Towards an Integrated Sustainability Evaluation of Energy Scenarios

with Automated Information Exchange

Jan S

oren Schwarz

, Tobias Witt

, Astrid Nieße

, Jutta Geldermann

, Sebastian Lehnhoff

and Michael Sonnenschein

Department of Computer Science, University of Oldenburg, Oldenburg, Germany

Chair of Production and Logistics, University of Goettingen, Goettingen, Germany

OFFIS - Institute for Information Technology, Oldenburg, Germany

Keywords:

Co-simulation, Energy Scenarios, Information Model, Multi-Criteria Decision Making (MCDM), Ontology,

Scenario Planning, Story-and-Simulation (SAS), Sustainability Evaluation.

Abstract:

To reshape energy systems towards renewable energy resources, decision makers need to decide today on how

to make the transition. Energy scenarios are widely used to guide decision making in this context. While

considerable effort has been put into developing energy scenarios, researchers have pointed out three require-

ments for energy scenarios that are not fulﬁlled satisfactorily yet: The development and evaluation of energy

scenarios should (1) incorporate the concept of sustainability, (2) provide decision support in a transparent way

and (3) be replicable for other researchers. To meet these requirements, we combine different methodologi-

cal approaches: story-and-simulation (SAS) scenarios, multi-criteria decision-making (MCDM), information

modeling and co-simulation. We show in this paper how the combination of these methods can lead to an

integrated approach for sustainability evaluation of energy scenarios with automated information exchange.

Our approach consists of a sustainability evaluation process (SEP) and an information model for modeling de-

pendencies. The objectives are to guide decisions towards sustainable development of the energy sector and to

make the scenario and decision support processes more transparent for both decision makers and researchers.

1 INTRODUCTION

The intended phase-out from nuclear and fossil

power, and the transition to renewable energy re-

sources in Germany pose new challenges. With the

EU and the federal government of Germany having

set targets for reducing energy demand and green-

house gas emissions until 2030 and 2050 respectively

(European Commission, 2014; Deutscher Bundestag,

2014; BMWi, 2010), decision makers in politics need

to initiate and project the transition process today to

achieve these targets. The goal of this transition pro-

cess is to reshape the energy infrastructure and related

planning and operation processes while also consid-

ering sustainable development. Thus, an approach

to evaluate and compare future scenarios and corre-

sponding transition paths regarding their sustainabil-

ity characteristics is needed to provide guidance to the

politically intended transition process.

Long-term energy scenarios are used to guide de-

cision making in this context (Grunwald et al., 2016),

and researchers have already put considerable effort

into developing these scenarios. For example, the

German database ”Forschungsradar Energiewende”

lists more than 920 publications on energy transi-

tion research in Germany from 2011 to 2016. This

database also includes some publications on EU and

worldwide levels, which are also relevant for the Ger-

man debate, e.g. the World Energy Outlook (Inter-

national Energy Agency, 2016). Based on these en-

ergy scenarios, different transition paths can be distin-

guished for the energy transition. If decision makers

make use of energy scenarios to decide upon strate-

gies, how the set targets of the energy transition can

be achieved, the development and evaluation of en-

ergy scenarios should meet some basic requirements

(see section 2 for more details):

• A concept of sustainability should be deﬁned and

operationalized with relevant dimensions in the

evaluation process.

http://www.forschungsradar.de/studiendatenbank.html

188

Schwarz, J., Witt, T., Nieße, A., Geldermann, J., Lehnhoff, S. and Sonnenschein, M.

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange.

DOI: 10.5220/0006302101880199

In Proceedings of the 6th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2017), pages 188-199

ISBN: 978-989-758-241-7

• The decision support process should be transpar-

ent to allow decision makers to rely on the results.

• The scenario development and decision support

processes should be replicable to allow repeated

evaluation of energy scenarios.

In the article at hand, we will show an integrated

approach for sustainability evaluation of energy sce-

narios with automated information exchange meeting

these requirements. Our approach is based on the sus-

tainability evaluation process (SEP) developed in the

research project NEDS (Blank et al., 2016). The SEP

combines qualitative future scenarios with quantita-

tive simulation and multi-criteria decision-making.

To cope with the requirements regarding replicabil-

ity, we introduce an information model to support the

automation of information exchange in the SEP.

The remainder of this article is structured as fol-

lows: In section 2, we elaborate on the requirements

for energy scenarios. In section 3, we deﬁne the term

”energy scenario”, review related methodologies for

their development and also review approaches for in-

tegrating energy scenarios with methods from deci-

sion analysis research. We furthermore describe the

applicability of information modeling in the context

of energy scenarios. In section 4, we combine the re-

viewed methodologies in our conceptual solution to

set up the SEP as an integrated process with contin-

uous tool support. After that, we give details on this

solution regarding the SEP in section 5, and regard-

ing the information model supporting the SEP in sec-

tion 6. In section 7, we show how the three above-

mentioned requirements are met by our solution and

discuss open questions and future work.

2 REQUIREMENTS FOR

ENERGY SCENARIOS

As stated in the introduction, the development and

evaluation of energy scenarios should satisfy three ba-

sic requirements: They should include a deﬁnition

and operationalization for sustainability, provide de-

cision support in a transparent way and be replicable

for other researchers. In this section, we elaborate

on these requirements in more detail and point out

that many energy scenarios do not satisfactorily ful-

ﬁll them.

2.1 Sustainability Deﬁnition and

Operationalization

Traditionally, a triangle of objectives including secu-

rity of supply, economic viability and environmental

compatibility is used to guide decision making in the

energy sector. For example, in Germany these ob-

jectives are stated in the ”Energiewirtschaftsgesetz”

(Energy Sector Act) (Deutscher Bundestag, 2005).

Meanwhile, the concept of sustainable development

is increasingly used to guide decisions in multiple

ﬁelds of politics. According to the triple-bottom-line

interpretation of sustainability, economic prosperity,

social justice, and environmental quality need to be

achieved simultaneously (Elkington, 2002).

