Optimisation of interaction parameters for CFD modelling of multiphase flow using NR method

摘要

Accurate modelling of fluid and particle flow is essential for reliable predictions of gas-solid reactions in the multiphase reactor. Due to the promising features such as short contact time, less back mixing of solids and low residence time, downer column has been used in industries than conventional riser configuration. However, the effective simulation and validation of the hydrodynamic characteristics have been scarce. A RANS model with an appropriate Eulerian-Eulerian approach poses a challenging task of choosing the hydrodynamic simulation parameters, such as i) particle-particle restitution coefficient, ii) particle-wall restitution coefficient and iii) specularity coefficient. The objective of the present study was to optimise these model parameters to obtain close validation with experimental data so that accurate downer hydrodynamics can be predicted. The error obtained from validating the experimental data with the CFD simulation output is our function which needs to be minimised. Newton Raphson multi-parameter root finding method was used for this purpose. The Jacobian evaluation required multiple CFD simulations at each iteration. A 61% reduction in error was achieved after 3 iterations. Thus, the optimised model parameters of hydrodynamics used for further modelling study in predicting hydrodynamic characteristics of a 200 mm diameter 8 m column.