See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/224142796

Software architecture for Smart Metering systems with Virtual Power Plant

Conference Paper · May 2010

DOI: 10.1109/MELCON.2010.5476237 · Source: IEEE Xplore

CITATIONS READS
20 179

4 authors, including:

Aleksandar Erdeljan Filip Kulic

University of Novi Sad University of Novi Sad
47 PUBLICATIONS   211 CITATIONS    53 PUBLICATIONS   363 CITATIONS   

SEE PROFILE SEE PROFILE

Slobodan Lukovic
University of Lugano
28 PUBLICATIONS   285 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

SCCER - FURIES View project

Swiss Competence Center for Energy Research on the Electrical Infrastructure (SCCER-FURIES) View project

All content following this page was uploaded by Slobodan Lukovic on 26 August 2014.

The user has requested enhancement of the downloaded file.

Journal of Scientific & Industrial Research
VUKMIROVIC et al : A NOVEL SOFTWARE ARCHITECTURE FOR SMART METERING SYSTEMS 937
Vol. 69, December 2010, pp. 937-941

A novel software architecture for smart metering systems

Srdjan Vukmirovic*, Aleksandar Erdeljan, Imre Lendak and Darko Capko
Computer and Control Department, University of Novi Sad, Trg Dositeja Obradovica 6, Novi Sad, Serbia, Europe

Received 29 June 2010; revised 24 September 2010; accepted 28 September 2010

For smart metering systems (SMSs), this study presents a novel architecture, which allows simple, secure and fast
integration of SMSs with other software products. Proposed architecture that combines better features of two models [Common
Information Model (CIM) and OPC’s Unified Architecture (UA)] has proven in practice as a good basis for a commercial Meter
Data Management System. In a case study, created using data from a pilot project, 7 experiments determined memory usage of OPC
UA and CIM servers, duration of exporting data out of CIM and importing them to OPC UA, and size of XML export file.

Keywords: Common Information Model (CIM), OPC Unified Architecture (UA), Smart metering

Introduction may go up to twice. Collaboration and integration in SMSs

Energy efficiency is extremely important in generation play an important role in choosing products2 . Existing
and transmission (Energy Management Systems – EMS), SMS architectures are either highly specialized and solve
power distribution (Distribution Management Systems – only a single problem or, they are high level architectures
DMS) and for each energy consumer. New, smart meters barely touching integrations while data models and
allow individual consumers to take part in process of integration interfaces are not defined3,4 . Apart from
energy saving. These meters, which can be turned on or monetary and operational benefits of SMSs, new
off remotely, become an important part of a complex architecture also brings some social and environmental
and general supervisory and control system1 . Primary gains 5,6 .
goal of smart metering systems (SMSs) is to provide billing IEC’s Common Information Model (CIM) was
information for power distribution utilities. Data gathered originally developed for EMS. There are studies on XML
could also be used to better inform consumers about based data exchange, which allow optimal speed and ease
consumption and price. Smart metering software should of extension7,8 . Studies9,10 present various ways to adapt
have latest security policies and standards incorporated, CIM for use with Web services and Enterprise Service
since it handles all consumer personal information data Buses (ESB)9,10. OPC’s specification on Data Access
that must be protected from potential misuse. Software (DA) has not only been used in process industry but also
is expected to work without interruption (twenty four in telecommunications 11,12 and simulations13 . With Unified
hours, seven days a week – 24/7); thus all components Architecture (UA), OPC Foundation defines a completely
must be exceptionally reliable and subject to fault new architecture and it bases communications on service-
tolerance. Further, software must exchange data with other oriented architectures. Although UA is a relatively new
subsystems and applications used by power utilities. specification, there are proposed methods for migrating
Although smart metering area is relatively new, there existing DA servers to UA14 . Most UA related studies15,16
are already reports on role of Advanced Metering deal with integration of UA with existing industrial
Infrastructure (AMI) systems as a key element providing standards, while others propose ways to enhance
savings in power distribution systems1 . When AMI architecture17,18 and security of communications 19 . To
systems are integrated with other subsystems, savings use CIM and OPC UA data together, one study20 proposes
wrapping of OPC UA functions around existing CIM
*Author for correspondence server, whereas other study21 suggests for CIM server to
E-mail: [email protected] use OPC UA server as a data source.
938 J SCI IND RES VOL 69 DECEMBER 2010