If both approaches are integrated, technical, eco-

nomic, environmental, and social criteria are all rele-

vant to operationalize sustainability for decision mak-

ing in the energy sector. However, the majority of

recent research on energy transitions focuses on se-

lected aspects and consequently fails to consider all

relevant aspects simultaneously - see (Keles et al.,

2011; Kronenberg et al., 2012; Hughes and Strachan,

2010) for reviews of considered aspects in Germany

and the UK. Therefore, these studies might not be

suitable to guide political decision making towards

sustainable development of the energy sector in the

deﬁnition given above.

2.2 Transparency of the Decision

Support Process

While many studies aim at providing decision support

for political decisions to promote sustainable develop-

ment of the energy system, they fail to conceptualize

this as a formal decision problem by using methods

from decision analysis research. For example, in their

review on 24 energy scenarios in Germany, (Kronen-

berg et al., 2012) point out that an integrated sustain-

ability assessment would increase the usefulness for

decision support, but is missing in most of the re-

viewed energy scenarios. This means that the process

of generating recommendations from energy scenar-

ios, i.e. the underlying decision support process, is

not transparent for decision makers (Grunwald et al.,

2016). Most energy scenarios do not specify decision

alternatives for decision makers and delineate them

from external uncertainties. Also decision makers’

preferences are not explicitly considered, so that in-

terpreting results for a decision towards a sustainable

energy system remains an implicit and manual task.

This is a problem, because most energy scenarios

are developed in principal-agent relationships (Grun-

wald et al., 2016). For example, if researchers do

not explicitly communicate the uncertainties associ-

ated with scenarios, decision makers may misinterpret

the results. To this end, transparency of the methods

and models used in the scenario process is needed.

To overcome these problems, (Grunwald et al., 2016)

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange

189

proposed to introduce standards for energy scenarios

in terms of scientiﬁc validity, transparency and open-

ness of the results.

2.3 Replicability

Replicability is crucial to achieve scientiﬁc validity of

energy scenario studies. It allows other researchers to

repeat calculations and simulations of scenarios with

the same parameters to replicate the results. There-

fore, all information about the scenarios should be

documented and published including input data, mod-

els and assumptions (Grunwald et al., 2016). This

enables researchers to take an external scenario and

vary the parameters or add new models to expand the

focus of the scenario. This way, replicability also al-

lows to reuse scenarios and to compare different sce-

nario studies. Furthermore, in combination with a

transparent decision making process it allows other

researchers to replicate scenarios while also consider-

ing further sustainability facets.

During scenario development, simulation and

evaluation, various data is exchanged between differ-

ent actors, models and software. Due to the complex-

ity of these processes (in particular the integration of

simulation models from various domains), the infor-

mation exchange is to a high degree error-prone. This

calls for tool-support and automation with all data

ﬂows and dependencies being well deﬁned and docu-

mented. This documentation of energy scenarios sig-

niﬁcantly improves the replicability (Grunwald et al.,

2016).

3 RELATED WORK

In this section, we introduce different methodologies,

which we combine to set up an integrated sustainabil-

ity evaluation process with automated information ex-

change: Firstly, as we focus on scenarios in the en-

ergy domain, concepts of scenarios and methods for

scenario design in this domain are described. Sec-

ondly, we discuss recent work from the area of de-

cision support and discuss the applicability of multi-

criteria decision making (MCDM)

in an energy sce-

nario context. Thirdly, we point out how information

models are used to support simulation and evaluation

processes.

Also called multi-criteria decision aiding/analysis

(MCDA).

3.1 Energy Scenarios

Energy scenarios are a tool to investigate possible

transformations of future energy systems. (Grunwald

et al., 2016, p. 9) deﬁne them as a description of ”a

possible future development (or a future state) of the

energy system.” An exemplary methodology to set up

these scenarios is the story-and-simulation (SAS) ap-

proach (Alcamo, 2008). While this approach stems

from environmental modeling, it is increasingly used

in the energy context (Weimer-Jehle et al., 2016).

SAS scenarios combine qualitative stories and quan-

titative data for simulation studies. With this two-fold

concept, SAS scenarios allow both the involvement of

decision makers and technical simulation.

The variation of quantitative attributes within the

SAS scenarios leads to adapted simulation model

parametrization. To this end, it should be clear, why

certain quantitative parameters of a simulation are

varied and others are not, based on a storyline.

For the sake of transparency a well-deﬁned pro-

cess for creating these stories is necessary, which can

be found e.g. in scenario planning

. This is an expert-

based management tool, which originates from strate-

gic planning on company level in the 1970s (van der

Heijden, 1996). Based on this prior work, (Gause-

meier et al., 1998) proposed the following process for

the developing future scenarios:

Firstly, participating domain experts discuss fac-

tors inﬂuencing a scenario and systematically iden-

tify the most important key factors. Afterwards, they

identify the key factors’ projections, which are pos-

sible developments up to a certain point in time to

span a broad range of possible future developments,

and describe them. These different projections are

checked by the experts for consistency and the re-

sults are recorded in a consistency matrix. Based on

this matrix a scenario software uses cluster analysis to

build projection bundles, which represent consistent

combinations of projections and thus possible future

scenarios. In the last step, the domain experts write

storylines, i.e. textual descriptions, for all future sce-

narios.

3.2 Integration of Scenario Planning

and MCDM

According to (Belton and Stewart, 2003, p. 2),

MCDM is ”an umbrella term to describe a collec-

tion of formal approaches, which seek to take ex-

plicit account of multiple criteria in helping individu-

als or groups explore decisions that matter”. We shall

Also called scenario management.

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

190

highlight three characteristics in this deﬁnition, which

show the applicability of MCDM for energy systems

planning: Firstly, as stated in section 2.1, technical,

social, environmental and economic aspects are rel-

evant for energy systems planning (”multiple crite-

ria”). Secondly, the decision for a future energy sys-

tem affects many stakeholders with conﬂicting ob-

jectives. In this context, MCDM can help to struc-

ture and inform debates (”individuals or groups”).

Thirdly, the energy system is fundamental for sustain-

ing modern societies (”decisions that matter”).

