Smart grid automation using IEC 61850 and CIM standards

A. Naumann a,n, I. Bielchev b, N. Voropai c, Z. Styczynski b

a Fraunhofer Institute for Factory Operation and Automation IFF Magdeburg, Germanyb Otto-von-Guericke-Universität-Magdeburg, Germanyc Irkutsk State Technical University, Russia

a r t i c l e i n f o

Article history:Received 21 August 2013Accepted 7 December 2013Available online 22 January 2014

Keywords:Smart gridPower system automationIEC 61850Common Information Model (CIM)Standard harmonizationAdaptive protection

a b s t r a c t

An overview of basic IEC standards for smart grid applications is given and some examples of feasibleinformation and communication technology for smart energy systems are shown. As ICT key standardsfor power grid automation, the two core standards IEC 61850 and IEC 61970 are presented in the paper.Protection automation relying on smart grid ICT technology is shown, and the hurdles to be overcome forthe realization of smart grid automation are discussed. Practical examples for are demonstrated. Oneapproach of making different standards work together is presented, which today is still not sufficientlysolved and is a main hurdle on the way towards a seamless smart grid automation system.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The expansion of the renewable energies that feed an ever-growing share of distributed generation in the medium and lowvoltage networks (MV and LV) increases the risk of an instablenetwork operation and the risk of a critical situation occurring inthe power supply. To handle these threats and ensure a stable powernetwork operation, power system protection needs to be extended towork not only at the transmission system operator0s (TSO) level, butalso at the distribution level in an automatized way, so that protectionconfiguration and protection execution can act immediately to preventsystem failures (Han, Xu, Suonan, Lu, & Wu, 2013).

One burden for a broad application of automated protection todayis the difficulty in handling the many new data arising from thenecessary protection and automation devices, which cannot bemanaged and configured manually. Automatic data collection, con-centration and calculation tools must be developed. The realization ofthese tools heavily depends on the application of open, uniform andstandardized data formats for the data exchange (Lo & Ansari, 2013).

When talking about smart grid realization, the combination ofprotection mechanism and system automation is an important aspect(Falahati, Fu, & Mousavi, 2013). Conventional protection conceptsalone may be suitable for ensuring that assets will not be damagedin case of failure, since they are configured to protect a special assetaccording to statically defined parameters. But they are not sufficient ifa fast system restoration and failure fixing is necessary, especially in

the distribution system. When using automated protection, protectiondevices for assets can be configured dynamically depending on theload flow situation enabling an improved efficiency and improvedfailure recognition accuracy. Here, data from multiple sources like thecontrol center, other protection devices and measurement equipmentis used to allow an automated online estimation of the state that theprotection equipment is meant to protect.

Such adaptive protection algorithms have become necessary, sincethe load flow structure in the energy system changes as fast as thewind and sun feed-in changes, and accordingly, the load situations inthe power network also change quickly. Additionally, automatedprotection allows for a fast system recovery by combining all availabledata to estimate the failure position as accurately as possible. An evenhigher degree of protection automation can be reached by installingappropriate remote control devices, so that failures can be automati-cally insulated very fast, and the supply of the unaffected network canbe reestablished as soon as possible.

The installation of an appropriate monitoring and controlinfrastructure for the many decentralized devices in the smartgrid is, of course, a prerequisite and will result in a huge amount ofmonitoring, control and protection devices. For all these devices anautomated configuration and operation must be given, since amanual configuration for so many devices will inevitably result inerrors. So, the realization of adaptive and automated networkprotection can only be done when data handling and communica-tion mechanisms are used that are powerful enough, ensureinteroperability between all devices and guarantee enough flex-ibility (Apostolov, 2011; Golmie et al., 2013).

The smart grid standards developed by several standardizationgroups, especially by the International Electrotechnical Commission

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/conengprac

Control Engineering Practice

0967-0661/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conengprac.2013.12.001

n Corresponding author. Tel.: +49 3914090784.E-mail address: andre.naumann@iff.fraunhofer.de (A. Naumann).

Control Engineering Practice 25 (2014) 102–111

(IEC), are an important part of the data management and so makesmart grid automation possible. They standardize the way data shouldbe acquired, transmitted, forwarded and managed at all levels in thepower system hierarchy. Some of them are shown in this paper. Theexisting standards are already able to describe many functions that arenecessary for smart grid automation, but there are still some difficul-ties to overcome, as the authors will show in this paper. In particular,the harmonization between standards treating different levels of thesmart grid is still a big challenge. Especially when trying to map databetween the communication level (IEC 61850) and the energymanagement level (IEC 61970) for protection automation, no generalalgorithm is available for that today. By means of a simple caseexample from power system protection the application of smart gridstandards is presented and one approach for a concept to realize anecessary data mapping is shown in this paper.

This paper starts with an overview of smart grid relevantstandards and gives some more details on IEC 61850 and IEC61970/61968. In the next part an example for protection in thepower system is shown and typical parameters are listed. Accord-ing to the identified parameters and functions, in the next part therelevant data models from IEC 61850 and IEC 61970 are identifiedand extended and a proposal for a mapping between them isshown, The paper finishes with a conclusion.