Connection between central MDM systems and different

Other
Other AMIs is implemented through specialized software
OMS CMS DMS WMS systems
systems components, so-called Head End Systems. Data collected
by meters become more and more important for other
Enterprise Integration Bus subsystems (Outage Management Systems – OMS,
Customer Management System - CMS, Distribution
Management System - DMS and Work Management
MDM
System WMS). Their integration with MDMs can be
implemented through an Enterprise Service Bus (ESB),
Head End ...
Head End which uses data model and interfaces proposed
System System architecture (Fig. 1).
AMI AMI
... Integration
AMI AMI One of the most complex areas in distributed systems
Infrastructure Infrastructure
is integration of components. Various software producers
develop components and produce them in different time
Concentrator Concentrator
frames. Therefore, it is extremely important to clearly
define interfaces and data formats used for communication.
... It is likely that data models and interfaces will often
undergo changes in future. At the same time, Smart
Metering is becoming an increasingly relevant part of new
Smart Smart
meter meter systems since (almost) every other service will consume
data gathered from these meters. Thus, for standardization
of this area, following two widely accepted industry
Fig.1 — Open smart metering architecture standards identified are: i) OPC UA, which covers a wide
industrial area and defines general interfaces; and ii) IEC’s
CIM, which defines a data model for power generation,
This study presents a novel way of integrating smart transmission and distribution. Data exchange is performed
metering solutions with other subsystems in power through specialized interfaces.
distribution systems using IEC’s CIM and OPC’s UA.
Data Model
Proposed Architecture Part 3 of UA specification defines an object model
SMSs gather measurement values from distant suitable for modeling a wide range of different industrial
measurement devices (meters) and integrate them into a systems 22 . Base class of this object model is Node, which
complex and unified system for billing. Data can also be contains identifiers, name, description and path. Apart from
used for acquisition and control of power distribution base class, other two additional classes of interest for
systems. By introducing new meters, it is now possible MDM systems are ObjectNode (base class for all physical
to control consumption per individual consumer. Such a objects) and VariableNode (base class for anything
progress makes distribution management systems more measurable).
complex. Thus, integration with other subsystems and a CIM, in comparison to UA, is a more specialized
clearly defined data flow between them are important system; it defines separate object models for power
issues. generation, transmission and distribution systems. Although
Since a single utility controlling a power distribution CIM has numerous classes for modeling various
system can purchase different meters from different components of a power system, this study focused on the
producers, it is necessary to develop central systems, classes (IdentifiedObject, MeterAsset, MeterReading,
which would gather data from various AMI systems. Reading and MeasurementValue) used in SMSs. Base class
Central systems are Meter Data Management (MDM) in CIM is IdenifiedObject, which is very similar to UA’s
systems and they not only gather data, but also perform Node; it contains a unique identifier, name, description
data processing (validation, estimation, editing – VEE). and path. CIM models meter objects by MeterAsset class,
 VUKMIROVIC et al : A NOVEL SOFTWARE ARCHITECTURE FOR SMART METERING SYSTEMS 939

Node

VariableNode +BrowseName: QualifiedName

+Description: LocalizedText
+ArraySize: int +DisplayName: LocalizedText
+AccessLevel: byte +Handle: object
+DataType: NodeId +NodeClass: int
+MinimumSamplingInterval: int +NodeId ObjectNode
+StatusCode: StatusCode +References
+Timestamp: DateTime +Text: string +TypeId: ExpandedNodeId
+TypeId: ExpandedNodeId +TypeId: ExpandedNodeId +IsAbstract: bool
+UserAccessLevel: byte
+Value: Variant
OPC UA

CIM
Asset
+category: String
+corporateCode: String
+utcNumber: String
+serialNumber: String
MeasurementValue +lotNumber: String
+timeStamp: AbsoluteDateTime +manufacturedDate: AbsoluteDateTime
+sensorAccuracy: PerCent +purchasePrice: Money
+critical: Boolean
MeterReading +application: String
+installationDate: AbsoluteDateTime
+valuesInterval: DateTimeInterval +acceptanceTest: AcceptanceTest
+initialCondition: String
0..* +initialLossOfLife: PerCent
0..*
+status: Status
+MeterReadings

+MeterReadings

0..*
EndDeviceAsset
Reading +Readings
+disconnect: Boolean
+value: Float +MeterAsset 0..1 +loadControl: Boolean
+reverseFlowHandling: Boolean
+demandResponse: Boolean
+Reading 0..1 +metrology: Boolean
MeterAsset
+outageReport: Boolean
+formNumber: String +relayCapable: Boolean
+ReadingQualities
0..* ReadingType +kR: Float +readRequest: Boolean
+kH: Float +dstEnabled: Boolean
+kind: ReadingKind +timeZoneOffset: Minutes
ReadingQuality +unit: UnitSymbol +amrSystem: String
+multiplier: UnitMultiplier
+quality: String +intervalLength: Seconds
+reverseChronology: Boolean
+defaultValueDataType: String
+defaultQuality: String
+dynamicConfiguration: String
+channelNumber: Integer