Researchers have already applied MCDM exten-

sively in sustainable energy planning, e.g. (Wang

et al., 2009) and (Oberschmidt et al., 2010) provide

overviews. Surprisingly, the integration with scenario

planning has not been in the focus of research. (Stew-

art et al., 2013, p. 682) identify four concepts for sce-

narios, which are more or less well-suited for integrat-

ing MCDM and energy scenarios:

1. external situations affecting consequences of pol-

icy actions,

2. exploration of future conditions or environments,

3. advocacy of particular courses of action,

4. representative sample of future states.

(Kowalski et al., 2009) integrate scenario planning

and MCDM and apply this to energy systems plan-

ning in Austria. However, they do not differentiate

between decision alternatives and external uncertain-

ties: They use the scenarios from scenario planning

directly as decision alternatives (see also (Madlener

et al., 2007)). This approach is consistent with the

scenario concept ”exploration of future conditions or

environments”, in the sense that the decision alter-

natives are deﬁned by a range of possible outcomes.

While scenarios were originally designed to deal with

external uncertainties in strategic planning, (Kowalski

et al., 2009) do not consider external uncertainties.

(Stewart et al., 2013) provide general guidelines

for integrating MCDM and scenario planning. To that

end, they point out that ”it is essential that the sce-

narios reﬂect external driving forces (events, states)

which are separated from the policies or actions un-

der consideration” (Stewart et al., 2013, p. 683).

This implies the interpretation of scenarios as ”ex-

ternal situations affecting consequences of policy ac-

tions”. However, they do not provide a detailed pro-

cess model for integrating MCDM and energy scenar-

ios.

3.3 Information Modeling

In the aforementioned development, simulation, and

evaluation of energy scenarios, various data ﬂows

have to be deﬁned. For example, the simulation of en-

ergy systems includes material ﬂows (e.g. coal or bio-

gas for power plants), information ﬂows (e.g. in smart

grid control strategies), and electric power ﬂows. Ad-

ditionally, the behavior of users should be included to

achieve valid results. Therefore, various simulation

models have to be coupled to represent this complex,

dynamic, sociotechnical system. This is a complex

task, so automating the exchange of data is impor-

tant. This is usually done by deﬁning an information

model, which is described in (Lee, 1999, p. 1) as ”a

representation of concepts, relationships, constraints,

rules, and operations to specify data semantics for a

chosen domain of discourse.”

According to the deﬁnition and operationalization

of sustainability (see section 2.1), experts of different

domains take part in the scenario development pro-

cess. Therefore, the challenge is not only to model

information but also to collect domain knowledge. A

single term can be used in different domains with dif-

ferent meanings. Therefore, the terms and relation-

ships between them should be deﬁned. A common

technology for this representation of domain knowl-

edge are ontologies, which ”have been developed to

provide a machine-processable semantics of informa-

tion sources that can be communicated between dif-

ferent agents (software and humans)” (Fensel, 2004,

p. 3).

Information models are widespread in industry

to allow interoperability, e.g. the Common Informa-

tion Model (CIM) in the energy domain (Uslar et al.,

2012). It contains a data model (domain ontology),

various interface speciﬁcations, technology-speciﬁc

instantiations of the ontologies (communication and

serialization) and allows automated communication

between components of smart grids.

The common processes and languages for infor-

mation modeling and ontology development require

the user to be experienced in their usage. To avoid this

barrier for experts from different domains, who most

likely do not have this expertise, there are some ap-

proaches to use concept maps for the instantiation of

ontologies (Castro et al., 2006; Simon-Cuevas et al.,

2009).

4 CONCEPTUAL SOLUTION

In section 2, we have deﬁned three main requirements

for the devolopment and evaluation process of future

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange

191

energy scenarios: A deﬁnition and operationalization

of sustainability, transparency of the decision support

process to allow decision makers to rely on the results,

and replicability to ensure comparability. To take into

account these requirements, we propose to use SAS

scenarios in combination with scenario planning for

the story and integrate the following three approaches:

Firstly, to increase the transparency of the deci-

sion support process, we propose to integrate SAS

scenarios with MCDM. (Stewart et al., 2013) pro-

vide an overview of approaches from the 1990s and

early 2000s that already integrated MCDM and sce-

nario planning to provide an integrated sustainability

assessment. However, today only few energy scenar-

ios build upon these approaches and differentiate be-

tween courses of action (we will call these decision

alternatives)

and framework conditions (external

uncertainties). One obstacle might be the lack of

a process model integrating these two approaches.

While there already exist some guidelines for con-

structing SAS scenarios (Alcamo, 2008; Weimer-

Jehle et al., 2016), these guidelines do not integrate

MCDM. In the research project NEDS, a novel pro-

cess model has been introduced for the integration

of MCDM and SAS scenarios for sustainability eval-

uation: the sustainability evaluation process (SEP)

(Blank et al., 2016). We will describe some parts of

this process model in more detail in section 5.

Secondly, the complexity of energy scenarios calls

for tool-support and automation of the process (see

also section 2.3). Therefore, an information model is

proposed to structure the data ﬂows and dependencies

in the scenario process, organize the communication

between experts from different domains and collect

their knowledge. To allow the participation of domain

experts without having them to learn new complex de-

scription languages and techniques in detail, the infor-

mation model uses a mind map for the representation

of knowledge. For integration in the process and the

simulation, we use the modeled information to instan-

tiate an ontology and make the information available

in a machine readable format.

Thirdly, we propose to use co-simulation for the

simulation part of SAS scenarios, to allow consider-

ing multiple dimensions of sustainability. Simulation

is typically done in one single simulation software

(e.g. Matlab). This makes it hard to integrate simula-

tion models from different domains, because most do-

mains use speciﬁc software and languages (Schloegl

et al., 2015). An approach for solving this issue is co-

simulation, which is deﬁned in (Bastian et al., 2011)

In the remaining sections of this paper, deﬁnitions re-

lated to the SEP are highlighted in bold typesetting at their

ﬁrst occurrence.

as ”an approach for the joint simulation of models

developed with different tools (tool coupling) where

each tool treats one part of a modular coupled prob-

lem.” A co-simulation framework with focus on the

energy domain is e.g. mosaik

(Lehnhoff et al., 2015;

Rehtanz and Guillaud, 2016). It allows coupling dif-

ferent simulators, provides an application program-

ming interface (API) for different programming lan-

guages and handles the scheduling and information

exchange between the simulation models during sim-

ulation.

