See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/267297793

Food process modeling

Article · January 2001

CITATIONS READS

4 3,478

2 authors, including:

Ashim Datta
Cornell University
138 PUBLICATIONS   6,325 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Food process modeling View project

All content following this page was uploaded by Ashim Datta on 10 April 2017.

The user has requested enhancement of the downloaded file.

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

FOOD PROCESS MODELING

Datta, Ashim K.
Department of Agricultural and Biological Engineering, Cornell University, USA

Rattray, Jeff
Department of Food Science, Purdue University, USA

Keywords: Analytical, numerical, empirical, heat and mass transfer, neural network,
design.

Contents

1. Modeling and its Various Uses

2. Types of Process Modeling

S
TE S
2.1. Analytical Models

R
AP LS
2.2. Numerical or Computational Models
2.3. Observational (Empirical) Models
3. Other Models Used in Food Plant Design and Operation
C EO
4. Computational or Numerical Models
4.1. Computer-based Engineering as a Design Tool
4.2. Typical Characteristics of Physical Processes in Food Processing
4.3. Typical steps in Numerical Modeling
4.3.1. Governing Equations and Boundary Conditions
E –
H

4.3.2. Mesh Generation

4.3.3. Material Property Data
PL O

4.3.4. Solution Technique

M SC

4.3.5. Post-Processing
4.4. Application Examples
4.4.1. Computational Electromagnetics: Microwave Heating
4.4.2. Computational Mechanics: Thermal Cracking during Freezing
SA NE

4.4.3. Computational Heat and Moisture Transfer: Combined Microwave and Infrared
Heating
4.4.4. Other Processes
U

4.5. Future Needs and Developments in Computational Modeling for Food Processing
4.5.1. Linkages between Commercial Software
4.5.2. Customized Software for Food Processing
4.5.3. Inclusion of Appropriate Physics
4.6. Resource Needs for Numerical Modeling for Design Purposes
4.7. A Cautionary Statement
5. Observational (Empirical) Models
5.1. Introduction
5.1.1. Biological Neurons
5.1.2. Artificial Neurons
5.2. Learning
5.3. Training
5.3.1. Initialization
5.3.2. Performance Tracking

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

5.4. Other Issues

5.5. Examples
Glossary
Bibliography
Biographical Sketches

Summary

Modeling, particularly numerical and observational models of food processes, has come
a long way and has become an integral part of research and design. Although analytical
models were the main mode of modeling before the days of faster computers, computer-
based numerical and empirical models are the primary types of models in development
today. Food processes are often complex, with variation in properties due to natural
biological variability and transformation of the materials during processing. The
numerical and empirical models are highly versatile and can cover a wide range of input

S
TE S
and output parameters and process types, thus providing flexibility in developing a more

R
AP LS
realistic description of the process. It is important to bear in mind the limitations of each
modeling method. In so doing, the judicious use of these models is expected to provide
improved efficiency of the food sector by providing a better understanding of the
C EO
processes and shortening the cycle for product and process design.

1. Modeling and Its Various Uses

A model is a mathematical analog of a physical process. An accurate model is expected

E –

to work just like a physical prototype, but its engine is mathematical rather than
H

physical. Models have two major uses. The most common use of a model is to provide a
PL O

better understanding of the physical process, by seeing the relationships between its
input and output parameters. This is referred to as the research objective of modeling.
M SC

For example, in modeling a sterilization process, one would need to know the effect of
steam temperature (input parameter) on the extent of lethality or bacterial death (output
parameter). As computing techniques and computational hardware and software
SA NE

underwent a revolution, the second major objective emerged: to use the model in design
or in checking the “what if” scenarios. As applied to the sterilization problem, one can
now easily check if the sterilization process improves in quality if the process
U

temperature (input variable) is varied in some way. Better yet, in the future,
optimization software is likely to become available that would provide a process
temperature profile for optimum quality. Product processes and equipment designs are
rapidly becoming the major driving force for the use of modeling. In this article, the
discussions refer to both research and the design use of models.

