Tomographic Identification of Bubbles in Two-Phase Flows Using a Genetic Algorithm

Genetic algorithm (GA)-based tomographic reconstruction algorithm is examined to use for nonintrusive optical or acoustic imaging of bubbles in two-phase flows. Individual bubble boundaries are described by elliptical basis functions whose major and minor axes specify the bubble shape and sizes. Genetic algorithm, a robust combinatorial non-linear optimization based on fittest survival principle, allows a regressive determination of the bubble center locations, shape and sizes simultaneously by maximizing the fidelity of guessed images compared with measured images. Preliminary results indicate a strong potential of GA-based tomography for accurate reconstruction of two-phase flow images.