.Energy Systems Series Editor: Panos M. Pardalos, University of Florida, USA .Josef Kallrath • doc

Energy Systems
Series Editor:
Panos M. Pardalos, University of Florida, USA
Josef Kallrath
•
Panos M. Pardalos
Steffen Rebennack
•
Max Scheidt
Editors
Optimization in the
Energy Industry
ABC
Editors
Prof. Dr. Josef Kallrath
Am Mahlstein 8
67273 Weisenheim
Germany

Steffen Rebennack
University of Florida
Department of Industrial & Systems
Engineering
303 Weil Hall, P.O.Box 116595
Gainesville FL 32611-6595
USA
steffen@ufl.edu
Prof. Panos M. Pardalos
University of Florida
Department of Industrial & Systems

Engineering
303 Weil Hall, P.O.Box 116595
Gainesville FL 32611-6595
USA
pardalos@ufl.edu
Dr. Max Scheidt
ProCom GmbH
Luisenstraße 41
52070 Aachen
Germany

Cover art entitled “WATER AND SOIL, THE PRIMEVAL ELEMENTS TO CREATION” is designed by
Elias Tyligadas.
ISBN: 978-3-540-88964-9 e-ISBN: 978-3-540-88965-6
Library of Congress Control Number: 2008942066
c
 2009 Springer-Verlag Berlin Heidelberg
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication
or parts thereof is permitted only under the provisions of the German Copyright Law of September 9,
1965, in its current version, and permission for use must always be obtained from Springer. Violations are
liable to prosecution under the German Copyright Law.
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws
and regulations and therefore free for general use.
Cover design: WMX Design GmbH, Heidelberg
Printed on acid-free paper
springer.com
To our families.

Preface
Today, the optimization of production planning processes by means of IT and
quantitative methods is a de-facto standard in the energy industry. Franch et
al. in Chapter 1 and Ikenouye in Chapter 2 give an introduction, overview, and
reasons for this. Furthermore, the energy problem now is not only a challenging
one but also one of the most important issues in the world from the political
and economical points of view. In every country, the government is faced with
the problem of how to adopt the system of ‘Cap and Trade.’ Especially energy
consuming industries, such as steel, power, oil and chemicals, are seriously
confronted with this problem.
VIII Preface
This is also the reason why the German Operations Research Society
(GOR) and one of its working groups, held a symposium with the title
“Stochastic Optimization in the Energy Industry.” During the 78th meeting
of the GOR working group “Praxis der Mathematischen Optimierung/Real
World Optimization” in Aachen at Procom GmbH on April 21/22, 2007, the
speakers with an application background explained their requirements for
stochastic optimization solutions based on practical experiences. The speakers
from the research side and the software system suppliers examined different
aspects of the whole subject – from the integration of wind energy, the chain
of errors in nuclear power plants and the scheduling of hydroelectric power
stations, and the risk assessment in trading activities to the various software
systems which support stochastic optimization methods.
The symposium offered an interesting overview which reflected the re-
quirements, possibilities and restrictions of “Stochastic Optimization in the
Energy Industry.” As the speakers came from all over the world (Brazil, USA,
The Netherlands, Norway, Switzerland and Germany) it was also an ideal
platform to exchange ideas across countries in the energy sector and beyond.
This book is partly based on the contributions the speakers made to the
workshop, but also contains chapters provided by other colleagues. The chap-