5 SUSTAINABILITY

EVALUATION PROCESS (SEP)

Having presented our general solution approach for

integrating SAS scenarios and MCDM, we now con-

cretize this in terms of a process model for the SEP.

An overview is given in ﬁgure 1. It is subdivided into

four parts that will be explained in the next sections.

The parts do not have to be performed in sequential

order, but should be done at least partly in parallel:

Two different entry points are given, on the left and

on the right side of ﬁgure 1.

5.1 Future Scenarios

In the ﬁrst step

, the scenario planning process (see

section 3.1) is used to develop future scenarios. These

are qualitative, i.e. textual, descriptions of possible fu-

tures and thus need to be transformed into quantita-

tive assumptions, which are used in simulation mod-

els and sustainability evaluation. Therefore, the next

step is the deduction and classiﬁcation of attributes

characterizing these assumptions. On the one hand,

the attributes are deduced from key factors of the fu-

ture scenarios. On the other hand, attributes can also

be deduced from results of the sustainability evalua-

tion phase (in form of transformation functions and

sustainability evaluation criteria, which we shall ex-

plain in section 5.4).

In future scenarios, the distinction between deci-

sion alternatives and external uncertainties is blurred

and therefore a classiﬁcation of attributes is needed.

Figure 2 provides on its left side an overview on the

different types of attributes. To allow differentiat-

ing between attributes characterizing decision alter-

natives and external uncertainties, we introduce a sys-

tem boundary. Naturally, the decision makers’ sphere

http://mosaik.ofﬁs.de

For better readability we use the term ”step” to describe

subprocesses of the SEP.

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

192

NEDS – Sustainability Evaluation ProcessNEDS – Sustainability Evaluation ProcessNEDS – Sustainability Evaluation Process

Future ScenariosFuture ScenariosFuture Scenarios Preparation of Evaluation Scenarios & AlternativesPreparation of Evaluation Scenarios & AlternativesPreparation of Evaluation Scenarios & Alternatives

Modelling

Simulation

Modelling

Simulation

Modelling

Simulation

Sustainability EvaluationSustainability EvaluationSustainability Evaluation

Citizen

involvement

(Symposium)

Model setup &

data preparation

Model execution

Definition of

sustainability

evaluation criteria

set boundaries

to relevant scenarios

Transformation

function

Scenario planning

[Gausemeier]

Future scenarios

Attribute

deduction &

classification

List of attributes

General

framework

conditions

Reduction of

scenario set

List of relevant

future scenarios

(subset)

Quantification of

attributes (ranges)

Quantified

scenario subset

Define evaluation

scenarios

Alternatives specific

for each evaluation

scenario

Evaluation

scenarios

(quantified)

endogenous

exogenous,

static/variable

Combine

Derived

attributes

Sustainability

evaluation

criteria (SEC)

Quantified

SEC for each

scenario/

alternative

combination

Define in detail for

individual models

Sustainability

facettes

NEDS

Sustainability

Evaluation

[MCDA]

combine with appropriate

quantified SEC

Sustainability

order of

alternatives per

evaluation

scenario

non-derived attributes are

passed directly

Reduce attribute

ranges to discrete

values / alternative

Evaluation

objects

(scenario +

alternative)

Quantified general

framework

conditions

needed for

all scenarios

Weighting

of SEC

Stakeholder

involvement

(workshop)

Sensitivity analysis

of SEC weights

process

document

data flow

data

illustrative:

data flow,

iterations or

additional

comments

entry point

endpoint

Figure 1: Sustainability evaluation process (blue legend icons give information on the semantics used within the diagram)

(Blank et al., 2016).

of inﬂuence provides this system boundary: If deci-

sion makers can decide upon the values of attributes,

they are classiﬁed as endogenous attributes. A com-

bination of values for all endogenous attributes conse-

quently constitutes a decision alternative. For exam-

ple, political decision makers can decide, if and which

renewable energy technologies are granted subsidies.

Thereby, they inﬂuence their respective shares in the

energy mix. In contrast, if decision makers cannot

decide upon values of attributes, they reﬂect external

uncertainties. For example, political decision mak-

ers in a federal state government cannot decide on

the development of prices for crude oil. We fur-

ther distinguish external uncertainties into scenario-

speciﬁc framework conditions and general frame-

work conditions. For example, prices for crude oil

might be a scenario-speciﬁc framework condition,

while the demographic development might be a gen-

eral framework condition. The decision, whether

framework conditions are interesting and therefore

scenario-speciﬁc, depends on the systematic selection

of key factors in the future scenarios. Since general

framework conditions are not scenario-speciﬁc, they

also do not have an impact on the decision between

alternatives.

The results of this part of the process are tex-

tual descriptions of future scenarios, deﬁned general

framework conditions and a list of the classiﬁed at-

tributes (see left column in ﬁgure 1).

5.2 Preparation of Evaluation Scenarios

and Alternatives

In a ﬁrst screening, the future scenarios are evaluated

against the general framework conditions, e.g. targets

for reducing greenhouse gas (GHG) emissions, which

set boundaries for identifying decision alternatives.

This is necessary, because the future scenarios are de-

signed to reﬂect a broad range of future projections.

In this step, future scenarios may be discarded, if it

is obvious that they will not comply with all general

framework conditions. For example, if there exists a

future scenario, in which the shares of renewable en-

ergy technologies stay on today’s levels and energy

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange

193

Derived

attributes

Transformation

functions

General

framework

conditions

Scenario-

specific

framework

conditions

Endogenous

attributes

exogenous

endogenous

scenario-static scenario-variable

Attributes

Quantified

sustainability

evaluation

criteria (SECs)

Figure 2: Overview of attribute classiﬁcation in different types and the data ﬂows of their values.

demand increases, it is quite obvious that GHG re-

duction targets cannot be met. However, this scenario

may be used as a reference scenario, depending on the

objective of the evaluation.

