Article from Journal of Food Engineering 153:63-72
Abstract
A novel method of reciprocating agitation for enhancing heat transfer during thermal processing of canned particulates in Newtonian fluid (glycerin) was investigated in this study. First, a lab-scale reciprocating agitation retort was developed by modifying an existing conventional still retort to include a reciprocating mechanism providing reciprocations at a frequency of 0-5 Hz and amplitude of 0-30 cm. Temperature distribution studies inside the modified retort and inside the reciprocating cans revealed better temperature uniformity during reciprocating agitation as compared to still mode. Heat penetration studies were subsequently conducted to evaluate the influence of reciprocating agitation on heat transfer coefficients (overall U, and fluid-to-particle hfp), process time and quality loss. Reciprocating agitation resulted in 2-7 times enhancement in U & hfp leading to 52-87% reduction in equilibration time of the cold spot. This enabled 46-62% reduction in process time resulting in 26-36% reduction in quality-deterioration index (cook-value/lethality) as compared to conventional still mode of thermal processing. Thus, reciprocating agitation was shown to have potential to deliver high quality products. The effect of container orientation on heat transfer during reciprocation agitation thermal processing was then studied by placing experimental cans in one of the three possible orientations viz. horizontally along axis of reciprocation (HA), horizontally perpendicular to axis of reciprocation (HP) & vertically (V). HA orientation provided most rapid heat transfer followed by HP and V orientations, respectively. Here, it was seen that a high reciprocation intensity was not required to decrease process time considerably and a mild reciprocation intensity was sufficient. HA orientation was recommended for processing liquid foods and HP or V orientation was optimal for liquid particulate food mixtures. Influence of various process variables (operating temperature, reciprocation frequency, reciprocation amplitude, container headspace, liquid viscosity/concentration, particle density & particle concentration) on U and hfp were estimated for two situations: i) cans filled with single particle (liquid-only situation) and ii) cans filled with multiple particles. In single particle study, both U and hfp were influenced significantly by reciprocation frequency, reciprocation amplitude, operating temperature and liquid viscosity. Headspace affected only U, while particle density affected only hfp. In general, higher U and hfp were obtained at lower liquid viscosity and at higher P a g e | v frequency, amplitude, temperature and headspace. In the multiple particle study, it was generally observed that the magnitude of U and hfp were larger as compared to single-particle (liquid-only) scenario due to higher level of turbulence/mixing inside the cans on account of collisions amongst particles. U & hfp followed the trend: Nylon > polypropylene > Teflon, while, the rate of temperature increase followed the trend: polypropylene > Nylon > Teflon. U & hfp were found to increase on increasing particle concentration, however, beyond an optimal concentration U & hfp decreased due to dominance of conduction based heating. Based on the results of these studies, simultaneous multi-objective optimization of heat transfer coefficients (U & hfp – to maximize heat transfer) and reciprocation intensity (RI – to minimize agitation losses) was conducted. For this, a composite model was developed using all available data and was then subjected to numerical optimization. High RI (37-45 ms-2 ) was recommended to maximize U & hfp, whereas lower RI (16-19 ms-2 ) was found optimal for simultaneous optimization of U, hfp & RI. Product composition containing low viscous liquids filled with 23-27% concentration of particles with density of 1134-1351 kg/m3 were found most desirable. Optimal conditions were also reported for different operating temperatures, liquid viscosities, particle concentrations and particle densities. Dimensionless correlations were then developed for predictive modeling and scale-up of reciprocating agitation thermal process using multiple non-linear regressions of significant dimensionless groups. Dimension of can along the axis of reciprocation was found to play a dominant role in the heat transfer phenomenon and was included in the characteristic length formulated for Nusselt number at can wall (from U). On comparing results of this study with literature, it was clear that forced convection effects and turbulence levels during reciprocating agitation were significantly larger than other modes of rotary agitation. Yet, effect of natural convection could not be ignored at lower RI. Finally, flow visualization studies were carried out to characterize particle motion/mixing behavior of liquid/particle mixtures by videotaping particle motion/mixing in transparent containers under various conditions. The nature of mixing and particle motion was found markedly different for different orientations and particle densities. Mixing time was affected by frequency, liquid viscosity, container orientation, particle concentration and particle density.