a survey on communication networks for electric system ...a survey on communication networks for...

Computer Networks 50 (2006) 877–897

www.elsevier.com/locate/comnet

A survey on communication networks for electricsystem automation

V.C. Gungor a,*, F.C. Lambert b

a Broadband and Wireless Networking Laboratory, School of Electrical and Computer Engineering,

Georgia Institute of Technology, Atlanta, GA 30332, United Statesb National Electric Energy Testing, Research and Applications Center, Georgia Institute of Technology,

Atlanta, GA 30332, United States

Received 18 January 2006; accepted 26 January 2006Available online 21 February 2006

Responsible Editor: I.F. Akyildiz

Abstract

In today’s competitive electric utility marketplace, reliable and real-time information become the key factor for reliabledelivery of power to the end-users, profitability of the electric utility and customer satisfaction. The operational and com-mercial demands of electric utilities require a high-performance data communication network that supports both existingfunctionalities and future operational requirements. In this respect, since such a communication network constitutes thecore of the electric system automation applications, the design of a cost-effective and reliable network architecture is cru-cial. In this paper, the opportunities and challenges of a hybrid network architecture are discussed for electric system auto-mation. More specifically, Internet based Virtual Private Networks, power line communications, satellite communicationsand wireless communications (wireless sensor networks, WiMAX and wireless mesh networks) are described in detail. Themotivation of this paper is to provide a better understanding of the hybrid network architecture that can provide heter-ogeneous electric system automation application requirements. In this regard, our aim is to present a structured frameworkfor electric utilities who plan to utilize new communication technologies for automation and hence, to make the decision-making process more effective and direct.� 2006 Elsevier B.V. All rights reserved.

Keywords: Electric system automation; Internet based Virtual Private Network; Power line communication; Satellite communication;Wireless sensor networks; Wireless mesh networks; WiMAX

1389-1286/$ - see front matter � 2006 Elsevier B.V. All rights reserved

doi:10.1016/j.comnet.2006.01.005

* Corresponding author. Tel.: +1 404 894 5141; fax: +1 404 8947883.

E-mail addresses: [email protected] (V.C. Gungor),[email protected] (F.C. Lambert).

1. Introduction

Electric utilities, particularly in urban areas, con-tinuously encounter the challenge of providing reli-able power to end-users at competitive prices.Equipment failures, lightning strikes, accidents, andnatural catastrophes all cause power disturbances

.

mailto:[email protected]

mailto:[email protected]

878 V.C. Gungor, F.C. Lambert / Computer Networks 50 (2006) 877–897

and outages and often result in long service inter-ruptions. Electric system automation, which is thecreation of a reliable and self-healing electric systemthat rapidly responds to real-time events with appro-priate actions, aims to maintain uninterrupted powerservice [6]. The operational and commercial demandsof electric utilities require a high-performance datacommunication network that supports both existingfunctionalities and future operational requirements.Therefore, the design of the network architecture iscrucial to the performance of the system.

Recent developments in communication technol-ogies have enabled reliable remote control systems,which have the capability of monitoring the real-time operating conditions and performance of elec-tric systems. These communication technologies canbe classified into four classes, i.e., Power Line Com-munication, Satellite Communication, WirelessCommunication, and Optical Fiber Communica-tion. Each communication technology has its ownadvantages and disadvantages that must be evalu-ated to determine the best communication technol-ogy for electric system automation. In order toavoid possible disruptions in electric systems dueto unexpected failures, a highly reliable, scalable,secure, robust and cost-effective communicationnetwork between substations and a remote controlcenter is vital [11,14]. This high performance com-munication network should also guarantee verystrict Quality of Service (QoS) requirements to pre-vent the possible power disturbances and outages[10].

When the communication requirements of elec-tric system automation are considered, Internetcan offer an alternative communication network toremotely control and monitor substations in acost-effective manner with its already existing com-munication infrastructure. However, Internet can-not guarantee very strict QoS requirements thatthe automation applications demand, since datacommunication in Internet is based on best effort

service paradigm [32]. Furthermore, when a publicnetwork like the Internet is utilized to connect thesubstations to a remote control center, security con-cerns arise. In this context, Internet based VirtualPrivate Network (Internet VPN) technologies,which are transforming the Internet into a securehigh speed communication network, constitute thecornerstone for providing strict QoS guarantees ofelectric system automation applications [7]. InternetVPN technologies offer a shared communicationnetwork backbone in which the cost of the network

is spread over a large number of users while simul-taneously providing the benefits of a dedicated pri-vate network. Therefore, Internet VPN technologyas a high speed communication core network canbe utilized to enable minimum cost and highlyreliable information sharing for automationapplications.

Although Internet VPN technologies can providethe necessary reliable communication for substa-tions in urban areas, this may not be the case forsubstations in remote rural locations where the highspeed communication core network, e.g., Internet,might not exist. Therefore, when the individualcommunication capabilities and locations of electricsystems are taken into account, it is appropriate toconsider the overall communication infrastructureas a hybrid network as shown in Fig. 1. This hybridnetwork consists of two separate parts:

• High speed communication core network: It can beeither a private network or public network. Dueto several technical advantages [32], Internetbased Virtual Private Network can be consideredas a cost-effective high speed communication corenetwork for electric system automation.

• Last mile connectivity: It represents the challengeof connecting the substations to the high speedcommunication core network. The communica-tion technologies for last mile connectivity canbe classified as: (i) Power line communication,(ii) Satellite communication, (iii) Optical fibercommunication, and (iv) Wireless communica-tion. Each possible communication alternativesfor last mile connectivity introduces its ownadvantages and disadvantages.

Many researchers and several international orga-nizations are currently developing the required com-munication technologies and the internationalcommunication standard for electric system auto-mation. In Fig. 2, the summary of these communi-cation system development activities is presented[14]. Despite the considerable amount of ongoingresearch, there still remains significantly challengingtasks for the research community to address bothbenefits and shortcomings of each communicationtechnology. Since a cost-effective data communica-tion network constitutes the core of the automationapplications, in this paper, the opportunities andchallenges of a hybrid network architecture aredescribed for automation applications. More specif-ically, Internet based Virtual Private Networks,

Fig. 1. The overall communication network architecture for electric system automation.

Fig. 2. Summary of communication system development activities for electric utilities.

V.C. Gungor, F.C. Lambert / Computer Networks 50 (2006) 877–897 879

power line communications, satellite communica-tions and wireless communications (wireless sensor

networks, WiMAX, and wireless mesh networks)are discussed in detail. The motivation of this paper


is to provide a better understanding of the hybridnetwork architecture that can provide heteroge-neous electric system automation applicationrequirements. In this respect, our aim is to presenta structured framework for electric utilities whoplan to utilize new communication technologiesfor automation and hence, to make the decision-making process more effective and direct.

