"Energy modelling with

System Dynamics"
Prof. Dr. Stefan Wolfgang Pickl
Prof. Dr. Bo Hu
Mag. Armin Leopold M.Sc.
Universität der Bundeswehr München
COMTESSA
 Agenda

 How System Dynamics started

 Limits to Growth
 The importance of Energy
 Energy Modelling with System Dynamics
 COMTESSA and System Dynamics
 Germany's Electricity Industry in 2025
 Data analysis for forecasting with SD
 Energy Foresight and the Chances for SD
System Dynamics

Introduced by Jay W. Forrester of MIT in 1958

System Dynamics is: “The study of information-

feedback characteristics of industry activity to show
how organizational structure, amplification
(in policies), and time delays (in decisions and actions)
interact to influence the success of the enterprise”
(Forrest 1958 & 1961)
System Dynamics is closely linked
to the development of the computer

"Models and simulations of any type are tools

for the scope with the reality; they are as old
as mankind itself."
System Dynamics offers just such tools "with
the single and important difference, here
models of thinking translated into computer

Bossel 2004
models and simulations with the computer
performed."
Jay W. Forrester –
founder of system dynamics (SD)

Born 1918 in Nebraska, USA

Studied electrical engineering

Headed initially different MIT programs for

implementation of simulation systems for the US Navy

Brought magnetic core memory into use

Joined 1956 the MIT Sloan School of Management

Solved with SD first management issues for

General Motors and Digital Equipment Corporation
System Dynamics

„System dynamics combines the theory,

methods, and philosophy needed to analyse
the behavior of systems in not only
management, but also in environmental
change, politics, economic behavior,
medicine, engineering, and other fields“
 The beginning

 WHIRLWIND I- MIT's first general-

purpose digital computer

 Aircraft flight simulator for the U.S.

Navy Jay Forrester: Director of the MIT
Digital Computer Laboratory

 Forrester invented coincident-current

random-access magnetic computer
memory
 DYNAmic MOdels
(DYNAMO)

 Developed by director Jay W. Forrester at

the MIT in the late 1950th
 Simulation language+accompanying
graphical notation
 developed within the system dynamics
analytical framework
 Used for Industrial Dynamics, Urban
Dynamics and World Dynamics
 The first models

“The study of information-feedback

characteristics of industry activity
to show how organizational
structure, amplification (in
policies), and time delays (in
decisions and actions) interact to
influence the success of the
enterprise”
 The first models

Urban Dynamics:

• The first major non-corporate

application of system dynamics.

• Illustrates why many well-known

urban policies are either ineffective
or make urban problems worse
 Success continues

In 1970, Jay Forrester was invited by the

Club of Rome to a meeting in Bern,
Switzerland
Forrester said that system dynamics could
be used to address the predicament of
mankind.
The model showed a collapse of the world
socioeconomic system sometime during
the twenty-first century, if steps were not
taken to lessen the demands on the earth's
carrying capacity.
 1971

Invitation by the Club of Rome

to the MIT to create
a so-called “World model”
The main intention of the World model was:
To model, to simulate and furthermore analyze the
main reasons and the major problems of population
growth, the industrial capitals, the world food
production, the world ressource consumption and
also the world pollution issues.
 "WORLD DYNAMICS"
WORLD Dynamics (Forrester 1971)
Shortage of natural resources
Extreme pollution crisis

Population:
6000 Mio (2030)
1000 Mio (2070)
 Future Challenges -
40 Years Limits to Growth
System dynamics models now
cover a wide range of areas

• public affairs, including public health,

• energy and the environment,
• social welfare,
• sustainable development,
• security and
• many other related areas.
THE IMPORTANCE OF
ENERGY
EU Commissioner for Energy
Günther Oettinger:

"When it comes to energy production and

consumption, Europe needs more than ever to be
on a sustainable path.
It is time to invest in higher energy efficiencies
and a wider use of renewable sources: this is the
best way forward to ensuring safe and
competitive energy for us and children.”
 Motivation

In 2020 reduction of greenhouse gas emissions

in Germany about 40% compared to 1990

In combination with the decision to phase out

nuclear power supply till the end of 2022

The reliability of energy supply must be kept

at a high and secure level
 Motivation (cont.)