2. Types of Process Modeling

For simplicity, modeling can be divided into three groups:

1. Analytical Models
2. Numerical Models
3. Observational (Empirical) Models

2.1. Analytical Models

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

Analytical models usually refer to models that have analytical or closed-form solutions,
which can often be developed without the aid of a computer. The heyday of analytical
solutions was generally before the advent of high-speed computers when large amounts
of memory became available. Analytical models continue to be of great importance
because of their potential power to provide a complete understanding of the system
when used. The cost of this power is in its limited applicability, as much of food
processing is simply too complicated to describe using such models. However,
approximate analytical models that do require computers are still used in active areas of
research. For example, pieces of analytical solutions can make numerical models
(described later) more effective. In food processing applications, until very recently,
modeling has mostly been based on simple analytical solutions. The best analytical
solution was that of Ball in 1923, applied to heat sterilization. It involved combining the
solution to partial differential equations describing heating and cooling with the first-
order kinetics of bacterial death, and obtaining approximate closed-form solutions.
Today, most sterilization of canned foods, pouches, etc., is based on this analytical

S
TE S
solution. Another widely used analytical solution in food processing is the solution of

R
AP LS
the mass diffusion equation, applied to moisture transport in drying and similar
processes. In this article, models of this type will not be covered in any further detail,
and the reader is referred to textbooks, the references for which are given in the
C EO
Bibliography.

2.2. Numerical or Computational Models

The analytical models described in the previous section have major limitations, such as
E –

the requirements of a simple geometry, constant processing conditions, constant

H

properties, and so forth. Numerical, computer-based solutions are highly versatile and
PL O

can avoid the limitations just mentioned. A numerical model typically involves solving
the set of partial differential equations that describe the physics of the model more
M SC

exactly than an analytical one. Access to fast computers with vast amounts of memory
has made the reliance on these types of models almost the default in recent years,
especially in food process modeling. Numerical models will be described in detail later
SA NE

in this article.

2.3. Observational (Empirical) Models

U

Analytical and numerical models assume that the model (process description) is known,
and tries to find the detailed behavior of the system. Frequently, finding the model in
the first place is the most difficult, so difficult that one begins inferring a model from
measured data. These models coming from introspection or observation (or both) can be
called observational or empirical models. Such models are used to characterize and
classify the data, to generalize from the measurements in order to make predictions
about new observations, or to learn something about the rules underlying the observed
behavior. An example of this type of model is the neural network model that will be
described in detail later. These models do not attempt to explain the workings of the
process represented on a chemical or physical level. They simply predict on the basis of
known data, without determining more about the underlying process. As such, they can
model complex processes easily, but also have certain inherent limitations.

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

3. Other Models Used in Food Plant Design and Operation

In addition to examples of models provided in this article, there are many other models
used in the design of large-scale industrial food processes. For an overview of
computer-aided engineering in the food industry, see the references given in the
Bibliography. There are process modeling software packages that can combine various
operations of a food process as modules and help identify the optimum combination of
operations for improved quality and reduced cost of the overall process. These models
provide the next higher scale of integration, as compared to the numerical or
observational model described later in this article, which are typically applied to a single
operation. The multi-operation modeling software packages mostly have their roots in
chemical processing. On the production floor, modeling is also used for inventory
control, statistical process control, machine vision, and expert systems for maintenance
and scheduling. These other types of models are beyond the scope of this article.

S
TE S
4. Computational or Numerical Models

R
AP LS
Computational or numerical models based on the detailed physics of the process have
gone through a revolution in recent years, primarily due to the availability of faster and
C EO
cheaper computers with vast amounts of memory. New computational branches of
existing disciplines have appeared, for example, computational solid and fluid
mechanics, computational heat transfer, and computational electromagnetics. In this
section, numerical models in a number of areas with application to food processing are
described. This section is intended for the food product, process, and equipment design,
E –

an application area that is rapidly increasing.

H
PL O

4.1. Computer-based Engineering as a Design Tool

M SC

Computer models of food processes have been developed mostly from research activity
in the past, starting with perhaps the earliest study by Teixeira et al. in 1969. Design
typically requires frequent computations with more frequent changes in parameters.
SA NE

Such intensive computations have only recently been practical due to the availability of
high-powered desktop workstations, PCs, and advanced, user friendly, software (see
Software for Food Engineering Applications). These advances have made computer
U

models a practical tool for product and process design. Consequently, modeling for
design purposes has been shown or suggested for food applications in research.

Use of computer modeling or computer-prototyping has several advantages for food

products and the process development environment:

1. Testing “what-if” scenarios is quick and inexpensive, thus shortening the design
cycle (quicker turnaround), which should result in reduced costs and increased
profits.

2. Modeling can provide insights into complex processes that are otherwise difficult to
understand. It provides a clearer understanding of the interactions between the
physical processes and their sensitivity to various operational parameters. This can
enable the designer to be more creative.