Fig. 2— Unified CIM and UA model for MDM systems

which is derived from Asset class and its functions are UA’s VariableNode. Since CIM IdentifiedObject and
metering, load management, connect/disconnect, UA Node are base classes with same attributes, they
accounting functions, etc 23 . MeterReading class groups can be merged into a single base class. Any additional
a set of values obtained from a meter. Reading class classes can then be derived from this common base class:
models a single measured value, which is derived from MeterReading directly and Asset and MeasurementValue
MeasurementValue and contains a value, precise time of through ObjectNode and VariableNode respectively.
measurement and its quality. Another potential way of merging CIM and UA is to
When merging class hierarchies of two models, UA create more generic classes, out of which both different
classes are placed higher on the scale as they are a lot standards’ classes could be derived. However, such an
more generic than their CIM counterparts (Fig. 2). All approach would result in a more complicated data model
physical entities are derived from UA’s ObjectNode entails and the gain would be minimal because UA classes are
that CIM’s Asset should also be derived from this class. very generic and only a few of those fields are not needed
Similarly, CIM MeasurementValue should be derived from in CIM; further, all fields in CIM are optional.
940 J SCI IND RES VOL 69 DECEMBER 2010

Table 1—Experimental results for CIM and OPC UA integration

Number Memory usage CIM export Message of OPC UA Memory usage

of meters of CIM server durations CIM export import durations of OPC UA
MB MB server, MB

1 5.43 0.98 0.64 0.99 5.84

10 6.92 1.62 6.42 1.62 7.38
100 18.9 9.78 64.1 11.1 23.3
500 92.5 36.6 320.1 37.8 88.1
1000 118 39.9 641.2 46.1 160
5000 543 162 3 208 259 782
10000 1082 526 6 415 736 1552

Interface Definition power consumption. Consumption values are read every

UA specifies data access interfaces for current and 15 min during one month. There are 96 readings per
historical values, model modification and searching. All meter per day, and 2880 per month. One month period is
these interfaces are designed to be secure and grouped chosen because most billing reports and validation rules
into service sets and only a selected few of these service are done based on a month period. Following series of
sets are allowed to be implemented. There are a few tests were performed to measure key aspects of system
important differences between UA and OPC integration: i) Amount of time needed to export meters
specifications: i) UA defines a unified data model and from CIM system and to import data to OPC UA system;
thereby simplifies design of UA based systems; and ii) ii) Size of XML based export file that is transferred
UA allows dynamic addition and deletion of measurements between systems; and iii) Memory usage of both CIM
(impossible in previous versions). Tags can be grouped and OPC UA servers to show a difference in memory
into views, and each view can have different access consumption. All experiments were conducted for a
controls. Thus, different parts of model are accessible to different number of meters (from a single meter to
different (data) consumers. Services sets for OPC UA 10 000 meters in the group). Based on results (Table 1),
(SecureChannel, Session, NodeManagement, View, following conclusions could be made: i) Memory
Query, Attribute, Method, MonitoredItem, Subscription, consumption is very similar in both CIM and OPC UA
MeterAsset, MeterReading, and Event) and their purpose servers, hence, there is a small performance impact in
are already reported10, 24, 25. merging the two models; and ii) This kind of integration
Integration server architecture for MDM combines cannot be used in a real-time system since export of large
the best of both models (CIM and UA), works with a scale systems take several minutes.
unified object model and allows effortless integration with
any system, which supports either of the two Conclusions
aforementioned standards. Also, by implementing both Proposed architecture by using UA and CIM for
Request/Response pattern and Event–Listeners approach, SMSs allows rapid development, scalability and easier
this architecture could find a wider range of application maintenance. UA specifications are generic and define a
and it could be easily configured to match small, medium wide array of different functions whereas CIM interfaces
and large power distribution utilities. Also, memory and data model are specifically developed for power
consumption (or other resource consumption) is not systems (which include Smart Metering as well). Data
significantly increased when two data models are merged. model shown in this paper combines better of the two
models (UA and CIM) and ensures that all objects
Results exchanged conform to two standards. Data exchange
Case Study interfaces are designed to conform to both UA and CIM,
A performance test of proposed extension model was allowing generic access as defined in UA to Smart
carried out on meter readings used in a pilot project. Metering data. Proposed architecture was fully
Meters are configured to read only one value – interval implemented within MDM system, which proved its
 VUKMIROVIC et al : A NOVEL SOFTWARE ARCHITECTURE FOR SMART METERING SYSTEMS 941