Quantifying the attributes for a certain point in

time in the future, say the year 2030 or 2050, is the

next step. This means that the different types of at-

tributes (see section 5.1) need to be speciﬁed in such

a way that a future scenario is reﬂected in this spec-

iﬁcation. To that end, attribute speciﬁcations can be

gained e.g. from related quantitative energy scenarios

in literature, i.e. scenarios that ﬁt to the assumptions

of the selected future scenarios. Using these different

scenario studies, for general framework conditions a

single value has to be deﬁned, while for scenario-

speciﬁc framework conditions values have to be de-

ﬁned for every future scenario. For the endogenous

attributes, different scenario studies are used to deter-

mine ranges of possible values in every future sce-

nario. After that, ﬁnal discrete attribute values are

chosen in such a way that they reﬂect decision alter-

natives. According to the SAS procedure proposed by

(Alcamo, 2008), fuzzy set theory can be used for this

step.

The results of this part of the process are quan-

tiﬁed decision alternatives and quantiﬁed evaluation

scenarios illustrating the associated uncertainties.

5.3 Modeling and Simulation

In the previous parts, attributes have been deduced,

classiﬁed and quantiﬁed with the help of related liter-

ature. But not all relevant attributes can be quantiﬁed

this way and therefore simulation is used to provide

derived attributes. As mentioned in section 4, co-

simulation is used to allow the integration of simula-

tion models from different domains. Since modeling

is done differently in every domain and this is not the

focus of this paper, only a rough speciﬁcation of this

part is shown here. More details on this will be given

in future publications.

Firstly, the different simulation models have to

be set up and the input data has to be prepared,

e.g. scaled to the scope of the simulation scenario. As

indicated by a feedback loop in the process, models

may have dependencies between each other that have

to be considered. We shall explain modeling these

dependencies in more detail in section 6.3. After sim-

ulation scenarios are set up, simulation can be started

and the derived attributes are provided.

5.4 Sustainability Evaluation

The last part of the process is the evaluation of sus-

tainability, which is assisted by MCDM. As a require-

ment for any MCDM, the researchers need to deﬁne

the criteria, in terms of which the performance of al-

ternatives is measured (Belton and Stewart, 2003).

Since the criteria should reﬂect the different aspects

of sustainability, we name them sustainability eval-

uation criteria (SECs). The values for the SECs rep-

resent the performance of a decision alternative under

a given scenario in terms of these criteria for a certain

year.

To guarantee that all relevant SECs are consid-

ered, we propose a two-step procedure: Firstly, re-

lated literature on MCDM in the energy context pro-

vides input for the SECs. SECs should also be gained

from public participation, e.g. by using questionnaires

in a public symposium (see entry point in the right

column of ﬁgure 1). Secondly, the researchers should

condense the collected SEC to avoid redundancies

and deﬁne them.

The SECs are structured hierarchically so that the

ﬁrst level of the hierarchy represents the overall objec-

tive, e.g. identifying a sustainable power system. The

second level represents the aspects as deﬁned in sec-

tion 2.1 (technical, economic, environmental, social).

The lower levels of the hierarchy represent sub-goals,

which are ultimately broken down into quantiﬁable

SECs. For example, climate protection and biodiver-

sity protection are sub-goals in the environmental do-

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

194

main, which can be broken down into greenhouse gas

potential, particulate matter formation, land use etc.

With the researchers having collected and struc-

tured the SECs, the decision makers need to assign

weights to reﬂect their preferences (Belton and Stew-

art, 2003). (Wang et al., 2009) provide an overview

on standard procedures for the assigning weights in

MCDM. While directly asking decision makers as

commissioners of energy scenarios is the straightfor-

ward approach, a multi-actor MCDM involving dif-

ferent stakeholder groups is also possible (Steinhilber,

2015).

The next step is to deﬁne a transformation func-

tion for each criterion to determine the quantiﬁed

SECs for each decision alternative (see the right side

of ﬁgure 2). This function transforms derived and

non-derived attributes into concrete values for the

SECs. For example, the installed capacities of dif-

ferent power plants and a speciﬁed demand curve for

a year, say 2050, are given as input for a simulation.

Then, the schedules of these power plants in 2050 are

calculated in the simulation by matching supply with

demand. From these schedules, the CO

-emissions

in 2050 can be calculated for this decision alternative

(installed capacities of power plants) with a transfor-

mation function (provided the CO

-emissions for spe-

ciﬁc power plants are also known and normalized to

a reference unit, e.g. [t CO

-eq/MWh]). In this ex-

ample, the demand curve could be a scenario-speciﬁc

framework condition and thus altered to reﬂect differ-

ent future scenarios.

To deﬁne the transformation functions, it is impor-

tant that the criteria ﬁt the simulation models or rather

that this ﬁt is established: In the best case, researchers

can choose, expand or design simulation models in

such a way that all SECs can be calculated. If this

is not possible, external studies might provide input

for the transformation functions. For example, stud-

ies on life-cycle assessments can provide input for the

environmental criteria.

Lastly, the actual evaluation of the decision al-

ternatives is performed by aggregating the quantiﬁed

SECs of the different alternatives with the weights as-

sociated with them. To that end, different aggrega-

tion methods exist in MCDM. A suitable method in

this context is PROMETHEE, since it is easily under-

standable and therefore also transparent for decision

makers (Brans and Vincke, 1985). The result of this

method is a (partial or total) ranking of the decision

alternatives. This order can then be used to generate

recommendations for the decision makers. Further-

more, sensitivity analysis of the SEC weights can be

used to check the robustness of the results.

The results of this part of the process are a rank-

ing of alternatives and recommendations for further

action for the decision makers.

6 INFORMATION MODEL

Having presented the SEP of energy scenarios, we

shall point out how an information model can help

to automate the information exchange during this

process. The objectives of the information model

are to structure the data ﬂows and dependencies in

the described SEP, organize the communication be-

tween experts from different domains and collect their

knowledge. The experts should be enabled to con-

tribute directly to the SEP without using description

languages and techniques as commonly used in infor-

mation modeling. To allow participation of experts

from all domains considered, an easily usable soft-

ware is necessary, which is found in mind mapping

tools.

Our approach for fulﬁlling the objectives with the

information model in a mind map is explained in the

following sections. In section 6.1 we describe the

information model’s general structure and concretize

this with an example in section 6.2. To allow the au-

tomated integration of the modeled information, we

show its machine readable representation as an ontol-

ogy in section 6.3 and explain the integration in the

SEP in section 6.4.

