Smart Grid and Renewable Energy, 2009, 8–12
Published Online September 2009 (http://www.SciRP.org/journal/sgre/).
Information Services for Smart Grids
ABSTRACT
Interconnected and integrated electrical power systems, by their very dynamic nature are complex. These multifaceted
systems are subject to a host of challenges-aging infrastructure, generation availability near load centers, transmission
expansion to meet growing demands, distributed resources, dynamic reactive compensation, congestion management,
grid ownership vs. system operation, reliability coordination, supply and cost of natural resources for generation, etc.
Other types of challenges facing the industry today include balancing between resource adequacy, reliability, econom-
ics, environmental constraints, and other public purpose objectives to optimize transmission and distribution resources
to meet the needs of the end users. The goal is to provide a vision for a comprehensive and systematic approach to
meeting the grid management challenges through new information services. These services will have as the heart of
their data streams a sensor web enablement that will make the grid a part of the semantic web.
Keywords: Protection and Control, Ontology, Information Semantics, Sensor Web
1. Introduction
Since it is not possible to completely prevent blackouts,
t hen effective and fast power system restoration is nec-
essary
t o minimize the impact of major disturbances.
This requires rapid decision in a data rich, but informa-
tion limited environmen
t . The streams of data from a
variety of sensors do not prov
i de system operators with
the necessary information to act on
i n the timeframes
necessary to minimize the impact of a disturbance. Even
if there are fast models that can convert
t he data into in-
formation, the system operator must deal with
t he chal-
lenge of not having a full understanding of the context of
the information and, therefore, the information content
canno
t be used with any high degree of confidence.
Some of the key elements for response in smart grids ar
e:
 Well-defined procedures that require over
all coordina-
tion within the affected area, as well as w
it h the
neighboring gr
i ds.
 Reliable and efficient software tools to aid opera t ors
and area coordinators in executing dynamic con
t ro l
procedures and in making the right decisions.
 Control solutions reducing the overload and instability
risks during recovery.
Today’s technology allows improved processes and
smar
t systems to aid in decision-making to minimize
impacts of outages (spatially and temporally). Standard
opera
ti ng procedures, based on pre-defined system con-
ditions and operating parameters, can be provided via a
set of power system information services. For example,
rapid restoration or minimizations of outages by selected
islanding are options for consideration in minimizing the
consequences of an outage t o a user.
Information services are focused on providing the righ t
information at the right moment to the right decision
maker. High-level operational information services (i.e.,
ac
ti onable intelligence) are often needed along with sup-
portive sensor data or trends to provide context. The in-
formation serv
i ces required by grid operators could vary
from scenar
i o development to estimates of socio-econo-
mic impacts of failures to quantitative statistics, trends
and forecasts. These services also must be available in a
geospatial context and a
t various temporal scales to sup-
port the needs of system operators, planners, and regula-
tory agencies. Inform
ati on services must be characterized
by a strong integration of gr
i d data with ancillary data
and information, and this will require a knowledge based
approach for capturing the best practices of utilities and
regulators. The complexity of these informa
t ion services
will require a network of partners who will contr
i bu t e to
the production of the services. To facilitate these services
it will be incumbent upon the power research community
t o develop tools to facilitate operational data acquisition
and handling in interoperable formats and to create i n-
form
ati on products through a coordinated process chain.
The successful conversion of power sensor data into ac-
tionable in
t elligence will require the integration of power
system expertise in modeling, data management and ser-
vice delivery to descr
i be the state of the grid and to pre-
dict responses to actual and potential change.
2. Information Semantics
Information semantics is an innovative approach for han-

Information Services for Smart Grids 9
d li ng complex systems. This approach can be used to
accompl
i sh knowledge discovery and to provide decision
support to gr
i d operators by focusing on making ma-
chines more
cl osely interact at human conceptual levels.
These software tools are based on semantics and ontolo-
gies that use web-addressable sensors, World Wide Web
Consortium (W3C) standards, such as the Ontology Web
Language (OWL), and standard
i zed markup languages
(e.g., sensorML).Information semant
i cs provides mean-
ings for systems, data, documents, agents and spans on-
tologies, knowledge representations, semantic web, natu-
ral language processing, and knowledge management.
