Proceedings of the 2003 Winter Simulation Conference
S. Chick, P. J. Sánchez, D. Ferrin, and D. J. Morrice, eds.
PROBLEMS OF VISUALIZATION OF TECHNOLOGICAL PROCESSES
Pavel Slavik
Marek Gayer
Department of Computer Science and Engineering,
Faculty of Electrical Engineering
Czech Technical University in Prague
Karlovo namesti 13,
121 35 Prague 2, CZECH REPUBLIC
Frantisek
Hrdlicka
Ondrej Kubelka
Department of Thermal and Nuclear Power Plants,
Faculty of Mechanical Engineering
Czech Technical University in Prague
Technicka 4,
166 07 Prague 6, CZECH REPUBLIC
ABSTRACT
The main characteristics of technological processes is
their dynamic behavior. As the data from most general ap-
plications can be visualized by means of static images the
appropriate methods for visualization of technological
processes should be developed. Some applications (will be
discussed later) attempt to visualize dynamic processes, but
the known approaches usually suffer from some disadvan-
tages. Moreover most of the commercial solutions avail-
able do not cover problems for specific applications. Our
attempt was to investigate problems of dynamic visualiza-
tion for specific applications. This allowed us to develop
and implement a specific approach where acceleration both
of simulation and visualization processes could be used
(which gave us an advantage over commercial solutions
based on AVS and other systems).
This paper deals with problems of visualization of
dynamic
phenomena. An effort to develop new visualization
schemes has been described. The main idea is to extend
approaches used in the case of visualization of phenomena
of static nature into an environment where dynamic phe-
nomena are investigated and visualized. We introduced the
“level of detail” approach in time scaling in the environ-
ment of dynamic processes where time plays a primary
role. In the case of visualization of dynamic phenomena
the users are looking for specific dynamic patterns that
should help them to understand in a better way the nature
of dynamic processes under investigation. A new approach
that should meet these requirements has been developed.
This approach has been verified by means of two systems
used for simulation and visualization of technological
processes that are of a dynamic nature.
The most important issue of this paper is definition of
an approach to control visualization of dynamic processes
– especially how to control time scale. This approach has
been tested on two applications where technological proc-
esses were visualized under our new approach. Our solu-
tions can serve as pilot solutions that will result in defini-
tion of research strategy for more general problems of a
dynamic nature.
1 INTRODUCTION
Scientific visualization has proved during its existence to
be superior in many applications. Without methods of sci-
entific visualization it would not be possible to understand
sets of data acquired either by measurements or by simula-
tions. The visual presentation of data allows the user to
process visually large amounts of data and thus makes it
possible to discover some hidden relationships. This ap-
proach is supported by many visualization techniques de-
veloped in the last years. As the use of scientific visualiza-
tion methods are more and more widely used – especially
in new applications – it is obvious that the need for devel-
opment of new methods for understanding data in visual
form should be developed.
2 VISUALIZATION OF DYNAMIC PROCESSES
One application where problems of visualization of dy-
namic phenomena has been solved is cartography - see
Morrison (1998). Besides visualization of standard geo-
graphical information visualized in the form of maps, the
following dynamic phenomena. should also be visualized:
• Flows (of people, money, information, water,
ice, lava)
• Events (forest fires, earthquakes, traffic accidents,
military battles)
In this paper we give an overview of our research in the
field of simulation and visualization of technological proc-
esses. We will show that the need for innovative methods in
the field of scientific visualization is urgent as new problems
related to visualizing and understanding data have emerged.
• Processes (global economic restructuring, global
warming, stream sedimentation).
Slavik, Gayer, Hrdlicka, and Kubelka
There has been an attempt to solve these sorts of prob-
lems in a systematic way that resulted in problem decom-
position into several sub problems that should be treated by
means of specific disciplines like cognitive sciences, hu-
man computer interaction, scientific visualization and oth-
ers. We have tried to modify the schemes used in cartogra-
phy to solve the above mentioned problems with schemes
applicable in our field (visualization of technological proc-
esses). These problems in the field are of complex nature.
On one hand it is necessary to determine how to represent
dynamic phenomena to users and on the other hand how to
develop methods for allowing the user to interact with
these representations.