• Moreover, it is aimed to established

an electric energy supply system
fully on the basis of renewable
resources till 2050.

• Supported by the public, however,

no substantial income loss or even
an economic down turn will be
accepted in this context.
ENERGY MODELLING WITH
SYSTEM DYNAMICS
Short overview
COMTESSA AND SYSTEM
DYNAMICS
GERMANY'S ELECTRICITY
INDUSTRY IN 2025: EVALUATION OF
PORTFOLIO CONCEPTS
Germany's Electricity
Industry until 2025

Resources like fossil fuels or the storage capacity of atmosphere for CO2 and other
Greenhouse Gas (GHG) emissions are finite.

Germany’s Energy Transition Plan is challenging

Reducing GHG emissions in Germany about 40% in 2020 compared to 1990
In combination with the decision to phase out nuclear power supply till the
end of 2022
The reliability of energy supply must be kept at a high level

SRU 2010, UBA 2010

No substantial income loss or even an economic down turn will be accepted
in this context

Target 2050: an electric energy supply system 100% on the basis of renewable
resources
Energy modeling with SD -
The research gap

“My experiences with energy industry modeling convinced me that

the ability to simulate the information feedback in the system is a
truly unique feature of the system dynamics approach.” (Ford:
1997, p.77)

• The publication “System Dynamics and the Electric Power

Industry” by Ford shows that SD is able to model and
simulate especially the interconnections in an energy grid.
• The comparison of different decision options for Germany
which are dimensioned on the same reliability level were
not discussed yet with SD.
Energy modeling with SD -
The research gap

“For markets in transition, where strategic imbalances exist, system

dynamics has a useful role to play in developing a better
understanding of processes which might shape their evolution.”
(Bunn: 1997, p.271)
• Bunn explains in “Modelling latent market power across gas and
electricity markets“ in details the challenges to make the transition of
an energy market more understandable.
• Nevertheless there is no publication yet, which compares and analyses
CO2 mitigation and production costs of eight portfolio concepts under
four price scenarios with SD, which are essential for the German
energy transition.
• Furthermore, there is no scientific publication yet, which shows the
impact of a concept of renewable load management with System
Dynamics.
Objective of this publication

Forrester 1994, Andersen 1997, Ghaffarzadegan 2011

We are developing highly aggregated System Dynamics models depicting the
development of energy market
Based on our previous work in the area of emissions trading under uncertainties

 Intention: SD makes the modeling process transparent and intuitive in

approaching the politically active public
 Dynamic behavior is key for understanding of electricity supply
 Resource consumption and reliability of electricity supply over the course of a
year are simulated
Objective of this publication

In our System Dynamics simulations the minimal Dispatchable Capacity which

still provides reliable electricity supply under given Load Profile and WDPV
Profile for an entire year was found interactively for each concept with its
characterizing key parameters like WDPV Capacity, SNG Capacity and so on.
A concept is considered as reliable if the cumulative energy shortage is less
than 2.6 TWh during the entire year or 0.3 GW in average
Shortages are displayed in red color

Different concepts to be compared with each other are dimensioned on the same
reliability level.
The total production cost which includes investment, operating, fuel costs and
emission permits was then calculated for each concept.
 Screenshot of our concept of
renewable load management with
System
Given WDPV capacity and
Dynamics
Shortage indicator

BMWT 2011, SRU 2010, UBA 2010, Ventana Systems 2009

storage it is to find the Supply shortage
minimal Dispatchable
Capacity which ensures
reliable electricity supply Filling level of storage