Figure 1 represents the overall data flow from meas-
uremen
t s (power flow, voltage) to actionable intelligence
(open breaker x) on the grid via the Semantic Web. The
Semantic Web makes the meaning of information acces-
sible not only
t o humans, but also to machines. A spe-
cific instance of
t he Semantic Web is a sensor web en-
ablement (SWE) that makes sensors addressable over the
web and provides an ex
t ens i ve monitoring and sensing
system that contributes
ti me l y, comprehensive, continu-
ous, and multi-mode observations for the grid. SWE fa-
cilitates flexible data discovery, access, tasking and pro-
viding alerts based on open standards. The semantic web
is an extension of the current web in wh
ic h information
is given well defined meaning, better enabl
i ng computers
and people to work in cooperation. Ontology is a knowl-
edge representation scheme that provides the glue
t o hold
everything together and defines the terms used
t o de-
scribe as well as represent an area of knowledge. It de-
fines the vocabulary and the meaning of the vocabulary
in contex
t . This discussion will focus on results from a
sensor web enablement project for power sensors and
how these are used as the basis for information services
for the system operator.
3. Timing Issues Related to Services
Figure 2 shows the timing of events in the electric power
grid and the reaction times commonly available for either
an automated or operator intervention. At one end of the
spectrum are the actions taken by the power system pro-
tection equipment to take an automated response based
on measured quantities (e.g., frequency, current). At the
other end of the spectrum, the operator controls the sys-
tem in a steady state mode using data acquired from a
host of sensors via a SCADA system. Actions may be
automated or are more often made based upon an opera-
tor’s visual interpretation of the data presented through a
variety of meters and display dev
i ces.
In steady state operations, an operator normally has
adequa
t e time to consider the data, consult text based
help guides, or seek another operator’s opinion before
having to make a decision. In between these two ends of
the spectrum is a
ti me in which an operator may have to
make a decision based simply on heuristics or past ex-
Figure 1. Process model to change data to action
Figure 2. Critical timing and reaction times for power grid
operations
periences.Obviously,
t hese actions often may not result
in the best consequence. This is the critical time period in
which immediate actions must be made by an operator to
prevent wide area collapses of the gr
i d.
To ensure the secure and stable operation of the power
system across the temporal spectrum, it is required to de-
velop and apply new decision support tools that provide
acti onable intelligence in the required timeframe. In the
aspects of secure and stable control, we need to think of an
automatic p
il o t power system concept, representing a trend
to improve Energy Management Systems (Figure 3).
In other aspects, such as emergency control, restora-
ti on control and etc., multiple services are required to be
harmoniously interconnected into a multi-agent system
t o
perform calculations, analyses, and be able to crea
t e ac-
tionable intelligence to support auto-pilot opera
t ion.
4. Information Semantics and O ntology
Applications in Power System of the F uture
The present power system operation environment is pr i-
mari l y composed of many distributed tools and compo
nents. The interfaces of the various components must be
so standard
i zed to work in a “plug and play” concept
similar to the hardware used in the substation automation.
For modeling and
t ools, there are two major tasks;

Information Services for Smart Grids 10
Figure 3. Steady state, transient state and dynamic state
representations of control on the grid
namely data sharing amongst applications, and integra-
tion of functions of various
t ools. Data sharing stream-
lines the process of validation and requ
i res a common
data format for exchanging data amongs
t appl icati ons.
Today, each application uses a different format or dif-
ferent system topology and level of details based on dif-
ferent applications. For example, short circuit programs
use intermediate busses to reflect physical construction
of line and tower geometry. Data exchange amongst dif-
feren
t applications such as (Power Flow to EMS), and
short circuit
t o power flow is paramount. Likewise, data
exchange be
t ween similar applications from different
power companies using different manufacturer products
(e.g.: Planning, Opera
ti on, Transient, and short circuit
program) is vital. Semant
i c heterogeneity is a major det-
riment to exchange of data among the various services
necessary for auto-p
il ot.
The next large scale task (that should start today) is to
integrate the functions of the individual services so that
t hey can perform more complicated functions. To realize
t he harmonious interconnection of these information ser-
vices, we must start from the data and information level,
and work in three aspe
ct s:
1) S
t andard i za ti on of the data model. This effort i s ad-
dressed by IEC 61970 standard. The standard addresses
requirements for transmitting
i nform ati on among EMS
software or different EMS products. EPRI also has fo-
cused on the data standardization re l a t ed with EMS ap-
pl icati ons.
