作者: Gudrun Thäter , Mathias J. Krause , Hermann Nirschl , Márcio Dorn , Robin Trunk
DOI: 10.3390/COMPUTATION9040040
关键词:
摘要: A detailed knowledge of the influence a particle’s shape on its settling behavior is useful for prediction and design separation processes. Models in available literature usually fit given function to experimental data. In this work, constructive data-driven approach presented obtain new drag correlations. To date, only considered parameters are derivatives axis lengths sphericity. This does not cover all relevant effects, since process arbitrarily shaped particles highly complex. work extends list by, e.g., convexity roundness evaluates relevance each. The aim find models describing coefficient velocity, based extended set parameters. data investigations obtained by surface resolved simulations superellipsoids, applying homogenized lattice Boltzmann method. closely study shape, equal volume, therefore range Reynolds numbers, limited [9.64, 22.86]. Logistic polynomial regressions performed quality investigated with further statistical methods. addition studied relation between number, dependency terminal velocity also investigated. found are, an adjusted determination 0.96 0.86, good agreement data, yielding mean deviation below 5.5% training test dataset.