A traditional approach that covers (from our point of
view to a certain extent) the need for investigation of dy-
namic processes is animation. The standard approach does
not allow the user to manipulate interactively the course of
animation presented to the user and thus the user can not
study all aspects of dynamic processes under investigation.
Besides this fact there are also some technical problems
like time consuming calculations (both of simulation and
visualization nature).
It is important to understand dynamic phenomena in
specific cases when research in a specific field is being
done. Another important aspect that is worth considering is
the field of university education. The purpose of animation
when used in an instructional setting is to serve as an aid to
student learning. In order to achieve this purpose, students
should be presented with the type of information that will
help them to develop a conceptual understanding of the
material. In our case (technological processes) the students
have an idea about the nature of the process simulated and
visualized from static descriptions in textbooks. They have
also practical experience as they visited some sites where
these processes run in reality. But this knowledge is
knowledge “of the first level” only.
To understand deeply the process (running with differ-
ent parameters in various environments etc.) it is necessary
to get a feeling of the nature of the dynamics of the process
by means of dynamic parameter and environment settings
by the user. This is the specific role of animation the new
visualization system should comply with. According to
Morrison (1998), we can adopt an approach that could be
described in the following way: Computer animation is go-
ing to become a standard tool of expression in the same
way alphabets became. Animation is a potentially more
complex form of illustration and as such the necessity to
teach people how to utilize it may be greater than the need
to teach the interpretation of illustrations.
3 SYSTEMS FOR VISUALIZATION OF
TECHNOLOGICAL PROCESSES
Research at CTU Prague has been carried on in order to
design and implement systems based on the above men-
tioned approach with the aim to develop visualization
methods that support visualization of dynamics in the
processes under investigations. As an example of results
achieved we will describe two systems that were imple-
mented in the (visualization) framework mentioned. These
systems are used in the area of power engineering. They
deal with problems of dynamic behavior of clean up filters
and combustion processes.
One of the main problems we had to solve was the
problem of creation of appropriate simulation model for
both cases where the speed of simulation was a very im-
portant criterion. Traditional models of dynamic processes
are mostly based on the use of differential equations or
their systems. The main disadvantage of this approach are
large time requirements. In case we would like to create an
animated sequence of appropriate length we have to limit
ourselves to a few examples of those sequences that can be
generated within time we have at disposal. This means that
we should understand the dynamics of processes on the ba-
sis of a few examples only. Because of the complex nature
of these processes we can have a situation where these ex-
amples do not cover the wide spectrum of cases that de-
scribe the dynamics of the process to its full extent.
The solutions used in our case were based on a finite
volume method approach and on the use of particle sys-
tems. Especially in the case of simulation and visualization
of combustion process the approach used has shown itself
as innovative because we used a combination of particle
systems in finite element method environment. Our ap-
proach in general allows us to speed up significantly the
simulation process and in such a way it makes it possible
to perform many more simulations (and visualizations)
than ever before. Some solutions will be less precise in
comparison with sophisticated commercial products (from
the point of view of analysis of single static images that are
visualizations of single simulation steps) – see in the fol-
lowing text. But the advantage of flexible investigations of
dynamics of processes significantly prevails.
4 SIMULATION AND VISUALIZATION
OF DYNAMIC BEHAVIOR OF
GAS FLUE FILTERS
New emerging technologies developed for efficient clean
coal processing require research of phenomena used in
these processes. In this paper we will deal with the prob-
lem of cleaning gas that was obtained by means of coal
gasification. Realization of this process has form of filters
with various properties. For in-depth coverage of these is-
sues, see simulation texts such as Smid, Kuo, Hsiau, Wang
and Chou (1997), Chou, Tseng, Smid, Kuo and Hsiau and
Tsai (1999) and Hrdlicka, Kubelka and Slavik (2002).
The material with filtration capability fills the body of
the filter, see Figure 6 (flow of granules) and Figure 7 (flow
of gas through granules) for an illustration. In the case of
Slavik, Gayer, Hrdlicka, and Kubelka
static filters the material is adsorbing the dust particles and
the adsorption capability quickly decreases. This is the rea-
son why the material should move in the filter in a down-
ward direction (as the result of gravity). At the bottom of the
filter the material with a high share of dust particles is re-
moved and is recycled (cleaned) outside the filter. The recy-
cled material is used repeatedly in the filter.
