A Comparison of Smart Grids Domain Ontologies

Jos

e Miguel Blanco

, Bruno Rossi

and Tom

s Pitner

Faculty of Informatics, Masaryk University, Brno, Czech Republic

Keywords:

Smart Grids, Ontologies, Semantic Web Reasoners, Performance.

Abstract:

Smart Grids (SG) represent one of the key critical infrastructures. Over time, several ontologies were deﬁned

in the SG domain to model aspects such as devices and sensors integration, and prosumers’ communication

needs. In this paper, we review the state of the art regarding semantic web reasoning in the domain of SGs. We

compare ﬁve main ontologies in terms of descriptive statistics (e.g., number of axioms), load time and reason-

ers runtime performance. Results show that not all the ontologies in the SG domain are readily available, and

that some of them might be more appropriate for deployment in devices with limited computational resources.

1 INTRODUCTION

Smart Grids (SG) have become one of the most

important Critical Infrastructures (CI) and Cy-

ber–Physical Systems (CPS) in the modern society.

The potential for the improvement of the energy sec-

tor has been seen as a key factor, driven by the smart

integration of users’ behaviour leading to a more sus-

tainable and secure electricity supply driven by re-

newable energy sources (Yu et al., 2011). How-

ever, as a representative of a typical CPS, the grid

has many complexities in the integration of cyber re-

sources (e.g., computing algorithms, communication

and control software) with physical parts (e.g., sen-

sors, smart meters) (Farhangi, 2009).

To tackle the complexity of SGs, the Smart Grid

Architecture Model (SGAM) (Bruinenberg et al.,

2012) was proposed, aiming at dividing the complex-

ity into several layers and their integration: a com-

ponent layer (the physical devices), a communication

layer (protocols), a data layer (information data mod-

els), a function layer (functionalities to be provided),

and a business layer (the business requirements). At

the data layer level, the adoption of domain ontolo-

gies aims at allowing reasoning over the many devices

and interactions with different goals like tracking in-

formation about reliability of devices or their integra-

tion (Zhou et al., 2012; Catterson et al., 2005; Gillani

et al., 2014; Schachinger et al., 2016).

https://orcid.org/0000-0001-9460-8540

https://orcid.org/0000-0002-8659-1520

https://orcid.org/0000-0002-2933-2290

Many ontologies have been deﬁned over time for

the energy sector (e.g., SEPA’s Smart Grid Ontology,

Open Energy Ontology, like we will discuss in sec-

tion 3). Yet, there is no comprehensive evaluation of

the proposed ontologies related to the current state of

the art. In this paper, we want to provide evaluation

of existing SGs ontologies: what is their availability

/ support, how much are they usable, and what is the

performance overhead considering the possibility of

deployment in devices such as smart meters with rel-

atively low computational power available.

In particular, we are interested into looking at the

impact of each different ontology in terms of per-

formance (execution time) of the reasoner — It has

been shown that different ontologies have better or

worse performance depending on the reasoner they

are loaded in (Kang et al., 2014; Kang et al., 2012).

We have one main Research Question (RQ) for

this paper: What is the state of different Smart Grids

domain ontologies that were deﬁned over the years?

We answer this RQ by looking both at the availability

of ontologies, some descriptive statistics of the on-

tologies representing the size, and at the performance

when used in a reasoner, taking into account the po-

tential limited resources of embedded devices such as

smart meters.

The paper is structured as follows. In Section 2,

we provide related works in the area of SGs ontolo-

gies and related semantic technologies, reasoners im-

provements, and ontologies performances. In Sec-

tion 3, we discuss existing ontologies for the SG do-

main. In Section 4, we analyze the main SGs ontolo-

gies identiﬁed, in terms of metrics related to size, and

Blanco, J., Rossi, B. and Pitner, T.

A Comparison of Smart Grids Domain Ontologies.

DOI: 10.5220/0010710000003058

In Proceedings of the 17th International Conference on Web Information Systems and Technologies (WEBIST 2021), pages 115-123

ISBN: 978-989-758-536-4; ISSN: 2184-3252

115

performance of the reasoners. In Section 5 we present

the main results of the analysis. In Section 6, we con-

clude the paper.

2 RELATED WORKS

There are several research articles that focus on the

topic of SG, ontologies, and semantic web reasoners.