ters of the first part of the book give a general introduction to the field.
The second part contains deterministic models, while the third part provides
methods and applications involving uncertain data. The fourth part includes
contributions which focus on pricing.
After opening the European markets for electricity, the energy supply
companies expect both new risks and new chances. The ex-ante uncertain
market price increasingly determines the amount of their self-generated en-
ergy. While the classic unit scheduling objective is the cost-optimal production
plan, in liberalized energy markets a holistic examination of the power-station
and trading portfolio results in multiple chances to improve the profit situa-
tion.
Borisovksy et al. in Chapter 3 consider the problem of constructing trading
hubs in the structure of electricity wholesale markets. The nodes of a trading
hub are used to calculate a reference price that can be employed by the mar-
ket participants for different types of hedging. The need for such a reference
price is the considerable variability of energy prices at different nodes of the
electricity grid at different periods of time. Hub construction is viewed as a
mathematical programming problem.
These changes in electric network infrastructure and government policies
have created opportunities for the employment of distributed generation to
achieve a variety of benefits. Fidalgo et al. in Chapter 4 propose a decisions
support system to assess some of the technical benefits, namely, voltage profile
improvement, power loss reduction, and network capacity investment deferral,
brought through branch congestion reduction.
Bulatov discusses in Chapter 5 three special energy problems which can
be solved in polynomial time, exploiting their convexity. These problems are:
Preface IX
Minimal shutdown during power shortages in a power supply system, search
for optimal states in thermodynamic systems and optimal allocation of water
resources.

The book covers several optimization issues for power plants. Kusiak &
Song discuss in Chapter 6 the improvement of combustion processes with
application in boiler performance. The modeling of nonlinear processes in
nuclear power plant cores is discussed by Yatsenko et al. in Chapter 7. Design
optimization of polygeneration energy systems are modeled via mixed-integer
nonlinear programs by Liu et al. in Chapter 8 and also by J¨udes et al. in
Chapter 9. Mathematical modeling of biomass-based power plants are dis-
cussed by Bruglieri & Liberti in Chapter 10 and by Lai et al. in Chapter 11.
Electric power systems are considered by Woolley et al. in Chapter 12 and by
Chiang et al. in Chapter 13.
Software systems geared to today’s market requirements are able to repre-
sent the whole portfolio consisting of both generating and trading components.
This increases the transparency of the whole planning process. At the same
time, risks become apparent and have to be supervised and validated.
Due to increased cost pressure on power generation and trading companies,
caused by operating under market conditions, a cost efficient management of
the risks becomes more important. As a result of the liberalization of the
markets for electrical energy, companies are exposed to higher uncertainties
in power generation and trading planning, e.g., the volatility of the prices
for electrical energy and for primary energies, especially natural gas. Risks
and uncertainties are normally not yet explicitly considered by today’s com-
mercial optimization systems. In a deterministic approach, all information is
considered to be certain. Actually, there are relative uncertainties in different
exogenous factors, e.g., the prices in spot and futures trading, in load forecast,
the expected input of wind energy, the water supply and the power stations’
availability. However, in the academic world there are a lot of activities on
that topic. The contributions of Eichhorn et al. in Chapter 14, Epe et al.
in Chapter 15, Heitmann & Hamacher in Chapter 16, Bl¨asig & Haubrich in
Chapter 17, Radziukynas & Radziukyniene in Chapter 18, and Weber et al. in
Chapter 19 are all related to risk minimization and stochastic programming.

To derive robust decisions, stochastic optimization operations are suitable
for mid- and long-term calculations although they generally take a long time
for the computing work. In the electricity industry the observed increases
of electricity price dynamics combined with the characteristic periodicity of
related decision processes have motivated the use of multistage stochastic pro-
gramming in recent years to provide flexible models for practical applications
in the sector. Especially in power generation and trading, the planning process
must obey highly complex interrelations between manifold influences. They
range from short term price fluctuations as observed in spot markets to long
term changes of fundamental influences. Not only changes in the electric sup-
ply system itself must be considered, but also the related availability and costs
of required fuels. This is outlined by Frauendorfer & G¨ussow in Chapter 20.
XPreface
Another example is the valuation of electricity swing option by Steinbach &
Vollbrecht in Chapter 21. The optimization and subsequent hedging of reser-
voir discharges for a hydropower producer is discussed by Fleten & Wallace
in Chapter 22.
This book can be read linearly, from beginning to end. This will give a good
overview of how rich the world of energy is for mathematical optimization and
especially optimization under uncertainty. The book covers a wide range of
techniques and algorithms. Those readers already familiar with the topic are
encouraged to visit directly the topics of their interest but we are sure they
will also detect many facets of a field which will have a large impact on the
future of mankind.
We would like to take this opportunity to thank the authors for their
contributions, the referees, and the publisher for helping to produce this book.
June 2008 Josef Kallrath
Panos M. Pardalos
Steffen Rebennack
Max Scheidt

