DEVELOPMENT OF A SYSTEMS SIMULATION PLATFORM TO ANALYSE
MARKET LIBERALISATION AND INTEGRATED ENERGY CONSERVATION
POLICIES IN COLOMBIA
Isaac Dyner' and Derek Bunn?
Universidad Nacional de Colombia and London Business School
ABSTRACT
With energy markets throughout the world moving rapidly into re-structuring,
liberalisation and privatisation, this presents a major challenge to traditional
modelling approaches. Econometric methods have limited applicability for a
new environment and optimisation techniques of “planning” are less relevant in
an era of market forces. This paper presents the use of system dynamics, in a
generalised way, to provide a platform for integrated energy analysis. Issues of
modularity and policy evolution are important in the design of the system to
facilitate its use, and reuse. In the 18 months since this work began for the
Colombian Ministry of Energy, it has had to evolve to support the analysis of a
number of changing perspectives and constraints. Hence the concept of a
platform rather than a model has to be implemented in a coherent way if it is
going to provide sustained value for on-going governmental policy support.
The history of System Dynamics (SD) models in energy analysis and planning can be
traced back to the early seventies when research conducted at MIT led to the creation of the
SD model COAL2. Improved versions of this model have been used to support all US
energy plans since 1977 and The National Energy Strategy 1991 (Naill, 1992). During the
last fifteen years Ford (Ford and Bull, 1989) has applied SD to a number of issues in the
electricity sector of the US.
In the UK, SD is being used to explore a diversity of issues related to the privatisation of the
electricity industry (Bunn et al., 1992 and 1996), and in Colombia this tool has been used to
study energy efficiency penetration and electricity substitution by gas in the residential and
industrial sectors (Dyner et al., 1995).
ENERGY POLICY ISSUES IN COLOMBIA
Despite the recent Public Services and Electricity laws, the transition towards a more
liberalised environment turns out to be a delicate affair as low prices, low reserve margin
and high subsidies, appear not to be the most favourable initial conditions. Thus a number
of open questions emerge concerning:
' Isaac Dyner. Facultad de Minas, Universidad Nacional de Colombia, A.A. 1027,
Medellin, Colombia. E Mail: idyner@perseus.unalmed.edu.co
? Derek W. Bunn. London Business School, Sussex Place, Regent’s Park, London NW1
4SA, UK. E Mail: dbunn@lbs.lon.ac.uk
45
e Evolution of prices
¢ Subsidies regime
© Supply reliability -
© Supply alternatives
e Demand side management
© Competition in the market of large consumers
« Competition in the retailed market
In this sense, there is not an obvious recipe to follow, as a number of crucial interlinked
problems arise: the selection of programmes, scheduling and prioritisation of projects
(seeking possible synergies), incentives for private investors and market regulation,
among many others. Therefore a framework for analysis capable of handling breath and
depth will be required. Breath in terms of being able to address a number of connected
policy issues. Depth in terms of supporting analysis at a detailed level, as in the case of
specific Demand- Side Management (DSM) programmes.
Moreover, an incrementalist approach needs to be followed, both, to analyse the
implementation of programmes and projects, as well as, for the purpose of policy making.
In many aspects, with respect to data not yet available in Colombia, the modelling
approach needs to make use of experiences from elsewhere. For example, in imposing
upon Colombia the UK-like power pool, can we incorporate, as “archetypes” re-
customised versions of the SD-type models that we have developed to understand the UK
system and thereby investigate whether similar behavioural properties will be exhibited in
Colombia? So much of the world-wide trend to re-structuring has been promoted by
analogies, that we need a platform to verify if the analogies will transfer to new contexts.
Similarly, can the insights from consumer choice models be assimilated and assessed into
this context?
As a context testing framework, the platform construction process has to evolve through
the adaptation of changing circumstances, including new components to address a diverse
number of complementary and dynamic issues. Furthermore, transferability of the
platform between different groups of modellers requires a modular and transparent
structure. End users also need to be trained to use their “‘own” platform in order to
develop it further as new policy issues emerge.
