EUROCON 2003 Ljubljana, Slovenia
Abstract--We present our educational system for interactive
education of combustion processes. The system is built on
several concepts used mainly in the computer graphics area
(fluid simulator, particle system) and combustion simulation
field (simplified combustion process model and heat transfer
engine). Together they combine unique and original concepts
that offer real-time simulation and visualization of the
combustion process. The user may have immediate interaction
during simulation and visualization – e.g. changing coal inlets
and combustible properties and other input parameters during
simulation. The system allows real-time monitoring of about 40
basic cell volume characteristics inside the boiler and 10
pulverized coal particle characteristics. All these features are
available immediately, without needing to wait hours for
complex calculations to finish. The system is especially suitable
for interactive education purposes in power-engineering.
Index Terms—Interactive Education, Power Engineering,
Coal Combustion, Real-time Simulation and Visualization, CFD
I.
I
NTRODUCTION
N recent years, modern high performance desktop and
workstations have made a revolution to the power
engineering industry and computer graphics area. Currently,
there is a vast amount of research projects, applications and
commercial products, which are able to simulate and
visualize many natural production and technological
processes. They are constructed for both the economical and
ecological reasons – the possibility of relatively quick and
precise analysis and design, which modern software tools
offer, can save much engineering effort and money that
would be needed for corresponding manual design and
calculations. In other words, they allow the designer to
experiment extensively with the model of the boiler designed
without the necessity to build a physical model of the boiler.
Although today’s commodity PC’s have multiplied their
performance, there are still many tasks for which the current
computing power is still insufficient. The simulation of
various fluid flow related tasks and combustion processes is a
typical example. For these we must use large simplifications
in the description of corresponding physical descriptions and
equations. But even with these simplifications, modeling and
solving complex tasks such as combustion processes in
today’s packages and commodity systems can take hours or
more. This is just the price for reaching acceptable precision
of computations needed for professional industry design. A
good example of the above is modeling of the combustion in
boilers using the well-known and widely used FLUENT
package [1]. General disadvantage of this approach,
especially in education, is the complexity of simulation which
results in very time-consuming calculations.
We are in a different situation when we can dispense from
the reliability and high precision required for industry and
production applications. The reason for such decisions can be
desire of unconditionally real-time, interactive combustion
simulation available on today’s commodity PC’s and
workstations. This can be especially useful when designing
tools for engineering education and training.
A. Visualization of simulation
In a simulation field, visualization has major importance in
presenting the simulated data in a natural, easily read,
understandable and expressive form. This is especially
required for real-time applications, when a suitable form of
visualization is absolutely essential for synoptic presentation
of the computed data.
B. Interactive education
Nowadays, common availability of high performance
PC’s, easy to use operating systems, high quality projectors
and slightly increasing general computer knowledge, leads to
choosing more effective, telling and visual forms of
Marek Gayer
1
, Pavel Slavík
2
and František Hrdlička
3
1,2
Department of Computer Science and Engineering, Faculty of Electrical Engineering
Czech Technical University in Prague
Karlovo náměstí 13, 121 35 Prague 2
Czech Republic
3
Department of Thermal and Nuclear Power Plants, Faculty of Mechanical Engineering
Czech Technical University in Prague
Technická 4, 166 07 Prague 6
Czech Republic
email: xgayer@fel.cvut.cz
Interactive Educational System for Coal
Combustion Modeling in Power Plant Boilers
I
EUROCON 2003 Ljubljana, Slovenia
education. It can start by choosing illustrative schemes and
pictures, or using common animation formats (like MPEG
and AVI files).
One of the most interesting and favorite forms of education
is interactive education (and derived forms, like distant
education). It is especially suited for complex, practical tasks,
when long, theoretical explanation would be ineffective or
could even lead to confusion of students.
An interactive form of education (if used in a proper and
meaningful way) offers “doing by learning” and “what if”
features. It can fulfill dynamic requirements of the teacher
and students. If it is used in an individual learning and
practicing part of education, it can motivate the students to
use their creativity and can easily answer some of their
questions. Furthermore it can motivate them in the learning
process, which now becomes more interesting and offers
more “fun”.
Usual problem of the seminar courses and lessons
efficiency is individual works of each of the students. An
interactive educational system for coal combustion modeling
is one of the options for more efficient education. A
challenging part of education in the field of combustion
processes is the dynamic behavior of burning coal particles.
