagation, these could be propagated to the final problem evaluated answer. Thus obtaining, at the end of the stochastic simulation process the usual confidence bands to the evaporation process of the n-heptane combustible droplet.