These papers can be classiﬁed in three different cate-

gories. First of all, we have papers that introduce new

ontologies or semantic technologies relative to the do-

main (Section 2.1); secondly we have papers that de-

scribe ways to improve the performance by deﬁning

new semantic web reasoners or updating already ex-

isting ones (Section 2.2). Finally, the third category

refers to papers that focus on measuring the metrics

of a number of ontologies, usually in more general

way and without focusing in a domain or another,

but providing an invaluable value for this article (Sec-

tion 2.3). With this in mind we proceed to describe

each paper in relation to these categories.

2.1 SG-related Ontologies and Semantic

Technologies

Regarding SG-related ontologies, (Zhou et al., 2012)

main aim is to introduce an ontology with the beneﬁt

of being extensible to other domains close to the SG

domain. Authors provide a case study on Complex

Event Processing (CEP) with the use of SPARQL,

that should be, or can be, as lightweight as possible.

In the same vein, (Cuenca et al., 2017) deﬁnes the

OEMA ontology created to deal with energy manage-

ment. The ontology has high modularity to allow for

very speciﬁc domain applications without wasting re-

sources.

Another ontology to be used in the SG domain

is presented in (Gillani et al., 2014). It describes a

series of cases for which it could be useful and the

whole ontology is divided into classes applicable to

many different infrastructures. Finally, authors in-

troduce inductive inferences going over patterns that

could be identiﬁable thanks to the ontology and a non-

monotonic approach. In the end, they propose the

ontology to be used for CEP. Similarly, (Hippolyte

et al., 2016) describes the implementation of an on-

tology for Multi Agent Systems (MAS) in SGs for

the automatic negotiation between different members

of a network, adding to the ﬂexibility of the system.

They base the work on a Java-based implementation,

providing a tool to convert the ontology into a code-

based implementation.

Yet another ontology for energy management in

SGs is introduced in (Schachinger et al., 2016). The

main aim of the work is to unify the domain of SGs.

Authors provide theoretical testing of the ontology

to showcase all the relations between classes and in-

stances of the ontology which ends up being shown

in a comprehensive diagram. Finally, (Li et al., 2019)

develops an ontology focused on key performance in-

dicators to show the elements that require more en-

ergy given their power consumption patterns. Fur-

thermore, the authors intend to use the ontology as a

way to interchange data between different stakehold-

ers of the network, thus enabling potential support to

Big Data scenarios.

Nevertheless, describing the provided ontologies

in the SG domain is not enough if there are no seman-

tic tools to work with them. In this sense, (Atanasov,

2015) describes in great detail a semantic model for

prepaid smart metering devices. They gloss over all

the data that is generated from the smart meters, and

deﬁne all the semantic labels required for the data

to be processed. The main strength of the work re-

sides in the fact that just from time and volume (of

power consumption) we can derive some effective se-

mantics. The work of (Santodomingo et al., 2015) is

devoted to standardizing the different data types that

might come into play in a SG. It offers an advantage

and usage scenario if the same ontology has not been

used in all the devices or if the sources are not based

on Web of Things (WoT).

The state of the semantic technologies can be seen

in (Dogdu et al., 2014), where authors reﬂect on the

state of the art implementation of semantic technolo-

gies in SGs, giving a precise overview of what they

are good for and how they behave better and are more

conveniently than usual data models. The work done

is based on a real-world implementation of seman-

tic technologies in SGs. Similarly, (Donohoe et al.,

2015) acts as a survey about which technologies are

available when dealing with the data of the SGs focus-

ing on the context-aware ones. They propose a fur-

ther direction on which a middleware platform could

be developed, listing the possible associated require-

ments and the challenges they might face.

In (Wemlinger and Holder, 2011), the authors fo-

cus on providing a semantic framework for smart en-

vironments. This could potentially lead to integrat-

ing all data generated by the smart devices into the

semantic web reasoner, therefore, strengthening the

conclusions obtained. Supposedly helping to trace

the patterns of malfunctioning appliances. As in the

previous cited paper, (Flores-Martin et al., 2019) pro-

poses another unifying method for different smart de-

vices, but with the advantage of being more general:

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

116

from smart environments to smart devices. They ex-

pand their results by being able to integrate wearables

in the detection of malfunctioning smart meters.