Contents
Conventions and Abbreviations 1
Part I Challenges and Perspectives of Optimization
in the Energy Industry
1 Current and Future Challenges for Production Planning
Systems
Torben Franch, Max Scheidt and G¨unter Stock 5
2 The Earth Warming Problem: Practical Modeling
in Industrial Enterprises
Susumu Ikenouye 19
Part II Deterministic Methods
3 Trading Hubs Construction for Electricity Markets
Pavel A. Borisovsky, Anton V. Eremeev, Egor B. Grinkevich,
Sergey A. Klokov and Andrey V. Vinnikov 29
4 A Decision Support System to Analyze the Influence
of Distributed Generation in Energy Distribution Networks
J.N. Fidalgo, Dalila B.M.M. Fontes and Susana Silva 59
5 New Effective Methods of Mathematical Programming
and Their Applications to Energy Problems
Valerian P. Bulatov 79
6 Improving Combustion Performance by Online Learning
AndrewKusiakandZheSong 131
XII Contents
7 Critical States of Nuclear Power Plant Reactors
and Bilinear Modeling
Vitaliy A. Yatsenko, Panos M. Pardalos and Steffen Rebennack 149
8 Mixed-Integer Optimization for Polygeneration Energy
Systems Design
Pei Liu and Efstratios N. Pistikopoulos 167
9 Optimization of the Design and Partial-Load Operation

of Power Plants Using Mixed-Integer Nonlinear Programming
Marc J¨udes, Stefan Vigerske and George Tsatsaronis 193
10 Optimally Running a Biomass-Based Energy Production
Process
Maurizio Bruglieri and Leo Liberti 221
11 Mathematical Modeling of Batch, Single Stage, Leach Bed
Anaerobic Digestion of Organic Fraction of Municipal Solid
Waste
Takwai E. Lai, Abhay K. Koppar, Pratap C. Pullammanappallil
and William P. Clarke 233
12 Spatially Differentiated Trade of Permits
for Multipollutant Electric Power Supply Chains
Trisha Woolley, Anna Nagurney and John Stranlund 277
13 Applications of TRUST-TECH Methodology in Optimal
Power Flow of Power Systems
Hsiao-Dong Chiang, Bin Wang and Quan-Yuan Jiang 297
Part III Stochastic Programming: Methods and Applications
14 Scenario Tree Approximation and Risk Aversion Strategies
for Stochastic Optimization of Electricity Production and
Trading
Andreas Eichhorn, Holger Heitsch and Werner R¨omisch 321
15 Optimization of Dispersed Energy Supply – Stochastic
Programming with Recombining Scenario Trees
Alexa Epe, Christian K¨uchler, Werner R¨omisch, Stefan Vigerske,
Hermann-Josef Wagner, Christoph Weber and Oliver Woll 347
16 Stochastic Model of the German Electricity System
Nina Heitmann and Thomas Hamacher 365
Contents XIII
17 Optimization of Risk Management Problems in Generation
and Trading Planning