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

3. Modeling allows front-end engineering before physical prototyping, making the

prototypes closer to the optimum and reducing the number of prototypes.

4. Modeling makes possible concurrent design and analysis, also shortening the design
cycle. While an experiment is underway, the results can be used simultaneously to
further optimize the process on the computer, also reducing the amount of
experimentation.

Computer-aided engineering (CAE), simulation-based engineering, and computer

prototyping all refer to the use of computers to build and test computer models of
products and processes to reduce the extent of physical prototypes. Computer
prototyping can help today's competitive product and process design by a) reducing
cost, b) reducing the time needed to market, and c) making more dramatic changes
possible.

S
TE S
Industries, such as automotive and chemical processing, have been exploiting the

R
AP LS
advantages of computer-aided engineering in a very significant way. The food industry
also stands to benefit from the use of this tool. Although the total application is still
somewhat small, use of computational software is also rapidly increasing in food
C EO
process engineering. The food industry, however, is generally behind other processing
industries in such CAE applications.

4.2. Typical Characteristics of Physical Processes in Food Processing

E –

Following are some of the unique aspects of food processing problems (see Engineering
H

Properties of Foods):
PL O

• In addition to temperature changes during a heating or cooling process, there are

M SC

biochemical (nutrient, color, flavor, etc.) or microbial changes that are important to
know.
SA NE

• The moisture in food is constantly undergoing either loss (due to evaporation,

especially when heated) or gain (from humid surroundings).

•
U

The properties of foods, such as density, thermal and electrical conductivity, specific
heat, viscosity, permeability, and effective moisture diffusivity are often a function
of composition, temperature and moisture content, and therefore keep changing
during the process. The system is also quite non-homogenous. Such detailed input
data are not available.

• The hygroscopicity of food materials can cause the food to shrink upon significant
loss of moisture or swell when gaining moisture.

• Often irregular shapes are present.

• Processes such as temperature, moisture and mechanical changes are often coupled.

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

To make a realistic physics-based numerical model, these characteristics should be

included as much as possible.

4.3. Typical steps in Numerical Modeling

The numerical computation or modeling process typically consists of three steps: Pre-
processing, Processing and Post-processing. Pre-processing typically defines the
geometry, computational grids, governing equations, boundary conditions, properties,
and methods used for solving. In the processing step, the computer solves the problem.
In the post-processing step, the solution is visualized by the user, using shaded contours,
movies, etc.

4.3.1. Governing Equations and Boundary Conditions

In this the most important pre-processing step, mathematical analogues (equations) of

S
TE S
the physical process, are developed in terms of a set of equations called the governing

R
AP LS
equations and the boundary conditions. The mathematical description is often an
intelligent simplification, especially in the initial stages. This step is often the most
important and most difficult. The goal is to keep as many details of the process as
C EO
possible, without creating unnecessary complexities. Modeling software is typically
able to solve a very general set of governing equations and boundary conditions,
covering a wide range of processes. In choosing a software program, availability of
necessary governing equations is only one of many factors. Each package has its
strengths and weaknesses, i.e., it can be more efficient in solving a specific type of
E –

problem.
H
PL O
M SC

4.3.2. Mesh Generation

In this pre-processing step, after the geometry has been defined, the geometry needs to
SA NE

be broken down into smaller pieces for a numerical solution. The more the geometry is
broken down, the more accurate the final solution, but the computation time can
increase (sometimes dramatically) and eventually make it unrealistic to compute. Thus,
U

this step is a careful balance between providing enough elements (or nodes) such that all
the essential physics are captured (but not too many).

4.3.3. Material Property Data

A very important step in simulation-based engineering is to use accurate material

property data. However, this is where food processes are also at a slight disadvantage,
compared to processes not using natural materials. For example, properties of steel are
more easily available and perhaps have a lot less variability than chicken soup. Data on
food properties can be obtained from measurements, computerized databases,
handbooks and prediction formulas. In simulation-based design, it is possible to relax
the restrictions on requiring very accurate data. By varying the property data around
their expected value in the computer model, one can bracket the properties and predict
the effect of a range of properties on the process. Sometimes this is more useful than

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

having the model predict output for a single property value since, in reality, food
properties often vary from batch to batch due to differences in formulations, etc. This
concept of bracketing can be extended to other input parameters, such as geometry, and
is generally referred to as the sensitivity analysis of the model. Such sensitivity analysis
can provide important insight into the model, as in which parameters are the most
critical. Being able to perform easy sensitivity analysis is one of the greatest advantages
of modeling.

