|Abstract or Summary
- Freezing is one of the important technologies for preservation of foods.
In this project, using surimi as a food model, thermophysical properties of frozen
foods were evaluated and the freezing process was simulated using a finite
To measure temperature-dependent thermal conductivity, a line-source
probe system was used. Effects of test conditions and sample history were
investigated. Thermal conductivity of Alaska pollock (Theragra chalcogramma)
surimi having 0, 4, 6, 8, and 12% cryoprotectant levels was measured in the
range of -40 to 30 ° C. Other thermal properties were analyzed using differential
scanning calorimetry (DSC) at the same cryoprotectant concentrations and in the
same temperature range. Each dynamically corrected DSC thermogram was
used to determine initial freezing point, unfreezable water (bound water),
apparent specific heat, enthalpy and unfrozen water weight fraction.
When water content of the sample is controlled, thermophysical
properties of surimi have a relatively weak dependence upon cryoprotectant level
in the unfrozen and fully frozen (-40° C) ranges. However, the initial freezing
point and the properties just below this point were significantly affected.
From measured data, the Schwartzberg thermal property models for
frozen foods were investigated. The models agreed well with experimental data.
However, possibility for further improvement is demonstrated by using DSC
analysis. This research additionally demonstrated the great potential of DSC for
measuring and modeling frozen food thermal properties.
Using the derived property models, a commercial PC-based finite element
package was used to simulate the process of freezing a food block in a plate
freezer. The capability of the program to handle temperature-dependent
thermal properties and time-dependent boundary conditions enabled a simulation
which accounted for measured changes in thermal properties, ambient
temperatures and overall heat transfer coefficient. Predicted temperature history
agreed well with measured data. Sensitivities of important model parameters,
which were varied within their experimental error range, were also investigated
using a factorial experimental design method. The result showed that in
decreasing order of influencing freezing time prediction, attention should be
given to apparent specific heat, block thickness, overall heat transfer coefficient,
ambient temperature, thermal conductivity, and density.