To conclude, (Hippolyte et al., 2018) deﬁnes the

structure and the rationale behind an ontology that

might be used to establish a relationship between

energy consumption, data visualization and decision

support. Authors focus on the web-based scalability

of the proposed structure.

2.2 Reasoners Improvements

There are several recent improvements in semantic

web reasoners that can help considering the poten-

tially limited computational power of devices such as

smart meters. (Tai et al., 2011) provides a reasoner

designed to be implemented in a resource-constrained

platform. This makes it interesting to be considered

for an implementation in a smart meter, given the

small overhead. The downside is that it is based on

OWL instead of OWL2, so an update might be needed

to work with the current ontologies. Similarly, (Ali

and Kiefer, 2009) introduces a light-weight semantic

reasoner for constrained resources. Nevertheless, the

approach is quite different from the previous one and

the performance is not as good.

Also, (Speiser et al., 2013) outlines the use of

Linked Data in SGs communication, reinforcing the

idea that it should be lightweight. Authors produce

an evaluation based on connection time. They use re-

source constrained devices by today’s standards and a

lightweight SPARQL engine programmed on C. The

obtained metrics are all under a second distance be-

tween them, so the added value regarding the SGs

domain is quite high, taking into account the goal of

real-time data processing.

2.3 Ontologies Performance

The works presented in this section deal with the per-

formance of ontologies, but all of them are built upon

wide domains, with no speciﬁc SGs ontologies con-

sidered, as there is no speciﬁc study that deals with

performance in the SG domain. In particular, (Kang

et al., 2012) measures the performance of 350 real-

world ontologies and 4 OWL2 reasoners. The goal is

to identify a set of metrics that can predict the perfor-

mance of a semantic web reasoner. Eight ontology-

level metrics are considered to be good predictors for

ontologies performance. As a conclusion they try to

use machine learning to see which metrics could be

the best predictors. More recently, (Kang et al., 2014)

uses an even larger number of ontologies, 450, as an

extension and with the same goal of the previous pa-

per. Authors also focus more on the identiﬁcation

of performance hotspots, identifying the bottlenecks

when designing an ontology. Similarly, (Wang and

Parsia, 2007) identiﬁes semantic web reasoners bot-

tlenecks for performance. Four case studies based

on generated and real-world ontologies are provided.

Authors identify challenges for ontologies engineer-

ing and implement a tool to support statistical analysis

of performance of ontologies.

Recently, (Pe

na et al., 2020) shows some of their

current work in which an enriched ontological model

is used to improve the effectiveness of a recommender

system. In particular, authors go through all the char-

acteristics that make an ontology effective, selecting

among them the ones more relevant for the task at

hand. Also, (Maarala et al., 2017) introduces a new

semantic technology reasoner to be applied to the In-

ternet of Things (IoT) domain. Their intention is to

make the reasoner and data collection as scalable as

possible with the idea of implementing the technol-

ogy into real-life scenarios. It is worth mentioning

that they focus on showing the performance data and

point out possible bottlenecks of the technology.

3 DOMAIN ONTOLOGIES FOR

SMART GRIDS

As a ﬁrst step, we collected existing ontologies in the

SG domain. We have focused on a series of SG do-

main ontologies reported in previous research as well

as some that are used in industry and are yet to be re-

ported in research venues. All these ontologies can be

seen in Table 1. Some of the ontologies are not avail-

able, despite their description and applicability being

reported in research papers — this is mostly due to

a fair share of broken URLs or missing references,

which hinders the adoption of the ontologies.

The OEMA ontology was introduced in (Cuenca

et al., 2017) to provide a modular approach ontol-

ogy for the SGs domain. In this case, we are using

the complete ontology so we consider it to be able

to tackle the network as a whole. The SEPA’s is the

ontology developed by the Smart Electric Power Al-

liance for controlling buildings; this includes regis-

tering energy usage, the base for any SG application.

The Open Energy Ontology is intended for modelling

the whole energy system with an interest in the open

source model. Facility Ontology was designed to con-

trol the production of any facility, and for this reason,

it provides control over the energy usage, crucial for

SGs. SSG, introduced in (Salameh et al., 2019), aims

at modeling the SG components, their features and

properties, allowing the achievement of the SG objec-

A Comparison of Smart Grids Domain Ontologies

117

Table 1: Smart Grids Ontologies.