An important task is to introduce and describe particle traces
and their changes (with respective heat release). Also it is
important to clearly demonstrate concrete power output
(specified volume load) and its changes, heat transfers into
the combustion chamber walls and temperature field into
furnaces. All these requirements can be met when using our
system, which will be described in the following text. The
model respects a concrete fuel system, its specific features
and the necessity to respect inlets of fuel and combustion air.
II. S
IMULATION OF COMBUSTION PROCESSES
We use the following components to form our education
system – the fast, structured fluid simulator and particle
system. Both of these parts will be further described and
explained in the following text.
A. The fluid simulator
Nowadays, simulation and visualization of various
physical and natural phenomena using fluid simulators and
solvers based on the Navier-Stokes equations has major
theoretical and practical importance in simulation and
especially the computer graphics field. These simulators and
solvers are widely used for various research projects and
practical applications such as animations of liquids and water
[2], fire [3], gas and smoke [4], and many others. Some of
them are used for special effects such as melting [5] and
animations in movies [6]. In most cases, the fluid simulators
are suited for specialized applications, but in most cases, with
certain effort, they can be modified and utilized in general
applications. For example, a fluid simulator originally
developed for animation of liquids could be modified for air
flow computation.
We can divide fluid simulators into two types. First are
complex and stable (but computationally more expensive)
methods such is in [7], that are able to determine the flow
progress independently of the time steps, but at some cost to
the computational speed of the single frames. The second are
unstable, where an acceptable time step is dependent on the
type of task we solve. Our simulator, as well as others [8], is
unstable. It is based on the principle of local simulation and
uses a 2D structured grid [9]. The simulated area is divided
into grid cells. In each step we calculate the new
characteristics (e.g. velocities, masses) for all grid cells. All
calculations are based on nearest neighbors of the calculated
cells (see Fig. 1). We periodically repeat these computations
in each time step of the simulation. The stability of unstable
solvers depends on proper selection of dimensions of the
selected grid cell, and time step in regards to speed of value
changes (such are velocities and mass) in cells during
simulation.
Fig. 1. Division of the boiler chamber into 2D grid cells. The cell values in
the next time step are computed from nearest neighbors only.
B. Coal particle system
Particle systems in computer graphics can clearly simulate
and visualize various natural phenomena such as dust, rain
and other models which could involve particle object
primitives. Particle object abstraction is also used for
industrial technology [10]. By its nature, particle systems also
represent a suitable way for modeling the pulverized coal
combustion process.
We use a coal particle system, enabling easy and fast
computation of the combustion processes. In our system, the
particle system allows us both the computation and
visualization of coal mass elements in the boiler. The
particles displayed and calculated do not correspond to the
real coal particles in the boiler. Instead, they represent the
corresponding mass of coal in the cell under investigation.
The quality and speed of both simulation and visualization
can be altered by increasing or decreasing the amount of
particles. Currently, the amount of particles being used for
simulation varies between thousands and ten thousands.
The combustion process of the pulverized coal and
resulting heat convection is a quite complex problem. Again
we are exploring some simplifications due to the need for fast
computation. We use a simple, statistical view of the
combustion process [11]. The combustion and heat transfers
EUROCON 2003 Ljubljana, Slovenia
and fluxes are being computed separately for single grid cells
and corresponding particles inside them.
Fig. 2. Example schematic interaction of coal particles with air mass during
the combustion process for the time dt in a selected grid cell
Our education system uses the industry standard OpenGL
platform for reliable and fast visualization. This means that
our system could be used on a standard low-cost graphics
accelerator. The values of characteristics are visualized using
color attribute of drawn graphics primitives (e.g. darker color
values correspond to greater values, see Fig. 3).
Fig. 3. Visualization of the combustion temperatures of the power-plant
boiler. The greatest temperatures are in the core of the flame tube
C. Volume characteristics
The selected local characteristics 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 concentrations and
several others (about 40 total) can be visualized. We use
OpenGL linear interpolated quads with the support of the
graphics hardware acceleration for visualization of the grid
cell characteristics (see Fig. 3). This concept allows easy,
real-time visualization which gives results similar to the
widely used isosurfaces technique.