The remainder of the paper is organized as fol-lows. In Section 2, the benefits and open researchchallenges of Internet based Virtual Private Net-works are discussed for electric system automation.In Section 4, both advantages and disadvantagesof alternative communication technologies aredescribed for last mile connectivity. In Sections 5and 6, the opportunities and challenges of wirelesssensor networks, wireless mesh networks andWiMAX are explained, respectively. Finally, thepaper is concluded in Section 7.

2. Internet based Virtual Private Networks

Recent advances in Internet technology andInternet-ready IEDs (Intelligent Electronic Devices)have enabled cost-effective remote control systems,which makes it feasible to support multiple automa-tion application services, e.g., remote access to IED/relay configuration ports, diagnostic event informa-tion, video for security or equipment status assess-ment in substation and automatic metering. Whiletraditional private Supervisory Control and DataAcquisition (SCADA) systems constitute the corecommunication network of today’s electric utilitysystems, the Internet based Virtual Private Network(Internet VPN) technology provides an alternativecost-effective high speed communication corenetwork for remote monitoring and control of theelectric system.

Specifically, Internet VPN technology is a sharedcommunication network architecture, in which thecost of the network is spread over a large numberof users while simultaneously providing both thefunctionalities and the benefits of a dedicated pri-vate network. Therefore, the main objective of anInternet VPN for electric system automation is toprovide the required cost-effective high performancecommunication between IEDs and a remote controlcenter over a shared network infrastructure with thesame policies and service guarantees that the electricutility experiences within its dedicated privatecommunication network. In order to achieve this

objective, the Internet VPN solution should providethe following essential performance attributes:

• Quality of Service (QoS): Internet technologyitself cannot guarantee very strict QoS require-ments that utility applications require, since datacommunication in the Internet is mainly based ona best effort service paradigm. In this respect,QoS capabilities of Internet VPN technologiesensure the prioritization of mission critical ordelay sensitive traffic and manage network con-gestion under varying network traffic conditionsover the shared network infrastructure.

• Reliability: The communication network shouldbe able to operate continuously over an extendedperiod of time, even in the presence of networkelement failures or network congestion. Toachieve this, the communication network shouldbe properly designed with the objective of nolosses in all working conditions and able to dealwith failure gracefully. Service providers supportService Level Agreements (SLAs), which definethe specific terms and performance metricsregarding availability of network resources andoffer the Internet VPN subscriber a contractualguarantee for network services and networkuptime. Therefore, Internet VPN technologyshould deliver data in a reliable and timelymanner for automation applications.

• Scalability: Since the number of substations andremote devices is large and growing rapidly, thecommunication system must be able to deal withvery large network topologies without increasingthe number of operations exponentially for thecommunication network. Thus, the designedhybrid network architecture should scale well toaccommodate new communication requirementsdriven by customer demands.

• Robustness: In order to avoid deterioratingcommunication performance due to changingnetwork traffic conditions, the dimensioning pro-cess to assign the bandwidth to the virtual linksof the Internet VPN should be based not onlyon the main bandwidth demand matrix, but alsoon other possible bandwidth demand matrices toprovide a safe margin in network dimensioningto avoid congestion [28]. In case the network con-gestion can not be avoided with the current net-work traffic, low priority non-critical datatraffic should be blocked so that the most criticaldata can be transmitted with QoS guarantees[26]. This way, additional bandwidth for high


priority data becomes available to enable thereal-time communication of critical data, whichis particularly important in case of alarms in elec-tric systems.

• Security: Security is the ability of supportingsecure communication between a remote controlcenter and field devices in order to make the com-munication safe from external denial of service(DoS) attacks and intrusion. When a public net-work like the Internet is utilized to connect thefield devices to a remote control center, securityconcerns can arise. Hence, Internet VPN has toprovide secure data transmission across an exist-ing shared Internet backbone and thus, protectsensitive data so that it becomes confidentialacross the shared network.

• Network management: In order to provide thecommunication requirements of automationapplications, electric utilities demand flexibleand scalable network management capabilities.The primary network management capabilitiesof Internet VPN include: (i) bandwidth provi-sioning, (ii) installing security and QoS policies,(iii) supporting Service Level Agreements, (iv)fault identification and resolution, (v) additionand removal of network entities, (vi) change ofnetwork functions, (vii) accounting, billing andreporting. In addition to these network manage-ment capabilities, Internet VPN technology canenable rapid implementation and possible modi-fications of the communication network at areasonable cost. Therefore, Internet VPN tech-nology with effective network managementapproaches provides a flexible cost-effective solu-tion that can be easily adapted to future commu-nication requirements that utility automationapplications demand.

Despite the extensive research in Internet VPNtechnologies [32], there are still several openresearch issues, e.g., efficient resource and routemanagement mechanisms, inter-domain networkmanagement, that need to be developed for automa-tion applications. In the current literature, twounique and complementary VPN architecturesbased on Multi Protocol Label Switching (MPLS)and IP Security (IPsec) technologies are emergingto form the predominant communication frame-work for delivery of high performance VPN services[32]. In Fig. 3, we compare both the advantages anddisadvantages of MPLS based VPN and IPSec VPNarchitectures in terms of performance attributes

described above. As shown in Fig. 3, each InternetVPN technology supports the performanceattributes to varying degrees and thus, the mostappropriate choice depends on the specific commu-nication requirements of the electric utilities.

In Fig. 4, a decision tree for choosing an appro-priate Internet VPN technology for electric systemautomation is illustrated. As shown in Fig. 4, if anelectric utility requires a high performance commu-nication network ensuring very strict Quality of Ser-vice (QoS) requirements, the next decision point inthe decision tree can be the size of the communica-tion network, i.e., the number of communicationentities that need to be interconnected. Electricutilities that need to connect a large number of sub-stations and a remote control center should prefercost-effective MPLS based Internet VPN technol-ogy, since they can reduce the communication costsignificantly compared to dedicated private leasedcommunication lines. If the number of sites is notlarge in the network, electric utilities can utilize ahybrid network including IPSec Internet VPN andlayer 2 technologies such as Frame Relay andATM for the automation applications. If there areno QoS communication requirements, the possibleoptions include either using public Internet whenno secure communication is required or using anIPSec Internet VPN when secure communicationis required in automation applications.

In fact, the actual selection of Internet VPN tech-nology depends on several factors such as the costof communication architecture, geographic cover-age of the communication architecture, the loca-tions of substations and a remote control center,service level agreements, network managementtypes, i.e., customer based or network based man-agement, etc. As a result, electric utilities shouldevaluate their unique communication requirementsand the capabilities of Internet VPN technologiescomprehensively in order to determine the bestInternet VPN technology for automationapplications.