Net load
Different decision options
are compared to each other

Portfolios: Different decision options and technologies, for example, SNG

(Synthetic Natural Gas), are compared to each other
Our reference portfolio concept "Rnw" ("renewable")
• facilitating 150 GW wind and solar power
• can be improved both economically as well as ecologically by a renewable
load management penetration of 10% ("RLM").
The total production cost can be reduced by up to 2.7 B€/year or 4.80 €/MWh in
2025 while improving mitigation of GHG emissions from 31.2% to 34.7%
(compared to 1990).
Under the assumption that 200 GW wind and solar power are installed a
renewable load management penetration of 20% on the consumer's side can
even improve CO2 mitigation from 38.4% to 44.7% while saving up to 7.0
B€/year or 12.4 €/MWh.
 Different decision options and
technologies, for example, SNG, are
compared to each other
Capital and operating costs are
included in the simulations

Costs for fuel and emission permits are

taken into account using three scenarios

WDPV= 500 GW or
200€/MWh e/200 €/tCO2

BMWT 2011, SRU 2010, UBA 2010

CO2 mitigation and production costs
of eight portfolio concepts under four
price scenarios

Annual capital cost

is calculated as the
IV sum of interest and
repayment within a
III period of 30 years.

BMWT 2011, SRU 2010, UBA 2010

II
I

RLM RLM+
 Conclusions

• System Dynamics not been used for optimizing asset

structure of energy portfolio yet
• Load management produces significant potential savings
on production costs at acceptable reliability and helps to
achieve the political goal of reducing GHG emissions.
• Data analysis is especially important for forecasting and
controlling with SD and can be employed for balancing the
grid
• Further research with SD is planned to estimate if the
savings on the side of electricity production are sufficient
to be used as incentives to motivate consumers to
implement the necessary load management.
IT-Based Decision Support for Turning on Germany’s energy
transition

THE IMPACT OF THE NORD.LINK

COMPLEX DECISION SUPPORT
WITH SYSTEM DYNAMICS
 Nord.Link “contributes to further
integration of the North-European power
markets”.

 Connecting Germany to
Norway’s huge capacity of
pumped storage (~60 GW, ~80

Prognos 2012, Seidler 2012, Statnett 2012

TWh)
 Planned capacity: 1.4 GW
 Expected to be completed:
2018
 Costs estimated: 1.6 (1.5-2.0)
B€
 Further competing projects
existing/expected: NorGer,
connections via Denmark
 The research gap

The research question: Does Nord.Link also

contribute to mitigate emissions of the German
electricity industry? And if yes, how efficient is
such an approach compared to others?

• Although the issue of energy storage by water storage was discussed

already in details in the last century.
• This specific project “Nord.Link” is one of the latest transnational
electricity grid related projects in Germany and therefore it was not
scientifically discussed yet.
• In particular System Dynamics was used for the first time for
analyzing the economic and ecologic effects of this project.
 System dynamics modeling and
simulation as an IT-based decision
support for electricity supply

Forrester 1994, Andersen 1997, Ghaffarzadegan 2011

Based on our previous work in the area of emissions trading under
uncertainties
Intention: System Dynamics makes the modeling process transparent and
intuitive in approaching the politically active public
Dynamic behavior is key for understanding of electricity supply
Resource consumption and reliability of electricity supply over the course
of a year are simulated
Compared to other decision options and technologies, for example, more
WDPV or SNG (Synthetic Natural Gas)
 Dynamic behavior is key for
understanding of electricity supply
Installed WDPV
Total load power in capacity
Germany

BMWT 2011, EXX 2012, Das Gelbe Forum 2012

Net load covered by dispatchable
(mainly thermal) power plans

Intermittent WDPV power. Average

output: 15.5% of installed capacity

 Increasing probability of WDPV power excess with

increasing installed WDPV capacity in the future; A
positive impact on emissions mitigation by storage
systems is then expected
 A high dispatchable power capacity is necessary for a high
reliability of electricity supply
Objective of this model