A SYSTEMS SIMULATION PLATFORM FOR COLOMBIA
Figure 1 shows the general platform structure along with its most relevant components
and the corresponding interrelationships. This design facilitates the analysis of energy
policies (DSM plans and programmes, the large scale gas plan and the indicative plan for
the expansion of power generation). The platform is system dynamics based with about
350 vector equations (24 of which are differential equations) and about 1000 vector
variables. Work started in 1995 as part of an 18 months development project for the
Colombian Ministry of Mines and Energy.
\qG
EXOGENOUS VARIABLES
SOCIO-ECONOMICS
ENDOGENOUS VARIABLES
a SOCIO-ECONOMICS
PRICES-PENETRATION
LI
ENERGY ENERGY
DEMAND | ee SUPPLY
Figure 1 Platform structure
POLICY EVALUATIONS
We explored the implication of pricing policies, when these coincide with programmes
and actions on DSM as well as on supply-side management. Simulation results in Figure
2 show the evolution of the increment on the average energy bill of domestic customers
and the additional revenues available to electricity utilities under specific pricing
scenarios.
mg
i
|
13°95 7 9 11 13 15 17 19 21
Time
Figure 2 Increments on: average energy bill (customer) and industry revenues.
Figure 3 illustrates the effect of a gradual pricing policy on additional sales and on
relative savings, and Figure 4 shows the average bill increment that customers may have
to be prepared to pay if tariffs are to follow market prices.
| |
i | 20 a
i Z _
i 15 v
' va {
i 10 a !
: {
| se :
| i i
Soviets et i | 1991988 2000 2002 2004 2006 2008 2010 |
i 1 Tir
oe eR Fag ' ms
Figure 3 Electricity sales and savings Figure 4 Average energy bill increment
\47
CONCLUSION
The modelling protocol of the platform exhibited in this paper is based upon the need of
implementing generic, modular, adaptable and transportable structures, capable of
supporting the process of system analysis for intervention. The platform developed here
therefore provides a context rich basis for investigating a number of different policy issues.
Although a similar framework has been undertaken in the US continuously since 1976, the
open literature gives limited information with respect to the modelling protocol that needs
to be followed if such a platform is to provide a sustained level of policy support through
changing circumstances.
As has been shown in this paper a System Dynamics based platform has the potential to
provide such evolutionary policy support under complex policy environments, although it
then needs to have a multifaceted perspective on its use. In particular, the application of
SD in these circumstances goes beyond the single-focus of providing insight into a pre-
selected policy issue. Yet, if the traditional approaches of optimisation and econometrics
are losing some of their relevance in the new era of market liberalisation, and SD is well
placed to fill the new modelling requirements, it seems that it must offer governmental
energy advisors some of the same features of longevity, modularity, adaptability and
detail that the other approaches have provided. The features then constitute requirements
for a clearer modelling protocol to be developed upon a systems simulation platform.
Acknowledgement
We wish to acknowledge the support of UPME (Colombian Ministry of Mines and
Energy), London Business School (University of London), Universidad Nacional de
Colombia and COLCIENCIAS during this research.
REFERENCES
Bunn, D. W. and E. Larsen. 1992. Sensitivity reserve margin to factors influencing
investment behaviour in the electricity market of England and Wales. Energy policy, pp.
420-429.
Bunn, D. W., I. Dyner and E Larsen. 1996. Strategic opportunities across the new UK
electricity and gas markets. (Submitted for publication)
Dyner, I., R. Smith and G. Pefia. 1995. System dynamics modelling for energy efficiency
analysis and management. J. Opl. Res., Vol 46, No 10, pp 1163-1173.
Ford, A. and M. Bull. 1989. Using system dynamics for conservation policy analysis in the
Pacific Northwest. System dynamics review, pp 1-16.
Naill, R.1992. An SD model for national energy policy planning, System Dynamics
Review, 8, pp 1-19.
14h
