Learning Dynamics and Control Using Remotely Tutored Simulation and Virtual Experiments

Abstract

The DynLAB project developed by an inter-national consortium aims at motivating young people to engineering study, and at improving engineering training using innovative didactic and technological approaches. The resulting web-based course is supported across the Internet by a software environment including a robust DYNAST simulation engine, publishing and monitoring tools, and a large collection of re-solvable examples including 3D virtual experiments. DYNAST solves nonlinear algebro-differential equations submitted in a textual form. For a system model submitted in a graphical form characterizing the system real configuration DYNAST formulates the underlying equations automatically. It is also capable of providing the system semisymbolic analysis in time- and frequency-domains. Besides this, DYNAST can be also used across the Internet as a modeling toolbox for the MATLAB control-design toolset.

Full text

Page 1

Learning Dynamics and Control Using Remotely Tutored Simulation and Virtual Experiments
Learning Dynamics and Control
Using Remotely Tutored Simulation
and Virtual Experiments

H. Mann, M. Sevcenko
Czech Technical University / Computing and Information Centre, Prague, Czech Republic


Abstract—The DynLAB project developed by an inter-
national consortium aims at motivating young people to
engineering study, and at improving engineering training
using innovative didactic and technological approaches. The
resulting web-based course is supported across the Internet
by a software environment including a robust DYNAST
simulation engine, publishing and monitoring tools, and a
large collection of re-solvable examples including 3D virtual
experiments. DYNAST solves nonlinear algebro-differential
equations submitted in a textual form. For a system model
submitted in a graphical form characterizing the system real
configuration DYNAST formulates the underlying equations
automatically. It is also capable of providing the system
semisymbolic analysis in time- and frequency-domains.
Besides this, DYNAST can be also used across the Internet as
a modeling toolbox for the MATLAB control-design toolset.

I.
II.

•
using computers to carry out old exercises without
radical modification of the curriculum to incorporate
computers in a way fully exploiting their contem-
porary capabilities
PROJECT DYNLAB
A project called DynLAB – Course on Dynamics of
Multidisciplinary and Controlled Systems in a Virtual Lab
has been developed by an international consortium [2]. The
emphasis and style of the course differs from most of the
existing courses by a number of innovative features:
•
exposing learners to a novel systematic and efficient
methodology for realistic modeling of multi-
disciplinary system dynamics applicable to electrical,
magnetic, thermal, fluid, acoustic and mechanical
dynamic effects in a unified way
•
introducing learners to the methodology through
simple, yet practical, examples to stimulate their
interest in engineering before exposing them to
rigorous theory and advanced mathematics
•
giving learners a better ‘feel’ for the topic by problem
on-line simulation, graphical visualization, and by
interactive virtual experiments
•
allowing different target groups to select individual
paths through the course tailor-made to their actual
needs and respecting their background
•
allowing both for self-study and remote tutoring with
investigative and collaborative modes of learning
•
integrating computers into the course curriculum
consistently and giving learners a hands-on
opportunity to acquire the necessary skills
•
exploiting the computers not only for equation
solving, but also for their formulation minimizing
thus learners’ distraction from their study objectives
•
giving learners the opportunity to benefit from
‘organisational learning’,
knowledge recorded during previous problem solving
both in academia and industry
The intended target groups of the DynLAB course are
students wishing to complement the traditional courses,
distance-education students at different levels of study,
practicing engineers as a part of their life-long learning as
well as teachers intending to innovate the courses they
teach.
Index Terms—control, dynamics, e-learning, Internet, remote
simulation, virtual experiments.
INTRODUCTION
The subject of dynamics and control underlies all
aspects of modern technology and plays the determining
role in the world-market competition of engineering
products. Its importance increases with the ever-growing
demands on operational speed, efficiency, safety,
reliability, or environmental protection of the products.
Nevertheless, national authorities and entrepreneurs in
many countries report lack of qualified engineers as well
as a critical overall decline of interest in engineering study
among young people. Professional associations call for
radical changes in the engineering curriculum and for new
innovative approaches to vocational training (e.g. [1]).
The existing courses on dynamics and control are
criticized namely for
•
discouraging young people from engineering study by
overemphasis on theory and mathematics at the
expense of practical engineering issues in the
curriculum
•
separating courses on dynamics analysis along the
borders between the traditional engineering discip-
lines despite the fact that most of the contemporary
engineering products are of multidisciplinary nature
•
presenting ‘textbook’ problems carefully engineered
to fit the standard ‘underlying’ theory without having
the students to undertake realistic modeling
i.e. from utilizing
iJOE International Journal of Online Engineering - www.i-joe.org