Name Aim URL

OEMA Ontology network Whole Network Ontology http://www.purl.org/oema/ontologynetwork

SEPA’s Smart Grid Ontology System to control buildings https://github.com/smart-electric-power-alliance/

Electric-Grid-Ontology

Open Energy Ontology Energy System Modelling https://openenergy-platform.org/ontology/

Facility Ontology SG Components https://github.com/usnistgov/facility

DABGEO Energy Management Applica-

tions

https://innoweb.mondragon.edu/ontologies/dabgeo/

index-en.html

SSG (Salameh et al., 2019) SG Components N.A.

Prosumer-oriented (Gillani et al.,

2014)

Prosumer Ontology N.A.

N.A. (Schachinger et al., 2016) SG Integration N.A.

N.A. (Zhou et al., 2012) SG Demand / Response Ap-

plications

N.A.

N.A. (Catterson et al., 2005) SG Health Monitoring N.A.

tives. The ontology of (Catterson et al., 2005) focuses

on monitoring the health of the components of a SG,

in particular, that of any given transformer. DABGEO

is the natural continuation of the OEMA Ontology:

introduced in (Cuenca et al., 2020) is aimed at pro-

viding complete control to energy management appli-

cations.

An explanation for the ontologies of (Gillani et al.,

2014), (Schachinger et al., 2016) and (Zhou et al.,

2012) can be found in Section 2.

4 ANALYZING THE

ONTOLOGIES

In this section we provide a description of the pro-

cess we adopted to obtain the metrics regarding the

description and performance of the main ontologies

from Table 1. For this task we have used the Prot

software in its 5.0.0 Linux version. As for the rea-

soners we are using HermiT, version 1.4.3.456, and

Pellet, version 2.2.0.

First of all, we begin by downloading the ontology

from the corresponding repository (Fig. 1). Once we

have acquired the ontology, we proceed to load it into

Prot

e. As it is loaded, it might occur that the on-

tology itself tries to import some other ontologies. If

these are readily available, Prot

e will load them au-

tomatically, but in the case they are not, we load them

manually. Once the main ontology and all the im-

ported ones are loaded into Prot

e, they are merged

into one ﬁnal ontology that represents what we will be

measuring. This allows us to work with a real-world

ontology as well as makes the workﬂow easier.

https://protege.stanford.edu/products.php

After the ﬁnal ontology is obtained, we annotate

the metrics of the ontology that are listed by Prot

In this case we are going for the metrics that are la-

belled as Axiom, Logical Axiom Count, Declaration

Axiom Count, Class Count, Object Property Count,

Data Property Count, Individual Count, and Anno-

tation Property Count. To this list we also add how

much time Prot

e has taken to load the ontology.

Once we have the metrics we proceed to start the

associated reasoner in Prot

e and get the reasoning

time. Afterwards it is the moment to extract any con-

clusions from the data collected.

4.1 Ontology Metrics

The results and the metrics of each of the selected

main SG ontologies that were analyzed can be

seen in Table 2. These metrics have been obtained

accordingly to the process described in Section 4. All

the measurements were done on an Intel i7-3537U

CPU @ 2.00GHz, with 8GB of RAM, running

Ubuntu 20.04.2 LTS. The metrics regarding time are

extracted by performing ﬁve different times the same

task in a completely new instance of Prot

e and

obtaining the mean.

We also measured the memory and CPU usage

for each ontology as it can be seen in Figures 3 to

7. To capture these plots we used a script, based on

Python’s libraries psrecord and matplotlib, that is:

psrecord $(pgrep process_name)

--interval 1 --plot file_name.png

Finally, a comparative of the times of each ontol-

ogy can be seen in Figure 2. It is important to men-

tion that the missing bars, indicate that the time is not

available for some or other reason.

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

118

Table 2: Metrics.