worth. Experimental results showed that merging of two 12 Hernández S, Ochoa P & Ramírez J, Bluetooth and OPC (OLE
for process control) for the distributed data integration, in Fourth
standards is feasible solution for development of MDM
Congress of Electronics, Robotics and Automotive Mechanics
systems that could be integrated with wide range of power (Morelos) 2007, 27-32.
distribution systems. 13 Santos R A, Normey-Rico J E, Gomez A M, Arconada & Moraga
C, OPC based distributed real time simulation of complex
References continuous processes, Simulation Modeling Practice & Theory,
1 Tram H, Technical and operation considerations in using smart 13 (2005) 525-549.
metering for outage management, in Transmission and 14 Hannelius T, Roadmap to adopting OPC UA, in IEEE Int Conf
Distribution Conf & Expo, ((IEEE PES, New Orleans, USA)) on Industrial Informatics (INDIN 2008) (DCC, Daejeon, Korea)
2008, 1-3. 13-16 July 2008.
2 Matheson D, Jing C & Monforte F, Meter data management for 15 Grossmann D, Bender K & Danzer B, OPC UA based field
the electricity market, 8th Int Conf on Probabilistic Methods device integration, in SICE Ann Conf (Tokyo) 2008 933-938.
Applied to Power Systems (Ames – Iowa) September, 2004. 16 Yamamoto M & Sakamoto H, FDT/DTM framework for field
3 Fiischer R & Schulz N, Information Management for an device integration, in SICE Annu Conf, 2008, 925-928.
Automated Meter Reading System (Michigen Technological 17 Schleipen M, OPC UA supporting the automated engineering of
University, Michigan) 1-5. production monitoring and control systems, ETFA, (2008)
4 LeMay M, Nelli R, Gross G & Gunter C, An integrated 640-647.
architecture for demand response communications and control, 18 Tan V, Yoo D & Yi M, A framework towards OPC web service
in Proc 41st Hawaii Int Conf on System Sciences – 2008 for process monitoring and control, in Int Conf on Advanced
(Waikooa, HI) 2008, 174. Language Processing and Web Information Technology
5 Neenan B & Hemphill R, Societal benefits of smart metering (Dalian Liaoning) 2008, 562-568.
investments, Electricity J, 21 (8) (2008) 32-45. 19 Braune A, Hennig S & Hegler S, Evaluation of OPC UA secure
6 Loveday D, Bhamra T, Tang T, Haines V, Holmes M & Green communication in web browser applications, in 6th IEEE Ind
R, The energy and monetary implications of the ‘24/7’ ‘always Informat Conf (Daejeon) 2008, 1660-1665.
on’ society, Energy Policy, 36 (2008) 4639-4645. 20 Rohjans S, Uslar M & Appelrath H, OPC UA and CIM:
7 McMorran A, Ault G, Elders I, Foote C, Burt G & McDonald J, Semantics for the smart grid, in Transmission & Distribution
Translating CIM XML power system data to a proprietary format Conf and Expo (IEEE PES, New Orleans, USA) 2010, 1-8.
for system simulation, IEEE Trans Power Syst, 19 (2004) 21 Cleveland F, Small F & Brunetto T, Smart Grid: Interoperability
229-235 and Standards (Utility Standards Board) September 2008,
8 Popovic D, Varga E & Perlic Z, Extension of the common 1-28.
information model with a catalog of topologies, IEEE Trans 22 OPC UA Specification: Part 3 - Address Space Model, Version
Power Syst, 22 (2007) 770-777. 1.00 (OPC Foundation, ) July 2006.
9 Mackiewicz R, The benefits of standardized web services based 23 Electrotechnical Commission (IEC) Task Group 13, IEC 61968-
on the iec 61970 generic interface definition for electric utility 11 Ed.1: Application integration at electric utilities - System
control center application integration, in IEEE Power Syst Conf Interfaces for Distribution Management – Part 11: Common
& Expo (Atlanta) 2006, 491-494. Information Model.
10 An introduction to the CIM for Integrating Distribution 24 OPC UA Specification: Part 4: Services. Version 1.01.05 (OPC
Applications and Systems (Electric Power Research Institute, Foundation) February 2007, 19-25.
California) November 2008. 25 Electric Power Research Institute, Enterprise Service Bus
11 Kapsalis V, Koubias S & Papadopoulos G, OPC-SMS: a wireless Implementation Profile, (CIM) Extensions for
gateway to OPC-based data sources, Compu Standards & Distribution, April 2009:http://my.epri.com/portal/
Interfaces, 24 (2002) 437-451. server.pt?space=CommunityPage&cached=true&parents.

View publication stats