Page 2

Learning Dynamics and Control Using Remotely Tutored Simulation and Virtual Experiments
III. MULTIDISCIPLINARY DYNAMICS
The engineering systems become more and more
complex with regards both to the number of their
components as well as to the variety of phenomena
affecting their dynamics, either in a useful or undesirable
way. The phenomena involved in system dynamics might
come simultaneously from different energy domains
treated traditionally by different engineering disciplines
(electrical, electronic, magnetic, mechanical, fluid,
acoustic, thermal, etc.). To be able to cope with the
contemporary systems, engineering students should be
introduced into an approach to system dynamics treating
phenomena from different domains in a unified way. Such
a unified approach into engineering study gives students a
more comprehensive view of the real world.
There are two additional advantages to the unified
courses on system dynamics besides the ability to cope
with the contemporary systems. First, a properly planned
curriculum which includes such a course avoids
unnecessary duplication. This allows introducing more
advanced material into the study program. Second, a
unified course is consistent with the growing tendency of
students to decide on their branch of engineering relatively
late in their study. Such a course is also appropriate as a
part of a program of continuing education for graduated
engineers who received their degrees earlier.
To understand and predict the dynamic behavior of such
systems as well as to design them, engineers resort to
computer-assisted modeling, simulation and analysis.
Introduction of these techniques allows engineers
considering a larger variety of designed systems and for
their thorough verification before the system prototypes are
constructed and tested experimentally. Testing by
simulation is the only option in cases where experimen-
tation is too expensive or dangerous. Simulation also helps
to maintain the systems, and in the case of a system failure
it can be used to diagnose the failure cause. As
computation techniques allow for introducing new products
of higher-quality into the market faster, they help in
acquiring higher profits.
Simulation is also a well proven learning tool helping to
facilitate learners’ comprehension of dynamics and control
principles. A modern course on engineering dynamics
should consider multi-level modeling. When designing a
complex system, engineers resort to dynamic models of
several levels of system abstraction and idealization. The
design process usually starts by the conceptual design
phase of the highest abstraction and aims towards the
technological phase in which an assembly or some other
way of system production is designed. Between these, also
two intermediate design phases, both concerned about
system dynamics, can be recognized.
In the functional design phase, interactions between the
system components are assumed to take form of physically
dimensionless signals, states and disturbances. Typically,
the design of automatic control, of digital circuitry
architecture, or of imbedded software is carried within this
phase. The physical design phase is concerned about
implementation of the system architecture, functions and
signals in terms of physical phenomena and quantities.
Both useful and undesired multi-domain physical effects
are considered here in terms of energy transfer,
accumulation and dissipation. The interrelationships of
these effects are governed by physical laws.
DYNAST SOFTWARE
IV.
A.
B.
C.
Efficient simulation
In the past, efficiency of simulation was evaluated with
regard to its demand of computer time only. Nowadays,
however, the computer time is so inexpensive that the cost
of simulation is dominated by the cost of personnel
required to prepare the input data, to supervise the
computation and to interpret the results. Therefore, an
efficient simulation software tool should minimize
demands on its users’ time and qualification. In the other
words, software should be sufficiently user-friendly and
computationally robust.
When evaluating simulation software one should take
into consideration its application area. Models of high
abstraction and idealization used in the conceptual design
of control, for example, can be conveniently represented by
block diagrams. Using such block diagrams for physical-
level models is, however, a cumbersome and error prone
task. It requires an involved manual formulation of the
underlying equations and, in addition, manual construction
of block diagrams representing the equations.
DYNAST simulation software
To comply with the above mentioned efficiency
requirements, the DYNAST software package was chosen
as the kernel tool for multidisciplinary system simulation
within the DynLAB course.
DYNAST provides there:
•
solution of nonlinear differential and/or algebraic
equations submitted in a natural textual form
•
simulation of real dynamic systems the models of
which are submitted in a graphical form resembling
the real system configuration (the underlying
equations are then formulated automatically)
•
semisymbolic-form transfer functions and responses of
automatically linearized system models
•
online support for simulation, virtual experiments, and
monitoring of submitted tasks
•
modeling toolbox for MATLAB and Simulink
Multipole modeling
The automatic formulation of equations for physical-
level modeling of multidisciplinary systems is based in
DYNAST on multipole modeling. Such a modeling
procedure starts with decomposition of the modeled
systems into disjoint subsystems. This idea is similar to free
body diagrams in mechanics or control surfaces in
thermodynamics, for example. A subsystem multipole
model approximates the subsystem energy interactions
with the rest of the system under the assumptions that
•
the interactions take place just in a limited number of
interaction sites formed by adjacent energy entries
into the subsystems (like fluid inlets, electrical
terminals, translating
connections, heat-transferring contact surfaces, etc.)
•
the energy flow through each such entry can be
expressed by a product of two complementary power
variables (force – velocity, torque – angular velocity,
volume flow rate – pressure, electrical current –
voltage, magnetic flux rate – magnetic voltage, or
entropy flow – temperature)
or rotating mechanical
iJOE International Journal of Online Engineering - www.i-joe.org