2. Fundamental IEC standards for smart grids

2.1. ICT requirements of smart energy systems

The realization of the smart grid and thus of smart energysystems makes a broad implementation of an appropriate ICTsystem necessary. The expansion of the ICT system to the dis-tribution level down to the low voltage level – including smallstations, energy producers and consumers and even storages – is anecessary step to enable smart functionalities so that measure-ment and status values can be communicated between thedifferent devices and control centers. The same holds true forcommands and set point values (Li, Yang, & Ishchenko, 2012).According to these aspects the main requirements for the ICTsystem of the energy system can be named (Miao & Chen, 2012):

� Availability on all levels of the energy system.� Standardized communication interfaces.� Plug0n run capability for new components.� Safe, reliable and secure data exchange (Wang & Lu, 2013).

When fulfilling these requirements an open architecture ofthe energy system can be built, independent of producer specificprotocols, allowing a simple installation of new devices and theintegration of all system participants. Fig. 1 illustrates this conceptand compares the situation today with the target situation for smartgrids. The left side shows the situation today, using different commu-nication protocols between different parts of the system. At the lowerlevel of the system, mainly connecting electrical loads to the energysystem, no ICT is used. In the future, however, there will be more smallgenerators and distributed storages (e.g. electric vehicles) at the lowvoltage level so the communication needs to be expanded to this level.The application of a common data exchange format for all participantsensures a working information interchange.

2.2. ICT protocol reference architecture

Several smart grid ICT standards mainly define communicationprotocols and interface specifications, but other aspects like cybersecurity (Hahn & Govindarasu, 2013; Miao & Chen, 2012) andfunction modeling are also described.

The IEC and its technical committee 57 (TC 57) is especiallyactive in this field. The graphic shown in Fig. 2 shows the referencearchitecture developed by the TC 57 organizing the communica-tion and interface related standards in a structure (IEC/TR 62357-1ed1.0, 2012). At present more than 100 IEC Standards have beenidentified as relevant to the smart grid.

The following list shows only some of them:

� IEC 61850: communication networks and systems for powerutility automation (IEC/TR 61850-1 ed2.0, 2013) (see Section 2.3)

� IEC 61970: energy management system application programinterface including the common information model (IEC 61970-1 ed1.0, 2005) (see Section 2.4)

� IEC 61968: system interfaces for distribution management.� IEC 61400-25: communications for monitoring and control of

wind power plants (IEC 61400-25-1 ed1.0, 2006).� IEC 62325: framework for energy market communication.� IEC 62351: standard for the data transfer security (IEC/TS

62351-1 ed1.0, 2007).� IEC 62056: data exchange for meter reading, tariff and load

control (IEC 62056-21 ed1.0, 2002).� IEC 61508: functional safety of electrical/electronic/program-

mable electronic safety-related systems (IEC 61508-1 ed2.0,2010).

� IEC 61131: programmable controllers (IEC 61131-1 ed2.0, 2003).

Fig. 1. The ICT structure of the power grid today and tomorrow (Naumann).

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111 103

� IEC 61334: distribution automation using distribution linecarrier systems (IEC/TR 61334-1-1 ed1.0, 1995).

� ISO/IEC 14543: home electronic system (HES) architecture (ISO/IEC 14543-2-1 ed1.0, 2006).

� IEC 61499: distributed control and automation (IEC 61499-1ed2.0, 2012).

Three of the main standards that are especially needed forsmart grid and protection automation are the IEC 61850 series, theIEC 61970 and the IEC 61968 series (Andren, Stifter, & Strasser,2013; Buchholz et al., 2010). A closer look at these three standardsis given in the following chapters to show the basic concepts andthe different approaches that they use.

2.3. IEC 61850 – power system communication

The IEC 61850 standard series consists of several parts, whichwere originally developed to standardize the substation automa-tion. Since these concepts proved capable not only for protectionbut for the whole power system automation as well, the standardhas been extended and revised accordingly to standardize thecommunication between power system devices (Darabi, Falahat, &Vakilian, 2013; Han et al., 2013; Ke, Huang, Hsu, & Huang, 2013;Sizu, Wei, & Fengying, 2013; Xiong, Liu, & Deng, 2013).

Several parts describe the whole procedure of managing energysystem automation projects (IEC 61850-4 ed2.0, 2011) and con-formance testing (IEC 61850-10 ed2.0, 2012), but the main partsdescribe the mechanisms of data exchange (IEC 61850-7-1 ed2.0,2011), provided services (IEC 61850-7-2 ed2.0, 2010), a uniformconfiguration description (IEC 61850-6 ed2.0, 2009) and also themapping of the defined communication procedures to existingdata exchange transport mechanisms (IEC 61850-8-1 ed2.0, 2011).

One of the advantages of the IEC 61850 compared to conven-tional protocols like IEC 60870 (IEC 60870-5-1, 1990) is thestandardized semantics of data exchange, whereas IEC 60870 only

describes the syntax of the data packets to be transmitted. Thisenables producer independent data exchange, since every datafield describes its meaning itself. For this, IEC 61850 uses an objectoriented and hierarchical data structure model, based on so calledlogical devices (LD), logical nodes (LN), data object (DO) and dataattributes (DA). Fig. 3 shows an example of such a data structurewhich may be provided by some monitoring equipment, measur-ing some power network parameters. Inside this equipment someLD can be defined, containing some LNs which represent somepart of the information that can be read or written. In Fig. 3 severalLNs (LPHD, LLN0, MMXU, MMTR, MFLK, MHAI) are shown con-tained in the LD, and it is possible to include several more.

The LN MMXU, which according to IEC 61850-7-4 (IEC 61850-7-4ed2.0, 2010) provides information about measurements in a threephase system, can contain some DOs. In the example in Fig. 3 DOs forreactive power (TotVar), frequency (Hz), phase to ground voltage andphase current are provided. Each of them contains the necessary DAsto provide the necessary information as presented in the example forthe voltage, having complex measurement values for each of thethree phases and the neutral. Each measurement in the IEC 61850format also contains a timestamp and a quality indicator.