6.1 Structure of the Information Model

The information model links future scenarios and

simulation scenarios to the sustainability evaluation.

While future scenarios give an overview of possi-

ble developments in the future, simulation scenar-

ios describe the conﬁguration of a concrete simula-

tion, which includes the used simulation models and

how they are connected and parametrized. The con-

nections from both of these scenarios to the evalua-

tion are implemented with transformation functions,

which provide the dependencies and mathematical

descriptions by mapping attributes and derived at-

tributes onto the SECs.

The structure of the information model is depicted

in ﬁgure 3. The left side represents the domains of in-

terest for the future scenarios and simulation scenar-

ios and consists of different levels ordered from left to

right. On the ﬁrst level the domains are listed. Each

domain is subdivided into domain objects, which rep-

resent objects of the real world. The domain objects

We chose the mind mapping tool XMind

(http://xmind.net), which can be extended with plug-

ins and is available as open source software.

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange

195

Figure 3: Structure of the information model.

consist of attributes

and derived attributes that de-

scribe them. Attributes have deﬁned units and are in-

stantiated with values in each evaluation scenario.

The derived attributes represent the results from

simulations and have to be connected to the corre-

sponding simulation model. The inputs of a sim-

ulation model are connected with incoming arrows,

which show the data ﬂows from the attributes to a sim-

ulation model. The outputs of a simulation model are

connected with reversed arrows (from a simulation

model to the derived attributes). Derived attributes

are also marked with cogwheel icons to allow to dis-

tinguish them from other attributes.

The right side of the information model represents

the evaluation part of the SEP. On the ﬁrst level, the

major objective is deﬁned – in our case sustainabil-

ity. The objective consists of different facets on the

second level (e.g. technical, economic, environmental

and social as introduced in section 2.1). Every facet is

subdivided into various SECs on the third level. These

SECs have to be deﬁned by transformation functions,

which may have attributes and derived attributes from

the left side of the information model as input.

6.2 Example

An exemplary segment of the information model

modeled in XMind is shown in ﬁgure 4. On the left

side, the two domains ”information and communi-

cation technology” (ICT) and ”energy” are depicted.

”Control systems” and ”power plants” are domain ob-

jects, which are described by some attributes.

One single simulation model is included in this

example representing a controller for the operational

planning of power plants. This (simpliﬁed) controller

uses the control strategy of the control system and the

power of the power plants as inputs. The results are a

The term attribute in the information model includes all

framework conditions and endogenous attributes described

in section 5.1 and ﬁgure 2.

schedule for all power plants and the period of use for

every single power plant.

Power, speciﬁc CO

-emissions and period of use

feed the transformation function for calculating the

-emissions of the whole system. In this case,

the transformation is an aggregation of data, which is

used to quantify the SEC CO

-emissions on the right

side of the information model.

6.3 Ontological Representation

In section 6.1, the structure of the information model

in a mind map has been described. To allow the au-

tomated integration of the modeled information in the

SEP it has to be made available in the form of a ma-

chine readable format. As described in section 3.3,

ontologies are a frequently used technology for a ma-

chine readable representation of knowledge. Thus,

we aim to allow the representation of the informa-

tion model’s content with an ontology. By doing so

it provides a structure for reasoning the information

to infer implicit knowledge and query the informa-

tion with languages like e.g. SPARQL to support the

development of simulation scenarios.

The capability to query data can be used to support

the user in the following ways. For example, when

the left and right sides of the information model have

to be coupled, the user can be supported in different

tasks: Firstly, querying allows to identify attributes

that are missing on the left side of the information

model as input for transformation functions to allow

the previous deﬁned evaluation (use it from the right

to the left side). This information helps to choose the

right simulation models or to include the right key fac-

tors in the scenario planning. Secondly, it allows to

identify evaluation criteria that can be added on the

right side of the information model to make sure that

all results from future scenarios and simulation are

used (use it from the left to the right side).

Additionally, the modeled information allows to

examine the dependencies between simulation mod-

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

196

Figure 4: Example of the information model in XMind.

els. More speciﬁcally, the data ﬂows between models

can be analyzed to ﬁnd mutual or cyclic dependen-

cies to get information about the order in which the

models have to be executed.

As the information model is implemented in a

mind map, a transformation to an ontology is needed.

To allow the ontological representation a base ontol-

ogy is implemented with the general structure of the

information model. Based on the modeled informa-

tion in the mind map, the base ontology is instantiated

with the concrete information by an extension of the

mind mapping software.

6.4 Integration in the SEP

The information model is integrated in the SEP (see

ﬁgure 1) and supports the following steps:

Deduction of Attributes: The results of the deduc-

tion of attributes (see section 5.1) are listed in

the information model. In parallel, the SECs and

transformation functions are deﬁned in the infor-

mation model (see section 5.4). As explained

in the previous section, the information model is

able to check the dependencies and identify miss-

ing attributes, SECs or transformation functions.

Quantiﬁcation of Attributes: Having listed the at-

tributes in the information model, experts need to

quantify them, so that is it possible to use them

as parameters in the simulation models and in the

evaluation with the MCDM. In this step, different

values are assigned to the attributes (to represent

the different scenarios). The information model

supports that by providing a (database) schema for

the data.

Modeling and Simulation: The information model

supports designing the simulation models and set-

ting up simulation scenarios (see section 5.3). As

it contains the connections of attributes and sim-

ulation models, it illustrates, which dependencies

have to be considered in modeling and simulation.

Sustainability Evaluation: In the sustainability

evaluation part of the SEP, the information

model is directly integrated, because it contains

the transformation functions, which deﬁne the

derived attributes and attributes as input of the

SECs. It also provides the data schema for

handling the results.

7 CONCLUSIONS AND FURTHER

WORK

In this article, we have shown how the SEP developed

in the project NEDS (Blank et al., 2016) integrates

SAS scenarios with MCDM and leads towards an in-

tegrated sustainability evaluation of energy scenarios.