Page 3

Learning Dynamics and Control Using Remotely Tutored Simulation and Virtual Experiments
Each energy entry into a subsystem is represented in its
multipole model by a pole associated with a pair of the
power variables. In graphical symbols of individual
multipoles, the poles are denoted by pins, i.e. by short line
segments sticking out of the symbol outlines. Dynamics of
a complete system is represented graphically by a multipole
diagram consisting from symbols of subsystem multipole
models. The sites of energy interaction between adjacent
energy entries are portrayed in the diagram by the diagram
nodes. The energy entries interacting mutually are
represented in the diagram by pins interconnected to the
same node by line segments called links. The links can be
viewed as idealized subsystem interconnections capable of
transferring energy in both directions without any
dissipation, accumulation or delay.




























In each energy domain, the most rudimental multipoles
represent pure twopoles like pure energy sources,
accumulators and dissipaters. Energy conversion from one
domain to another one can be modeled by pure transducers.
Only four different pure transducers are needed to model
all kinds of electro-magnetic, electro-mechanical, magneto
mechanical, fluid-mechanical, rotary-rectilinear, and other
energy conversions. Multipole models of real subsystems
like electronic or fluid devices, various mechanisms,
motors or sensors, heating or cooling units, etc., can be
build up from the pure multipoles. The multipoles can be
combined also with blocks or equations.
Fig. 1 shows the cross-section of a copying lathe and the
corresponding multipole model representing dynamics of
the lathe mechanical and hydraulic components.
The most important advantage of multipole diagrams
over block diagrams or bond graphs is in the isomorphism
between their structure and the geometric configuration of
the modeled real systems. In fact, multipole diagrams are
mappings of real system representations from the
geometric onto the topological space. The practical
consequence of the isomorphism is that the multipole
diagram can be set up in a kit-like fashion in the same way
in which the real system is assembled from its subsystems.
Such a modeling procedure can be based on mere
inspection of the modeled real system. Recollect that a
block diagram is just a graphical representation of a set of
equations. The line segments interconnecting blocks are
associated with just one variable that can propagate in one
direction only.
(b)
(a)
Figure 1. (a) Copying lathe, (b) its multipole model.
D.
V.
A.
The DynLAB course is delivered within a web-based
learning environment supporting
collaboration and their communication with a tutor.
Investigative learning is encouraged by a large collection
of solved problems and virtual experiments. The examples
can be resolved and modified in an interactive way across
the Internet. This gives the learners a hands-on oppor-
tunity to acquire the necessary skills in solving real-life
problems.
Submodel libraries
DYNAST is accompanied by libraries of submodels for
electronic and fluid-power devices, electro-mechanical
transducers, mechanism parts, control units, etc. The
submodel dynamics can be described by a combination of
multipoles, blocks, and equations nested in a hierarchical
way. The libraries are open for easy addition of user-
defined submodels and their symbols. Each DYNAST
submodel description is encapsulated in an independent
file. The default values of submodel parameters can be
overridden by values specified in terms of constants or
symbolic expressions. Fig. 1 shows the dialog used to
specify parameters of a hydraulic cylinder making a part of
a machine.
Using the multipole approach is also of several other
important advantages:
•
multipole models can be developed, debugged, tuned
up and validated once for ever for the individual
subsystems independently of the rest of the system,
and once they are formed they can be stored in
submodel libraries to be used any time later
•
this job can be done for different types of subsystems
(e.g., fluid power devices, electronic elements,
electrical machines, mechanisms, etc.) by specialists
•
the submodel dynamics can be represented by
different descriptions each of them suiting best to the
related engineering discipline (lagrangian equations in
mechanics, circuit diagrams in fluid power or
electronics, block diagrams in control, etc.)
•
modeling refinement or subsystem replacement (e.g.,
replacement of an electrical motor by a hydraulic one)
can be taken into account by a submodel replacement
without interfering with the rest of the system model
LEARNING ENVIRONMENT
Distributed simulation system
learners’ mutual
iJOE International Journal of Online Engineering - www.i-joe.org