The connection of several logical nodes from different physicaland logical devices allows the realization of functions like mon-itoring, control or protection (Zhu, Wang, & Shi, 2013). The dataexchange between different logical nodes and devices is per-formed according to the services specified in IEC 61850-7-2 (IEC61850-7-2 ed2.0, 2010).

These self-explaining structures and the provided services are onemajor advantage of IEC 61850. In combinationwith the other featuresmentioned above, this standard allows the realization of the plug0nrun capability in the electric grid (Naumann; Sizu et al., 2013). Theapplication of the standardized configuration according to IEC 61850-6 ed2.0 (2009) complements this capability since every new devicecan be configured independent of the hardware or producer specificconventions, and the self-explaining propagation of the configured

Fig. 2. TC 57 reference architecture for smart grid standards (Strategiekreis Normungsroadmap E-Energy/Smart Grid in der DKE, 2010).

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111104

IEC 61850 data model makes an automated connection to othersystem components possible.

2.4. IEC 61968/61970 and the common information model

While the IEC 61850 standard is mainly focused on thecommunication between several single devices in the powersystem, the IEC 61970 and IEC 61968 standards concentrate onthe interfaces in energy management systems. The aim of thesestandards is to provide a data model to exchange complex datathat represents information about many aspects of the powersystem, ranging from simple topology and asset data to economicaspects. IEC 61968 extends the IEC 61970 to the aspects of thedistribution system (Zhang, L., Gao, Yao, Cao, & Yang, 2013).

The core of IEC 61970/61968 is the so called common informa-tion model (CIM) which mainly is an object oriented modelconsisting of many classes, which are associated with each otherand represent prototypes for the objects whose information needsto be stored (IEC 61970-301 ed4.0, 2013). This object orientedmodel is designed as a class diagram conforming to the UMLspecification (Pilona, 2005; Seemann & von Gudenberg, 2006).Accordingly, the specified classes contain attributes and theinheritance of classes is an applied concept. The associationbetween classes furthers semantics to the whole model. To clusterthe whole model, which today consists of several hundred classes,

each class is put into one package. There are packages describingthe basic aspect, network topology, measurement aspects andassets, just to name a few.

So the questions arise, why we need to use the IEC 61970 andIEC 61968 and especially the CIM for protection automation? Is theusage of IEC 61850 not sufficient? And the answer is, no it is notsufficient because we are aiming towards advanced and auto-mated protection mechanism, and not all necessary informationmay be available from IEC 61850 capable devices. A lot of dataneeds to be managed in a central control center, especially whentaking into consideration a complex network where each elementhas its own parameters. The CIM has the ability to cover manymore aspects of the power system, while the IEC 61850 mainly hasparameters that are directly connected to the process (Ling,Hongyong, & Xia, 2013). Take for example a power transformer:with CIM you are able to save data like impedance of windings,which may be relevant for the calculation of optimal protectionsettings. With the IEC 61850, however, you do not have data fieldsfor the impedance of a transformer. What you need is the resultingprotection configuration, but these values can only be calculatedwhen we have the appropriate data in CIM.

An example, of how power system information can beexchanged according to the CIM format is presented in Fig. 4.Here a simple two winding transformer shall be mapped to a CIMconformed representation. In the real world a transformer consists

Fig. 3. Example data structure according to IEC 61850 data model.

Fig. 4. Example data in CIM representation.

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111 105

of the transformer “backbone” including iron core and housingand of two or more separate windings. All of these parts havespecific parameters which may be more or less important for theelectric network. In CIM these parts are represented by theappropriate classes and the parameters are mapped to the definedattributes. The CIM format allows the data exchange betweendifferent applications in a common, standardized way, since theused classes, their attributes and also the relations between classesare defined (Rohjans, 2013). Additionally user-specific andapplication-specific extensions are possible, which may be rele-vant for future aspects that the CIM does not yet model.

2.5. Other standards

As already shown in chapter 2.2 there are many more standardsthan IEC 61850 and IEC 61970/61968 that are relevant for smartgrid realization. Although this paper focuses on these three, theconcepts shown here are also relevant for the application of otherstandards and the mapping between different existing standards.The integration of low voltage level devices, in particular, mainlydepends on the successful mapping of smart metering and homeautomation protocols to the energy system automation standards,although this is not that relevant for protection automation.Standards for cyber-security like IEC 62351 (IEC/TS 62351-1ed1.0, 2007) are far more relevant here, since an attack on theICT system of a utility may result in a fatal black-out. Also thestandards IEC 61499 (IEC 61499-1 ed2.0, 2012) and IEC 61131 (IEC61131-1 ed2.0, 2003), which are already widely-used for systemautomation mainly in the industrial area, are very important forprotection automation.

2.6. Bridging gaps between standards

The different smart grid relevant standards standardize alldifferent aspect of the electrical energy system. Nevertheless, theyare connected in some points, especially when talking about

communication standardization. Standards for smart meteringand home automation need to have some connection to standardsfor system management or power system automation, sincemetering and home automation data is used for grid management(Lee, Park, Jeong, Kang, & Park, 2013). This is the point where someharmonization becomes necessary to enable gateway technolo-gies. The same is true in the area of protection automation where amapping between IEC 61850 and the CIM is necessary (Zhang, Z.J.,Luo, & Du, 2013; Ling et al., 2013). Since both standards usedifferent approaches and data structures, gateways need to bequite sophisticated (Naumann; Naumann, Buchholz, Komarnicki, &Brunner, 2011).