Boris Blaesig and Hans-J¨urgen Haubrich 387
18 Optimization Methods Application to Optimal Power
Flow in Electric Power Systems
Virginijus Radziukynas and Ingrida Radziukyniene 409
19 WILMAR: A Stochastic Programming Tool to Analyze
the Large-Scale Integration of Wind Energy
Christoph Weber, Peter Meibom, R¨udiger Barth and Heike Brand 437
Part IV Stochastic Programming in Pricing
20 Clean Valuation with Regard to EU Emission Trading
Karl Frauendorfer and Jens G¨ussow 461
21 Efficient Stochastic Programming Techniques
for Electricity Swing Options
Marc C. Steinbach and Hans-Joachim Vollbrecht 485
22 Delta-Hedging a Hydropower Plant Using Stochastic
Programming
Stein-Erik Fleten and Stein W. Wallace 507
Index 525
List of Contributors
R¨udiger Barth
Institute for Energy Economics
and the Rational Use of Energy
University Stuttgart
70565 Stuttgart
Germany
ruediger.barth@ier.
uni-stuttgart.de
Boris Blaesig
Institute of Power Systems
and Power Economics
Schinkelstrasse 6, 52056 Aachen

Germany

Pavel A. Borisovsky
Omsk State Technical University
11 Prospect Mira, 644050 Omsk
Russia

Heike Brand
Institute for Energy Economics
and the Rational Use of Energy
University Stuttgart
70565 Stuttgart
Germany
heike.brand@ier.
uni-stuttgart.de
Maurizio Bruglieri
INDACO, Politecnico di Milano
Via Durando 38/a, 20158 Milano
Italy

Valerian P. Bulatov
Melentiev Energy Systems
Institute of SB RAS 130
Lermontov Strasse
Irkutsk, 664033
Russia

Hsiao-Dong Chiang
School of Electrical and Computer
Engineering

Cornell University, Ithaca
NY 14853
USA

William P. Clarke
School of Engineering
The University of Queensland
Brisbane, Qld 4067
Australia

XVI List of Contributors
Andreas Eichhorn
Humboldt-University Berlin
Department of Mathematics
10099 Berlin
Germany
/>~
eichhorn

Alexa Epe
Ruhr-Universit¨at Bochum
Universit¨atsstraße 150
44801 Bochum
Germany

Anton V. Eremeev
Omsk Branch of Sobolev
Institute of Mathematics
SB RAS
13 Pevtsov St., 644099 Omsk

Russia

J.N. Fidalgo
INESC Porto and Faculdade
de Engenharia da
Universidade do Porto
Rua Dr. Roberto Frias
4200-465 Porto
Portugal

Stein-Erik Fleten
Norwegian University of Science
and Technology
Department of Industrial
Economics and Technology
Management, Alfred Getz v. 1
7491 Trondheim
Norway

Dalila B.M.M. Fontes
LIAAD - INESC Porto L.A.
and Faculdade de Economia da
Universidade do Porto
Rua Dr. Roberto Frias
4200-464 Porto
Portugal

Torb en Franch
ProCom GmbH
Luisenstr. 41, 52070 Aachen

Germany


Karl Frauendorfer
Institute for Operations Research
and Computational Finance
University of St. Gallen
Switzerland

Egor B. Grinkevich
Administrator of Trade System
for United Energy System of Russia
12 Krasnopresnenskaya
Naberezhnaya, 123610 Moscow
Russia

Jens G¨ussow
Institute for Operations Research
and Computational Finance
University of St. Gallen
Switzerland

Thomas Hamacher
Max-Planck-Institut f¨ur
Plasmaphysik, Gruppe f¨ur
Energie und Systemstudien
Boltzmannstrasse 2 Garching
Germany

List of Contributors XVII

Hans-J¨urgen Haubrich
Institute of Power Systems
and Power Economics
Schinkelstrasse 6, 52056 Aachen
Germany

Nina Heitmann
Max-Planck-Institut f¨ur
Plasmaphysik, Gruppe f¨ur
Energie und Systemstudien
Boltzmannstrasse 2
85748 Garching
Germany

Holger Heitsch
Humboldt-University Berlin
Department of Mathematics
10099 Berlin
Germany
/>~
heitsch