We have described the process and added details to

different parts. Additionally, we have introduced an

information model, which leads towards an automa-

tion of information exchange in the SEP. In this in-

formation model, future and simulation scenarios are

linked to the sustainability evaluation via attributes,

transformation functions and SECs. We shall high-

light how the three requirements for the development

and evaluation of energy scenarios – integrate a sus-

tainability deﬁnition, add transparency to the decision

support process and allow for replicability (see sec-

tion 2) – are met by this integrated approach:

Firstly, co-simulation integrated in SAS scenarios

simpliﬁes the consideration of various sustainability

facets in simulations, because it allows coupling sim-

ulation models from different modeling tools and pro-

gramming languages. Thereby, not only technical or

economic, but also environmental and social criteria

can be considered simultaneously in energy scenar-

ios. The information model facilitates the integration

of different dimensions of sustainability by model-

ing the data ﬂows between different models, software

and actors in the SEP. Overall, it allows to handle

the complexity of multi-domain co-simulation scenar-

ios and supports the communication between experts

from different domains.

Secondly, integrating MCDM into energy scenar-

ios makes the decision support processes more trans-

parent for decision makers. Integrating MCDM into

energy scenarios facilitates problem structuring: The

structured nature of MCDM approaches challenges

decision makers to think about and make their own

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange

197

preferences explicit. Additionally, the proposed SEP

fosters the communication of uncertainties by differ-

entiating explicitly between decision alternatives and

external uncertainties. Thereby, decision makers are

presented with the impacts of their decisions in highly

uncertain environments. This way, decision alterna-

tives can be identiﬁed and evaluated.

Thirdly, the information model addresses both the

transparency and replicability of the development and

evaluation of energy scenarios by modeling the data

ﬂows and dependencies between different models,

software and actors in the SEP. The ontological rep-

resentation of information leads towards an automa-

tion of scenario deﬁnition, simulation preparation and

evaluation of the results in the context of energy sce-

narios. This is crucial to handle the complexity of

energy scenarios.

While the proposed solution satisﬁes the require-

ments for energy scenarios in theory, this is only a

starting point and its applicability needs to be vali-

dated empirically. To that end, the SEP is currently

applied in the project NEDS to analyze the transition

of the energy system of the German federal state of

Lower Saxony up to 2050. Researchers from business

administration, computer science, economics, electri-

cal engineering, and psychology work together with

stakeholders to identify and evaluate future states of

the energy system in Lower Saxony. Additionally, the

following steps will be undertaken in NEDS:

1. Reﬂection of transition paths: To evaluate not

only target years within the given future scenarios

but include the transition to these in the sustain-

ability evaluation is important for a holistic view

on the problem domain. Possible methodical ap-

proaches are multi-period MCDM, e.g. by adapt-

ing the PROMETHEE method to handle multi-

period decision problems.

2. Information model usage for data manage-

ment: To allow an automation of evaluation in

the SEP the direct integration of the information

model in the simulation process has to be estab-

lished. The deﬁnition of data in the information

model will also be integrated in the data manage-

ment and used for semantic analysis and informa-

tion retrieval in the results.

Further work beyond the NEDS project might add

handling of model induced uncertainties in the simu-

lation process: As every simulation model is a sim-

pliﬁcation of the real world, simulation naturally in-

troduces uncertainties, which have to be dealt with.

There is some work done about quantifying uncertain-

ties in the context of co-simulation framework mosaik

(Steinbrink and Lehnhoff, 2016), which can be inte-

grated in the SEP and the information model in future

work.

ACKNOWLEDGEMENTS

The research project ’NEDS Nachhaltige Energiev-

ersorgung Niedersachsen’ acknowledges the support

of the Lower Saxony Ministry of Science and Cul-

ture through the ’Niederschsisches Vorab’ grant pro-

gramme (grant ZN3043).

REFERENCES

Alcamo, J. (2008). The SAS Approach: Combining Qual-

itative and Quantitative Knowledge in Environmental

Scenarios. In Alcamo, J., editor, Environmental Fu-

tures: The Practice of Environmental Scenario Anal-

ysis, volume v. 2 of Developments in integrated envi-

ronmental assessment, pages 123–150. Elsevier, Am-

sterdam and Boston.

Bastian, J., Clauß, C., Wolf, S., and Schneider, P. (2011).

Master for Co-Simulation Using FMI. Proceedings

of the 8th International Modelica Conference, pages

115–120.

Belton, V. and Stewart, T. J. (2003). Multiple criteria deci-

sion analysis: An integrated approach. Kluwer Aca-

demic Publishers, Boston, Mass., 2. print edition.

Blank, M., Blaufuß, C., Gl

otzel, M., Minnemann, J.,

Nieße, A., Pothen, F., Reinhold, C., Schwarz, J. S.,

Stahlecker, K., Wille, F., Witt, T., Rendel, T., Eggert,

F., Engel, B., Geldermann, J., Hofmann, L., H

ubler,

M., Lehnhoff, S., Paech, N., and Sonnenschein, M.

(2016). Process for simulation-based sustainability

evaluation of smart grid future scenarios. http://neds-

niedersachsen.de/uploads/tx tkpublikationen/

Whitepaper-SEP-V1.pdf, version 1.0, accessed

2017-02-21.

BMWi (2010). Federal Ministy of Economics and Tech-

nology. Energiekonzept f

ur eine umweltschonende,

zuverl

assige und bezahlbare Energieversorgung.

https://www.bmwi.de/Redaktion/DE/Downloads/E/

energiekonzept-2010.pdf, accessed 2017-02-21, in

German.

Brans, J. P. and Vincke, P. (1985). Note–a preference rank-

ing organisation method: (the promethee method for

multiple criteria decision-making). Management Sci-

ence, 31(6):647–656.

Castro, A. G., Rocca-Serra, P., Stevens, R., Taylor, C.,

Nashar, K., Ragan, M. a., and Sansone, S.-A. (2006).

The use of concept maps during knowledge elicita-

tion in ontology development processes–the nutrige-

nomics use case. BMC bioinformatics, 7:267.

Deutscher Bundestag (2005). German Bundestag. Gesetz

uber die Elektrizit

ats- und Gasverordnung (En-

ergiewirtschaftsgesetz): EnWG. http://www.gesetze-

im-internet.de/enwg 2005/BJNR197010005.html, ac-

cessed 2017-02-21, in German.