One general approach of mapping data between both standardsis drawn in Fig. 5. This mapping approach sketches the building ofCIM objects from the data structures (LNs and DOs) provided bysome IEC 61850 capable device. Since the LN XCBR means theexistence of some load breaking switch the instantiation of theappropriate class in the CIM is necessary. The same thing is truefor the MMXU LN, which means some measurement creates a CIMobject for analog measurements. The created CIM objects can evenbe specified in more detail by setting the appropriate attributes.

For the measurement, a DO named Hz means that somefrequency measurement exists. Accordingly, the CIM object “Analog”can be parameterized as frequency measurement using the string“Frequency” for the attribute “measurementType” (see (IEC 61970-301 ed4.0, 2013) for more details).

The example above only shows a small aspect of the mappingbetween different smart grid standards. To handle the wholecomplexity of mapping, an intelligent approach is needed. Onepossible solution might be the application of ontology mapping(Mathias, 2010). Ontology mapping can be seen as some kind ofdictionary, not only containing simple tables of what attributefrom standard A corresponds to which attribute from standard B,but also includes some more complex algorithms like data typeconversion, building of complex attributes and mapping strategiesand can also evaluate the semantics of the used data model. To

Fig. 5. Creating CIM objects from IEC 61850 description.

Fig. 6. Using ontologies to map between IEC 61850 and CIM.

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111106

realize such a system, first each of the standards must be analyzedand the according ontology for each of them must be defined, like adictionary describing the rules of e.g. how IEC 61850 data structuresare built. Then, by the combination of the different standard-specificontologies, so called mediator ontologies can be derived. Thismediator ontology is the thing the gateways needs for convertingdata from one format to the other (Kim et al., 2013). This approach isshown in Fig. 6. The figure shows a simple example of mappingtimestamps between IEC 61850 and CIM. The information is named“timestamp” and is stored in an integer format for the “seconds sinceepoch” and the “fraction of second” and this information must thenbe converted to a date string named “DateTime”.

3. Protection system realization

3.1. Protection applications

To demonstrate the application of smart grid IEC standards forprotection automation in this paper two examples are shown. Bothrely on the concept of overcurrent protection. The first concept isalready in use today in electrical energy systems as sequencedprotection schemes, shown in Section 3.2. This protection ensuressome backup protection mechanisms, in case one or more break-ers should fail.

Although the application of this protection today is quitecommon, most of them are realized independent from IEC 61850and the CIM and use instead producer specific protocols.

The second example is more advanced and allows a fast systemrecovery if there is a line sustaining an earth failure or short circuitin an open loop distribution system; it is explained in Section 3.3.The conventional way of recovering such a failure is to have somemaintenance staff check each station in the loop manually fortripped short circuit detectors and perform the necessary switchoperations inside the station to isolate the failure. By applyingremote controlled monitoring and control equipment, this proce-dure can be performed completely automated, and so it only takesa few minutes until normal system operation is reestablished.

The examples here are intentionally kept simple concerningthe protection mechanisms. They represent some easy to handlefunctions but still rely on data that is available in IEC 61850 andshould also be available in CIM. Since this paper is focused on thenecessary data models and the correct mapping between them,the complexity of the protection functionality is not the mostimportant thing here. Other examples and functionalities, whichmay require more complex data models and mapping algorithms,are distance and differential protection application and theircombination. Further the application of adaptive distance protec-tion is even more complex and will require advanced data modelsin the future.

3.2. Basic sequenced protection

This example shows the operation of a sequenced overcurrentprotection in a simple power network feeding some loads (heresome motors) as shown in Fig. 7. This network could sustain linefailures anywhere, but this example shows cases with failures atthe line ends, since these cases are relevant for the protectionsettings in the conventional system. This scheme contains thefeeding network (N), a transformer (T), transmission lines (L1–L5),power breakers (B1–B5, BM1-2) and electric loads (M1, M2).

To ensure safe operation and the correct behavior of theprotection devices, the tripping characteristics of each protectiondevice needs to be calculated to fulfill the following requirements:

– If the current at one overcurrent relay is above the definedmaximum current, the corresponding breaker should trip asfast as possible (fastness).

– If one breaker fails to trip, the breaker one step closer to thefeeding network shall take the function of tripping(redundancy).

– In case of failure, the tripped parted of the network shall be assmall as possible (selectivity).

According to these requirements the trip delay of each over-current relay can be calculated according to Eqs. (1) and (2), sothat the fastest trip will occur close to the end of the line and the

Fig. 8. Short circuit in an open loop topology.

Fig. 7. Part of an exemplary power network structure with short circuits at the endof a line.

Table 1Selectivity steps and time delay for breaker configuration.

Breaker name Selectivity step [s] Time delay [s]

BM1 0 0B1 0.235 0.235B2 0.235 0.470B3 0.235 0.705B4(1) 0.235 0.940B4(2) 0.235 0.470BM2 0 0B5 0.235 0.235

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111 107

biggest delay has to be configured close to the feeding network.

tB_n ¼ tB_n�1þΔtB_n ð1Þ

ΔtB_n ¼ tPB_n�1þtEPB_n�1þtEPB_nþtinertPB_n�1þtSP ð2Þwhere n is the number of breaker according to scheme in Fig. 7;tB_n is the breaker time delay of breaker n; ΔtB_n is the selectivitystep; tPBis the proper operation time of breaker; tEPB is the protec-tion error; tinertPB is the inertia time of protection; tSP is the supplytime of protection.