4.3.4. Solution Technique

Once the equations and properties are defined, the next step is to select a solution
method (choice of time discretization, matrix solution method, etc.) that provides the
most efficient solution to the set of equations describing the food process. Today, there
is a large quantity of software available in any given area, such as mechanics or heat
transfer. For example, over 75 commercial computational fluid dynamics (CFD)

S
TE S
software packages are currently available. The advantages of using available software

R
AP LS
are that often one can choose a solution method without having to be very
knowledgeable about the details of such methods. Once the solution methods are
specified, the computer takes over the solution procedure. This is the processing step.
C EO
The equations are discretized and large linear systems of algebraic equations (matrices)
are formed and solved using well-known procedures.

4.3.5. Post-Processing
E –

Post-processing is the important step of visualizing the results and making further
H

computations from raw data generated by the solution. Most commercial CFD packages
PL O

are able to show nice contour, history, vector, and other plots. These spatial or temporal
profiles in whole 3-D region can provide insight and understanding of the process,
M SC

which is not possible using experimentation. For example, the spatial distribution of
moisture or oil during a drying or a deep-frying process (see Frying) can be obtained
computationally, however it requires elaborate and expensive experimentation using the
SA NE

MRI. Such visualization is one of the greatest advantages of modeling.

4.3.6. Trusting Computational Results

U

Computational results should never be trusted blindly and the software should never be
used as a blackbox. However, it is also important to accept that there is no foolproof
way to confirm the computational results. Several steps can be taken to minimize the
chances of obtaining a wrong solution. Some of these include checking for mesh
convergence, checking input file for accuracy of problem definition, using common
sense about the process physics, comparing the experimental data, and checking against
the results of a simpler problem.

4.4. Application Examples

Application of numerical modeling in food processing can include computational fluid

dynamics (CFD) and heat transfer to solve flow/heating/cooling problems,
computational mechanics to solve rheological and stress related problems,

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

computational electromagnetics to solve microwave and other heating problems, and so

on. Examples of some food applications are provided below.

-
-
-

TO ACCESS ALL THE 22 PAGES OF THIS CHAPTER,

Visit: http://www.eolss.net/Eolss-sampleAllChapter.aspx

Bibliography

S
TE S
Bhagat P. (1990). An introduction to neural nets. Chemical Engineering Progress 86(8), 55-60.
[Information on neural network training].

R
AP LS
Choi Y.S., Whittaker A.D., and Bullock D.C. (1996). Predictive neuro-fuzzy controller for multivariate
process control. Transactions of the ASAE 39(4), 1535-1541. [Example of real-life neural network
applications].
C EO
Cowan J.D. (1967). A mathematical theory of central nervous activity. Ph.D. thesis, University of
London, London, UK. [Early work on making mathematical models analogous to nervous system
elements].
E –

D’Amico S., Vega P., and Alonso L. (1995). Neural network based modeling and control of a non-linear
H

process. Proceedings of the Eleventh IASTED International Conference on Applied Informatics, 56-59.
PL O

[Example of real-life neural network applications].

Datta A.K. (1998). Computer-aided engineering in food process and product design. Food Technology
M SC

52(10), 44-52. [This article discusses the generic steps needed to develop a food process model on the
computer. It shows a list of processes that can be modeled with current commercial software, together
with the limitations of current software and data (and thus shows the research needed)].
SA NE

Demuth H. and Beale M. (1994). Neural network toolbox user’s guide. Natick, Mass: The Math Works,
Inc. [Software manual that contains useful theory and background].

Fravolini M., Ferroni A., Ficola A., and LaCava M. (1997). Neural networks for modeling and
U

identification of the dough rising process inside an industrial proofing chamber. 1997 IEEE International
Conference on Neural Networks Conference Proceedings, 1, 127-132. [Example of real-life neural
network applications].

Garcia L.A., Argüeso F., Garcia A.I., and Diaz M. (1995). Application of neural networks for controlling
and predicting quality parameters in beer fermentation. Journal of Industrial Microbiology 15(5), 401-
406. [Example of real-life neural network applications].

Gershenfeld N. (1999). The nature of mathematical modeling. Cambridge, UK: Cambridge University
Press. [An overview of the three general modeling types—analytical, numerical, and observational].