The design of filters of this type is a rather compli-
cated task and many parameters should be taken into ac-
count. The most important parameters are: the type of fil-
tration material, size of filtration granules, internal filter
configuration and the speed of the gas that should be
cleaned – all these parameters should be visualized in some
form in our system.
It is possible to say that the filtration process has sev-
eral components:
• Physical component where the physical behavior
of filtration granules is investigated (the speed the
granules are moving downwards, creation of stag-
nant zones etc.). This behavior is influenced by
the size of granules, the type of filtration material
and the shape of the filter including the internal
arrangement
• Adsorption component where the adsorption ca-
pability of the filtration material plays the decisive
role. Also the speed of the gas plays an important
role in this component.
• Chemical component where possible chemical re-
actions between the gas and flue gas into active
surface of the filtration material are investigated.
4.1 Simulation Model
In our case we have developed a model where only the first
two components (physical and adsorption ones) will be
taken into account. This approach covers the technologies
most widely used at the present time. From the point of dy-
namic behavior of the filter the chemical component does
not play as important a role as the other two components, so
we could afford to neglect it. Moreover this component re-
quires development of a specific methodology which will be
a topic for further research.
Due to the complicated nature of the problem it was
necessary to solve our problem stepwise:
• The first step was development of the model that
describes the mechanical behavior of filtration
granules in filters with simple configuration
• The second step was development of a model that
deals with the behavior of gas in the filter
• The third step was modification of the results
achieved in the first step in such a way that the
model deals with complex configurations of the
filter.
The model developed was a 2D one in order to be able
to compare results obtained from this model with the re-
sults obtained from the physical model, which is similar to
concept of Hsiau and Tsai (1999) where a 2D physical
model was used (Figure 6). Another reason for choosing
the 2D case was the fact that the implementation of this
model is considerably less demanding than the 3D one.
Such an approach substantially reduces the computation
time and thus allows the user to perform much more ex-
periments within given period of time. The use of 2D mod-
els gives results that correspond very well with a real 3D
filter – see Chou, Tseng, Smid, Kuo and Hsiau (1999),
Smid, Kuo, Hsiau, Wang and Chou (1997).
The granules have the form of spheres with a typical di-
ameter from 1 to 10 mm. These granules are able to adsorb
dust and gas pollutants from the polluted gas that flows
through the filter. Each granule was represented by an ele-
ment in the filter bed. The principal assumption was that the
movement of granules in the bed is the result of forces acting
on each granule in the lower part of the filter. The nature of
these forces is simply the weight of granules in layers above
the granule in question. The solution to the problem (creat-
ing a satisfactory model) was to deal with the distribution of
forces caused by the weight of granules in the filter bed. Due
to the large number of granules the calculation of forces dis-
tribution was a rather time consuming process.
The model used for simulation of the movement of
granules has been based on the partition of the filter vol-
ume into a finite set of volume elements. The elements
have cubical form and are of the same size. This approach
allowed us to deal uniformly with all elements in the filter
volume which simplified the computational process. The
state of a volume element is given by several parameters,
like the number and positions of granules in the element.
From the point of view of the adsorption process, we are
interested in the parameter characterizing concentration of
pollutants that is assigned to each element. This concentra-
tion is stepwise modified during the gas movement through
the filter due to particle pollutants adsorption. The gas flow
is modeled as the transition of concentration of pollutants
between individual volume elements. The change in con-
centration of pollutants is determined by the number of
granules in the volume element and by their adsorption ca-
pability. The adsorption capability is given by a formula
(known as Dubinin linearized formula). For a more de-
tailed overview and explanation, see Hrdlicka, Slavik and
Kubelka (2001).