OEMA Ontology net-

work

SEPA’s SG Ontology Open Energy Ontol-

ogy

Facility Ontology DABGEO

Axiom 24738 5491 8563 15248 11678

Logical Axiom Count 9577 2426 1614 11904 4349

Declaration Axioms Count 5686 111 1215 691 2590

Class Count 3502 250 928 431 1964

Object Property Count 941 70 83 107 270

Data Property Count 712 28 N.A. 43 204

Individual Count 47 234 104 1722 69

Annotation Property Count 32 117 95 52 43

Loading time 7241.2ms 12284.6ms 1296.2ms 3726ms 1973.2ms

Reasoning time (HermiT) N.A. 1106ms 331.8ms N.A. 56051.2ms

Reasoning time (Pellet) 26406.6ms 788.2ms N.A. N.A. N.A.

4.2 Issues

Despite our best work, obtaining all the metrics af-

ter analyzing all the ontologies has been challenging,

and have encountered multiple issues that we proceed

to list below. These issues are something to take into

account when analyzing the ontologies, but not some-

thing that makes impossible to work with them. They

are, therefore, questions that need to be taken into ac-

count when deploying any of these ontologies in a

real-world scenario.

The issues that have aroused when analyzing the

ontologies are:

• OEMA cannot be loaded as it was intended. Dur-

Download the Ontology from the corresponding repository

Load the ontology into Prot

Import the missing ontologies

Merge all the imported ontologies into one

Annotate the metrics of the resulting ontology

Start the integrated reasoner of Prot

Measure the reasoning times of the corresponding ontology

Analyze and extract conclusions regarding the obtained data

Figure 1: Process Model.

ing the process of loading it, it tries to import cer-

tain ontologies that are no longer available online.

• The loading time of OEMA when tracking its use

gets too big (287516 ms) without any reason.

• OEMA does not reason with HermiT because the

reasoner does not support built-in atoms, some-

thing that the ontology has.

• Open Energy Ontology does not reason with Pel-

let because of issues with the transitive rules.

• Facility Ontology does not reason with HermiT

nor Pellet. For the ﬁrst it states that there is a Non-

simple property at Cardinality restriction. For

the latter, it ﬂags an error related to the Transi-

tiveObjectProperty axiom. Let us state that it ac-

tually reasons with ELK (with an average time of

2576.4ms) after several warnings.

• For DABGEO ontology an error happens while

running the Pellet reasoner and the process has to

be killed. It is due to trying to use a literal as an

individual.

5 TECHNICAL RESULTS

As the ﬁgures and tables show, there is a trend be-

tween the ontologies OEMA and SEPA, as OEMA

0 1 2 3 4

·10

OEMA

SEPA

Open Energy

Facility

DABGEO

7,241.2

12,284.6

1,296.2

13,077.2

1,973.2

1,106

331.8

56,051.2

26,406.6

788.2

Time (in ms)

Loading Time

Reasoning Time (HermiT)

Reasoning Time (Pellet)

Figure 2: Performance comparison of ontologies.

A Comparison of Smart Grids Domain Ontologies

119

appears to be the one with the highest count of Ax-

ioms and Classes, while SEPA is the one with the

lowest number of those. Nevertheless, there is no rea-

son to believe that these numbers actually affect per-

formance in any way as, particularly, SEPA appears

as the one with the highest loading time. The only

two metrics that are higher for SEPA in comparison

with OEMA are the Individual Count and the Anno-

tation Property Count. Interestingly enough, the Open

Energy Ontology has a higher Individual Count than

SEPA (actually, the highest of them all), but at the

same time, the Open Energy Ontology has the lowest

loading time of all the ontologies. Therefore, if we

were to link the high loading time of SEPA to any-

thing, it, necessarily, would be linked to the Annota-

tion Property Count, where SEPA has the highest of

all the ontologies.

Figure 3: OEMA Performance.

When looking to the reasoning time while using

HermiT we found that DABGEO has the highest time

by a wide margin, while the Open Energy Ontol-

ogy has the lowest. When comparing the metrics of

ontologies we discover that DABGEO has a higher

count on all categories except for Individual Count

and Annotation Property Count. We can assume that

having a higher value in most of the metrics affects the

performance of DABGEO within the reasoner. The

fact that aside these two ontologies only SEPA is able

to reason with HermiT sheds some light into the An-

notation Property Count issue that we pointed out for

SEPA and its high loading time. It comes to show that,

despite taking longer to load, its performance with re-

spect to the reasoner is way better than that of DAB-

GEO. We could conclude that having a high Annota-

tion Property Count damages the performance while

loading the ontology, but not as much while reasoning

with it.