• Our system dynamics model describing the

dynamic behavior of an electricity supply system
• One possible pathway to facilitate a renewable-
centered electricity supply is thus to establish a
storage subsystem.
A general issue:
• How to transfer the mainly wind power overflow
from Germany with the help of the new trans-
national cable to the Norwegian water reservoirs.
 A system dynamics model
describing the dynamic behavior of
an electricity supply system

The main objective: Dispatchable Storage

minimizing consumption power subsystem
of chemical fuel and the
total cost of production
while keeping a certain
reliability of electricity
supply

WDPV power
with time profile

Arto 2012
Total load with
time profile
Answering the research
question:

“Does Nord.Link also contribute to mitigate emissions

of the German electricity industry? And if yes, how
efficient is such an approach compared to others?
 No contribution below 100 GW installed WDPV
 Each Nord.Link causes additional TPC of
approx.1€/MWh …
 0.8 – 2.3% mitigation pro €/MWh can be achieved by
WDPV
 … and 0.3% mitigation at 200 GW WDPV
The main result: Nord.Link has a rather
low contribution to emissions mitigation
compared to WDPV
THE PEAK –
0.8 – 2.3% mitigation
100% No contribution 210
pro €/MWh can be
below 100 GW
95% achieved
195 by WDPV
installed WDPV
90% 180

Total Production Cost (€/MWh)

CO2 Emission (1990=100%)

85% 165

80% 150
Each Nord.Link causes
75% 135
additional TPC of
approx.1€/MWh …
70% 120

65% 105

60% 90

55% 75

50% 60
20 40 60 80 100 120 140 160 180 200 220
… and only 0.3%
Installed Wind and Photovoltaic Electricity Generation Capacity (GW)
mitigation at 200
Emission 0 x NL Emission 1 x NL Emission 4 x NL Emission 16 x NL
TPC 0 x NL (Low Price) TPC 1 x NL TPC 4 x NL TPC 16 x NL GW WDPV
TPC 0 x NL (High Price) TPC 1 x NL TPC 4 x NL TPC 16 x NL
 Conclusion (1)

System Dynamics modeling helps to understand the fact that:

 Nord.Link contributes more to reliability
 but not strongly to economic and ecologic efficiency
 Nord.Link can be regarded as a starting point for further investments
 It strengthens the security of energy supply

Water reservoirs in Norway seem to be a gigantic “battery for the future”

 the low interconnector’s transmission capacity seems to be the
bottleneck
 severe costs for the construction of this electricity interconnector
 Conclusion (2)

„It enables governments and utilities to evaluate the long-term

consequences of the market structure and regulatory framework and to
focus on problems in the transition.” (Bunn: 1997, p.287)

Our System Dynamics modeling also enables the policy

maker to communicate this part of the German energy
transition process in a more understandable and objective
way to the public.

In our upcoming research we try to analyze the potential of other

technologies and evaluate their future importance for the German
energy transition.
DATA ANALYSIS FOR
FORECASTING AND
CONTROLLING WITH SD
 Data analysis for
forecasting and controlling

From the operational point of view:

 Transmission System Operators have to
master complexity and uncertainties of
electricity consumption in real-time grid
operation
 High uncertainties caused by the weather
dependent renewable electricity
production
 Forecasts are made for the consumption
as well as for the renewable electricity
production.
 If an electricity provider ordered too much
on the future market (it went long) these
quantities can be sold on the spot market
and, vice versa.
 Data analysis for
forecasting and controlling

Need for short-term trading increases with declining forecast quality= exposes the trader to a
higher price risk
TSOs have to ensure the stability of the grid
Actually regulating power is reserved for failure of power plants, transformers or power
lines to ensure n-1 criteria
If the forecast is not accurate regulating power is also used to correct these inaccuracies for
which it is not primarily intended
Example: December 2011, Western Europe
The market was obviously long. This led to the situation that the costs for negative control
energy (reducing production) per a specific quarter hour raised up to more than 500
€/MWh.
This figure shows the prices per quarter hour in December 2011.
The polynomial dashed trend line shows the balanced amount of regulating power of the
considered month which was negative in the most time.
Long positions of the providers' portfolios led
to high rates for regulating power

amprion.net 2012]
 Data analysis for forecasting
and controlling (cont.)