D. Particle characteristics
The coal particles are visualized using the standard
OpenGL capability of drawing pixels (see Fig. 4). We utilize
built-in hardware support for visualization of smoothed
pixels and alpha blending, available and supported in today’s
commodity available graphics accelerators. Thus, even at
higher zoom levels and/or used particle sizes we obtain
visually acceptable representation of the particles (see Fig. 5).
Fig. 4. Visualization of thousands of virtual coal particle characteristics
together with selected cell grid characteristic (drawn on the background).
Fig. 5. Visualizing particles using full hardware accelerated, combined
smoothing and blending of pixels gives an acceptable visual quality even
with a high zoom level.
We can visualize in this way the particle diameters, mass
of particles, the time and distance particles spend in boiler,
the distance it arrived inside the boiler chamber, combustible
part of the particle and a few more (about 10 total).
Utilizing the advantage of the particle system concept, we
can easily construct the particle traces. We produce this effect
by keeping the previous particle positions and characteristics
in main memory. After that we can draw the particles in
current time step together with the kept ones. The particle
traces can clearly indicate the velocities and direction of
movement of coal particles, which are visible even on the
static state picture (see Fig. 6).
t = 0.00 seconds:
T = 600K (above ignition),
O
2
concentration = 60%
Coal particle
Partially burned particle
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
B
C
C
C
EUROCON 2003 Ljubljana, Slovenia
Fig. 6. Real-time visualization of partial particle traces helps in determining
particle speed, direction and dynamics even in static images. However, the
visually best overview of dynamics is gained in real-time mode of our
system.
E. Particle and volume statistics
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 grid
characteristics or values of any described characteristics of
particles. We can measure and visualize the value distribution
in the grid cells and particles for all the above described
characteristics. The sample visualization output is shown in
Fig. 7.
Fig. 7. Sample statistics of coal particle diameters distribution inside the
boiler chamber.
III. I
NTERACTIVE FEATURES OF OUR SYSTEM
We can divide the interactive features of our system into
the two following parts – interactive simulation and
interactive visualization of results.
A. Interactive simulation
During the simulation, we can interactively modify any of
about 30 simulation parameters. We can for example, on the
fly, increase the number of generated particles (improving the
simulation precision at a certain cost of visualization and
simulation speed, see Fig.
8
) or change the mass of average
coal particles flowing from the inlets of the boiler.
Fig. 8. The count of new generated coal particles is being modified and
immediately applied to the simulation computation on the fly
Moreover we can interactively change the parameters of
the inlets – completely change velocities, position, direction,
and angle of spread, air and coal masses, diameters and
coefficients of air and coal flowing to the boiler chamber.
The interactive modification of boiler inlet is shown in Fig. 9.
Fig. 9. Interactive modification of coal inlet parameters. The state of the inlet
and particles before change is on the left side. The changed state after the
next 5 seconds of interactive simulation is shown on the right side.
Furthermore we can select any one of the particles
(representing corresponding mass of the coal under
investigation), change it’s parameters on the fly, including
position and then watch how it behaves after applying such
changes (see Fig. 10).
Fig. 10. Tracking and monitoring of characteristics of the selected particle
(highlighted as the light gray disc) inside the boiler chamber. We can lock the
viewing area to its position and watch the coal particle as it flows through the
boiler with any zoom level.
B. Interactive visualization of results
We can select any combination of the particle
characteristics, grid cell characteristics and statistics. We can
EUROCON 2003 Ljubljana, Slovenia
change visualization parameters, e.g. size of the particles,
select area in the boiler under investigation, change the
visualization method of drawing pixels and set zoom level.
We can also adjust color palettes and color ranges for
visualization, setup brightness or invert the screen colors
(choosing a color scheme of light background and dark
particles and grid cells) suitable for printing and many others.
IV. FEEDBACK
FROM
STUDENTS
Response of students to this new way of education was a
very positive one. This new way allows students to perform
their own experiments individually and in a practical way
answer questions emerging from theoretical foundations of
the theory of combustion. The system allows for individual
experience based on concrete choice of correct and incorrect
parameters (like size and other properties of particles, ratio
between amount fuel and air, velocities of the air flow,
positions and directions of inlets etc.), their influence on the
combustion process and watching corresponding dynamic
behavior. This possibility to gain an individual experience
based on experiments has been valued very much by our
students. This fact is especially important because the subject
where this system has been used belongs to a very basic
subject in the power-engineering study track.