Page 4

Learning Dynamics and Control Using Remotely Tutored Simulation and Virtual Experiments











Figure 2: Environment for remote modeling, simulation and virtual experiments.
As shown in Fig. 2, the kernel of the distributed
simulation system is formed by the DYNAST Solver that
the learners can access across the Internet in several ways
including even e-mail. Setting up multipole and block
diagrams directly on a web page is enabled by the
schematic editor DYNCAD formed by a Java applet.
DYNCAD converts diagrams into the DYNAST input
language and sends the data to the DYNAST Solver across
the Internet. After the computation results are sent back
and plotted on the client-computer screen.
Remote tutoring is enabled by the software tool called
DYNAST Monitor that is linked to the DynLAB server
across the Internet. It allows tutors to observe textual data
and graphical-form diagrams submitted by learners to the
DYNAST Solver. Tutors can not only monitor learners’
activities, but they can also communicate with them, assist
them in solving their problems and correct their errors if
necessary. DYNAST Monitor appeared to be very useful
even for tutors sharing the same computer room with
learners.
C.
Support for control design










































(a)
B.
User-friendly simulation environment
Even more comfortable and user-friendly mode of
access to DYNAST Solver provides DYNAST Shell. This
mode requires, however, downloading and installing this
free software on client computers with MS Windows.
DYNAST Shell has been designed for a wide variety of
tasks in a way suitable to users of different levels of
qualification and experience. All operations are intuitive
and they are supported by a context-sensitive help system.
A built-in syntax analyzer is continuously checking the
submitted data. Dialog windows (wizards) in DYNAST
Shell allow for submitting data without knowledge of the
input language (though DYNAST input language is very
user-friendly and sounds natural to engineers). The input
data is directly interpreted without any compilation delay.
DYNAST Shell includes graphical editors for multipole
and block diagrams as well as for submodel symbols. A
special dialog box for each new submodel is formed
automatically as shown in the screenshot of the DYNAST
Shell graphical interface shown in Fig.1b.
DYNAST Shell can also communicate with the server-
based DYNAST Publisher. It is a documentation system
for automated publishing reports on simulation experi-
ments and descriptions of library submodels using LaTeX.
The systems extracts automatically the relevant parts of the
input data and captures the submitted multipole or block
diagrams as well as the resulting output plots and includes
them into the documents. The documents can be converted
by the server software into PostScript, PDF and HTML
formats.
(b)
(c)
(d)
Figure 3: (a) Inverted pendulum, (b) multipole model,
(c) analog PID control, (d) digital PID control.
iJOE International Journal of Online Engineering - www.i-joe.org