Susumu Ikenouye
Ike Ltd.
112-0012, 6-12-2-304, Otsuka
Bunkyoku, Tokyo
Japan

Quan-Yuan Jiang
School of Electrical Engineering

Zhejiang University, Hangzhou
P.R. China

Marc J¨udes
Institute for Energy Engineering,
Technische Universit¨at Berlin
Marchstrasse 18, 10587 Berlin
Germany

Sergey A. Klokov
Omsk Branch of Sobolev Institute
of Mathematics SB RAS
13 Pevtsov St., 644099 Omsk
Russia

Abhay K. Koppar
Department of Agricultural
and Biological Engineering
University of Florida
Gainesville, FL 32607
USA

Christian K¨uchler
Humboldt–Universit¨at zu Berlin
Unter den Linden 6, 10099 Berlin
Germany

Andrew Kusiak
The University of Iowa
Department of Mechanical

and Industrial Engineering
3131 Seamans Center, Iowa City
IA 52242-1527
USA

Takwai E. Lai
School of Engineering
The University of Queensland
Brisbane, Qld 4067
Australia

Leo Liberti
LIX, Ecole Polytechnique
F-91128 Palaiseau
France

XVIII List of Contributors
Pei Liu
Centre for Process Systems
Engineering
Department of Chemical
Engineering
Imperial College London
London SW7 2AZ
UK

Peter Meibom
Risø National Laboratory
for Sustainable Energy
Technical University of Denmark

Roskilde
Denmark

Anna Nagurney
Department of Finance
and Operations Management
Isenberg School of Management
University of Massachusetts
Amherst, MA, 01003
USA

Panos M. Pardalos
Department of Industrial
and Systems Engineering
Center for Applied Optimization
University of Florida, Gainesville
FL 32611, USA

Efstratios N. Pistikopoulos
Centre for Process Systems
Engineering
Department of Chemical
Engineering
Imperial College London, London
SW7 2AZ
UK

Pratap C. Pullammanappallil
Department of Agricultural
and Biological Engineering

University of Florida
Gainesville, FL 32607
USA

Virginijus Radziukynas
Lithuanian Energy Institute
Laboratory of Systems Control
and Automation
Lithuania

Ingrida Radziukyniene
Vytautas Magnus University
Faculty of Informatics
Lithuania

Steffen Rebennack
Department of Industrial
and Systems Engineering
Center for Applied Optimization
University of Florida, Gainesville
FL 32611
USA

Werner R¨omisch
Humboldt-University Berlin
Department of Mathematics
10099 Berlin
Germany
/>~
romisch

Max Scheidt
ProCom GmbH
Luisenstrasse 41, 52070 Aachen
Germany


List of Contributors XIX
Susana Silva
ALERT - Life Sciences Computing
S.A.
Rua Antnio Bessa Leite
1430, 2
o
4150-074 Porto
Portugal

Zhe Song
The University of Iowa
Department of Mechanical
and Industrial Engineering
3131 Seamans Center, Iowa City
IA 52242-1527
USA

Marc C. Steinbach
Leibniz Universit¨at Hannover
IfAM Welfengarten 1
30167 Hannover
Germany

www.ifam.uni-hannover.de/
~steinbach

G¨unter Stock
Meischenfeld 11, 52076 Aachen
Germany

John Stranlund
Department of Resource Economics
College of Natural Resources
and the Environment
University of Massachusetts
Amherst, MA 01003
USA

George Tsatsaronis
Institute for Energy Engineering
Technische Universit¨at Berlin
Marchstrasse 18, 10587 Berlin
Germany

Stefan Vigerske
Humboldt–Universit¨at zu Berlin
Unter den Linden 6, 10099 Berlin
Germany
/>~
stefan

Andrey V. Vinnikov
Administrator of Trade System

for United Energy System
of Russia Joint Institute
for Nuclear Research
12 Krasnopresnenskaya
Naberezhnaya, 123610 Moscow
Russia