3. Last mile connectivity for electric utilities

In this section, both advantages and disadvan-tages of possible communication technologies forlast mile connectivity are explained in detail. Thecommunication technologies evaluated for last mileconnectivity are: (i) Power line communication, (ii)Satellite communication, (iii) Optical fiber commu-nication, (iv) Wireless communication.

Fig. 3. Comparison of MPLS based Internet VPN and IPSec Internet VPN for electric system automation applications.

Fig. 4. A basic Internet VPN decision tree for electric system automation.



3.1. Power Line Communication

Power Line Communication (PLC) is transmis-sion of data and electricity simultaneously overexisting power lines as an alternative to construct-ing dedicated communications infrastructure.Although PLC has been in operation since the1950s as low data rate services such as remote con-trol of power grid devices, it has become moreimportant in recent years due to developments intechnology, which enable PLC’s potential use forhigh speed communications over medium (15/50 kV) and low (110/220 V) voltage power lines[5]. However, there are still several technical prob-lems and regulatory issues that are unresolved.Moreover, a comprehensive theoretical and practi-cal approach for PLC is still missing and there areonly a few general results on the ultimate perfor-mance that can be achieved over the power linechannel. As a result, commercially deployable, highspeed, long distance PLC still requires furtherresearch efforts despite the fact that PLC might pro-vide an alternative cost-effective solution to the lastmile connectivity problem. In the following, weexplain both advantages and disadvantages ofpower line communication technologies for auto-mation applications.

3.1.1. Advantages

• Extensive coverage: PLC can provide an extensivecoverage, since the power lines are alreadyinstalled almost everywhere. This is advanta-geous especially for substations in rural areaswhere there is usually no communicationinfrastructure.

• Cost: The communication network can be estab-lished quickly and cost-effectively because itutilizes the existing wires to carry the communi-cation signals. Thus, PLC can offer substationsnew cost-saving methods for remotely monitor-ing power uses and outages.

3.1.2. Disadvantages

• High noise sources over power lines: The powerlines are noisy environments for data communi-cations due to several noise sources such as elec-trical motors, power supplies, fluorescent lightsand radio signal interferences [27]. These noisesources over the power lines can result in high

bit error rates during communication whichseverely degrade the performance of PLC.

• Capacity: New technological advances haverecently enabled a prototype communicationmodem which achieves a maximum total capacityof 45 Mbps in PLC [1]. However, since powerline is a shared medium, the average data rateper end user will be lower than the total capacitydepending on coincident utilization, i.e., thenumber of users on the network at the same timeand the applications they are using. Thus, possi-ble technical problems should be comprehen-sively addressed with various field tests beforethe PLC technology is widely deployed.

• Open circuit problem: Communication over thepower lines is lost with devices on the side ofan open circuit [14]. This fact severely restrictsthe usefulness of PLC for applications especiallyinvolving switches, reclosers and sectionalizers.

• Signal attenuation and distortion: In power lines,the attenuation and distortion of signals areimmense due to the reasons such as physicaltopology of the power network and load imped-ance fluctuation over the power lines. In addi-tion, there is significant signal attenuation atspecific frequency bands due to wave reflectionat the terminal points [12]. Therefore, the com-munication over power lines might be lost dueto high signal attenuation and distortion.

• Security: There are some security concerns forPLC arising from the nature of power lines [22].Power cables are not twisted and use no shieldingwhich means power lines produce a fair amountof Electro Magnetic Interference (EMI). SuchEMI can be received via radio receivers easily.Therefore, the proper encryption techniquesmust be used to prevent the interception ofcritical data by an unauthorized person.

• Lack of regulations for broadband PLC: In addi-tion to technical challenges, fundamental regula-tion issues of PLC should be addressed forsubstantial progress to be made. The limits oftransmitted energy and frequencies employedfor PLC should be determined in order to bothprovide broadband PLC and prevent the interfer-ence with already established radio signals suchas mobile communications, broadcastingchannels and military communications. In thisrespect, the Institute of Electrical and ElectronicsEngineers (IEEE) has started to develop a stan-dard to support broadband communications


over power lines [18]. The standard is targeted forcompletion in mid 2006.

1 Satellites can be classified into Geostationary Earth Orbit(GEO) satellite, Middle Earth Orbit (MEO) satellite, Low EarthOrbit (LEO) satellite according to the orbit altitude above theearth’s surface [9].

3.2. Satellite communication

Satellite communication can offer innovativesolutions for remote control and monitoring of sub-stations. It provides an extensive geographic cover-age, and thus, can be a good alternativecommunication infrastructure for electric systemautomation in order to reach remote substationswhere other communication infrastructures such astelephone or cellular networks might not exist. Inpractical applications, Very Small Aperture Termi-nal (VSAT) satellite services are already availablethat are especially tailored for remote substationmonitoring applications [33]. Furthermore, withthe latest developments in electric system automa-tion, satellite communication is not only used forremote control and monitoring of substations butalso used for Global Positioning System (GPS)based time synchronization, which provides micro-second accuracy in time synchronization [38]. Inaddition, satellites can be used as a backup for theexisting substations communication network. Incase of congestion or link failures in communica-tion, critical data traffic can be routed throughsatellites [8]. In the following, we present bothadvantages and disadvantages of satellite communi-cation technologies.

3.2.1. Advantages

• Global coverage: Satellite communication sup-ports a wide geographical coverage (includingremote, rural, urban and inaccessible areas) inde-pendent of the actual land distance between anypair of communicating entities. In case no com-munication infrastructure exists, especially forremote substations, satellite communication pro-vides a cost-effective solution.

• Rapid installation: Satellite communication offersclear advantages with respect to the installationof wired networks. A remote substation can joina satellite communication network by onlyacquiring the necessary technical equipmentwithout the need for cabling to get high-speedservice [20]. Cabling is not a cost-effective nor asimple job when the substation is located in aremote place. Due to economical reasons, someutilities have already installed satellite communi-cation for rural substations monitoring [33].


• Long delay: The round-trip delay in satellite com-munication, especially for Geostationary EarthOrbit (GEO) satellites,1 is substantially higherthan that of terrestrial communication links.The transport protocols developed for terrestrialcommunication links such as TCP are not suit-able for satellite communication, since necessarydata rate adjustments of TCP can take a longtime in high-delay networks such as satellite net-works [16]. On the other hand, it is possible toreduce the round-trip delay by using satellites inlower orbits. Particularly, LEO satellites offersignificantly reduced delay, which is comparableto that of terrestrial networks.