V. C
ONCLUSION
Our educational coal combustion system is based on a
simple fluid simulator. The fluid simulator allows real-time
computation of the air flowing inside the boiler. We selected
particle systems for maintaining visualization of combustion
process dynamics. By their nature, they can even be utilized
in the simulation and computation part.
The high speed of the fluid simulator and combustion
powered by the particle system and simplified combustion
engine allows real-time visualization of the results (using
OpenGL graphics interface). The system has been
implemented in 2D grid cell space, and regarding the
methodology used, can be easily extended to 3D space. But
even with the 2D grid version, the system is fully sufficient
for educational purposes, where clarity and real-time
interactivity demonstrate both the universality and
preciseness of computation. However, conversion of all our
concepts and architecture of the system to 3D grid cell space
is also planned.
The most powerful and new feature of our system is the
simulation and visualization interactivity, which is available
during real-time computation of the combustion process,
without needing to stop or restart the system. Tens of input
parameters, including coal inlets setup can be modified on the
fly.
Our results make it possible to get a good preview of the
dynamics of combustion processes in a boiler. The students
and developers of the combustion boilers could now test
many configurations and modifications of the pulverized coal
boilers interactively with an immediate response. The system
by itself can in an interactive, efficient and attractive form,
give an overview of how power-plant boilers work, with an
overview of fundamental boiler parameters. The interactive
way of modeling can bring the student basic knowledge and
policies of constructing performance and efficient boiler
solutions and more. We recommend it as an introduction
application for a combustion processes overview when
studying power plants.
Currently, our system is implemented in Microsoft Visual
C++ and runs at interactive frame rates even on a commodity
PC equipped with only AMD Athlon 1333 Mhz and nVidia
GeForce2 MX based graphics card.
The system has been used in the educational process in the
Faculty of Mechanical Engineering at the CTU, Prague. The
result of practical usage of our interactive system can be
described as a very successful one. It is a transition from the
traditional way of education based on the mediation of
knowledge (from professors to students) to the education
based on gaining an individual experience. In such a way
students can master new knowledge in a more interesting
way and much deeper than in a traditional approach.
VI. ACKNOWLEDGMENT
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 communications).
R
EFERENCES
[1] Fluent Inc. - Flow modeling software and services. Corporate website.
http://www.fluent.com/
[2] D. Enright, Stephen Marschner and Ronald Fedkiw. Animation and
rendering of complex water surfaces, Proceedings of the 29th annual
conference on Computer graphics and interactive techniques, pp. 736-
744, ACM Press, 2002,
[3] D. Nguyen, R. Fedkiw and H. Wann Jensen. Physically based modeling
and animation of fire. Proceedings of the 29th annual conference on
Computer graphics and interactive techniques, pp. 721-728, ACM
Press, 2002.
[4] R. Fedkiw, J. Stam, and H. W. Jensen. Visual Simulation of Smoke,
Proceedings of the 28th annual conference on Computer graphics and
interactive techniques, pp. 15-22, ACM Press, 2001.
[5] M. Carlson and P. Mucha and R. Brooks Van Horn, III G. Turk.
Melting and flowing. Proceedings of the ACM SIGGRAPH symposium
on Computer animation, pp. 167-174, ACM Press, 2002.
[6] P. Witting. Computational Fluid Dynamics in a Traditional Animation
Environment, Proceedings of the 26th annual conference on Computer
graphics and Interactive Techniques, pp. 129-136. ACM
Press/Addison-Wesley Publishing Co., 1999.
[7] J. Stam, Stable Fluids. In Computer Graphics Proceedings Annual
Conference Series (Los Angeles, USA), pp. 121–127, ACM Press,
1999.
[8] N. Foster and D. Metaxas, Realistic Animation of Liquids, Graphical
Models and Image Processing: GMIP, 58(5), pp. 471–483, 1996.
[9] M. Gayer, P. Slavík, F. Hrdlička - Interactive System for Pulverized
Coal Combustion Visualization with Fluid Simulator, Proceedings of
the 2nd IASTED International Conference: Visualization, Imaging, and
Image Processing, pp. 288-293. Anaheim: Acta Press, 2002.
[10] M. Rhodes,: Introduction to particle technology, John Wiley & Sons
Ltd., 1998.
[11] J. Tomeczek. Coal Combustion. Krieger Publishing Company, 1994.