Haykin S. (1994). Neural networks: a comprehensive foundation. New York, NY: Macmillan College
Publishing Company. [Concepts of neural networks].

Hebb D.O. (1949). The organization of behavior: a neuropsychological theory. New York, NY: John
Wiley and Sons. [Early work on neural nets].

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

Heldman D.R. and Singh R.P. (1981). Food process engineering. Westport, CT: AVI Publishing Co., Inc.
[Introductory textbook that shows how simple models can be built for some of the basic food processes].

Hertz J., Krogh A., and Palmer R.G. (1991). Introduction to the theory of neural computation: lecture
notes volume 1. Redwood City, CA: Addison-Wesley Publishing Company. [Information on preparing
data for neural network training].

Hinton G.E., McClelland J.L., and Rumelhart D.E. (1986). Distributed representation. In Parallel
Distributed Processing, Vol. 1 (eds. D.E. Rumelhart and J.L. McClelland et al.), Chap. 7. Cambridge,
MA: MIT Press. [Development of terminology describing how neural nets store information].

Irudayaraj J. (2000). Food process operations modeling: Design and analysis. New York: Marcel Dekker,
Inc. [An edited compilation of modeling papers from a number of food processes].

Jain A.K. and Mao J. (1996). Artificial neural networks: a tutorial. Computer 29(3), 31-44. [Information
on neural network training].

Kaminski W., Sturmillo P., and Tomczak E. (1996). Genetic algorithms and artificial neural networks for

S
TE S
descriptions of thermal deterioration processes. Drying Technology 14(9), 2117-2133. [Example of real-
life neural network applications].

R
AP LS
Khankari K. and Pantankar S.V. (1996). Computer simulation of a conveyer dryer. Cereal Foods World
41(8), 681-685. [As the title suggests, this gives the details of a numerical model for a food dryer].
C EO
Kohonen T.K. (1988). An introduction to neural computing. Neural Networks 1, 3-16. [Information on
neural network training].

Kumar A. and Swartzel K.R. (1993). Selected food engineering problems and their solutions through
E –

FEM. In: FED-Vol. 171, Advances in Finite Element Analysis in Fluid Dynamics (Eds. M.N. Dhaubhadel,
H
M.S. Engelman, and W.G. Habashi). American Society of Mechanical Engineers. [A compilation of
models (using finite element method, a popular numerical method) in a number of food processes].
PL O

Lin C. and Lee C.S.G. (1996). Neural fuzzy systems: A neuro-fuzzy synergism to intelligent systems.
M SC

Upper Saddle River, NJ: Prentice-Hall, Inc. [An overview of the use of neural networks for control
systems].

Linko P., Uemura K., and Eerikäinen T. (1992). Neural networks in fuzzy extrusion control. Institute of
SA NE

Chemical Engineers Symposium Series 126, 401-410. [Example of real-life neural network applications].

McCulloch W.S. and Pitts W. (1943). A logical calculus of the ideas immanent in nervous activity.
Bulletin of Mathematical Biophysics 5, 115-133. [One of the first proposals for neural network
U

computing].

Melm G., Tung P, and Pringle A. (1995). Mathematical modeling of beer foam. MBAA Technical
Quarterly 32, 6-10. [Example of real-life neural network applications].

Moritomo T., Purwanto W., Suzuki J., and Hashimoto Y. (1997). Optimization of heat treatment for fruit
during storage using neural networks and genetic algorithms. Computers and Electronics in Agriculture
19, 87-101. [Example of real-life neural network applications].

Puri V.M. and R.C. Anantheswaran. (1990). Finite element method in food processing: A review. Paper
90-6523, ASAE, St. Joseph, MI. [Discusses the application of the finite element method in modeling a
number of food processes].

Quarini J. (1995). Applications of computational fluid dynamics in food and beverage production. Food
Science and Technology Today 9(4), 234-237. [Another compilation of examples of models in food
processing].

©Encyclopedia of Life Support Systems (EOLSS)

FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

Rattray J.J. (1998). Neural network predictive process modeling: Applications to food processing. Ph.D.
thesis, Purdue University. [Development of empirical methods for neural network optimization].

Ruan R., Almaer S., and Zhang J. (1995). Prediction of dough rheological properties using neural
networks. Cereal Chemistry 72(3), 308-311. [Example of real-life neural network applications].

