Just Suspended Speed Simulation in Torus Reactor Using Multiple Non-Linear Regression Model

In the chemical and water treatment industries, it is necessary to achieve maximum contact between solid liquid phase, thus promoting mass heat transfer, obtain a homogeneous solution. Increasing stirring speed most recommended solution in different types of reactors: stirred tank, column, tubular. However, this inadvertently increases energy consumption industry. Determination minimum speed, labeled just suspended (Njs) crucial attaining homogeneity, has been widely investigated. Numerous studies have carried out assess formulas for determining particle various reactor types. Given limitations existing formulations based on generalization unique equation computing Njs all soil classifications, appears that can only approximate complex phenomena depend several parameters. A novel formula was developed, results given paper demonstrate effectiveness generating significant uncertainties estimation Njs. The purpose study elaboration experiment-based data-driven calculate size classes. Nonlinear multiple regression (MNLR) models were used generate new formulas. gradient descent optimization algorithm employed solve hyperparameters each equation, utilizing supervised learning. comparison data indicated presented outperformed empirical provide useful means lowering consumption, while increasing transfer torus type reactors.