Hans-Joachim Vollbrecht
Fachhochschule Vorarlberg
FZ PPE S¨agerstrasse 4
6850 Dornbirn
Austria
www.staff.fh-vorarlberg.
ac.at/hvhans-joachim.voll

Hermann-Josef Wagner
Ruhr-Universit¨at Bochum
Universit¨atsstraße 150
44801 Bochum
Germany

Stein W. Wallace
Chinese University
of Hong Kong
Shatin NT, Hong Kong
China and Molde
University College
P.O. Box 2110
6402 Molde
Norway

XX List of Contributors
Bin Wang
School of Electrical
and Computer Engineering
Cornell University
Ithaca, NY 14853
USA

Christoph Weber
Management Sciences
and Energy
Economics Universit¨at
Duisburg-Essen
Universit¨atsstraße
2, 45141 Essen
Germany
Christoph
Weber@uni-
duisburg-essen.de
Oliver Woll
Universit¨at Duisburg-Essen
Universit¨atsstraße 2, 45141 Essen
Germany
Oliver.Woll@uni-duisburg-
essen.de
Trisha Woolley
Department of Finance and
Operations Management
Isenberg School of Management

University of Massachusetts
Amherst, MA, 01003
USA

Vitaliy A. Yatsenko
Space Research Institute NASU
and NSAU 40 Prospect Academica
Glushkova 03680 Kyiv
Ukraine

Conventions and Abbreviations
The following table contains in alphabetic order abbreviations used in at least
two chapters of the book.
Abbreviation Meaning
cf. Confer (compare)
CHP Combined heat and power
CVaR Conditional value-at-risk
e.g. Exempli gratia (for example)
EEX European energy exchange
GHG GreenHouse gas
HRSG Heat recovery steam generator
i.e. Id est (that is)
ISO Independent system operator
LP Linear programming
MIP Mixed integer (linear) programming
MINLP Mixed integer nonlinear programming
NLP Nonlinear programming
OPF Optimal power flow
PSO Particle swarm optimization
s.t. Subject to

SLP Successive linear programming
SQP Successive quadratic programming
1
Current and Future Challenges for Production
Planning Systems
Torben Franch, Max Scheidt, and G¨unter Stock
Summary. This article elaborates on the coming challenges production planning
departments in utilities are facing in the near and remote future. Firstly, we will
motivate the complexity of production planning, followed by a general solution ap-
proach to this task. The development of a new generation of energy management
tools seems necessary to fulfill the need to handle uncertainty and eventually cover
stochastic processes in energy planning. These new energy management systems
have to include complex workflows and different methods and tools into the plan-
ning process.
Key words: Energy management, Uncertainty in energy planning
1.1 Introduction
Energy planning can be complicated. Due to its techno-economic nature it
was already complex in monopolistic times and has gone from ‘complex’ to
‘very complex’ thereafter.
First of all, it is important to explain what production planning in the en-
ergy industry or energy planning, respectively, means. Production planning is
the commercial and technical organization that uses power plants to generate
income. It is the key organizational function that translates production capac-
ity into commercial value. In a nutshell, this means that without production
planning, power plants are not generating any income.
The objective for production planning is clearly to maximize the profits
that can be created by running power plants. As power plants inherently
produce more than electricity, the maximization of profits is typically subject
to a number of restrictions. These restrictions are particularly heat supply
but also technical restrictions and ancillary service commitments. Experience