A complex method for simulation of behavior of gran-
ules in the filter was designed and implemented. Part of the
granules have a different color in order to follow various ve-
locities in various locations of the filter. The simulation re-
sults were verified by means of experimental results ob-
tained from tests on a real model of the filter, see Figure 6
where comparison between results from physical tests and
our simulations has been done (the simulation results match
Slavik, Gayer, Hrdlicka, and Kubelka
very well with physical ones). The matching has been done
visually – in case of need of exact comparison, correspond-
ing static images acquired from both processes could be
matched by methods used in the field of computer vision.
Sample outputs of simulation and visualization of ad-
sorption process can be seen in Figure 7 where the degree
of gas purification in various locations inside the filter can
be seen visualized by different colors.
5 HANDLING DYNAMIC FEATURES
OF THE FILTER BEHAVIOR
The solution used has several interesting features that are
based on the use of methods typical for computer graphics
in a simulation environment. An example is the use of Bre-
senham algorithm – see e.g. Foley, Dam, Feiner and
Hughes (1990) for dealing with non-horizontal gas flow
(the non-horizontal trajectories were composed from hori-
zontal chunks calculated by means of Bresenham algo-
rithm). Also the general trajectories of the gas flow in
complex situations were calculated not by flow equations
but by means of Fergusson curves – see e.g. Foley, Dam,
Feiner and Hughes (1990) that are applicable for this spe-
cial case of gas flow in the dense (granules) environment.
From the point of view of the use of dynamic visuali-
zation it is necessary to point out, that approaches used in
“static” visualization applications should be extended into
the dynamic environment. The main framework for the use
of visualization is the well known “visualization mantra”
defined by Shneiderman (1998): “Overview, zoom, details-
on-demand”. In the case of dynamic visualization we have
to control the time scale for dynamic visualization. At first
we try to identify some interesting dynamic patterns in the
dynamic process as a whole (with high speed of animation)
and in case that such a pattern has been found (in case of
dynamic filter behavior we are e.g. interested in the speed
of flow of granules in particular locations - like stagnant
zone where granules move with very low velocity) we can
in detail investigate this pattern with lower speed. This ap-
proach has been to some extent implemented in the above
given system for filter simulation.
The most important problem when dealing with dy-
namic visualizations is the speed both of simulation and
the visualization. In our projects we concentrated ourselves
on both of these aspects. In the first case (filters) we de-
veloped a special approach to simulation based on the use
of a finite element method, see Hrdlicka, Slavik and
Kubelka (2001) and Hrdlicka, Kubelka and Slavik (2002).
The movement of granules and their behavior from the
point of gas adsorption was simulated in volume elements.
The visualization (the dynamic aspects) part of the system
allowed us to change the speed of visualization.
6 VISUALIZATION OF DYNAMICS OF
COMBUSTION PROCESSES
The second system developed (combustion process simu-
lation and visualization in power-plants boilers) allows
more complex investigations. As the combustion process
is more complex than the previous one we can assume
that the volume of data generated by simulation will be
rather extensive. This situation negatively influences the
speed of visualization process and also handling exten-
sive data files in general.
6.1 Classical Modeling and Simulation
The simulation of various fluid flow related tasks and
combustion processes in power-plant boilers are generally
computationally expensive. This is a reason why we must
use large simplifications in the description of correspond-
ing physical descriptions and equations. But even with
these simplifications, modeling and solving complex tasks
such as combustion processes in today’s packages and
commercially available systems like FLUENT (Fluent
2003), can take hours or more. This is just the price for
reaching acceptable precision of computations needed for
professional, industrial design.
6.2 Dynamics Simulation and Visualization
A general disadvantage of the approach mentioned is the
complexity of simulation which results in very time con-
suming calculations. This has negative influences or in
most cases completely disables the possibilities of dynamic
presentation and visualization of results. Nevertheless, dy-
namics of the combustion is an interesting part of combus-
tion process study, namely at the start of the process.
We therefore make certain dispensations from the high
precision and reliability required for industry and produc-
tion applications. Our main goal is to construct compo-
nents and methods allowing dynamics study.
7 THE
BASIC
COMPONENTS
We use the following components to form our system – the
fast, structured fluid simulator and the particle system.
Both of these parts are described and explained more fully
in the following text.
