Modelling convective heat transfer to non-spherical particles

A powder is a composition of many unique particles in almost all cases. The grains have different shapes and sizes which influence drag and heat transfer, so in multiphase flows, each particle also interacts differently with the continuous phase. In this work, a practical approach to calculate the convective in-flight heating of differently shaped, non-spherical particles, is presented. A model from literature, which uses the same parameters for the calculation of particle drag coefficients and Nusselt numbers, was selected. The model parameters were determined by simply measuring the settling velocity of multiple particles in still air. A novel method for particle classification, which has already been proven to be essential for precise drag calculations of powders with differently shaped grains, was extended for heat transfer calculations. The approach was validated by numerical calculations and experimental data of a simple test rig where the particles were heated from room temperature in free fall through hot air. After a drop height of 1.7 m, the mean particle temperature was measured with a specially developed device. The particle temperature in the test rig increased in the order of 300 K. The commercial CFD-code Ansys Fluent 17.0 in combination with an Euler-Lagrangian approach for multiphase flow calculations was used for the simulations. The numerically and experimentally determined particle temperatures agreed very well in the case of spherical and as well for non-spherical particles. In comparison to the standard model of the CFD code, which only considers spherical particles, the error in calculating the mean particle temperature was reduced from −59 K to just 2 K. Moreover, it was shown that the standard deviation of the particle temperature distribution increases significantly when the different particle shapes are included in the calculation. It was concluded that the particle classification improves the calculation of maximum and minimum particle peak temperatures. The proposed approach is fully independent of the particle type and it can be used to determine the heat transfer characteristics of various different powders. Further, the customized model can be added to nearly every CFD-code on the market and improve the results considerably while computational cost remains low due to the numerically efficient Euler-Lagrangian approach.