Sablani S.S., Ramaswamy H.S., and Prasher S.O. (1995). A neural network approach for thermal
processing calculations. Journal of Food Processing and Preservation. 19, 283-301. [Example of real-life
neural network applications].

Scott G. and Richardson P. (1997). The application of computational fluid dynamics in the food industry.
Trends in Food Science and Technology 8, 119-124. [Yet another compilation of models in food
processing].

Singh R.P. and Medina A.G. (1989). Food properties and computer-aided engineering of food processing
systems. Boston, MA: Kluwer Academic Publishers. [A large volume of papers resulting from an
international workshop on all aspects of modeling of food processes].

S
TE S
Syu M.J. and Tsao G.T. (1994). Neural network modeling for predicting brewing fermentation. Journal of
the American Society of Brewing Chemists 52(1), 15-19. [Example of real-life neural network

R
AP LS
applications].

Teixeira A.A., Dixon J.R. Zahradnik J.W., and Zinsmeister G.E. (1969). Computer optimization of
C EO
nutrient retention in the thermal processing of conduction-heated foods. Food Technology 23(6), 137-142.
[Possibly the first use of the computational modeling technique in food process engineering, applied to
the sterilization of canned food].

Torok D.F. (1991). Computational thermofluid modeling in the food processing industry. Presented at the
E –

Conference on Food Engineering, Chicago, Illinois. Mar. 10-12. [An early introduction to modeling
H
possibilities in food processing using commercial software].
PL O

Trelea I.C., Courtois F., and Trystram G. (1997). Dynamic models for drying and wet-milling quality
degradation of corn using neural networks. Drying Technology 15(3&4), 1095-1102. [Example of real-
M SC

life neural network applications].

Werbos P.J. (1974). Beyond regression: new tools for prediction and analysis in the behavioral sciences.
Ph.D. thesis, Harvard University, Cambridge, MA. [Information on neural network training].
SA NE

Werbos P.J. (1989). Backpropagation and neurocontrol: a review and prospectus. International Joint
Conference on Neural Networks 1, 209-216. [Information on neural network training].
U

Zhang H. and A.K. Datta. (2000). Electromagnetics of microwave heating: Magnitude and uniformity of
energy absorption. Submitted to: Handbook of Microwave Technology for Food Applications (eds. A.K.
Datta and R.C. Anantheswaran) New York, NY: Marcel Dekker, Inc. [Discusses in detail the steps
involved in the modeling of microwave heating of foods in terms of how the electromagnetic fields are
distributed inside the food and microwave oven].

Websites

http://www.cfd-online.com/ For over 75 commercial computational fluid dynamics (CFD) software

packages that are available.

Biographical Sketches

Jeff Rattray is currently Manager of Web-based Instructional Technology for Purdue University School
of Pharmacy and Pharmaceutical Sciences. His duties include the design of instructional systems, and
research on learning paradigms and methodology. Previously, Dr. Rattray was on the staff of the

©Encyclopedia of Life Support Systems (EOLSS)

 FOOD ENGINEERING – Vol. IV - Food Process Modeling - Datta, Ashim K., Rattray, Jeff

Department of Food Science, Purdue University. His research centered on the application of artificial
intelligence to food processing and food-related issues.

Ashim K. Datta is Professor in the Department of Biological and Environmental Engineering, Cornell
University, Ithaca, New York. His research and teaching interests are generally in fluid flow, heat
transfer, and mass transfer in biological systems, particularly food processing systems. In teaching, he has
developed a fundamental course covering these areas of biological transport processes. He also teaches
how physics-based computer models in these areas can be used in both research and design of biomedical
and food processes. The goal is often to obtain better insight into biological processes and to reduce
prototype development. In his research, his groups goal is to develop computer models to understand
complex food processes such as baking, microwave heating, and rapid freezing in engineering terms so
that they can be optimized. Dr. Datta is a member of the American Society of Agricultural Engineers, the
Institute of Biological Engineering, and the American Institute of Chemical Engineers, among others. He
received a B. Tech. degree (1979) from the Indian Institute of Technology, Kharagpur, India, an M.S.
degree (1982) from the University of Illinois at Urbana-Champaign, and a Ph.D. degree (1985) in
Agricultural Engineering from the University of Florida, Gainesville.

S
TE S
R
AP LS
C EO
E –
H
PL O
M SC
SA NE
U

©Encyclopedia of Life Support Systems (EOLSS)

View publication stats