Measurement and Prediction of Freezing Rates

Abstract

This research has developed an experimental method to measure frozen phase thickness of liquids, based on the density differences between the frozen and unfrozen phases and compared experimental freezing rates with rates predicted by theoretical methods, including those of Planck, Neumann and Goodman. Freezing rate experiments produced ice thicknesses from 6 mm to 30 mm of distilled water, grapefruit juice and 5 and 10 per cent sodium chloride solutions with initial temperatures between 3.5 and 25 C, coolant temperatures between -4 and -16 C and with heat transfer coefficient values of 2000, 1700, 900 and 56.8 W/m2C, Accurate values of the heat transfer coefficient were determined by initial experimental work. A literature survey, although revealing many theoretical methods for predicting freezing times of liquids, alloys and watery foodstuffs, found that only in a few cases had comparison between theoretical and experimental results been made end in these cases agreement was generally poor. The theoretical methods used restrictive assumptions to simplify the moving boundary and heat conduction equations and showed wide variations (e.g. 100%) in predicted freezing times which precluded the choice of any as obviously accurate. A finite difference method of Vasil'ev and Uspenskii was modified to predict to within 8% the freezing times of distilled water. The freezing rates of electrolyte solutions, under conditions of low freezing rates (0.0012 mm/s), were found to be less than those of distilled water due to the formation of a layer of high solute concentration adjacent to the interface which lowered the solution freezing point. Incorporation of this effect into the Vasil'ev and Uspenskii method predicted electrolyte solution freezing rates to within 15%.