The values of the time delays and selectivity steps for eachbreaker of the sample network can be determined and theresulting setting values are shown in Table 1.

Time settings of the protection devices that are positioned atbreakers BM1, B1–B3 were calculated for short circuits at the loadM1, while the settings for the breakers BM2 and B5 werecalculated for short circuits at the second load M2. The settingsof breaker B4 were determined for the possible variants of theshort circuit at both loads.

3.3. Advanced protection and self-fixing capability

The example from Section 3.2 describes the basic protectionfunctionality. The obvious extension of this concept is the applica-tion of automated failure position detection and direct tripping ofthe corresponding breaker to avoid the trip delays necessary forsequenced protection. This will be possible when all integratedprotection relays are connected via communication technologywith each other or with a central station (e.g. a control center), sothat the parallel measurement of currents make it possible todetermine the location of the failure.

Today0s distribution systems quite often include open loopnetwork topology structures, which may especially benefit from

the automated advanced protection concept, since only a verysmall amount of time until complete supply reestablishment isnecessary. The scheme in Fig. 8 shows a sample open loopdistribution network. Many of these structures can be found inthe distribution network today.

Normally this loop is operated with breakers B91 and B92 open,and only in case of failure is the loop closed to reestablishconnection. The failure could be a short circuit on line L6 so thatbreakers B61 and B62 need to be opened to clear the failure andB91 and B92 can be closed.

The normal sequence for the scenario in Fig. 8 is

1. Occurrence of short circuit or earth failure on L6.2. Disconnecting the whole failure branch at B51.3. Disconnecting B61 and B62.4. Closing B91 and B92.5. Closing B51.

Depending on the size of this open loop, the conventionalprocedure of finding and disconnecting the damaged line, closingthe loop and reestablishing the supply may take several hours,since today some maintenance staff has to check each substationfor short circuit detection and according to the findings, executethe disconnect and connect switch actions. The usage of remotemonitored and controlled devices in the substations can reducethe time of lagging power supply to a few minutes, if not less. Thisof course requires the installation of appropriate devices in thestations, the availability of necessary communication infrastruc-ture and the usage of standardized data models all devices canunderstand and which are able to transmit and provide thenecessary data, like measured current values and breaker state.

4. Realizing a smart protection system

4.1. Using IEC 61850 for protection

As shown in the previous section the application of remotemonitoring and control offers a lot of benefits for protectionfunctions and is a prerequisite for protection automation. Mappingthe above examples of overcurrent detection, protection andautomated sequencing of breaker opening and closing to the IEC61850 world, shows which LNs to use in order to represent andtransmit all necessary data. Fig. 9 shows the LNs (IEC 61850-7-4ed2.0, 2010) to be applied and the assignment to the appropriateobjects of the real power system. The LN TCTR represents theparameters of the installed current transformer, while the LNMMXU provides the measured values of the current. The LN PTOCrepresents the overcurrent protection relay which uses the mea-sured values from MMXU and provides the information regardingwhen the appropriate breaker should trip. The breaker is repre-sented by the LN XCBR. Additionally, the protection relay (PTOC)sends information about the state to the human machine interfacein the control center, which is represented by the LN IHMI andwhich allows the control center operator to be informed andperform necessary actions.

All of the transmitted information contains a timestamp and aquality indicator to check if the received data is valid. Each of theLNs can be configured according to the needs of the protectionfunctionality. In particular, the LN PTOC needs to be configuredwith a tripping characteristic, which can be quite simply aconstant current and time value but can also be tripping curve,representing the tripping time dependent on the current value.Such a curve data structure is also defined in the IEC 61850. Tobuild a whole automated protection and self-fixing system in anopen loop the necessary LNs need be present for each of the usedFig. 10. CIM classes used for overcurrent protection.

Fig. 9. Logical nodes used for overcurrent protection.

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111108

measurement points, breakers and protection relays. The combi-nation of the data from LNs from different stations in the controlcenter enables the correct evaluation of the acquired data and theoperation of the correct breakers.

4.2. Modeling protection with CIM data models

Automated protection algorithms do not necessarily needadditional information from the distribution system control cen-ter, as long as the protection and automation algorithms are quitesimple and do not depend on information from other resourceslike other measurement points or states of energy sources. Foradvanced functions the provision of further information is neces-sary, which may update some protections characteristics depend-ing, for example, on the actual wind feed-in situation.

The use of CIM makes it possible to keep the necessary dataavailable in a central place and, additionally, it may be completedby more asset and network specific parameters that are notavailable in IEC 61580. Necessary configuration parameters forIEC 61850 capable devices then may be derived from theseparameters available from CIM. So the correct configuration andother data depend on the correct usage of data from CIM.

CIM defined classes that are to be used for overcurrent protec-tion functionality as described in this paper are shown in Fig. 10.The shown classes are assigned to the appropriate elements of thenetwork as well as to the corresponding configuration elementslike characteristic curves and also to measurement points. Havinga look at Fig. 10 shows that the representation of the network andits elements differs from the conventional one, since there areconnectivity nodes (the gray circles) and terminals (the small blackcircles) which are used to connect different elements like breakers,

lines, generators and bus bars with each other. Each of theseelements has a corresponding class in the CIM, which is named asshown in Fig. 10 “BusbarSection”, “Terminal” “ACLineSegment” etc.Measurements of physical units are mapped to the “Analog” class,so that current values can be provided, too.

