85-4

H. RAMASWAMY, Food Science and Agricultural Chemistry, McGill University, Macdonald Campus, 21111 Lakeshore Road, Ste Anne de Bellevue, QC H9X 3V9, Canada and M. R. Zareifard, Food Science and Agricultural Chemistry, McGill University, Macdonald Campus, 21111 Lakeshore Road, Ste Anne de Bellevue, QC H9X 3V9, Canada.

Availability of data on fluid-to-particle heat transfer coefficient (hfp) is important for the process design of particulate food in continuous flow systems. Experimental determination of hfp is a difficult task. Hence, available data on hfp should be modeled through dimensionless correlations to expand their utilization for scale-up conditions.

The objective of this research was to develop dimensionless correlations for relating heat transfer coefficients for spherical particles under forced convection tube-flow conditions.

Two different techniques were employed to evaluate hfp under tube-flow conditions: (1) particle oscillatory motion and 2) a calorimetric approach previously developed in the laboratory. About 600 experiments were carried out with spherical particles of different materials (Nylon and aluminum) and sizes (12.7-25.4 mm). Model particles were subjected to heat transfer under different experimental conditions (carrier fluid 0-1% CMC; fluid temperature 50-90^oC; particle linear velocity 0.06-0.21 m/sec using the particle oscillatory motion method; and fluid flow rate 9-19 L/min using the calorimetric method). A multiple regression and backward elimination procedure was used to include only statistically significant dimensionless groups.

The material properties, particle size, and velocity were found to have significant effects on the hfp and hence the Nusselt number. A series of correlations were developed in the form of Nusselt number as a function of other influencing dimensionless numbers such as Reynolds and Prandtl numbers. Introducing some nonconventional ratios such as the diffusivity ratio, defined as the ratio of fluid to particle thermal diffusivity, particle-to-tube diameter ratio (d/D), and particle-to-fluid velocity ratio (Vp/Vf) was found to improve the model performance. Coefficients of determination (R²) ranging from 0.88 to 0.99 were obtained for different correlations between the Nusselt number and other dimensionless group numbers.

Dimensionless correlations were developed to relate values of hfp under tube-flow conditions. Model predictions fitted the experimental data with good coefficients of determination. Dimensionless correlations, like the ones developed in these studies, are useful in generalizing the influence of process variables and scale-up conditions.