• To improve the prognosis for the model in 2025 it is

necessary to make analysis to the realistic profiles to get
information about dependencies to weekdays, relevant
events like holidays, political decisions, etc. .
ENERGY FORESIGHT AND THE
CHANCES FOR SD
Typical objectives of
Foresight
Energy foresight initiatives

Atomic energy sector:

• envisioning to define the most promising technologies
• their R&D needs to achieve deployment within the next 30 years

In 2000: The Generation IV International Forum

identify, evaluate and develop new systems of nuclear energy with the
possibility of licensing, constructing and generating electricity at
competitive prices while complying satisfactorily to nuclear safety,
minimization of waste generation, proliferation resistance and public
acceptance requirements.

In 2001: International Atomic Energy Agency

INPRO (International Project on Innovative Nuclear Reactors and Fuel Cycle).
• to assure that nuclear technology is permanently available in order to help the countries
all over the world during the 21st century to meet its energy needs contributing to their
sustainable developments.
Technology forecasting

• Began in 1950s in the United States defence sector

• Developing some of the principal tools of technology forecasting
• Delphi questionnaire survey and scenario analysis
• Large forecasting during the 1960s by the United States Navy and
United States Air Force.
• Later technology forecasting started also in the energy sector

Term “foresight” comes from Japan

• In 1970 Japan launched a 30-year national forecast exercise on the
future of S&T.
Aim was:
• giving advises for both public and private decision-makers
• by a broader direction-setting based on deep analysis of long-term
trends.
Foresight approach with
System Dynamics

Identification of Change Drivers

environmental scanning helps to identify change drivers
Strategic Network Modelling
to deduce key factors with important impact
Dynamic System Analysis
evaluate the impact of the key factors
Scenario Development
to foster double-loop-learning
Network Scenario Maps
System dynamics models serve as a strong basis for
communication
 Energy foresight

10 billion people by
2050 with energy
needs which come
mainly from fossil
fuels

Challenges facing
the global economic
and trade systems:
• Demographic composition of the population
strongly influences the consumption trend of
energy
• Hydrocarbon supply can not be replaced with
renewable energy  there will be up to 9
trillion barrels of oil-equivalent reserves
• Climate change is a human-induced factor
 more serious issue than simply
securing oil and gas
Applied System Dynamics projects
CONCLUSION
Disadvantages of System
Dynamics

 System thinking is necessary/ essential

 Only a limited number of parameters are possible to
include (see example: larger models have to be split into
sub-models)
 Validating a System Dynamics model is often difficult
 Some scientist say that the System Dynamics method is
old-fashion
Advantages of System
Dynamics

+ Not high mathematics necessary

+ Attractive visualisation of complex systems
+ Possible to communicate complex systems with the help of
System Dynamics model to the decision maker
+ More and more consultants use System Dynamics
+ System Dynamics strengthens the common understanding
of an issue
+ Network thinking process (feedback-loops)
Key aspects of System
Dynamics
 Thank you for your attention!

„System Dynamics modelling can be regarded as a powerful modelling

tool for creating shared understanding and managing both-
complexity and uncertainty“

