Application of the REA to Describe Droplet Drying Kinetics
The REA has been used to model drying of a range of food materials such as pulped kiwifruit leather, whey protein concentrate, lactose, skim milk powder, whole milk powder, cream, and mixtures of sugars (Chen et al., 2001; Chen and Lin, 2005; Lin and Chen, 2005, 2007). Results showed that this approach models moisture content and temperature profile versus drying time very accurately. For example, modeling of drying of aqueous lactose solution droplets showed that the average absolute difference of weight loss profile was about 1% of the initial weight, while that of the temperature profile was about 1.2°C. Moreover, application of the REA to model drying of cream and whey protein concentrate showed average absolute errors of weight profiles of 1.9% and 2.1% respectively, while errors for temperature were about 3°C and 1.9°C respectively (Lin and Chen, 2007). Modeling of skim milk and whole milk powder by the REA was also accurate (Chen and Lin, 2005).
The REA has been implemented in computational fluid dynamics (CFD)-based simulations of spray dryers to couple the dispersed phase (droplets dried) and the continuous phase (drying air) (Woo et al., 2008b; Jin and Chen, 2009a,b, 2000). CFD simulations using the REA can predict outlet air temperature and outlet particle moisture content reasonably well compared with the experimental data. In addition, the REA was also implemented to predict the evaporation zone, drying rate, trajectory of particles, and deposition of particles in spray dryers (Woo et al., 2008b). The application of CFD in conjunction with the REA to describe the performance of industrial-scale spray dryers in 2-D and 3-D was conducted (Jin and Chen, 2009a,b). Patel et al. (2009a) have extended the “single solid component” approach of the REA to a composite REA model - drying kinetics for mixtures of “non-interacting” solutes (maltodextrin and sucrose). The activation energy of the mixture was determined based on the mass fraction of each solute and their corresponding activation energies. It was shown that the average relative error between experimental and calculated data was below 1.5% for droplet weight and below 3% for droplet temperature (Patel et al., 2009a).