Although the CIM today is quite complex and offers a lot of datamodel definitions, it lags a sufficient description of protectionfunctions and devices. That is the reason why two classes in Fig. 10are drawn in red. While the green classes are already defined, thered classes have been defined on our own to extend the CIM forthe purpose of modeling protection functionality (Zhang, L. et al.,2013; Zhang, Z.J. et al., 2013. The class diagram of the newlycreated and the already existing classes of the CIM are shown inFig. 11. You can see that there are two classes for extending ageneral protection equipment class. One represents a directionalovercurrent relay with hard current limits used for sequenced overcurrent protection. The other one represents a definite time overcurrent delay, which has one (or more) current–time–character-istic curve associated with it. The current–time curve itself extendsan already existing general “curve” CIM class, which is a basis forevery type of curve data representation. So a special type of thiscurve for overcurrent protection needs was created, to be able tomap the typical overcurrent protection relay characteristic.

Applying this concept to other protection functions like dis-tance and differential protection makes the building of appropriateclasses necessary, which extend the existing model. These classeswill be more complex, containing more attributes and associa-tions, since their protection functionality also depends and moreparameters and more complex algorithms.

4.3. Mapping data between IEC 61850 and CIM

The data models of IEC 61850 and CIM use two differentapproaches, which need to be harmonized, to make automatedprotection work. On the one hand, there are logical nodes and datastructures that provide information. On the other hand, objectsand their attributes, instantiated from class prototypes are used.The data mapping for the example of overcurrent protection shallbe presented. The focus is on the mapping of configuration andmeasurement values. The mapping of the network topologydescription is neglected since the knowledge of the topology inthe IEC 61850 domain is not absolutely necessary, and it is onlyused once when the system is initiated. The operational data is themore interesting part here.

The mapping of measurement data from IEC 61850 to CIM,especially for current supervision, is done according to the map-ping scheme in Table 2. This table considers only one phase (PhaseA), since for the other phases the scheme is identical, except thatin the IEC 61850 path “phsA” would be written “phsB” and “phsC”.For every measurement value the value itself is transmitted as wellas the timestamp and a quality indicator for each measurement.The first column of Table 2 names the attribute path usedaccording to the IEC 61850 data model and the second column

Table 2Mapping of measurement data between IEC 61850 and CIM.

IEC 61850 CIM

LN, DO and DA Type Class and attribute Type

MMXU.A.phsA.cval. mag.f FLOAT32 AnalogValue.value Float

MMXU.A.phsA.cval.t. SecondsSinceEpoch INT32U AnalogValue. timestamp (from MeasurementValue) String (DateTime)MMXU.A.phsA.cval.t. FractionOfSecond INT24U

MMXU.phsA.q BOOLEAN (List) Quality61850 Boolean (List)

Fig. 11. Extension of the CIM for overcurrent protection functionality.

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111 109

shows the data type to be used. The third and fourth columnsshow the corresponding classes, attributes and data types from theCIM. The class “AnalogValue” is a specialized form of “Measur-ementValue” so that a timestamp can be associated with each“AnalogValue”, too. Since an “AnalogValue” object only representsa single measured value, many of them can be associated with an“Analog” object, which represents the point of measurement onsome equipment, as shown in Fig. 10.

The result is that mapping of measured values between IEC61850 and CIM is feasible, although several type conversions andsemantic mappings are necessary. Attributes that are contained inone single attribute structure in IEC 61850 need to be separatedand mapped to different CIM objects. The same is true the otherway round, so that combinations of different attributes from IEC61850 are necessary to get one single attribute in CIM (e.g. thetimestamp).

Mapping the configuration data of the overcurrent protectionrelay between IEC 61850 and CIM is not so easy at the moment,because there are no standardized data models in CIM availablethat could be used. When using the extensions shown in Fig. 11 aquite straight forward mapping is possible. The characteristictripping curve is available in IEC 61850 (common data classesCURVE, CSG, CSD according to IEC 61850-7-3 ed2.0 (2010)) and canbe mapped to the curve modeled in Figs. 10 and 11 named“OvercurrentProtectionCharacteristic”. Both data models containa set of data points describing a characteristic curve, showing therelationship between time and current.

Once the overcurrent relay picks up a signal and starts workingthe need for a trip can occur. If this happens, an appropriate signalmust be sent to the breaker, which should trip immediately andthe event must be recorded in the control center. So both datamodels, IEC 61850 and CIM, need to have these information fieldsthat contain information of when a trip occurred. Table 3 showsthe data mapping between IEC 61850 and CIM for the necessaryinformation for breaker tripping.

On the left side, the data attributes for trip signal, its timestampand the quality indicators from the IEC 61850 domain is shown.The LN PTOC contains the “Operate” DO with the attributes“general” (tripped or not tripped), the timestamp and the quality.On the CIM side, the class “Command” is used, which in the CIM isassociated with a discrete measurement that represents the stateof a of breaker (see (IEC 61970-301 ed4.0, 2013) for details). Thus, acommand is always connected to some discrete measurement.Additionally, the command also contains attributes for timestampand quality similar to the measurement shown in Table 2. Here,again, the type conversions are necessary to realize a correctmapping.