Jay W. Forrester
Annex A

SYSTEM DYNAMICS
LITERATURE
 COMTESSA and SD

Bo Hu, Armin Leopold, and Stefan Pickl: IT-Based Decision Support for Turning on Germany’s energy
transition The Impact of the Nord.Link: Complex Decision Support with System Dynamics. Operations
Research Proceedings 2012 - Selected Papers of the Annual International Conference of the German
Operations Research Society, to be published
B. Hu, A. Leopold, S. Pickl: Transition Towards Renewable Energy Supply—A System Dynamics Approach.
J. Crespo Cuaresma et al. (eds.): Green Growth and Sustainable Development, Dynamic Modeling and
Econometrics in Economics and Finance 14, 217-226, Springer, Berlin Heidelberg, 2013
Klaus Arto, Irina Dolgopolova, Bo Hu, Armin Leopold, Stefan Pickl: Germany's Electricity Industry in 2025:
Evaluation of Portfolio Concepts. The 2012 International Conference of the System Dynamics Society,
Proceedings, St. Gallen, 22.-25.07.2012
Bo Hu, Thomas Kuhn, Armin Leopold, Stefan Pickl, Jörg Wellbrink: Operation Contracting for Training
Facilities - a System Dynamics Modeling Approach. Ninth Annual Acquisition Research Symposium,
Proceedings, Vol. II: 381-389, Naval Postgraduate School, Monterey CA, 30 April 2012
Bo Hu, Armin Leopold, Stefan Pickl: Why Do We Hesitate to Take Action Toward Implementation of
Sustainable Technology? GOR-Workshop "Simulation und Optimierung komplexer Systeme - Fuzzy
Systeme, Neuronale Netze und Künstliche Intelligenz, München, 23.-24.02.2012
Bo Hu, Armin Leopold, Stefan Pickl, Hans-Rolf Vetter: Is Flexibility in a Professional Career Being
Rewarded? An Agent-Based Simulation. Proceedings of the 4th Symposium on Multiagent Systems,
Robotics and Cybernetics: Theory and Practice, 13-18, Pub. IIAS, Canada, 2011
Bo Hu, Armin Leopold: Web-based Participatory System Dynamics Modelling – Concept and Prototype
Development. The 2011 International Conference of the System Dynamics Society, Proceedings,
Washington DC, July 24 – 28, 2011
 Introduction to
System Dynamics

Sterman, John (2000): Business Dynamics:

Systems Thinking and Modeling for a
Complex World. Boston; London:
Irwin/McGraw-Hill.
Warren, Kim (2002): Competitive Strategy
Dynamics. Chichester: Wiley.
Morecroft, John (2007): Strategic Modelling
and Busines Dynamics. Chichester: Wiley
 Scientific Journals

System Dynamics Review

Publisher: John Wiley & Sons
Editor: Brian Dangerfield, University of
Salford
Systems Research and Behavioral Science
Publisher: John Wiley & Sons
Editor: Michael C. Jackson, Hull University
Business School
 Some classics

Forrester, Jay (1958): Industrial dynamics: a major breakthrough for decision makers, In: Harvard
Business Review 36. Jg., Nr. 4, S. 37.
Forrester, Jay (1961): Industrial Dynamics. Portland, Or.: Productivity Press.
Forrester, Jay (1969): Urban Dynamics. Cambridge, Mass.,: M.I.T. Press.
Meadows, Donella (1974): The Limits to Growth. A Report for the Club of Rome’s Project on the
Predicament of Mankind. 2d ed. Herausgegeben von Club of Rome und Potomac Associates.
New York: Universe Books.
Meadows, Donella/Meadows, Dennis/Randers, Jørgen (1992): Beyond the Limits. Global Collapse
or a Sustainable Future. London: Earthscan Publications.
Meadows, Donella/Meadows, Dennis/Randers, Jørgen (2006): Grenzen des Wachstums – Das 30-
Jahre-Update. Signal zum Kurswechsel; Stuttgart: Hirzel
Milling, Peter (1984): Leitmotive des System-Dynamics-Ansatzes, In:
Wirtschaftswissenschaftliches Studium 13. Jg., Nr. Heft 10, S. 507–513.
Richardson, George/Pugh, Alexander (1981): Introduction to System Dynamics Modeling with
DYNAMO. Cambridge, Mass.; London: MIT Press.
Sterman, John (1989): Misperceptions of Feedback in Dynamic Decision Making, In:
Organizational Behavior & Human Decision Processes 43. Jg., Nr. 3, S. 301–335.