2) Sharing of clear and precise data and i nform ati on
among software. This approach leads to higher work
i ng
efficiency and expansib
ilit y.
3) Standardize the semantics (terminology) and de-
ve
l op taxonomies for ontology drives decision support
t ools for the electric gr i d.
4.1 Semantics Driven Knowledge Discovery
Know l edge discovery (features, complex relationships,
and hypotheses that describe potentially interesting regu-
larities) from la rge heterogeneous networks of observ-
tions and
i nform ati on products generated from modeling
efforts are essential for protection and control of the elec-
tric grid. Domain sp
eci f i c knowledge building is required
before it can be discovered and shared. Data from vari-
ous sources (e.g. PMUs, SCADA,
et c.) are transformed
into information at different appl
i ca ti on domain data
analysis centers (research centers, universi
ti es, etc.) and
eventually into actionable intelligence. However, to
achieve this, middleware is required that provides tools
t o browse and access the data resources for resolv i ng
heterogeneity problems.
Middleware, by its most general definition, is any pro-
gramming that serves to “glue together” or med i a t e be-
tween separate and often already existing programs. Th i s
layer is introduced in systems to hide the heterogeneity
of
t he underlying components or applications and pro-
vide un i form access to their functions. In short, middle-
ware fac ilit a t es interoperability by hiding low-level ac-
cess and prov i d i ng standard services. To provide the ser-
vices necessary for an electric grid decision support sys-
tem middleware will need
t o be developed that provides
functionalities for on
t o l ogy management, storage, query,
and inference services. It w
ill also need to be designed to
enable resource discovery and
t o create semantic meta-
data. Such an ontology midd
l eware system serves as a
flexible and extendable platform for knowledge man-
agement solu ti ons.
Domain specific knowledge building is achieved
t hrough ontological modeling that provides functional-
ities for capturing knowledge. Ontology is “a shared,
form al conceptualization of a domain” [1]. The ontology
m i ddleware system serves as a flexible and extendable
platform for knowledge management solutions, Figure 4.
The m i ddleware also serves to hide the ontology sources
from domain specif i c application cli ents.
The seamless access to real time or near real time sen-
sor data is constrained by varying characteristics (physi-
cal/logical) of the sensor networks. This results
i n :
 Data sets stemming from the same data-source w it h
unequal updating per
i ods.
 Data sets represented in the same data-model, bu t ac-
quired by different opera
t ors.
 Data sets which are stored in similar, but not iden ti ca l
data-mod el s.
 Data sets from heterogeneous sources (across geo-
graphical boundaries), which differ in da
t a-model i ng,
scale, thematic content, contexts, meaning, et c.
The resolution of such conflicts depends on the recon-
c
iliati on of both syntactic and semantic heterogeneities in
the da t a. Syntactic standardization (interoperability) has
long been proposed and a number of metadata standards
(Feder
al Geospatial Data Committee, International S t an-
dards Organization, etc.) have been developed world-w-

Information Services for Smart Grids 11
Figure 5. Application services require metadata interpret-
able through ontologies for resource discovery
oped as a layer on top of the existing syntactic meta-
data s
t andard.
Figure 4. Ontology and information semantics in power
system applications
The importance of resolving semantic differences has
recent
l y gained wide attention resulting in the next gen-
era
ti on semantic web efforts, which is largely due to the
progress in techniques to model, capture, represent and
reason about semantics; gradual progress in attention
from data
t o information, and increasingly towards
knowledge acquisi
ti on and management.
ide dur i ng the last decade, which are now widely accepted
as
t he standardized models for both data and metadata.
Each of these standards originated in one particular com-
munity and was quickly adopted in a variety of domains.
Although
t he metadata standards support to a large extent
t he interoperability of the data, a problem that is still no t
completely solved is diversity (heterogeneity) in the proc-
ess of converting this data into information and ac ti onab l e
intelligence. In general, the data heterogeneity problems
can be divided into three categories [2].
Semantic interoperability requires resolving various
con
t ex t -dependant incompatibilities. The context refers to
the knowledge that is required to reason about the system
for
t he purpose of answering a specific query [2].