In regards to the reasoning time using Pellet we

found that OEMA has the highest reasoning time

while SEPA has the lowest. This comes in accordance

Figure 4: SEPA’s Ontology Performance.

to the data that we have regarding the metrics of the

characteristic of each ontology: The biggest ontology

takes longer to reason and the smallest takes the less

to reason (as it would seem intuitively before the anal-

ysis). This points out to the possible conclusion that

the size of the ontology actually matters when setting

up an ontology in a time-constrained situation. This is

even more important if we take into account that the

difference in time is an order of magnitude 33 times

larger.

Figure 5: Open Energy Ontology Performance.

Finally, when taking a look at the memory and

CPU usage metrics of each ontology, we found out

that OEMA appears as the more demanding ontol-

ogy. Nevertheless, we are not able to take out any

strong conclusions as there were issues when trying

to plot the OEMA performance. With that in mind,

if we leave OEMA aside, we have found that the

highest memory and CPU consumption is attributed

to DABGEO. This might be enough to point out that

a higher reasoning time, could potentially lead to a

higher resource cost. Let us remember that DABGEO

has the highest reasoning time of any ontology on any

reasoner. Therefore, when dealing with a resource-

constrained platform, we need to take into account the

reasoning time so less resources are required.

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

120

Figure 6: Facility Ontology Performance.

On the other hand, the less resource-consuming

ontology is the Open Energy Ontology. This might

come across as unexpected, as the ontology only met-

rics that are to be highlighted are the Data Property

Count, it has none, and the reasoning time. The only

conclusion that we could draw is that, as we pointed

above, the reasoning time is linked directly to the re-

source usage.

Figure 7: DABGEO Performance.

6 CONCLUSIONS

In this paper we have shown the state of the art regard-

ing semantic web reasoning in the domain of SGs.

After ﬁnding the most representative ontologies of

the domain, we have analyzed those that are avail-

able and established a comparison between them, try-

ing to point out possible bottlenecks and points of im-

provement. We have found that a higher loading time

seems to be linked to a higher reasoning time. In the

same way, a high reasoning time seems to be tied to a

higher resource consumption. These conclusions are

specially important in the domain because of their in-

tended implementation in resource-constrained plat-

forms such as smart meters.

About the different ontologies analyzed in the pa-

per, we can point out that the most resource consum-

ing and higher metrics in general is OEMA, while in

the other side of the spectrum we ﬁnd Open Energy

Ontology as the less resource-consuming ontology,

and SEPA as the ontology with the lowest metrics

overall. Despite that, regarding loading times, SEPA

is the one with the highest and Open Energy Ontology

the one with the lowest time. The highest time when

reasoning is that of DABGEO, while the lowest is the

one of the Open Energy Ontology.

Furthermore, it is worth noting that, despite an al-

most overﬂowing amount of references on the seman-

tic web technologies in the domain of the SGs, most

of these technologies are not readily available: There

are many broken URLs, ontologies that are no longer

available, outdated platforms. It is even more evident

when, after ﬁnding the few available ontologies, these

will not work properly with some staples of the se-

mantic web reasoning such as Prot

At this point, there are many possible lines on

which to expand the work already done. We intend

to provide a solid common platform, based on the

Jena

framework, to extend the comparison beyond

the readily available tools. We also want to pro-

vide approaches to improve the loading and reasoning

times as well as the resource load of the available on-

tologies. Finally, it is our intention to also provide a

comparison of the performance of the ontologies and

reasoning methods in a real-world environment such

as a resource-constrained device.

ACKNOWLEDGEMENTS

The research was supported from ERDF/ESF ”Cy-

berSecurity, CyberCrime and Critical Informa-

tion Infrastructures Center of Excellence” (No.

CZ.02.1.01/0.0/0.0/16 019/0000822).

REFERENCES

Ali, S. and Kiefer, S. (2009). µOR – A Micro OWL DL

Reasoner for Ambient Intelligent Devices. In Abden-

nadher, N. and Petcu, D., editors, Advances in Grid

and Pervasive Computing, Lecture Notes in Computer

Science, pages 305–316, Berlin, Heidelberg. Springer.