7.1 The Fluid Simulator
Nowadays, simulation and visualization of various physi-
cal and nature phenomena using fluid simulators and
solvers based on the Navier-Stokes equations has major
theoretical and practical importance in simulation and es-
pecially in the computer graphics field. These simulators
Slavik, Gayer, Hrdlicka, and Kubelka
and solvers have been widely used for various research
projects and practical applications such as animations of
liquids and water (Foster and Metaxas 1996), fire (Nguyen,
Fedkiw and Jensen 2002), gas and smoke (Fedkiw, Stam
and Jensen 2001), effects in movies (Witting 1999) and
many others.
The combustion process of the pulverized coal and re-
sulting heat radiation and transfer is a quite complex prob-
lem. Again we are using some simplifications due to the
need for speeding up the computation. Instead of simula-
tion of these processes using the classical complex differ-
ential equations approach, we use a simple, statistical view
of the combustion process. The combustion and heat trans-
fers and fluxes are being computed separately for single
grid cells and corresponding particles inside them. This is
shown in Figure 2.
Our fluid simulator is based on the principle of local
simulation and uses a 2D structured grid (Gayer, Slavik
and Hrdlicka 2002). The simulated area is divided into grid
cells. It is necessary to take into account that the cell has
some volume (we can speak about volume elements) and
thus the model is in fact of 2.5 nature where the third di-
mension (depth) is also considered in contradiction to
models of purely 2D nature (like a simulation option in
FLUENT). In each step we calculate the new characteris-
tics (e.g. velocities, masses) for all grid cells. All calcula-
tions are reduced on nearest neighbors of the calculated
cells, see Figure 1. We periodically repeat these computa-
tions in each time step of the simulation. Such an approach
has acceptable requirements on computation speed allow-
ing for study of dynamics of the combustion process.
t = 0.00 seconds:
T = 600K (above ignition),
O
2
concentration = 60%
Coal particle
Partially burned particle
C
C
C
C
C
t = 0.01 seconds:
T = 605K (increased)
O
2
concentration = 58%
Partially burned coal particles
Coal particle transformed to
burned gas particle
C
C
B
Figure 2: Example Schematic Interaction of Coal Particles
with Air Mass During the Combustion Process for the
Time dt in a Detected Grid Cell
7.3 Saving the Pre-Calculated
Data Sets and States
In many cases, direct simulation of complex tasks with
subsequent real-time visualization cannot used. For main-
taining unconditionally real-time dynamics study even in
these complex tasks, we use pre-calculated data sets and
fluid simulator states (FSS) extensions. They in general al-
low speed-up of the simulation and visualization by either
storing all computed data or storing only partially com-
puted data with subsequent computations. The incorpora-
tion of these extensions into the simulation scheme of our
system is shown in Figure 3.
Figure 1: Division of the Boiler Chamber to 2D Grid Cells
7.2 Coal Particle System
By their nature, particle systems also represent a suitable
way for modeling the pulverized coal combustion process.
In our system, the particle system allows us both the com-
putation and visualization of coal mass elements in the
boiler. The particles displayed and calculated do not corre-
spond to the real coal particles in the boiler. Instead, they
represent a corresponding mass of coal in the cell under in-
vestigation. Thus we have a set of virtual particles in each
cell. From this point of view we can neglect some physical
phenomena like mutual particle collisions. The quality and
speed of both simulation and visualization can be altered
by increasing or decreasing the amount of particles.
The disk requirements for pre-calculated states are in
orders less than the ones needed for storing the full data
sets. However, the full data sets are important for maintain-
ing easy time navigation and increasing the replaying
speed of results availability of the visualization speed
change. Also with full data sets, we can for example save
only every 10th or 100th frame and considerably lower the
disk space requirements, at the cost of losing availability of
small time steps selections (for slowing the simulation)
when replaying of the combustion process.