• Satellite channel characteristics: Different fromcabled and terrestrial network communications,satellite channels characteristics vary dependingon the weather conditions and the effect of fad-ing, which can heavily degrade the performanceof the whole satellite communication system[16]. Therefore, these communication challengesshould be taken into account while evaluatingthe communication technologies for electricsystem automation.

• Cost: Although satellite communication can be acost-effective solution for remote substations ifany other communication infrastructure is notavailable, the cost for operating satellites (theinfrastructure cost and monthly usage cost) forall substation communication networks is stillhigher than that of other possible communicationoptions. High initial investment for satellitetransceivers is one of the limitations of satellitecommunication.

3.3. Optical fiber communication

Optical fiber communication systems, which werefirst introduced in the 1960s, offer significant advan-tages over traditional copper-based communicationsystems. In electric system automation, an opticalfiber communication system is one of the technicallyattractive communication infrastructures, providingextremely high data rates. In addition, its ElectroMagnetic Interference (EMI) and Radio Frequency


Interference (RFI) immunity characteristics make itan ideal communication medium for high voltageoperating environment in substations [14]. Further-more, optical fiber communication systems supportlong distance data communication with less numberof repeaters2 compared to traditional wired net-works. This leads to reduced infrastructure costsfor long distance communication that substationmonitoring and control applications demand. Forexample, the typical T-1 or coaxial communicationsystem requires repeaters about every 2 km whereasoptical fiber communication systems require repeat-ers about every 100–1000 km [13].

Although optical fiber networks have severaltechnical advantages compared to other wired net-works, the cost of the optical fiber itself is stillexpensive to install for electric utilities. However,the enormous bandwidth capacity of optical fibermakes it possible for substations to share the band-width capacity with other end users which signifi-cantly helps to recover the cost of the installation.In this respect, optical fiber communication systemsmight be cost-effective in the high speed communi-cation network backbone since optical fibers arealready widely deployed in communication networkbackbones and the cost is spread over a large num-ber of users. As a result, fiber optic networks canoffer high performance and highly reliable commu-nication when strict QoS substations communica-tion requirements are taken into account. In thefollowing, we describe both advantages and disad-vantages of optical fiber communication for auto-mation applications.

3.3.1. Advantages

• Capacity: Extremely high bandwidth capacity ofoptical fiber communication can provide highperformance communication for automationapplications. Current optical fiber transmissionsystems provide transmission rates up to 10 Gbpsusing single wavelength transmission and40 Gbps to 1600 Gbps using wavelength divisionmultiplexing3 (WDM). In addition, very low bit

2 In long distance communications, it is necessary to introducerepeaters periodically in order to compensate for the attenuationand distortion of the communication signals.

3 Wavelength division multiplexing (WDM) is an effectiveapproach to exploit the bandwidth capacity available in opticalfiber. In WDM, multiple wavelengths are used to carry severaldata streams simultaneously over the same fiber.

error rates (BER = 10�15) in fiber optic commu-nication are observed. Due to high bandwidthcapacity and low BER characteristics, opticalfiber is used as the physical layer of Gigabitand 10 Gigabit ethernet networks.

• Immunity characteristics: Optical fibers do notradiate significant energy and do not pick upinterference from external sources [13]. Thus,compared to electrical transmission, opticalfibers are more secure from tapping and alsoimmune to EMI/RFI interference and crosstalk.


• Cost: Although fiber optic networks possess sev-eral technical advantages, the cost of its installa-tion might be expensive in order to remotelycontrol and monitor substations. However, fiberoptic networks might be a cost-effective commu-nication infrastructure for high speed communi-cation network backbones, since optical fibersare already widely deployed in the communica-tion network backbones and the cost is spreadover a large number of users.

3.4. Wireless communication

Several wireless communication technologies cur-rently exist for electric system automation [14].When compared to conventional wired communica-tion networks, wireless communication technologieshave potential benefits in order to remotely controland monitor substations, e.g., savings in cablingcosts and rapid installation of the communicationinfrastructure. On the other hand, wireless commu-nication is more susceptible to Electro MagneticInterference (EMI) and often has limitations inbandwidth capacity and maximum distances amongcommunication devices. Furthermore, since radiowaves in wireless communication spread in the air,eavesdropping can occur and it might be a threatfor communication security. Electric utilities explor-ing wireless communication options have twochoices; (i) utilizing an existing communicationinfrastructure of a public network, e.g., public cellu-lar networks, (ii) installing a private wirelessnetwork.

Utilizing an existing communication infrastruc-ture of a public network might enable a cost-effec-tive solution due to the savings in required initialinvestment for the communication infrastructure.On the other hand, private wireless networks enable


electric utilities to have more control over their com-munication network. However, private wireless net-works require a significant installation investmentas well as the maintenance cost [14]. In electric sys-tem automation, wireless communication technol-ogy has already been deployed. Recently, ShortMessage Service (SMS) functionality of the digitalcellular network has been applied in order to remo-tely control and monitor substations [34]. The con-trol channel of the cellular network is also utilized insome alarm-based substation monitoring cases [23].However, both of these communication technolo-gies are suited to the applications that send a smallamount of data and thus, they can not provide thestrict Quality of Service (QoS) requirements thatreal time substation monitoring applicationsdemand. In the following, we describe both advan-tages and disadvantages of wireless communicationtechnologies.

3.4.1. Advantages

• Cost: Utilizing an existing wireless communica-tion network, e.g., cellular network, might enablea cost-effective solution due to the savings inrequired initial investment for the communica-tion infrastructure. In wireless communication,cabling cost is also eliminated.

• Rapid installation: The installation of wirelesscommunication is faster than that of wired net-works. Wireless communication provides moreflexibility compared to wired networks. Withinradio coverage, communication entities can startto communicate after a short communicationinfrastructure installation.


• Limited coverage: Private wireless networks pro-vide a limited coverage, e.g., the coverage ofIEEE 802.11b is approximately 100 m [21]. Onthe other hand, utilizing existing public wirelessnetwork, e.g. cellular network, or WiMAX tech-nology can support much more extensive cover-age compared to wireless local area networks.However, some geographical areas, e.g., remoterural locations, may still not have any wirelesscommunication services.

• Capacity: Wireless communication technologiesprovide typically lower QoS compared to wiredcommunication networks. Due to limitationsand interference in radio transmission, a limited

bandwidth capacity is supported and high biterror rates (BER = 10�2–10�6) are observed incommunication. In addition, since wireless com-munication is in a shared medium, the applica-tion average data rate per end user is lowerthan the total bandwidth capacity, e.g., maxi-mum data rate of IEEE 802.11b is 11 Mbps whilethe average application data rate is approxi-mately 6 Mbps [21]. Therefore, each level in thecommunication protocol stack should adapt towireless link characteristics in an appropriatemanner, taking into account the adaptive strate-gies at the other layers, in order to optimize net-work communication performance.