Atanasov, I. I. (2015). Structure of semantic information for

prepaid smart metering. In 2015 12th International

Conference on Telecommunication in Modern Satel-

lite, Cable and Broadcasting Services (SIKS), pages

293–296.

https://jena.apache.org/index.html

A Comparison of Smart Grids Domain Ontologies

121

Bruinenberg, J., Colton, L., Darmois, E., Dorn, J., Doyle, J.,

Elloumi, O., Englert, H., Forbes, R., Heiles, J., Her-

mans, P., et al. (2012). Cen-cenelec-etsi smart grid

co-ordination group smart grid reference architecture.

CEN, CENELEC, ETSI, Tech. Rep, pages 98–107.

Catterson, V. M., Davidson, E. M., and McArthur, S. D.

(2005). Issues in integrating existing multi-agent sys-

tems for power engineering applications. In Proceed-

ings of the 13th International Conference on, Intel-

ligent Systems Application to Power Systems, pages

6–pp. IEEE.

Cuenca, J., Larrinaga, F., and Curry, E. (2017). A Uniﬁed

Semantic Ontology for Energy Management Applica-

tions. In WSP/WOMoCoE@ISWC.

Cuenca, J., Larrinaga, F., and Curry, E. (2020). Dab-

geo: A reusable and usable global energy ontology

for the energy domain. Journal of Web Semantics, 61-

62:100550.

Dogdu, E., Murat Ozbayo

glu, A., Benli, O., Akınc¸, H. E.,

Erol, E., Atasoy, T., Gurec¸, O., and Ercin, O. (2014).

Ontology-centric data modelling and decision support

in smart grid applications a distribution service oper-

ator perspective. In 2014 IEEE International Confer-

ence on Intelligent Energy and Power Systems (IEPS),

pages 198–204.

Donohoe, M., Jennings, B., and Balasubramaniam, S.

(2015). Context-awareness and the smart grid: Re-

quirements and challenges. Computer Networks,

79:263–282.

Farhangi, H. (2009). The path of the smart grid. IEEE

power and energy magazine, 8(1):18–28.

Flores-Martin, D., Berrocal, J., Garc

ıa-Alonso, J., Canal,

C., and Murillo, J. M. (2019). Enabling the Inter-

connection of Smart Devices Through Semantic Web

Techniques. In Bakaev, M., Frasincar, F., and Ko, I.-

Y., editors, Web Engineering, Lecture Notes in Com-

puter Science, pages 534–537, Cham. Springer Inter-

national Publishing.

Gillani, S., Laforest, F., and Picard, G. (2014). A Generic

Ontology for Prosumer-Oriented Smart Grid. CEUR

Workshop Proceedings, 1133.

Hippolyte, J. L., Howell, S., Yuce, B., Mourshed, M.,

Sleiman, H. A., Vinyals, M., and Vanhee, L. (2016).

Ontology-based demand-side ﬂexibility management

in smart grids using a multi-agent system. In 2016

IEEE International Smart Cities Conference (ISC2),

pages 1–7, Trento, Italy. IEEE.

Hippolyte, J. L., Rezgui, Y., Li, H., Jayan, B., and How-

ell, S. (2018). Ontology-driven development of web

services to support district energy applications. Au-

tomation in Construction, 86:210–225.

Kang, Y.-B., Li, Y.-F., and Krishnaswamy, S. (2012). Pre-

dicting Reasoning Performance Using Ontology Met-

rics. In Cudr

e-Mauroux, P., Heﬂin, J., Sirin, E.,

Tudorache, T., Euzenat, J., Hauswirth, M., Parreira,

J. X., Hendler, J., Schreiber, G., Bernstein, A., and

Blomqvist, E., editors, The Semantic Web – ISWC

2012, Lecture Notes in Computer Science, pages 198–

214, Berlin, Heidelberg. Springer.

Kang, Y.-B., Pan, J. Z., Krishnaswamy, S., Sawangphol,

W., and Li, Y.-F. (2014). How long will it take? ac-

curate prediction of ontology reasoning performance.

In Proceedings of the Twenty-Eighth AAAI Confer-

ence on Artiﬁcial Intelligence, AAAI’14, pages 80–

86, Qu

ebec City, Qu

ebec, Canada. AAAI Press.