Page 5

Learning Dynamics and Control Using Remotely Tutored Simulation and Virtual Experiments
In control design, the modeling efficiency of DYNAST
can be combined to a great advantage with the control-
design power of the MATLAB toolsets. Using DYNAST, a
model of the plant to be controlled can be easily set up in a
graphical form and then used to validate the open-loop
model. At the same time, DYNAST is able to compute the
required plant transfer-function poles and zeros and to
export them to MATLAB in an M-file. When the control
feedback loop is designed in the MATLAB environment
and added to the plant model in DYNAST, the complete
nonlinear control system is verified in DYNAST. In the
case of digital control design, the control system
configuration is implemented in Simulink. The block
representing there the controlled plant remains, however, in
DYNAST and communicates across the Internet with the
rest of the diagram using the Simulink S-function.
Let us consider analog-PID control design for the plant
in the form of the inverted pendulum given in Fig 3a. As it
is shown in [4], a considerable number of manual
operations is necessary before MATLAB can be exploited
for computation of transfer functions for such a plant.
DYNAST allows for avoiding all these tedious manual
operations. Fig. 3b shows a multipole model of the
pendulum, which has been set up using the web-based
schematic editor DYNCAD. After verification of the plant
open-loop responses using DYNCAD, an M-file with the
plant transfer functions is exported across the Internet to
MATLAB installed on a client computer. The plant PID
control design by means of MATLAB can than proceed as
described in [4]. Then the resulting feedback loop can be
added to the plant model in DYNCAD as shown in Fig. 3c.
Finally, the complete nonlinear control system is verified
in DYNCAD.
Also in the case of digital-PID control design the
transfer-function data for the plant model is first exported
to MATLAB. Then the digital control design can proceed
as described in [4]. To verify the design in this case, the
digital feedback loop is implemented in SIMULINK as
shown in Fig. 3d. The large square block represents there
the plant multipole model in DYNAST shown already in
Fig. 3b. The communication across the Internet between

(c)
(d)
(b)
Figure 4. Three tank virtual experiment.
(a)
this model remaining in DYNAST and the rest of the
diagram implemented in SIMULINK is enabled by the
SIMULINK S-function available for downloading at [2].
D. Virtual experiments
To stir up learners’ interest in dynamics and control as
well as to enhance their understanding of the topics, the
course text is augmented by 3D virtual experiments, most
of them interactively controllable. So far, the following
experiments are available at the project website: Carriage
& pendulum, Two tanks, Ball and beam, Gyro pendulum,
Three tanks, VTOL (Vertical Take Off and Landing)
aircraft emulator, Chemical reactor, Hydraulic cylinder,
Optical tracker, Two-link planar robot. The motions of
objects in all the experiments are driven by DYNAST
across the Internet. The only software the learners need to
download and install on their computers to be able to
observe the experiments, is the Cortona freeware VRML
browser.
As an example, Fig. 4a shows the three tank virtual
experiment. By clicking the mouse over the screen of their
computer students can adjust the red level marks on tanks 1
and 2, open or close any of the 6 valves interconnecting the
tanks with each other or with an outlet, and they can switch
on the pumps. This allows students to try to control the
system manually in such a way that the levels in tanks 2
and 3 reach the level marks as soon as possible and stay
there. Then they can go to the automatic control exploiting
the default PID control, or they can test a control algorithm
of their own design. The default control is illustrated by
Fig. 4b, Fig. 4c shows the physical-level multipole model
of the controlled plant. Submodels of motors, pumps,
valves and open tanks are included in a lower hierarchical
modeling level.
Fig. 5a shows a 3D virtual geometric model of a robot
the motion of which is governed by DYNAST simulation
across the Internet. The simulation utilizes the model of
robot dynamics given in Fig. 5b. Besides the robot motion,
learners can observe plotted responses of various robot
variables like the arm trajectory shown in Fig. 5c, for
example.





















iJOE International Journal of Online Engineering - www.i-joe.org