• Security: Wireless communication poses serioussecurity challenges since the communication sig-nals can be easily captured by nearby devices.Therefore, efficient authentication and encryp-tion techniques should be applied in order toprovide secure communication.

Note that with the recent advances in wirelesscommunications and digital electronics, hybrid net-work architectures have enabled alternative scalablewireless communication systems, which can providestrict quality of service (QoS) requirements of auto-mation applications. The details of these recentwireless technologies, i.e., wireless sensor networks,WiMAX and wireless mesh networks, are describedin the following sections.

4. Wireless sensor networks for automation

In this section, we explain the opportunities andchallenges of wireless sensor networks (WSNs) andpresent design objectives and requirements of WSNsfor electric system automation applications. In gen-eral, wireless sensor networks are composed of alarge number of low cost, low power and multifunc-tional sensor nodes that are small in size and com-municate un-tethered over short distances [4]. Theever-increasing capabilities of these tiny sensornodes enable capturing various physical informa-tion, e.g., noise level, temperature, vibration, radia-tion, etc., as well as mapping such physicalcharacteristics of the environment to quantitativemeasurements. The collaborative nature of WSNsbrings several advantages over traditional sensingincluding greater fault tolerance, improved accu-racy, larger coverage area and extraction of local-ized features. In this respect, wireless sensornetworks enable low cost and low power wireless


communication for electric system automationapplications, especially in urban areas.

Furthermore, in the area of electric utility mea-surement systems, WSNs are used in wireless auto-matic meter reading (WAMR) systems, which candetermine real-time energy consumption of the cus-tomers accurately. WAMR systems are importantfor electric utilities, since they can reduce opera-tional costs and enable remotely controlled flexiblemanagement systems based on real-time energy con-sumption statistics. Therefore, WSNs provide analternative real-time monitoring system for electricutilities with the potential to improve business per-formance and technical reliability of various electricutility operations.

In WSNs, the architecture of the networkdepends on the purpose of the application. Basedon the application requirements, the sensor nodesare scattered in a sensor field as shown in Fig. 5.Each of these scattered sensor nodes has the capa-bility to collect data and route data back to the sinknode in a multi-hop manner [3]. In this architecture,the role of the sink node is to monitor the overallnetwork and to communicate with the task man-ager, e.g., control center in the power utility, inorder to decide the appropriate actions. The sinknode can communicate with the task manager viaInternet or satellite.

4.1. Benefits of wireless sensor networks for

automation

Wireless Sensor Network (WSN) technology hascreated new communication paradigms for real-timeand reliable monitoring requirement of the electric

Fig. 5. An illustrated architecture

systems. Some of the benefits that can be achievedusing WSN technology are highlighted as follows:

• Monitoring in harsh environments: The sensors inWSNs are rugged, reliable, self-configurable andunaffected by extreme ambient conditions, e.g.,temperature, pressure, etc. Thus, WSNs canoperate even in harsh environments and elimi-nate the cabling requirement in electric systems.

• Large coverage: WSNs can contain a large num-ber of physically separated sensor nodes that donot require human intervention. Although thecoverage of a single sensor node is small, denselydistributed sensor nodes can work simulta-neously and collaboratively so that the coverageof the whole network is extended. Therefore,the coverage limitations of traditional sensingsystems can be addressed efficiently.

• Greater fault tolerance: The dense deployment ofsensor nodes leads to high correlation in thesensed data. The correlated data from neighbor-ing sensor nodes in a given deployment areamakes WSNs more fault tolerant than conven-tional sensor systems. Due to data redundancyand the distributed nature of WSNs, adequatemonitoring information can be transported tothe remote control center even in the case ofsensor and route failures.

• Improved accuracy: The collective effort of sensornodes enables accurate observation of the physi-cal phenomenon compared to traditional moni-toring systems [17]. In addition, multiple sensortypes in WSNs provide the capability of monitor-ing various physical phenomena in the electricsystem.

of wireless sensor networks.


• Efficient data processing: Instead of sending theraw data to the remote control center directly,sensor nodes can locally filter the sensed dataaccording to the application requirements andtransmit only the processed data. Thus, only nec-essary information is transported to the remotecontrol center and communication overheadcan be significantly reduced.

• Self configuration and organization: The sensornodes in WSNs can be rapidly deployed anddynamically reconfigured because of the self-con-figuration capability of the sensor nodes. The adhoc architecture of WSNs also overcomes thedifficulties raised from the predetermined infra-structure requirements of traditional communi-cation networks. More specifically, new sensornodes can be added to replace failed sensor nodesin the deployment field and existing nodes canalso be removed from the system without affect-ing the general objective of the monitoringsystem of the electric utility.

• Lower cost: WSNs are expected to be less expen-sive than conventional monitoring systems,because of their small size and lower price as wellas the ease of their deployment.

4.2. Wireless sensor network applications for

automation

WSN technology can enhance the performanceof electric utility operations by enabling wirelessautomatic meter reading and real-time and reliablemonitoring systems for electric utilities. In thefollowing, WSN applications for electric systemautomation are described in detail.

4.2.1. Wireless automatic meter reading (WAMR)

Currently, traditional manual electricity meterreading is the most common method for the electricutilities. These systems require visual inspection ofthe utility meters and do not allow flexible manage-ment systems for the electric utilities. In addition,network connections between traditional metersand data collection points are basically non-exis-tent; thus, it is impossible to implement a remotelycontrolled flexible management system based onenergy consumption statistics by using traditionalmeasurement systems.

With the recent advances in Micro Electro-Mechanical Systems (MEMS) technology, wirelesscommunications and digital electronics; the devel-

opment of low cost smart sensor networks, thatenable wireless automatic meter reading (WAMR)systems, has become feasible. As the de-regulationand competition in the electric utility marketplaceincrease, so does the importance of WAMR sys-tems. Wireless collection of electric utility meterdata is a very cost-effective way of gathering energyconsumption data for the billing system and it addsvalue in terms of new services such as remote deac-tivation of a customer’s service, real-time price sig-nals and control of customers’ applications. Thepresent demand for more data in order to makecost-effective decisions and to provide improvedcustomer service has played a major role in themove towards WAMR systems.

WAMR systems offer several advantages to elec-tric utilities including reduced electric utility opera-tional costs by eliminating the need for humanreaders and real-time pricing models based onreal-time energy consumption of the customers.Real-time pricing capability of WAMR systemscan also be beneficial for the customers. For exam-ple, using the real-time pricing model, the electricutility can reward the customers shifting theirdemand to off-peak times. Therefore, the electricutility can work with customers to shift loads andmanage prices efficiently by utilizing WAMR sys-tems instead of once a month on-site traditionalmeter reading.