5. Conclusion

The authors of this paper show what the information andcommunication needs are for protection automation and where

existing, and still under development, standards are required tofulfill these needs. In particular, the application of IEC 61850 isquite appropriate for protection automation (Chen, Liu, & Xu,2013), although most applications still rely on IEC 60870. In IEC61970/61968 (CIM) things are a little different: the groundwork fornetwork operation exists, but protection specific data models aremissing. Here developers are building their own models, like theexample shown in Section 4.2. For the simple example of over-current protection and self-fixing open-loop architectures this isnot such a hard task. However with more and more adaptive andadvanced protection mechanisms, more detailed and sophisticatedmodels are necessary. Especially the mapping between CIM andIEC 61850 is much more complex for advanced protection. Never-theless, the approach shown in this paper for modeling andmapping of protection functionality is also valid for complexfunctions. Future work to be done here is the extension of theCIM for a standardized protection. Furthermore, a standard-basedharmonization between CIM and IEC 61850 not only for protectionmust be accomplished. Although this work is ongoing in the IEC,big efforts are still needed to accomplish this task.

References

Andren, F., Stifter, M., & Strasser, T. (2013). Towards a semantic driven frameworkfor smart grid applications: Model-driven development using CIM, IEC 61850and IEC 61499. Informatik-Spektrum, 36(1), 58–68.

Apostolov, A. (2011). IEC 61850 based adaptive distribution protection. In 21stInternational Conference on Electricity Distribution CIRED. Frankfurt (Main).

Buchholz, B. -M., Styczynski, Z. A., Gurbiel, M., Riis, H., Vale, Zita A., Gelfand, A. M.,et al. (2010). Towards the wide implementation of standards IEC 61968/70(CIM) and IEC 61850 in the distribution system. CIGRE. Invited paper No.C6.105.

Chen, Y., Liu, D., & Xu, B. (2013). Travelling wave fault location equipment modelingbased on IEC 61850. Dianli Xitong Zidonghua/Automation of Electric PowerSystems, 37(2), 86–90.

Darabi, Z., Falahat, B., Vakilian, M., (2013). Application of IEC61850-enabled SAS inCBF protection of transformers. In IEEE PES Innovative Smart Grid TechnologiesConference, ISGT, art No. 6497864.

Falahati, B., Fu, Y., & Mousavi, M. J. (2013). Reliability modeling and evaluation ofpower systems with smart monitoring. IEEE Transactions on Smart Grid, 4(2),1087–1095 (art. No. 6423864).

Golmie, N., Scaglione, A., Lampe, L., Yeh, E., Smith, S., & Tong, L. (2013). Guesteditorial: smart grid communications. IEEE Journal on Selected Areas in Com-munications, 31(7), 1169–1171 (art No. 6547826).

Hahn, A., Govindarasu, M., 2013. Model-based intrusion detection for the smart grid(MINDS). In ACM International Conference Proceeding Series, art No. 27.

Han, G. -Z., Xu, B. -Y., Suonan, J. -L., Lu, G. -X., & Wu, J. -M. (2013). Communicationtechnology for distribution automation based on IEC 61850. Dianli Xitong Baohuyu Kongzhi/Power System Protection and Control, 41(2), 62–66.

IEC 60870-5-1, (February 1990). Telecontrol equipment and systems. Part 5:Transmission protocols – Section One: Transmission frame formats. ed. 1.0.

IEC 61131-1 ed2.0, (May 2003). Programmable controllers. Part 1: Generalinformation.

IEC 61400-25-1 ed1.0, (December 2006). Wind turbines – Part 25-1. Communica-tions for monitoring and control of wind power plants – Overall description ofprinciples and models.

IEC 61499-1 ed2.0, (November 2012). Function blocks. Part 1: Architecture.IEC 61508-1 ed2.0, (April 2010). Functional safety of electrical/electronic/program-

mable electronic safety-related systems. Part 1: General requirements.IEC 61850-10 ed2.0, December 2012). Communication networks and systems for

power utility automation. Part 10: Conformance testing.

Table 3Mapping of a trip command between IEC 61850 and CIM.

IEC 61850 CIM

LN, DO and DA Type Class and attribute Type

PTOC.Op.general. BOOLEAN Command.value Integer

PTOC.Op.t. SecondsSinceEpoch INT32U Command.timestamp (from Control) String (DateTime)PTOC.Op.t. FractionOfSecond INT24U

PTOC.Op.q BOOLEAN (List) Quality61850 Boolean (List)

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111110

IEC 61850-4 ed2.0, (April 2011). Communication networks and systems for powerutility automation. Part 4: System and project management.

IEC 61850-6 ed2.0, (December 2009). Communication networks and systems forpower utility automation. Part 6: Configuration description language for com-munication in electrical substations related to IEDs.

IEC 61850-7-1 ed2.0, (July 2011). Communication networks and systems for powerutility automation. Part 7-1: Basic communication structure – Principles andmodels.

IEC 61850-7-2 ed2.0, (August 2010). Communication networks and systems forpower utility automation. Part 7-2: Basic information and communicationstructure – Abstract communication service interface (ACSI).

IEC 61850-7-3 ed2.0, December 2010. Communication networks and systems forpower utility automation. Part 7-3: Basic communication structure – Commondata classes.

IEC 61850-7-4 ed2.0, (March 2010). Communication networks and systems forpower utility automation. Part 7-4: Basic communication structure - Compatiblelogical node classes and data object classes.