shows that production planning becomes very complex as soon as power plants
produce more than just straight power.
6T.Franchetal.
1.2 Production Planning – History and Present
A good example for how complex production planning really is and what
significant commercial impact it can have is depicted in Fig. 1.1. The pro-
ducer’s every day production capacity of his power plants is offered to the
Nord Pool exchange. When it is profitable, production is sold. The set of
assets consists of a number of smaller and larger production units using dif-
ferent fuels. Furthermore, heat is supplied to a stretched-out heat grid and
different steam grids. This example of production planning shows very clearly
that even small improvements in performance can have a significant impact on
results. Moreover, small planning mistakes can have very serious commercial
and operational consequences.
In Fig. 1.1 actual hourly production in December 2004 is depicted. At first
glance, it can be difficult to understand how this can be an optimal production
plan. However, there are some good explanations. The variation in production
is a function of many factors such as weekend stops, ancillary services delivery,
and commercial production. In the chart, one can see the ‘coal-minimum’ and
the ‘oil-minimum’ situations where reserves are delivered automatically and
manually. On closer examination, it is even possible to see that different on-
duty crews have different views of what is maximum and minimum production
capacity.
The deregulation of energy markets has had a very significant impact
on production planning: Firstly, the purpose of planning has changed from
minimizing cost of delivery to maximizing profits. Secondly, new markets
have emerged, like spot power, gas, and CO
2
. Thirdly, the roles of market
1-12-2004

2-12-2004
3-12-2004
4-12-2004
5-12-2004
6-12-2004
7-12-2004
8-12-2004
9-12-2004
10-12-2004
11-12-2004
12-12-2004
13-12-2004
14-12-2004
15-12-2004
16-12-2004
17-12-2004
18-12-2004
19-12-2004
20-12-2004
21-12-2004
22-12-2004
23-12-2004
24-12-2004
25-12-2004
26-12-2004
27-12-2004
28-12-2004
29-12-2004
30-12-2004
31-12-2004

Time (day)
0
50
100
Electrical production (MW)
150
200
250
300
Fig. 1.1. Production planning in practice
1 Current and Future Challenges for Production Planning Systems 7
participants have changed. Consequently, as a result of this, the production
planning workflow has changed as well.
In order to understand where production planning and production plan-
ning tools are today, it makes sense to look at the historical framework. The
European energy markets have been deregulated in the past 10 years and this
had a considerable impact on how energy companies behave in the market
and organize themselves, see [1]. Firstly, deregulation meant that the purpose
of an energy company changed. Today, companies very much strive to make
profits for their owners whereas prior to deregulation, the objective was to
minimize delivery costs to consumers. In the past, very often the result of
a year was decided when the annual budget was drawn up. Secondly, dereg-
ulation has opened new markets. Today, it is possible to trade spot power
and gas, imbalances and CO
2
emission rights – all products that were not
even known a few years ago. Lastly, deregulation changed the roles of market
participants. In some countries, this led to new players entering the markets,
yet in other countries, this resulted in the emergence of a few and very large
energy giants.

To illustrate how much all these factors have influenced production plan-
ning, taking a look at an illustration of production planning work processes
prior to deregulation makes sense.
Prior to deregulation, production planning consisted of the forecasting of
load and later the computation of the optimal production plan, see Fig. 1.2.
While this looks like a relatively simple task, it can be a difficult calcula-
tion, especially if the production system is complex. Previously, the focus of
attention was mostly on technical power plant availability and how to meet
production requirements. In those days fuel prices were relatively stable and
hence there was no need for daily calculations. Instead, calculations were made
weekly or even less frequently. For shorter periods, a prioritization of produc-
tion units was sufficient. Deregulation and the emergence of new markets
changed all this radically.
Fig. 1.2. Production planning before deregulation
8T.Franchetal.
Technical plant
availability
Optimisation
calculator
Sales strategy
Production
plan
Sales plan Market
Optimisation
calculator
Fuel prices
Heat load
forecast
CO2 prices
Power price