Slavik, Gayer, Hrdlicka, and Kubelka
Visualization
Visualization
Interaction
Interaction
Store Full
Data Sets
Store Full
Data Sets
Fluid sim ulator
Fluid sim ulator
Com bustion &
heat transfer
engine
Com bustion &
heat transfer
engine
Particle
system
Particle
system
G rid cells
G rid cells
Store FSS
extension
Store FSS
extension
Figure 3: Incorporation of Storing Fluid Simulator
States (FSS) and Data Sets Extensions to Architecture
or Our Simulation System
Figure 4: Real-Time Visualization of the Combustion
Temperatures Grows Inside the Power-Plant Boiler
8 DYNAMICS
VISUALIZATION
8.3 Dynamic Particle and Volume Statistics
The system built on the above concepts allows comprehen-
sive study of the dynamics of the combustion process. This
concerns mainly situations when the sudden change of
some parameters occurs (like amount of oxygen, amount of
fuel etc.). The visualization module deals with visualiza-
tion of volume characteristics changes, flowing coal parti-
cles, dynamics statistics, zooming and modifying of time
position and speed.
Another way of presenting the computed values is utilizing
the statistics feature offered by our system. The inputs for
statistics are either values of any selected cell grids charac-
teristics or values of any above described characteristics of
particles. We can measure and visualize the values distribu-
tion in the grid cells and particles for all the above described
characteristics. The sample visualization output is shown on
Figure 5. The statistics are available in real-time and imme-
diately react on the dynamic changes inside the boiler.
8.1 Volume Characteristics Changes
The selected local characteristics changes in the grid cell,
such as the total temperature, mass values of combustibles
and air, local wattages, and heat fluxes, heat radiations,
pressures, burned mass, released heat, oxygen concentra-
tions and others (total about 40) and their changes in se-
lected time steps can be visualized, see Figure 4. In Real-
time Mode, the Screen Immediately Reflects the Change of
the Parameters (Such as Temperature)
8.2 Characteristics of Flowing Coal Particles
Simply using moving points representing flowing (virtual)
coal particles, we can visualize particle diameters, mass of
particles, the time and distance particles spend in the boiler,
the distance it arrived inside the boiler chamber, combustible
part of the particle and a few more (total about 10).
Figure 5: Sample Statistics of Coal Particle Diameters Dis-
tribution Inside the Boiler Chamber
Slavik, Gayer, Hrdlicka, and Kubelka
2 mi n 6 mi n 10 mi n 14 mi n 18 min
R e a l te s t
1,5 s 4,5 s 7,5 s 10,5 s 13,5 s
C om pute r s im ula tion
Figure 6: Comparison between Reality and Computer Simulation in Different Time
Moments (the Time Scales for Real Test and Simulation were Different – the Matching
Image Pairs are Placed Vertically)
Figure 7: Simulation and Visualization of Adsorption Process in Filter
8.4 Zooming Features and Time Navigation
We also plan to add time parameters for studying the
changes of the statistics in time step (adding time to the Z
axis), allowing for clearly finding the extremes and varia-
tions of values and grow-up of these parameters. Also, de-
pendence of one parameter on the other (e.g. diameter
changes of the particles in regards to particle burnout)
would be interesting to be studied this way.
We can select any zoom level and set the current position
to any area of the boiler chamber under investigation. With
utilizing the pre-calculated datasets, we can furthermore
arbitrary seek to selected time position, easily change the
replaying speed (slower, faster) of the simulation (similarly
as described in gas flue filters part) or even back-reverse
the combustion process. Without pre-calculated data sets,
increasing of the playing speed can be emulated, but with
Slavik, Gayer, Hrdlicka, and Kubelka
considerable slowdown of the system speed. The reversing
of replaying is not available at all in such a case.
8.5 Evaluation of Results Achieved
Our experiments have shown that our set of results differ
from results obtained by means of FLUENT (that was cho-
sen as a reference platform) in the following way: about
60% - 80% of volume elements have their attribute values
(temperature, pressure etc.) different from values obtained
by FLUENT by less than 30% (Gayer, Slavik and Hrdlicka
2002). This error is considered to be acceptable in this type
of application – see Stastny, Ahnert and Spliethoff (2002).
Another criteria for evaluation of our results could be de-
rived from parameters describing the global state of the
boiler (like global boiler fuel and combustion air input,
heat output etc.). In this case our values of almost all
global parameters differ from values obtained by FLUENT
by less than 25% - see Gayer, Slavik and Hrdlicka (2002).