However, the real-time pricing model of electricutilities requires reliable two-way communicationbetween the electric utility and customer’s meteringequipment. WSN technology addresses this require-ment efficiently by providing low cost and lowpower wireless communication. In Fig. 6, a wirelessautomatic meter reading system using sensor net-work technology is illustrated. As shown in Fig. 6,the sensed data from the meter is collected by theutility control center through multi-hop wirelesscommunication. This monitoring system can alsoprovide flexibility to the electric utility so that utilitypersonnel or mobile utility controller can monitorthe system locally when it is required, e.g., in caseof alarm situations.

In summary, wireless automatic meter readingsystems can provide the following functionalitiesfor electric systems:

• Automatic meter reading functionalities: WSNsenable real-time automatic measurement ofenergy consumption of the customers. The

Fig. 6. An illustration of WAMR system using wireless sensor network technology.


automatic meter measurements can also be classi-fied as individual meter measurements, clustermeter measurements and global meter measure-ments. Here, the objective is to provide flexiblemanagement policies with different real-timemonitoring choices for electric utilities.

• Telemetry functionalities: The electric utility con-trol centers can obtain real-time data from smartsensor nodes and control some elements locatedat selected points of the distribution network,e.g. control of the status of the switches [25].Thus, distributed sensing and automationenhance electric utility services by reducingfailure and restoration times.

• Dynamic configuration functionality: In electricsystem automation applications, reliability ofthe measurements should be ensured even in caseof route failures in the network [31]. Thus, it isextremely significant to dynamically adjust theconfiguration of the network, e.g., dynamic rout-ing, in order to provide reliability requirementsof the applications. In this respect, the self-con-figuration capability of WSNs enables dynamicreconfiguration of the network.

• Status monitoring functionality: Monitoring thestatus of the metering devices, which are embed-

ded by smart sensors, is another functionality ofWAMR systems. This functionality can be veryhelpful to determine sensor node failures in thenetwork accurately and timely. In addition, sta-tus monitoring functionality can be utilized incase of tampering with metering devices. Forexample, if someone tries to vandalize a meteringdevice, the system can notify the police automat-ically [24]. This reduces the considerable costs ofsending service crews out to repair vandalizedmetering devices.

As advances in WAMR technologies continue,these systems will become less expensive and morereliable. Most utility and billing companies haverecognized that with the invention of low-cost, lowpower radio sensors, wireless RF communicationis, by far, the most cost-efficient way to collectutility meter data.

4.2.2. Electric system monitoring

Equipment failures, lightning strikes, accidents,and natural catastrophes all cause power distur-bances and outages and often result in long serviceinterruptions. Thus, the electric systems should beproperly controlled and monitored in order to take


the necessary precautions in a timely manner [36]. Inthis respect, wireless sensor networks (WSNs) can aprovide cost-effective reliable monitoring system forthe electric utilities [29]. An efficient monitoringsystem constructed with smart sensor nodes canreduce the time for detection of the faults andresumption of electric supply service in distributionnetworks.

In addition, electricity regulators monitor theperformance of the electricity distribution networkoperators utilizing a range of indices relating to cus-tomer service. Distribution network operators havetargets and incur penalties based on the length oftime of service interruptions, i.e., both outage fre-quency and duration [30]. Continuity of electricityservice is also crucial in today’s competitive electricutility marketplace from the perspective of customersatisfaction.

In order to evaluate the performance of the elec-tric system, several Quality of Service (QoS) indicescan be obtained utilizing WSN technology. Forexample, average duration of service interruptionand average repair time can be computed. Typically,for densely deployed urban areas, these perfor-mance indices are correlated with the time forremote or manual switching of supply circuits. Inthis context, smart sensor nodes deployed in theelectric utility can provide rapid identification ofservice interruptions and timely restoration of theelectric utility services. Therefore, WSNs can helpelectric utilities maintain regulatory targets for theperformance indices.

4.3. Wireless sensor network design considerations

When wireless sensor network technology forelectric system automation applications is consid-ered, there exist two key design elements which arecritical to develop cost-effective wireless sensor net-work to support both existing functionalities andnew operational requirements of the future electricsystems. These key elements are described in thefollowing.

4.3.1. Network topology and architecture

requirements

The topology of a sensor network has significantimplications on several network aspects, includingnetwork lifetime, routing algorithms, communica-tion range of the sensor nodes and etc. The networkarchitecture requirements contain the physical and

logical organization of the network as well as thedensity of the sensor nodes. In general, the objectiveof sensor networks is to efficiently cover the deploy-ment area. The logical and hierarchical organizationof the network also impacts energy consumptionand the selection of communication protocols. Inaddition, based on topology requirements, sensornetworks can have a distributed organization or aclustered organization, where selected nodes canhandle data forwarding. The network topologyand architecture requirements for electric utilitiescan be determined by answering the followingquestions:

• What type of network topology best fits theapplication? (Is it one to one, one to many, manyto one or many to many?)

• How will the monitoring network work? (Is itmaster–slave, point-to-point, point-to multipointor peer to peer?)

• What are the worst case ambient conditions inthe coverage area?

• How many substations should be controlled andmonitored including both current and futurerequirements of the electric system?

• Are there any known potential interferenceproblems due to physical obstructions, RF inter-ference from power lines or large inductionmotors?

4.3.2. Application requirements

The required information that is to be relayedthrough the sensor network for electric utilitiesshould be classified and quantified [2]. Theserequirements can be achieved by a comprehensiveanalysis of the electric system automation applica-tions. Based on the application requirements, theproperties of individual sensor nodes can also beidentified which impact network modelling andcommunication protocol choices. The followingquestions can help electric utilities to determinethese requirements:

• What are the QoS requirements of the applica-tion? (Does it require real-time monitoring ordelay tolerant monitoring?)

• Does the system continuously poll for the infor-mation (periodic monitoring) or is it generatedby exception (event-based monitoring)?

• What is the type of the sensor data, i.e., video,voice, data?


As a result, electric utilities should determine thenetwork topology, architecture and applicationrequirements comprehensively in order to establishthe best fit wireless sensor network for their applica-tions. Full consideration of the different sensor net-work options and how will they fit the electric utilityapplication is critical for a successful implemen-tation.

4.4. Design challenges of wireless sensor networks

Although WSNs bring significant advantagesover traditional communication networks, the prop-erties of WSNs also impose unique communicationchallenges. These challenges can be described asfollows:

• Limited resources: The design and implementa-tion of WSNs are constrained by three types ofresources: (i) energy, (ii) memory and (iii) pro-cessing. Constrained by the limited physical size,sensor nodes have limited battery energy supply[15]. For this reason, communication protocolsfor WSNs are mainly tailored to provide highenergy efficiency. It is also important to note thatin electric systems, the batteries of the sensors canbe charged by the appropriate energy supplies. Inaddition, the collaborative effort of sensor nodescan handle the problems of limited memory andprocessing capabilities of the sensor nodes.