IEC 61850-8-1 ed2.0, (June 2011). Communication networks and systems for powerutility automation. Part 8-1: Specific communication service mapping (SCSM) –

Mappings to MMS (ISO 9506-1 and ISO 9506-2) and to ISO/IEC 8802-3.IEC 61970-1 ed1.0, (December 2005). Energy management system application

program interface (EMS-API). Part 1: Guidelines and general requirements.IEC 61970-301 ed4.0, May 2013. Energy management system application program

interface (EMS-API). Part 301: Common information model (CIM) base.IEC 62056-21 ed1.0, (May 2002). Electricity metering – Data exchange for meter

reading, tariff and load control. Part 21: Direct local data exchange.IEC/TR 61334-1-1 ed1.0, (November 1995). Distribution automation using distribu-

tion line carrier systems. Part 1: General considerations – Section 1: Distributionautomation system architecture.

IEC/TR 61850-1 ed2.0, (March 2013). Communication networks and systems forpower utility automation. Part 1: Introduction and overview.

IEC/TR 62357-1 ed1.0, (October 2012). Power systems management and associatedinformation exchange. Part 1: Reference architecture.

IEC/TS 62351-1 ed1.0, (May 2007). Power systems management and associatedinformation exchange – Data and communications security. Part 1: Commu-nication network and system security – Introduction to security issues.

ISO/IEC 14543-2-1 ed1.0, (September 2006). Information technology – Homeelectronic system (HES) architecture. Part 2-1: Introduction and devicemodularity.

Ke, Y. -K., Huang, P. -H., Hsu, C. -Y., & Huang, B. -Y. (2013). Study on communicationprotocol application in smart grid. Applied Mechanics and Materials, 284–87,3235–3239.

Kim, D. -K., Yang, H., Jang, H., Hong, D., Falk, H., Kim, S., et al., (2013). Ametamodeling approach to unifying IEC 61850 and IEC 61970. In IEEE PESInnovative Smart Grid Technologies Conference, ISGT, art No. 6497888,.

Lee, B., Park, S., Jeong, D., Kang, B., Park, S., (2013). Design of energy-efficient datacommunication between a smart meter and home appliances. In Proceedings ofthe 6th IASTED Asian Conference on Power and Energy Systems (pp. 147–151).AsiaPES.

Li, Z., Yang, F., & Ishchenko, D. (2012). The standardization of distribution gridcommunication networks. IEEE Power and Energy Society General Meeting (artNo 6345482).

Ling, L., Hongyong, Y., Xia, C., (2013). Model differences between IEC 61970/61968and IEC 61850. In Proceedings of the 2013 3rd International Conference onIntelligent System Design and Engineering Applications, ISDEA (pp. 938–941),art No. 6456213.

Lo, C. -H., & Ansari, N. (2013). Decentralized controls and communications forautonomous distribution networks in smart grid. IEEE Transactions on SmartGrid, 4(1), 66–77 (art No. 6395792).

Mathias, Uslar, (2010). Ontologiebasierte Integration heterogener Standards in derEnergiewirtschaft (1st ed.). OlWIR Verlag für Wirtschaft, Informatik und Recht.

Miao, X., Chen, X., (2012). Cyber security infrastructure of smart grid communica-tion system. In China International Conference on Electricity Distribution, CICED,art No. 6508410.

Naumann, A. Leitwarte im Smart Grid (Vol. 47). Otto-von-Guericke UniversitätMagdeburg. Magdeburger Forum zur Elektrotechnik. isbn:978-3-940961-81-5.

Naumann, A., Buchholz, B. -M., Komarnicki, P., Brunner, C., (2011). Seamless datacommunication and management over all levels of the power system. In 21stInternational Conference on Electricity Distribution CIRED. Frankfurt (Main).

Pilona, Dan (2005). UML 2.0 in a Nutshell (1st ed.). O0Reilly Media, ISBN: 978-0-596-00795-9.

Rohjans, S., (2013). A standard-compliant ICT-architecture for semantic data serviceintegration in smart grids. In IEEE PES Innovative Smart Grid TechnologiesConference, ISGT, art No. 6497820,.

Seemann, Jochen, & von Gudenberg, JürgenWolff (2006). Software-Entwurf mit UML2 – Objektorientierte Modellierung mit Beispielen in Java (2nd ed.). Berlin:Springer-Verlag.

Sizu, H., Wei, L., Fengying, K., (2013). IEC61850 gateway SCL configuration docu-ment research. In Proceedings of the 2013 3rd International Conference onIntelligent System Design and Engineering Applications, ISDEA (pp. 824–827),art No. 6455739.

Strategiekreis Normungsroadmap E-Energy/Smart Grid in der DKE, (2010). DieDeutsche Normungsroadmap E-Energy/Smart Grid. Version 1.0, 28.03.

Wang, W., & Lu, Z. (2013). Cyber security in the Smart Grid: Survey and challenges.Computer Networks, 57(5), 1344–1371 (Cited 2 times).

Xiong, J., Liu, C., & Deng, F. (2013). Centralized protection and monitoring devices ofsmart substation. Dianli Xitong Zidonghua/Automation of Electric Power Systems,37(12), 100–103.

Zhang, L., Gao, Z., Yao, J., Cao, Y., & Yang, S. (2013). Coordination between CIM modeland IEC61850 model based on CIM/E language. Dianli Zidonghua Shebei/ElectricPower Automation Equipment, 33(4), 148–154.

Zhang, Z. J., Luo, Y., & Du, Y. (2013). Study of smart distribution integrationevolution. Applied Mechanics and Materials, 321–324, 2009–2012.

Zhu, L., Wang, P., & Shi, D. (2013). Status monitoring information model andconfiguration description of communication network in smart substations.Dianli Xitong Zidonghua/Automation of Electric Power Systems, 37(11), 87–92.

A. Naumann et al. / Control Engineering Practice 25 (2014) 102–111 111