The comparison between different simulated configura-
tions of boiler furnace (with immediate gain of results and
possible interactivity) corresponds to a similar situation ob-
tained by FLUENT. This type of correspondence is impor-
tant in practical use of the model.
Taking into account that generation of static images by
means of FLUENT takes hours and our system is able to
generate dynamic visualizations in real time (this means at
least 25 images per second) we can conclude that the loss
in accuracy is sufficiently compensated by the gain in
speed. The main goal of our research was to develop new
visualization methods. The simulation system was a sort of
test bed for development of new visualization methods. In
principle it would be possible to substitute our simulator
(in our visualization environment) by another simulator
with higher precision. Nevertheless our simulator generates
results that describe the simulated phenomena with satisfy-
ing accuracy – especially in applications like power engi-
neering education etc.
9 CONCLUSION AND FUTURE WORK
Design and implementation of systems allowing the user
investigation of dynamic phenomena was described. Visu-
alization of dynamic phenomena offers a new quality in
visualization of various processes and as such deserves
careful investigation both from the point of technical as-
pects of visualization and from the point of human percep-
tion. There are a lot of unsolved questions (e.g. another
kind of pollutant in the filter) that should be investigated in
the future. Both of our systems show that these questions
are of great practical use.
The first system covered some basic problems of dy-
namic visualization while the other one showed a way to a
potential solution of a more complex approach to dynamic
visualization. The first system been used for the design of
filters that will be used in practice. The second system is
currently used for students in power-engineering at CTU
Prague.
Due to the fact that the visualization part is separated
from the simulation part and the results of the simulation
part are stored in an interpretable form, we can interac-
tively control the visualization process. Besides the possi-
bility of the change of the time scale we have also the pos-
sibility to change the region of interest in various scales.
The combination of both of these approaches gives the user
a good opportunity to investigate dynamic phenomena both
in various time scales and also in various levels of detail.
This combination offers a new quality for visualization of
dynamic phenomena and gives a good base for further re-
search in the future.
Last, but not least, it is necessary to point out that
questions regarding visual perception of dynamic problems
are subject to intensive research in the field of psychology.
This situation offers good opportunity for development of
new visualization methods that will offer more deep insight
in data describing various phenomena. We see the main
contribution of this paper in modification of “visualization
mantra” for the dynamic environment.
ACKNOWLEDGMENTS
This project has been partially supported by CTU grant
CTU0210513. This project has been partially supported by
the Ministry of Education, Youth and Sports of the Czech
Republic under research program No. Y04/98: 212300014
(Research in the area of information technologies and com-
munications) and research program GACR 101/01/0955.
REFERENCES
Chou, C.S., C.Y. Tseng, J. Smid, J.T. Kuo. and S.S. Hsiau.
1999. Numerical Simulation of Flow Patterns of Disks
in the Symmetric moving granular Filter Bed. Techni-
cal Paper. National Taiwan University, Taipei.
Fedkiw, R., J. Stam, and H. W. Jensen. 2001. Visual Simu-
lation of Smoke. In Proceedings of the 28th annual
conference on Computer graphics and interactive
techniques, 15-22. ACM Press.
Fluent Inc. 2003. Flow modeling software and services.
Corporate website. Available via
<http://www.
fluent.com>
[accessed April 4, 2003].
Foley, J.,D., A. van Dam, S.K. Feiner, J.F. Hughes. 1990.
Computer Graphics: Principles and Practice, Addi-
son-Wesley, Reading, MA.
Foster, N. and D. Metaxas, 1996. Realistic Animation of
Liquids, Graphical Models and Image Processing:
GMIP, 58(5), 471-483. Canadian Human-Computer
Communications Society.
Gayer, M, P. Slavik and F. Hrdlicka. 2002. Interactive Sys-
tem for Pulverized Coal Combustion Visualization
Slavik, Gayer, Hrdlicka, and Kubelka
with Fluid Simulator. In Proceedings of the 2nd
IASTED International Conference: Visualization, Im-
aging, and Image Processing, 288-293. Anaheim:
Acta Press.
Hsiau, S.S and H.H.Tsai. 1999. Granular Flow in Dorfan
Impingo Filter for Gas Cleanup. In Proceedings of the
24th International Technical Conference on Coal
utilization & Fuel Systems, 659-667. ASME Fact U.S.
