Abstract

The aim of this work is to develop a mathematical model to estimate the batch drying curve of coriander seeds in an impingement dryer and to study the axial movement of a seed in a transparent prototype impingement dryer. The apparatus is a horizontal acrylic transparent cylinder with a slight slope to induce the axial and rotational movement of particles. Gas enters tangentially downwards through a narrow slot arranged all along the dryer, flows in a counterclockwise circular motion in the chamber - in crossflow with respect to the solids - and is discharged through an upper lengthwise expansion chamber. As a result of gas drag, the particles advance in a rotational-helicoidal motion between feed and discharge. Velocity and temperature profiles for gas in 2D turbulent flow were simulated using commercial software from Fluent Inc. Maximum velocities are shown to be located close the walls; most of the gas is recirculated, and the rest is exhausted. It is assumed that particle trajectories also follow a circular motion near the walls, as observed in the transparent reproduction of the dryer operating with ambient air for small batch of solids and/or a single particle. Air velocities along this trajectory are estimated from the simulated flow field. Particle motion, heating, and drying along this path are described by unsteady momentum, heat, and mass balances when subjected to gas drag and gravity forces. With respect to the axial trajectory of a coriander seed, for an inlet air velocity of 20 m/s at the slot the average experimental time for a complete circular cycle is 0.18 s and the simulated time is 0.21 s, whereas average experimental residence time is 1.53 s and the simulated time is 0.94 s. Differences between experimental results and simulations are due to air instability, leading to nonhomogeneous air velocity profiles along the equipment. The mathematical model is based on the assumption that air velocity profiles are homogeneous. Experimental observations indicate that the particle does not move along the equipment but sometimes moves backward (or erratically) or spins out advancing, due to an uneven air speed profile, and impacts against the wall. Finally, the drying model gives results that adjust to the batch experimental data, taking into account the deviations found with respect to the axial trajectory from a seed. This is because the model was devised exactly to predict the conduct of the system in batch operation for a particle bed, obtaining results that show the macrocospic response of the equipment (velocity and average temperature of the air). As it happens in this type of phenomenon, the drying rate in the constant period is a function of the adimensional Reynolds number.