t = ( W⁄ARc )(Xc − Xe ) ln ( X1 − Xe ⁄X2 − Xe ) (2.

5)

where, Xe is the equilibrium moisture content.

Characteristics of drying materials are very important to understand the drying kinetics. In most

cases, the constant and falling rate periods both exist. If the drying materials contain high

moisture content, all the drying may occur in the constant rate period but for slow drying

materials most of the drying may occur in the falling rate period. In case of solid with hard

permeable skin, it is necessary to retard evaporation and bring it in step with the rate of moisture

movement from the interior part of solid that could be achieved by increasing the relative

humidity of the drying air. Air temperature, humidity, material thickness and bed depth are

important for the drying materials with low moisture content and required extremely low final

moisture content. Drying time of these materials are long and air velocities are important only

to the extent at which they enhance heat transfer rates. When long drying periods are required,

the drying rate of falling rate period varies with the square of the material thickness that

indicates the need of using agitation or using thin layers during drying of materials.

Drying of food materials improve food stability, since it decreases the water activity of the food

material and the microbiological activity and, minimises physical and chemical changes during

storage. At the same time, it has some negative consequences in the food market because of

some physical changes that food material faces during drying. Particle shrinkage, bulk density

and porosity are important physical properties characterising the texture, quality.

Particle shrinkage, mechanism of shrinkage, factors affecting the magnitude of shrinkage are

discussed in the following section.

2.4 Shrinkage of food particle

The most important physical changes that a food particle suffers during drying is its external

volume change or particle shrinkage. Shrinkage determines the structural properties and the

47
quality of the product. The water loss and cellular stress for heating lead the structure change

or the decrease in dimension. Changes in shape and loss of volume usually increase the

hardness of the dehydrated product which causes negative impression at consumer level except

some dried fruits that have had traditionally a shrunken aspect like raisins, dates and so on

(Mayor & Sereno, 2004).

2.4.1 Mechanism of shrinkage

Mechanism of particle shrinkage during drying is described here according to Mayor and

Sereno (2004). The food systems are highly heterogeneous materials which might be

considered as a three-dimensional solid network or matrix holding large quantities of a liquid

phase, in most cases moisture. Biopolymers are the common structural elements of the solid

food matrix but in some cases a composite structure might be formed by incorporating the

additional structural elements (Aguilera, 1992). The size and shape of a food material is

determined by the particular structure of that material and the mechanical characteristics of its

elements. At the drying period, water is removed from the food particle that causes the

imbalance of pressure between the inner and the external pressure of the material generating

contracting stresses that lead to material shrinkage or collapse and changes in shape.

2.4.2 Factors affecting the shrinkage

(a) Volume of removed water; Shrinkage of food materials mostly relies on the volume of water

removed during drying, since the water removal causes the contracting stresses in the materials.

In some cases, when shrinkage of the material equals volume of removed water the mechanical

equilibrium is observed.

(b) Mobility of the solid matrix; Mobility of a solid food particle closely related to its physical

state for example high mobility corresponds to a viscoelasticity similar to a rubbery state and

low mobility corresponds to an elasticity similar to a glassy state. The mobility of the solid

48
matrix is described as a dynamic process with the rates that depend on the difference of the

temperature of the sample undergoing dehydration and the glass transition temperature by Levi

and Karel (1995). At a high moisture content, the food material usually stays at rubbery state

and particle shrinkage almost entirely compensates for moisture loss, and volume of the

material decreases linearly with moisture content. On the other hand, at low moisture contents,

the material to pass from rubbery to glassy state as the glass transition temperature increases

and the rate and extension of particle shrinkage decreases significantly. This deviations of

particle shrinkage from linearity observed by several authors (for example (Achanta et al.,

1997; Wang & Brennan, 1995)). For a low moisture content material where phase transition

from rubbery to glassy state occurs, rigidity of the material stops shrinkage and parallel pore

formation may observe.

(c) Drying rate: Particle shrinkage also depends on the drying rate. If drying rate is low, the

rate diffusion of water from inner to outer part of the food material is same as the evaporation

rate from the surface that indicates no sharp moisture gradients in the material resulting the

uniform shrinkage until the last stage of drying. This type of behaviour is mentioned by a few

authors (for example; (Litchfield & Okos, 1991) during drying of pasta and (Wang & Brennan,

1995) during drying of potato). On the other hand, if the drying rate is high, intense moisture

gradients through the material are observed. In that case, a rubber–glass transition may induced

by the low moisture content of the external surface leading to the formation of a porous outer

rigid crust or shell that complicates the relative shrinkage of the inner part of the food which is

still rubbery (Mayor & Sereno, 2004). The formation of a shell was observed experimentally

by (Wang & Brennan, 1995) during the drying of potatoes.