forecast
Reserves
commitments
Power load
forecast
Wind power
production
forecast
Fiscal regime
Fig. 1.3. Production planning work process today
Today, however, the amount of input data is not only much larger but
inputs are also much more volatile, see Fig. 1.3. This means that production
planners have to work very efficiently day in and day out to compile infor-
mation, do the necessary analysis and planning and then submit these to
the exchanges before noon. That means they have complex workflows, many
methods, lots of data and less time for it all. At the same time, the new
deregulated environment called for the development of new systems for ef-
fective data management and shorter calculation time for optimization. The
good news is that power load forecasting is no longer a task for production
planning. Today, this is the task of the retail manager. Furthermore, there are
now several new trading platforms, like exchanges, over the counter trading,
cross border trading and intraday trading. This is why sales strategies play
an important role. All in all, nowadays, production planning has very much
become a task of optimizing sales in an environment of volatile power and fuel
prices.
1.3 The Coming Challenge: Handling Uncertainty
“It’s hard to predict, especially the future”. This well-known saying attributed
to Winston Churchill proves to be valid in production planning as well. In fact,
production planning is very much exposed to risks and uncertainties, although
not much attention has been paid to this aspect for quite some time. One of

the most volatile commodities in the world is power, even more volatile than
fuel oil prices. As a comparison, in the period April 2006-March 2007, the fuel
1 Current and Future Challenges for Production Planning Systems 9
oil price has varied from US $50 to 75 per barrel, while the Nord Pool price
showed much larger variations and German EEX prices have been even more
volatile. This makes it very difficult to predict power prices a day ahead.
Fig. 1.4 depicts the base load prices for 2006 in the Nord Pool area DK2.
But it is even harder to predict hourly prices and profiles, which is shown
in Fig. 1.5 for Nord Pool DK2.
While hourly spot prices are so difficult to predict, they are one of the
most important parameters in a production plan. Wrong forecasts of spot
prices can lead to wrong decisions. If you base heat planning on a wrong
spot price profile, you could end up with power production in low price hours
and heat production in high price hours. Generally, you have to optimize the
combined heat, steam and power production portfolio regarding your forecasts
Day
EUR/MWh
0
10
20
30
40
50
60
70
80
90
100
1/1/06
1/2/06

1/3/06
1/4/06
1/5/06
1/6/06
1/7/06
1/8/06
1/9/06
1/10/06
1/11/06
1/12/06
Fig. 1.4. Spot prices (Nord Pool DK2)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
EUR/MWh
Mon 28/8/06
Mon 21/8/06
Mon 14/8/06
Mon 7/8/06
0
10
20
30
40
50
60
70
80
90
100
Fig. 1.5. Hourly spot prices (Nord Pool DK2) on 4 Mondays in August 2006

10 T. Franch et al.
of district heating, steam production and spot prices. This is naturally always
prone to errors resulting in imbalances between your day-ahead planning and
the required and delivered customer load.
While it is yet impossible to forecast exact values, in fact sometimes it is
possible to forecast the direction of imbalances. One example can be found in
the field of wind power forecasting.
The graph in Fig. 1.6 shows the forecasting of wind power production at
a Baltic Sea wind farm and the actual production curve. It shows that the
prediction for wind power production a day-ahead is very accurate.
However, the problem is that predictions are not always as good. As can
be seen in Fig. 1.7, which shows said wind farm on another day. This time,
the forecast results in notable imbalances which are priced with different im-
balance costs for each hour. The graph illustrates also the commercial risk
attached with such a wrong prediction regarding the exact time of the wind
load curve.
Forecasts of power prices and wind power production are by far not the
only sources of uncertainty and of commercial risks. There is uncertainty in
heat load forecasts, fuel prices, unit failures and many more. Basically, uncer-
tainty cannot be avoided. Uncertainty about input parameters leads to im-
balances – and even wrong decisions. This is especially true for virtual power
plants, see [5]. Also, one can forecast some effects in a short time horizon. The
key to this problem is handling the risks effectively. This is important because
the commercial implications can be very substantial. So, how do you do pro-
duction planning under uncertainty? One approach is to ignore it, because
1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Hour
MWh/h
Imbalance
Forecast

Actual
−60
−40
−20
0
20
40
60
80
100
120
140
160
180
Fig. 1.6. Forecasting wind production and little imbalances