• Dynamic topologies and environment: The topol-ogy and connectivity of the network may varydue to route and sensor node failures. Further-more, the environment, that sensor nodes moni-tor, can change dramatically, which may causea portion of sensor nodes to malfunction or ren-der the information they gather obsolete. Thus,the developed communication protocols forWSNs should accurately capture the dynamicsof the network [35].

• Quality of service concerns: The quality of service(QoS) provided by WSNs refers to the accuracybetween the data reported to the control centerand what is actually occurring in the environ-ment. In addition, since sensor data are typicallytime sensitive, e.g., alarm notifications for theelectric utilities, it is important to receive the dataat the control center in a timely manner. Datawith long latency due to processing or communi-cation may be outdated and lead to wrong deci-sions in the monitoring system. Therefore, the

developed communication protocols for WSNsshould address both real-time and reliablecommunication simultaneously.

5. WiMAX and wireless mesh networks for

automation

Recall that the proposed hybrid network archi-tecture consists of various types of networks includ-ing Internet VPN, wireless sensor networks,WiMAX and wireless mesh networks. In the previ-ous sections, we described the details of InternetVPN technologies (see Section 2) and wireless sen-sor networks (see Section 4) for electric utilities. Inthis section, we focus on wireless mesh networksand WiMAX technology for electric system auto-mation applications.

In Fig. 7, an illustration of the hybrid networkarchitecture utilizing WiMAX technology and wire-less mesh networks is shown. In this hybrid architec-ture, a set of electric utility subscribers is clusteredinto wireless mesh domains, where each domainhas a smaller dimension compared to the global net-work. Hence, each wireless mesh domain can be eas-ily managed by the centralized communicationentities, which are called as local control centers.Furthermore, in this architecture, each wirelessmesh cluster is monitored by a remote control cen-ter using WiMAX. Therefore, with the integrationof wireless mesh networks and WiMAX, electricutilities can fully exploit the advantages of multiplewireless networks. The main components of the pro-posed hybrid network architecture are brieflydescribed as follows:

• Wireless mesh domains: In the proposed hybridnetwork architecture, wireless mesh domainsconstitute a fully connected wireless networkamong each electric utility subscriber. Differentfrom traditional wireless networks, each wirelessmesh domain is dynamically self-organized andself-configured. In other words, the nodes in themesh network automatically establish and main-tain network connectivity. This feature bringsmany advantages for electric utilities, such aslow up-front cost, easy network maintenance,robustness, and reliable service coverage. In addi-tion, with the use of advanced radio technologies,e.g., multiple radio interfaces and smart anten-nas, network capacity can be increased signifi-cantly. Moreover, the gateway and bridge

Fig. 7. An illustration of hybrid network architecture using WiMAX and wireless mesh networks.


functionalities in mesh routers enable the integra-tion of wireless mesh domains with various exist-ing wireless networks such as wireless sensornetworks, wireless-fidelity (Wi-Fi), and WiMAX[3]. Consequently, through an integrated wirelessmesh network, electric utilities can take theadvantage of multiple wireless networks.

• WiMAX backbone: The necessary long distancecommunication (up to 31 miles) between localcontrol centers and a remote control center isprovided utilizing worldwide inter-operabilityfor microwave access (WiMAX) technology.With the integration of WiMAX technology,the capacity of the network backbone can beincreased up to 75 Mbps. In addition, WiMAXoffers a standardized communication technologyfor point-to-multipoint wireless networks, i.e.,IEEE 802.16 standard [37]. This enables interop-erability between different vendor products,which is another important concern for electricutilities. Furthermore, different from traditionalpoint-to-multipoint networks, WiMAX technol-ogy also supports non-line of sight communica-tion. Hence, electric systems suffering fromenvironmental obstacles can benefit fromWiMAX technology to improve the performanceof their communication system. WiMAX tech-

nology, particularly the IEEE 802.16e standard[37], also focuses on low latency handoff manage-ment, which is necessary for communicationswith users moving at vehicular speeds.

5.1. Benefits of hybrid network architecture using

WiMAX and wireless mesh networks

With the recent advances in wireless communica-tions and digital electronics, hybrid network archi-tectures have enabled alternative scalable wirelesscommunication systems, which can provide strictquality of service (QoS) requirements of electricsystem automation applications in a cost-effectivemanner. Some of the benefits of hybrid networkarchitectures are highlighted as follows:

• Increased reliability: In wireless mesh domains,the wireless backbone provides redundant pathsbetween the sender and the receiver of the wire-less connection. This eliminates single point fail-ures and potential bottleneck links within themesh domains, resulting in significantly increasedcommunications reliability [3]. Network robust-ness against potential problems, e.g., node fail-ures, and path failures due to RF interferencesor obstacles, can also be ensured by the existence


of multiple possible alternative routes. Therefore,by utilizing WMN technology, the network forelectric utilities can operate reliably over anextended period of time, even in the presence ofa network element failure or network congestion.

• Low installation costs: Recently, the main effortto provide wireless connection to the end-usersis through the deployment of 802.11 basedWi-Fi Access Points (APs). To assure almost fullcoverage in a metro scale area for electric systemautomation, it is required to deploy a large num-ber of access points because of the limited trans-mission range of the APs. The drawback of thissolution is highly expensive infrastructure costs,since an expensive cabled connection to the wiredInternet backbone is necessary for each AP.Installation of the required cabling infrastructuresignificantly increases the installation costs aswell as it slows down the implementation of thewireless network. As a result, the deployment ofAPs for wireless Internet connection is costlyand unscalable for electric system automationapplications. On the other hand, constructing awireless mesh network decreases the infrastruc-ture costs, since the mesh network requires onlya few points of connection to the wired network.Hence, WMNs can enable rapid implementationand possible modifications of the network at areasonable cost, which is extremely importantin today’s competitive electric utility environ-ment.

• Large coverage area: Currently, the data rates ofwireless local area networks (WLANs) have beenincreased, e.g., 54 Mbps for 802.11a and 802.11g,by utilizing spectrally efficient modulationschemes. Although the data rates of WLANsare increasing, for a specific transmission power,the coverage and connectivity of WLANsdecreases as the end-user becomes further fromthe access point. On the other hand, WiMAXtechnology enables long distance communicationbetween local control centers and a remote con-trol center without any performance degradation.As a result, the WiMAX backbone in the hybridnetwork can realize high speed long distancecommunication that automation applicationsdemand.