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

198

Deutscher Bundestag (2014). German Bundestag. Gesetz

ur den Ausbau erneuerbarer Energien (Erneuerbare-

Energien-Gesetz): EEG 2014. http://www.gesetze-

im-internet.de/eeg 2014/BJNR106610014.html, ac-

cessed 2017-02-21, in German.

Elkington, J. (2002). Cannibals with forks: The triple bot-

tom line of 21st century business. Capstone, Oxford,

reprint edition.

European Commission (2014). Communication from

the European Commission: A policy frame-

work for climate and energy in the period from

2020 to 2030. http://eur-lex.europa.eu/legal-

content/EN/ALL/?uri=CELEX:52014DC0015,

accessed 2017-02-21.

Fensel, D. (2004). Ontologies: A silver bullet for

Knowledge Management and Electronic-Commerce.

Springer-Verlag, Berlin.

Gausemeier, J., Fink, A., and Schlake, O. (1998). Scenario

management: An approach to develop future poten-

tials. Technological Forecasting and Social Change,

59:111–130.

Grunwald, A., Dieckhoff, C., Fischedick, M., H

ofﬂer, F.,

Mayer, C., and Weimer-Jehle, W. (2016). Consult-

ing with energy scenarios: Requirements for scientiﬁc

policy advice. Series on Science-Based Policy Ad-

vice. acatech/Leopoldina/Akademienunion (Eds.).

Hughes, N. and Strachan, N. (2010). Methodological re-

view of UK and international low carbon scenarios.

Energy Policy, 38(10):6056–6065.

International Energy Agency (2016). World En-

ergy Outlook 2016: Executive Summary.

http://www.worldenergyoutlook.org, accessed

2016-12-01.

Keles, D., M

ost, D., and Fichtner, W. (2011). The devel-

opment of the German energy market until 2030—A

critical survey of selected scenarios. Energy Policy,

39(2):812–825.

Kowalski, K., Stagl, S., Madlener, R., and Omann, I.

(2009). Sustainable energy futures: Methodological

challenges in combining scenarios and participatory

multi-criteria analysis. European Journal of Opera-

tional Research, 197(3):1063–1074.

Kronenberg, T., Martinsen, D., Pesch, T., Sander, M.,

Fischer, W., Hake, J.-F., Kuckshinrichs, W., and

Markewitz, P. (2012). Energieszenarien f

ur Deutsch-

land: Stand der Literatur und methodische Auswer-

tung. In Bruhns, H., editor, Energiewende - As-

pekte, Optionen, Herausforderungen, pages 132–168.

Deutsche Physikalische Gesellschaft - Arbeitskreis

Energie, Berlin, in German.

Lee, Y. T. (1999). Information modeling: From design

to implementation. Proceedings of the second world

manufacturing congress, pages 315–321.

Lehnhoff, S., Nannen, O., Rohjans, S., Schlogl, F., Dal-

hues, S., Robitzky, L., Hager, U., and Rehtanz, C.

(2015). Exchangeability of power ﬂow simulators in

smart grid co-simulations with mosaik. In Model-

ing and Simulation of Cyber-Physical Energy Systems

(MSCPES), 2015 Workshop on, pages 1–6.

Madlener, R., Kowalski, K., and Stagl, S. (2007). New ways

for the integrated appraisal of national energy scenar-

ios: The case of renewable energy use in Austria. En-

ergy Policy, 35(12):6060–6074.

Oberschmidt, J., Geldermann, J., Ludwig, J., and Schmehl,

M. (2010). Modiﬁed promethee approach for assess-

ing energy technologies. International Journal of En-

ergy Sector Management, 4(2):183–212.

Rehtanz, C. and Guillaud, X. (2016). Real-time and co-

simulations for the development of power system

monitoring, control and protection. In 2016 Power

Systems Computation Conference (PSCC), pages 1–

20.

Schloegl, F., Rohjans, S., Lehnhoff, S., Velasquez, J.,

Steinbrink, C., and Palensky, P. (2015). Towards a

Classiﬁcation Scheme for Co-Simulation Approaches

in Energy Systems. International Symposium on

Smart Electric Distribution Systems and Technolo-

gies. IEEE/IES, pages 2–7.

Simon-Cuevas, A., Ceccaroni, L., Rosete-Suarez, A., and

Suarez-Rodriguez, A. (2009). A Formal Modeling

Method Applied to Environmental-Knowledge Engi-

neering. International Conference on Complex, Intel-

ligent and Software Intensive Systems.

Steinbrink, C. and Lehnhoff, S. (2016). Quantifying

probabilistic uncertainty in smart grid co-simulation.

In 2016 Workshop on Modeling and Simulation of

Cyber-Physical Energy Systems (MSCPES), pages 1–

Steinhilber, S. (2015). Exploring options for the harmoni-

sation of renewable energy support policies in the EU

using multi-criteria decision analysis. Dissertation,

Karlsruher Institut f

ur Technologie.

Stewart, T. J., French, S., and Rios, J. (2013). Integrating

multicriteria decision analysis and scenario planning -

Review and extension. Omega, 41(4):679–688.

Uslar, M., Specht, M., Rohjans, S., Trefke, J., and Gonzalez

Vazquez, J. (2012). The IEC Common Information

Model. Springer, Berlin.

van der Heijden, K. (1996). Scenarios: The art of strategic

conversation. John Wiley & Sons, Chichester, Eng-

land and New York.

Wang, J.-J., Jing, Y.-Y., Zhang, C.-F., and Zhao, J.-H.

(2009). Review on multi-criteria decision analysis

aid in sustainable energy decision-making. Renewable

and Sustainable Energy Reviews, 13(9):2263–2278.

Weimer-Jehle, W., Buchgeister, J., Hauser, W., Kosow, H.,

Naegler, T., Poganietz, W.-R., Pregger, T., Prehofer,

S., von Recklinghausen, A., Schippl, J., and V

ogele,

S. (2016). Context scenarios and their usage for the

construction of socio-technical energy scenarios. En-

ergy, 111:956–970.

Towards an Integrated Sustainability Evaluation of Energy Scenarios with Automated Information Exchange

199