There are some process conditions as well that may effect on drying and the volume change of

the materials during drying like drying temperature, velocity of the air, relative humidity of the

air and most importantly the different drying systems. The choice of appropriate drying

49
systems/ design of dryers for foodstuffs is a complex problem due to the peculiar characteristics

of food materials. The dimensions and transport properties of individual food particles can be

modified by both the shape and size changes during drying and also the thickness and porosity

of the packed bed in the dryer (Ratti, 1994). The different drying systems with selection criteria

for the suitable dryer is discussed in the following section 2.5.

2.5 Drying systems

The most common and traditional drying method for food is the open-air-sun-drying. The

industrial drying system has been classified in different points of view. The simplest

classification divided all type of drying into two different categories: direct and indirect drying.

This classification is according to the method of heat delivery to the material being dried. A

very elaborate discussion regarding this manner are reviewed in several articles (Jamaleddine

& Ray, 2010; Mujumdar, 2014b). The food particle need to dry is a major determinant of the

dryer type used. A general overview of the different type of drying systems used in food drying,

dryer characteristics and the type of food particle usually used are summarised in Table 2.1.

The cost effective way of improving the existing drying systems is the simulation of the system

by computational approaches. A prior design of drying process can be predicted by numerical

modelling in a virtual way without the cost and effort of constructing a full scale setup and the

time to perform experiments. Drying kinetics with different drying stages can also easily be

described by some numerical modelling approaches. Several numerical approaches have been

set forth to model the drying systems with heat, mass and momentum transfer during drying of

a food material. The analytical approaches to establish the drying model and some advanced

numerical modelling methods of food drying are discussed in the following part.

50
 Table 2.1 Different drying systems used for different food products (adapted from (Jamaleddine & Ray, 2010))

Drying systems Drying feature Food products

Flash (Pneumatic/ (1) Inlet temperature range: 200–1400°F(100–650°C) agricultural products, fish products,

Convective) (2) High thermal efficiency dairy products, starch, flour, proteins

(3) Ability to handle continuous operation

Rotary (1) Inlet temperature range: 750–950°F(399–510°C) powders, cakes, granules, flakes,

(2) Ability to handle continuous operation pastes, gels, slurries

Belt (1) Ability to handle wet material with surface and bound moisture granular, fibrous, preformed, extruded

(2) Characterized by uniform gentle drying wet food products

(3) Low power demand

(4) High thermal efficiency

Fluidized Bed (1) High thermal efficiency carbohydrates, beverage products

Spouted Bed (1) Controlled residence time coarse products (grain), pastes, slurries,

(2) Good mixing with minimum attrition heat-sensitive foods

Spray (1) Can handle heat-sensitive, non-sensitive, and heat-resistance fluids dairy products, whey, coffee, tea, yeast

(2) Produce dry material of controllable size, shape and moisture content

(3) Can handle the drying process without the risk of explosion or fire

(4) High installation cost and low thermal efficiency

Impinging Jet (1) Complex fabrication and high air-handling costs coarse granules (cat or dog foods)

(2) Not recommended for heat-sensitive material

(3) Recommended for removal of unbound moisture

Pulse Combustion (1) Ability to increase productivity food products, granular material

(2) Ability to reduce pollutant emissions

(3) Significant fuel savings

Indirect drying (paddle, (1) The absence of oxygen that eliminate chance of combustion an entire range of food except liquids

tray, multiple-effect (2) No emissions

evaporation and disc (3) High heat transfer coefficients resulting in high evaporation capacity

dryers) (4) Controlling the drying rate

52
2.6 Analytical approaches for modelling of drying process and related shrinkage

2.6.1 Modelling of drying process

There are three main general approaches to formulate drying models (Chen & Putranto, 2013).

(a) Characteristic drying rate curve model or CDRC models. Different drying stages can be

described by these models such as, the constant drying rate period and falling rate period.

(b) The distributed-parameter models which include the models based on volume averaging

concept. These models employ coupled heat and mass diffusion equations with heat

conductivities and mass diffusivities. A number of mathematical models established in this

category including effective liquid diffusion, capillary flow, evaporation–condensation, dual

(temperature, water content gradient) and triple (temperature, water content and pressure

gradient) driving-force mechanisms (Chen & Putranto, 2013; Luikov, 1986). The detailed

transport equations for macro- and micro-scale structures of biological materials has been

proposed where the three phase (solid, liquid and vapour) conservations and their local volume-

averaged behaviours were considered but the small scale phenomena like shells, local pore,

pore channels was avoided (Whitaker, 1999). Pore-network models have become popular

nowadays with expand of multi-scale and multi-physics concepts of particles (Perré, 2011).