DOE and CSTA.
Hrdlicka, F., O. Kubelka and P. Slavik. 2002. Computational
Model for Efficient Filter Design. In Proceedings of the
27th International Technical Conference on Coal
Utilization & Fuel Systems. Gaithersburg : Coal
Technology Association, 539-548.
Hrdlicka F., P. Slavik and P. Kubelka., O. 2001. Simula-
tion of adsorption processes in gas cleanup filters, Air
Pollution IX, 385-394. WIT Press Boston.
Morrison, J.B. 1998. Lecture notes for Workshop on Cog-
nitive Models of Dynamic Geographic Phenomena
and Their Representations. Pensylvania State Univer-
sity. Available via
<http://www2.sis.pitt.
edu/~cogmap/ncgia/morrison.html>
[ac-
cessed April 4, 2003]
Nguyen, D., R. Fedkiw and H. Wann Jensen. 2002. Physi-
cally based modeling and animation of fire. Proceed-
ings of the 29th annual conference on Computer graph-
ics and interactive techniques, 721-728. ACM Press.
Shneiderman, B. 1998. Designing the User Interfaces. Ad-
dison Wesley.
Smid, J., J.T.Kuo, S.S.Hsiau, C.Y.Wang and C.S.Chou.
1997. Flow of Granules in Moving Bed Filter. In Pro-
ceedings Symposium on Transport Phenomena and
Applications, 289-294. National Taiwan University,
Taipei.
Stastny,M., Ahnert,F., Spliethoff,H. 2002. Three-
Dimensional Combustion Modelling of a Biomass
Fired Pulverized Fuel Boiler. In Heat Transfer, 2002,
119-126. WIT Press, Boston.
Witting, P. 1999. Computational Fluid Dynamics in a Tra-
ditional Animation Environment, In: Proceedings of
the 26th annual conference on Computer graphics and
Interactive Techniques, 129-136. ACM
Press/Addison-Wesley Publishing Co.
AUTHOR BIOGRAPHIES
PAVEL SLAVIK is an Associate Professor of Computer
Science at Czech Technical University in Prague -
Czech Republic. His professional interests cover user
interfaces and computer graphics (especially scientific
visualization). He works as a member of the Computer
Graphics Group at Department of Computer Science and
Engineering. The group is the largest group of its kind in
the country. He is an author of several textbooks used at
the university. He has also written several dozens of pa-
pers published on conferences round the world. He is
member of Eurographics, ACM, ACM SIGGRAPH and
ACM SIGCHI. Every year is a member of several
program committees of conferences concerning user in-
terfaces or computer graphics. His email address is
<slavik@cslab.felk.cvut.cz>
MAREK GAYER is a PhD student at the Department of
Computer Science and Engineering at Czech Technical
University in Prague – Czech Republic. His professional
interest is simulation and visualization of technological
processes. He is focused on development and implementa-
tion of new methods that should speed up both the simula-
tion and visualization of these processes. His email address
is
<xgayer@fel.cvut.cz>
.
FRANTISEK HRDLICKA is an Associate Professor at
Faculty of Mechanical Engineering, Czech Technical Uni-
versity in Prague. His professional interests include theory
of combustion processes, environmental problems and the
use of biomass. He is an author and co-author of several
textbooks in the field. Besides other activities he serves as
an advisor of Grant Agency of Czech Republic, Advisor
of Chemical Engineering Science, member of Supervision
Council of Czech Chamber of Civil Engineers. He regu-
larly publishes his research results in journals and confer-
ence proceedings. He is a member of international profes-
sional organizations like ASME. His email address is
<hrdlicka@fsid.cvut.cz>
ONDREJ KUBELKA graduated in Computer Science
from Czech Technical University in Prague. His profes-
sional interests include computer graphics, simulation and
network applications. He participated in several research
projects of this type. His e-mail address is
<ondrej.
kubelka@post.cz>
Document Outline
- MAIN MENU
- PREVIOUS MENU
- ---------------------------------
- Search CD-ROM
- Search Results
- Print