ELECTRICITE DE FRANCE
Direction des Etudes et Recherches
Maurice ROCHE
Service IPN - Département PROVAL
SYSTEM DYNAMICS APPLICATIONS
IN THE SECTOR OF ELECTRICITY
INTRODUCTION
As for as we know there are relatively few examples of application of
System Dynamics in the electric power Industry. Some of them are
nonetheless of high interest. We can mention the program of the
Bonneville Power Administration that had been presented in a
previous Conference of the System Dynamics Society. Our purpose at
Electricité de France was to scan the areas in which System Dynamics
could be helpfull to our company
MAIN FEATURES OF THE ELECTRIC SYSTEM
Before any further discussion, we need a comprehensive view of the
electric power industry.
Clearly, there is no question of drawing up a complete causal diagram
of a system of such enormous complexity; At this stage, we are forced
to rely on a few highly aggregated variables and, for the time being, to
overloook certain aspects that are rather more than mere details.
459
Economic
activity
Present
demand
Expected
demand
Clientele
Production ns |
Investments le L Fuels
Production
capacity
In this diagram, a variable such as "production" covers transport and
distribution, and is made up of quantities as well as prices. The other
variables are similarly complex. Economic, financial and tariff aspects
are not specifically mentioned, but are nonetheless omnipresent and
essential to the operation of the mechanisms that regulate the system.
This diagram could certainly apply to many industries. The
characteristics of the electricity system that we have to mention are:
- The high cost of transmission
- The technical difficulties to stock electricity
- The cost of investments.
Honever electric power Industries have developped around the world
under different regulations, the electric utilities have generally got
more or less a monopoly in there areas. At least they got a monopoly
as suppliers but very often they are in charge of nearly all their
production.
In many respects, decision making is an optimization problem and is
based on few mathematical tools, mainly mathematical programming
or statistical models.
The phenomenas involved at this level of the diagram are those
included in the large-scale models that have been developed by
electricity specialists for forecasting load curve, investment planning,
unit management optimization, etc. The loops involved in this diagram
are fairly insignificant, relate to relatively long time constants and are
reasonably well controlled by one actor.
Running mathematical models in series generates good results
“measured according to widely accepted criteria. The phenomena
develops slowly enough for it to make possible to avoid high-level
iterative procedures. The process can be controlled by revising
objectives on a yearly basis. Above all, system dynamics could play an
important role as a forecasting tool in the study of long-term
developments or in analysis of the main disturbances that can occur in
an electric power system.
At.a more macroscopic level, other loops appear, but this time the size
of the enterprise becomes insufficient. In this way, our fossil fuel
purchasing policy has only a negligible effect on the price of these
commodities. A dynamic model of the fuel market following the
example of those used by oil companies would be a useful tool but
would be limited in that it would not provide a basis for influencing
the development of the system in any significant manner. In such
cases, system dynamics is a particularly useful way of understanding
the mechanisms behind developments and helping to make forecasts,
but it does not play a direct role in. preparing the implementation of a
plan of action.
APPLICATIONS OF SYSTEM DYNAMICS
Finally, it would be of interest to make a brief analysis of some of the
myriad specific questions that would be amplified out of all
proportion in the diagram above. This type of question was tackled in
the studies on the regulation of the load curve by remote control, or
on the developpment of bi-energy applications.
Régulation of the load
The problem addressed here is to examine how undesirable
consumption peaks develop between tariff periods, and to determine
how to monitor and reduce these peaks through application of
appropriate tariffs combined with 188 Hz remote control. This
problem is a good example of system dynamics featuring a major
feedback effect :
- The tariff has an influence on consumer behavior;
- Consumer behavior determines the load curve and peaks in
that curve;
- The load curve is used to set the tariff.
461
Topics of this type are fraught with difficulties. The effects that we
are seeking to analyze are partly results of the large-scale
mathematical models above mentionned. The system dynamics
approach is interesting, but in this specific problem it very quickly
becomes necessary to take sophisticated mathematical models into
account in our analysis, and this requires a great deal of skill. It would
not be fair to claim that the subject is unsuitable for a system
dynamics approach, but it is true that it is too complicated to be used
as an introduction to the subject.
Bi-energy applications
The use of electricity through bi-energy processes, ie those which can
alternatively use electricity or another source of energy is obviously
interesting in some cases.
In the french electrical system for instance, some of these devices can
be operated seasonally, using electricity in the summer and a fossil
fuel when electricity cost is higher during other seasons.
This kind of opportunity can be used for many appliances, such as
boilers producing hot water and steam, which still play an important
part in the industrial equipment devoted to energy conversion.
The advantages that the bi-energy approach to electricity offers users
depend to a very large extent on the price of competing fuels (fuel oil
or coal). In certain periods, fluctuations in the price of fuel oil can
bring the development of the bi-energy approach to a virtual
standstill, and can discourage boiler manufacturers from remaining in
such an unstable market. Is it possible, therefore, to define a tariff
policy that attenuates the risks to the user of investing in bi-energy,
at the same time as ensuring the survival of bi-energy equipment
manufacturers ?
This topic is somewhat less difficult to deal with than the load curve
regulation example discussed above, but it is also less interesting as a
test bed for system dynamics in that it is considerably less complex
and feedback effects have turned out to have less influence that was
throught at the outset. It has shown that when we are dealing with
parameters that are merely appendages to the basic variables in the
overall diagram, the feedback effects are diminished and rarely play
a crucial role.
462
CONCLUSION
The system dynamics approach has nonetheless made it possible to
reach a more operational stage and to address practical questions
related to decision-making aids. A lot remains to be done before we
can make a clear distinction between the questions that can be
analyzed through a dynamic approach and those that it is still possible
to examine with non-retroactive models. Nevertheless, problems that
are less complex than the electric power system as a whole will
provide the basis for using system dynamics as a decision-making aid.
CC BY-NC-SA 4.0