Li, Y., Garc

ıa-Castro, R., Mihindukulasooriya, N.,

O’Donnell, J., and Vega-S

anchez, S. (2019). Enhanc-

ing energy management at district and building levels

via an EM-KPI ontology. Automation in Construction,

99:152–167.

Maarala, A. I., Su, X., and Riekki, J. (2017). Semantic Rea-

soning for Context-Aware Internet of Things Applica-

tions. IEEE Internet of Things Journal, 4(2):461–473.

Conference Name: IEEE Internet of Things Journal.

na, P., Trillo-Lado, R., Hoyo, R., Rodr

ıguez-Hern

andez,

M., and Abad

ıa, D. (2020). Ontology-quality Evalu-

ation Methodology for Enhancing Semantic Searches

and Recommendations: A Case Study:. In Proceed-

ings of the 16th International Conference on Web

Information Systems and Technologies, pages 277–

284, Budapest, Hungary. SCITEPRESS - Science and

Technology Publications.

Salameh, K., Chbeir, R., and Camblong, H. (2019). SSG:

An Ontology-Based Information Model for Smart

Grids, pages 94–124. Springer Berlin Heidelberg,

Berlin, Heidelberg.

Santodomingo, R., Uslar, M., Rodr

ıguez-Mond

ejar, J. A.,

and Sanz-Bobi, M. A. (2015). Rule-Based Data Trans-

formations in Electricity Smart Grids. In Bassiliades,

N., Gottlob, G., Sadri, F., Paschke, A., and Roman,

D., editors, Rule Technologies: Foundations, Tools,

and Applications, Lecture Notes in Computer Science,

pages 447–455, Cham. Springer International Pub-

lishing.

Schachinger, D., Kastner, W., and Gaida, S. (2016).

Ontology-based abstraction layer for smart grid inter-

action in building energy management systems. In

2016 IEEE International Energy Conference (ENER-

GYCON), pages 1–6.

Speiser, S., Wagner, A., Raabe, O., and Harth, A. (2013).

Web technologies and privacy policies for the Smart

Grid. In IECON 2013 - 39th Annual Conference of

the IEEE Industrial Electronics Society, pages 4809–

4814. ISSN: 1553-572X.

Tai, W., Keeney, J., and O’Sullivan, D. (2011). COROR:

A COmposable Rule-Entailment Owl Reasoner for

Resource-Constrained Devices. In Bassiliades, N.,

Governatori, G., and Paschke, A., editors, Rule-Based

Reasoning, Programming, and Applications, Lecture

Notes in Computer Science, pages 212–226, Berlin,

Heidelberg. Springer.

Wang, T. D. and Parsia, B. (2007). Ontology Performance

Proﬁling and Model Examination: First Steps. In

Aberer, K., Choi, K.-S., Noy, N., Allemang, D., Lee,

K.-I., Nixon, L., Golbeck, J., Mika, P., Maynard, D.,

Mizoguchi, R., Schreiber, G., and Cudr

e-Mauroux, P.,

editors, The Semantic Web, Lecture Notes in Com-

puter Science, pages 595–608, Berlin, Heidelberg.

Springer.

Wemlinger, Z. and Holder, L. (2011). The COSE Ontology:

Bringing the Semantic Web to Smart Environments.

WEBIST 2021 - 17th International Conference on Web Information Systems and Technologies

122

In Abdulrazak, B., Giroux, S., Bouchard, B., Pigot,

H., and Mokhtari, M., editors, Toward Useful Ser-

vices for Elderly and People with Disabilities, Lecture

Notes in Computer Science, pages 205–209, Berlin,

Heidelberg. Springer.

Yu, X., Cecati, C., Dillon, T., and Simoes, M. G. (2011).

The new frontier of smart grids. IEEE Industrial Elec-

tronics Magazine, 5(3):49–63.

Zhou, Q., Natarajan, S., Simmhan, Y., and Prasanna, V.

(2012). Semantic Information Modeling for Emerg-

ing Applications in Smart Grid. In 2012 Ninth Inter-

national Conference on Information Technology - New

Generations, pages 775–782.

A Comparison of Smart Grids Domain Ontologies

123