Overall heat transfer coefficient (U) and fluid to particle heat transfer coefficient (h$ sb{ rm fp}$) in canned liquid/particle mixtures, subjected to end-over-end processing, were experimentally evaluated. A methodology was developed to measure heat transfer coefficients, while allowing the movement of the particle inside the can, by attaching it to a flexible fine wire thermocouple. Heat transfer coefficients were evaluated under two conditions: (1) with a single particle in the can, (2) with multiple particles in the can. / In the single particle experiments, a spherical particle was used to evaluate the influence of system parameters and product parameters on associated U and h$ sb{ rm fp}$. Increasing values of all four system variables improved the heat transfer coefficients. The effects of rotational speed and headspace were most significant. The particle density had no significant effect, but liquid viscosity and rotational speed had significant effects on U. The h$ sb{ rm fp}$ values were influenced by rotational speed, liquid viscosity and particle density, with particle density most significant. / With multiple particles (Nylon) in cans, the associated convective heat transfer coefficients (U and h$ sb{ rm fp}$) were evaluated. Heat transfer coefficients increased with decreasing particle diameter, increasing particle concentration and increasing sphericity. / Flow visualization studies were carried out. With a single particle, liquid viscosity, particle density and rotational speed influenced particle motion. With multiple particles, motion/mixing was influenced by particle concentration, size, and shape, liquid viscosity and rotational speed. / Dimensionless correlations were developed for the predictive modeling of convective heat transfer coefficients. For U with a single particle, Nusselt number (Nu) was correlated to Reynolds number (Re), Prandtl number (Pr) and relative can headspace while with multiple particles, Re, Pr, ratio of particle to liquid concentration, relative particle diameter and particle sphericity were found to be significant. For h$ sb{ rm fp}$ with a single particle in the can, three different correlations, one each for a sphere, a cylinder and a cube were developed and the Nu was correlated to Re, Pr, density simplex, relative can headspace and the ratio of the sum of the diameter of rotation and diameter of the can to the can diameter. With multiple particles Nu was correlated to Re or Peclet number, ratio of particle to liquid thermal conductivity, particle to liquid concentration and particle sphericity. / A multi-layer artificial neural network (ANN) was used to model heat transfer parameters under conduction and convection heating conditions. The network was found to predict optimal sterilization temperatures with an accuracy of $ pm$0.5$ sp circ$C, and other responses such as process time and integrated heating time for quality attribute with less than 5% associated errors. Multi-layer neural network models were trained based on the experimental values of U and h$ sb{ rm fp}$, was obtained. The trained network was found to predict U and h$ sb{ rm fp}$ with less than 3% and 5% errors, respectively. The neural network models were more accurate than the dimensionless number models, for predicting U and h$ sb{ rm fp}$. (Abstract shortened by UMI.)
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.40437 |
| Date | January 1996 |
| Creators | Sablani, Shyam Swaroop. |
| Contributors | Ramaswamy, H. S. (advisor) |
| Publisher | McGill University |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Format | application/pdf |
| Coverage | Doctor of Philosophy (Department of Food Science and Agricultural Chemistry.) |
| Rights | All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
| Relation | alephsysno: 001538844, proquestno: NN19770, Theses scanned by UMI/ProQuest. |
Page generated in 0.0023 seconds