| dc.description.abstract | The broad objective of this research was to predict bulk potato temperatures at vanous
locations and at different times of storage in boxes under free natural ventilation using
both theoretical and physical experimental approaches. Mathematical modelling and
simulation were done using available empirical and static models namely Burton (1989)
and Nyaanga (1991) after the necessary modifications and a semi-dynamic and theoretical
model [named Computational Thermal Prediction (CTP)] that was developed during the
research. Experiments using potato storage boxes, thermocouple sensors and data loggers
were conducted.
The Computational Thermal Prediction (CTP) semi-dynamic model was developed as a
compromise between the specific, static and less precise empirical prediction expressions
and the generalisable, very precise and highly dynamic models such as Computational
Fluid Dynamics (CFD/CFX by AEA, 1996) which demand a lot of Central Processing
Unit (CPU) memory and time; after the two were found inadequate for free natural
ventilation which is neither fully static nor that dynamic as is the case of forced and other
forms of controlled potato storage. CTP mathematical relationships involving the various
parameters of the storage environment such as temperature; rate and amount of heat of
respiration generated by the potatoes into the system and the amount lost by convection;
and the physical characteristics of the bulk such as dimensions, porosity and density were
developed based on the theories of free natural convection and results from experiments
with appropriate assumptions. Finite element method of numerical analysis was used to
determine the temperature at various locations (mainly along the vertical axis) within the
bulk and at various time intervals.
Experiments were conducted on physical models using potato storage boxes and logging
the data after being sensed by thermocouples. The data was used to test and verify the
mathematical models and determine the temperature variations which the models could
not reveal such as lateral temperature gradients.
The results of the research indicate that it is possible to predict the bulk potato
temperature using empirical mathematical expressions using the modified Burton (1989)
and Nyaanga (1991) models; develop a more generalisable Computational Thermal
Prediction (CTP) model for the simulation of bulk potato temperatures; and
experimentally determine various bulk potato temperatures in free natural ventilation
potato storage. The empirical model predictions were not as accurate as the experiments
but they gave accurate trends (history) of the bulk potato temperatures with the time of
storage. The modified Burton (1989) model shows the temperature gradients in a bulk of
potatoes although with slightly higher magnitudes and variations compared to the
observed. The Nyaanga (1991) model simulations gave nearly the same values for all the
major temperatures within a bulk of potatoes that is, the temperature at the bottom, centre
and top were the same.
The Computational Thermal Prediction (CTP) simulations were more accurate than the
empirical models and the computer program gives a wide range of options. The
individual temperature trends, by various CTP options and at the different points of
monitoring were the same at a probability of 1% using general linear model. Generally,
the CTP predicted values are slightly higher than the observed. The small deviations
could have resulted from assumptions and constants used in the mathematical model such
as the major modes of heat and air exchange and their quantification. The CTP computer
program using user information about the free ventilation potato storage, gives the user
information and recommendations about the specified storage system with respect to bulk
size, store temperature and time of storage. However, all the simulations showed that the
bulk potato temperature varied with store specifications and ambient temperatures as
expected. | en |