Therefore, it is important to provide contextual knowl-
edge of the pro
t ec ti on and control domain applications in
order to ensure seman
tic interoperability. Each informa-
tion source serves as a con
t ex t for the interpretation of
the information contained
t her ei n. This view implies that
an information entity can only be completely understood
within its context and we need to f
i nd ways to preserve
the contextual information in the
t ran sla ti on process.
 Syntactic heterogeneity is caused by different l ogica l
models (e.g. relational vs. object oriented) or due t o
different geometric representations (raster vs. vec t or).
 Schematic heterogeneity occurs because of differen t
conceptual data models (e.g. objects in one da t abase
considered as properties in another, differen t generali-
zation hierarch i es).
 Semantic heterogeneity raises most serious informa-
t ion integration problems. It occurs because of the
differences in meaning, interpretation or usage of the
same or re lat ed data.
Assuming that ontologies are used to capture the
context of the information entities, then as we move from
one context
t o another there is a requirement to integrate
ontologies. A hybrid ontology approach consisting of a
global shared ontology that encompasses all the local
application le ve l ontologies for a domain of interest (e.g.,
protection)
i s proposed. Recent studies [3] have sug-
gested the advan
t ages of this approach to be :
One approach to the problem domain is by modeling
t he semantics of the data, instead of just relying on the
syntac ti c and structural representations. Our proposed
approach i s distinguished from other existing approaches
in the fo ll ow i ng manner :  New sources can be added easily without the need of
modif
icati on.
 Heterogeneous sensor data sets integrated t hrough on-
tology based approaches and intelligent reasoning over
the acquired knowledge base that enables access t o
content instead of just keyword based searches.
 Supports acquisition and evolution of on t o l ogy.
Understanding of the complex interrelationships within
t he electrical power system necessitates the exploration
of strategies for innovative acquisition, integration, and
da
t a exploitation technologies for fully interchangeable,
ti me l y, and accurate data analysis and creation of ac ti on-
able intelligence. Sharing of the generated datasets,
informa
t ion, and results, between geographically distrib-
uted organiza
ti ons often proves to be challenging. This is
due to the comp
lic a t ed steps involved in data discovery
 Use of real or near real time data derived from
sensor networks (e.g., SCADA).
 Semantics and content-based data extraction and inte-
gration from a host of sensors (sensor web).
 Involve industry and expert groups to propose and
evolve a Power and Control Semantic Metadata Stan-
dard (Figure 5), which we envision would be devel-

Information Services for Smart Grids 12
Figure 6. Architecture of the power grid sensors integration through semantics enabled middleware
and conversion that resul
t from the problems of syntactic,
structural, and semant
i c heterogeneity in the datasets. The
syntactic heterogenei t y problems have been addressed to
some extent by
t he standardization of metadata as advo-
cated by mu
lt ip le organizations. However, the lack of
sufficient description of the meaning of the data along
with a context may lead to
t he misinterpretation of data
by users who are not involved in
t he original data acqui-
sition process. Thus, semant
i c reconciliation is necessary
to guarantee meaningful da
t a sharing. Figure 6 shows a
high level visualization of t he layered architecture for
electric grid protection and contro
l utilizing a sensor web
enablement approach.
5. Conclusions
Some of the concepts suggested in this paper about u til-
iz
ing information semantics and integrating data from a
myriad of sensors will be required to maintain social and
env
i ronmen t a l obligations for the electric utility industry.
The protection and control will be instrumental to achieve
the re
lia b ilit y, efficiency, and financial aspects of the 21st
century gr i d.
REFERENCES
[1] T. R. Gruber, “A translation approach to portable ontol-
ogy specifications,” in Knowledge Acquisition, Vol. 5,
pp. 199–220, 1993.
[2] S. Ram, J. Park. Semantic Conflict Resolution Ontol-
ogy(SCROL): An ontology for detecting and resolving
data and schema-level semantic conflicts, IEEE Transac-
tions on Knowledge and Data engineering, Vol.16, No.2
February, 2004.
[3] H. Wache, T. Vögele, U. Visser, H. Stuckenschmidt,
G.Schuster, H. Neumann, and S. Hübner, “Ontol-
ogy-based Integration of Information-A Survey of Exist-
ing Approaches,” Proc. of the IJCAI–01 Workshop on
Ontologies and Information Sharing, Seattle, WA, pp.
108–117, 2001.