(c) The empirical model based on simple and multivariate regression methods. There are

many models in categories (a) and (c) and they may considered as “lumped drying models” as

they do not need to solve spatially distributed moisture content and temperature.

Reaction engineering approach (REA). The newer approach after the three general

approaches is the REA for drying model. This model was proposed first time in 1997 (Chen &

Chen, 1997) and this idea was partially inspired by a published paper (Gray, 1990) and a long-

established CDRC models (Chen & Putranto, 2013; Keey, 1978, 1992b; Van Meel, 1958). REA

was initially proposed as a lumped parameter model to describe local evaporation and

53
condensation, and then extended as a spatial REA model to model non-equilibrium multiphase

drying. A comparison of different approaches (general approaches and REA) in terms of their

concepts, advantages and limitations are summarised in Table 2.2. To model a range of food

materials such as pulped kiwifruit leather, whey protein, lactose, skim milk powder, and

mixtures of sugars drying, REA has been used with very accurate moisture content and

temperature profile along with drying time (Lin & Chen, 2005, 2006, 2007). REA was

implemented to develop CFD using a spray dryer for coupling the dispersed phase and the

continuous phase (Jin & Chen, 2010; Woo et al., 2008). Outlet air temperature and outlet

particle moisture content of a drying process can be predicted by CFD using REA. Moreover,

the evaporation zone, drying rate, trajectory of particles and deposition of particles in a spray

dryer can also be predicted by REA approach (Woo et al., 2008).

There are numbers of mathematical modelling found for drying process of food materials. The

previously developed drying models mainly concentrate on the physics of the drying process

and rarely used to discuss other aspects for example, energy efficiency, consumption and

recuperation, product loss and, biochemical and organoleptic quality as nutritional content of

food, colour (Defraeye, 2014). To consider these problems, some advanced modelling methods

for drying of food materials have been emerged.

54
 Table 2.2 Concepts of different analytical approaches of drying, their advantages and

limitations

Analytical Concepts Advantages Limitations

approaches

characteristic based on -recognises different drying -considered as ‘lumped

drying rate characteristic stages e.g. constant rate drying models’ as they no

curve model drying rate period and falling rate period need to solve spatially

curve - used to describe distributed moisture

concepts evaporation and content and temperature

condensation -obtains fair quality of

results

- drying conditions

determine the critical water

content

distributed- mostly based -involved in continuum type -overlapping of time scales

parameter on volume modelling (e.g. effective for not attaining local

models averaging liquid diffusion, capillary thermodynamic

concepts flow, evaporation– equilibrium

condensation, dual and triple -multi-scale configuration

driving-force mechanisms) -challenging for the great

- involved in pore network demand in computational

modelling power and mathematics

empirical obtained by -used to simply correlate the -considered as ‘lumped

models simple or weight loss over time (e.g. drying models’

multivariable Page model) - have little physics

regression - some accurate explained and difficult to

methods time functions can be used to extract fundamental

fit the data information

points to generate accurate

drying rate data sets

research based on the - simple, ordinary - need accurate

engineering expression of differential equation with experimental data to

approach the chemical time determine model

reaction rate -incorporated conventional parameters, accurate

approaches to express mass equilibrium

transfer at the boundary isotherm and surface area

- lumped and spatial REA measurement

are established.

- describes local evaporation

and condensation as well as

multiphase transport

phenomena

56
2.6.2 Modelling of particle shrinkage during drying

Food drying is a complex process of simultaneous heat and mass transfer in a system that passes

through different changes in it such as chemical composition, structure and physical properties.

For simplicity, sometimes shrinkage is considered negligible in the modelling of drying but in

reality it is barely negligible in food systems (Mayor & Sereno, 2004). Several authors

compared the results of two mathematical models for drying of foods, with and without

shrinkage and suggest that models with shrinkage lead to better predictions of values of

moisture content profiles and effective diffusivity (Deff) during the drying process (Balaban,

1990; Park, 1998; Rossello et al., 1997). Two different types of approaches are used to model

particle shrinkage during drying; (a) the empirical fitting of experimental shrinkage data as a

function of moisture content and (b) Fundamental models based on the physical interpretation

of the particles which tries to predict geometrical changes based on conservation laws of mass

and volume. In both cases, linear and non-linear models exist to describe shrinkage behaviour

during drying. Table 2.3 adopted from Mayor and Sereno (2004) provides some example of

shrinkage models.

57