• Automatic network connectivity: In the proposedhybrid network architecture, wireless meshdomains are dynamically self-organized andself-configured. In other words, the nodes in the

mesh network automatically establish and main-tain network connectivity, which enables seam-less multi-hop interconnection service for theelectric utilities. For example, when new nodesare added into the network, these nodes utilizetheir meshing functionalities to automaticallydiscover all possible routers and determine theoptimal paths to the control centers [3]. Further-more, the existing mesh routers reorganize thenetwork considering the newly available routesand hence, the network can be easily expanded.The self-configuration feature of wireless meshnetworks is so crucial for electric system automa-tion applications, since it enables electric utilitiesto cope with new connectivity requirementsdriven by customer demands.

5.2. Design challenges of hybrid architecture using

WiMAX and wireless mesh networks

Hybrid network architectures can provide aneconomically feasible solution for the wide deploy-ment of high speed wireless communications forelectric system automation applications. Some com-panies already have some products for sale and havestarted to deploy wireless mesh networks andWiMAX towers for various application scenarios.However, field trials and experiments with existingcommunication protocols show that the perfor-mance of hybrid network architectures is still farbelow what they are expected to be [3]. Therefore,there is a need for the development of novel commu-nication protocols for hybrid network architecturesand thus, many open research issues need to beresolved. Some of these research issues are describedas follows:

• Harsh monitoring environment: In substations,wireless links exhibit widely varying characteris-tics over time and space due to obstructionsand extremely noisy environment caused bypower lines and RF interferences. To improvenetwork capacity and limit radio interferences,advanced radio technologies, such as multiple-input multiple output (MIMO) techniques, mul-tiple radio interfaces and smart antennas, shouldbe exploited while developing communicationprotocols.

• Optimal placement of WiMAX towers: In theproposed hybrid architecture, it is importantto design an efficient and low cost network


infrastructure, while meeting the deadlines of thetime-critical monitoring data. Thus, the WiMAXtowers, equipped with expensive RF hardware,should be optimally placed in the deploymentfield in order to both reduce infrastructure costsand meet QoS requirements.

• Mobility support: Low latency handover manage-ment algorithms are required to support the com-munication services of mobile utility controllers.This way, mobile utility controllers can also mon-itor the system locally, when it is necessary, e.g.,in case of alarm situations.

• Integration of heterogeneous networks: Existingnetworking technologies have limited capabilitiesof integrating different wireless networks. Thus,to increase the performance of the hybrid net-work architectures, the integration capabilitiesof multiple wireless interfaces and the corre-sponding gateway/bridge functions of networkrouters should be improved.

• Scalability: In today’s competitive dynamic mar-ket environment, electric utilities might be able todeploy new substations and provision large ser-vice requests rapidly. In this respect, the designedhybrid network architecture should scale well toaccommodate new communication requirementsdriven by customer demands.

• Coordinated resource management: Distributiveand collaborative network resource managementis required to effectively respond to systemchanges due to wireless channel characteristics,contention and traffic patterns. This way, sys-tem-wide fairness and self-configuration of thenetwork can be realized.

• Security: Denial of service attacks in the networkmay cause severe damage to the operation of thedeployed hybrid network. Using efficient encryp-tion and cryptography mechanisms, securityproblems can be solved.

To solve all of these existing problems of hybridnetwork architectures, the protocol stack fromphysical to application layers needs to be improvedor re-invented. In this regard, a cross-layer design isrequired to jointly optimize the main networkingfunctionalities and to design communication proto-col suites that are adaptive to the dynamic charac-teristics of the wireless channel. This way, thehybrid network architecture can provide rapid iden-tification of service interruptions and timely restora-tion of the electric utility services.

6. Conclusion

Electric utilities, especially in urban areas, con-tinuously encounter the challenge of providing reli-able power to the end-users at competitive prices.Equipment failures, lightning strikes, accidents,and natural catastrophes all cause power distur-bances and outages and often result in long serviceinterruptions. In this regard, electric system auto-mation, which is the creation of a highly reliable,self-healing electric system that rapidly respondsto real-time events with appropriate actions, aimsto maintain uninterrupted power services to theend-users. However, the operational and commer-cial demands of electric utilities require a high-performance data communication network thatsupports both existing functionalities and futureoperational requirements. Therefore, the design ofa cost-effective and reliable network architecture iscrucial.

As the individual communication capabilitiesand locations of electric systems are taken intoaccount, it is appropriate to consider the overallcommunication infrastructure as a hybrid networkarchitecture. This hybrid network architecture con-sists of various types of networks such as Internet,wireless sensor networks, WiMAX and wirelessmesh networks. In this hybrid architecture, the com-munication network can be dynamically self-config-ured. This brings significant advantages for electricutilities, such as low up-front cost, easy networkmaintenance, robustness, and reliable service cover-age. Furthermore, with the integration of differentnetworks, electric utilities can fully exploit theadvantages of multiple wireless networks. Forexample, while low power and low range wirelesssensor networks can be utilized for urban areas,WiMAX technology, which enables reliable longdistance communication, can be used for ruralareas. As a result, the proposed hybrid networkarchitecture enables a fully connected communica-tion network for electric system automation appli-cations, such as real-time grid and equipmentmonitoring and wireless automatic meter readingsystems.

In this paper, the opportunities and challenges ofhybrid network architecture are discussed for elec-tric system automation applications. More specifi-cally, Internet based Virtual Private Networks,power line communications, satellite communica-tions and wireless communications (wireless sensor


networks, WiMAX, wireless mesh networks) aredescribed in detail. The motivation of this paper isto provide a better understanding of the hybrid net-work architecture that can provide heterogeneouselectric system automation application require-ments. Consequently, our aim is to present a struc-tured framework for electric utilities who plan toutilize new communication technologies for auto-mation and hence, to make the decision-makingprocess more effective and direct. Based on ourcomprehensive research, we make the following rec-ommendations for the electric utilities:

• Internet-ready IEDs: Recent advances in digitalelectronics and communication technology haveenabled the development of Internet-ready Intel-ligent Electronic Devices (IEDs). Internationalstandards are being developed (IEC 61850) topromote rapid configuration and integration intothe utility automation system [19]. Integratingthese IEDs into electric systems can offer variousbenefits, e.g., remote access to IED/relay config-uration ports, diagnostic event information, andvideo for security or equipment status assess-ment. To make sure these benefits are fullyexploited, there is a need for the appropriate dig-ital simulators in order to test and evaluate theperformance of multi-vendor IEDs and makemore informed decisions.

• Novel communications protocols: Although thehybrid network architecture offers many oppor-tunities for electric utilities, field trials and exper-iments with existing communication protocolsshow that the performance of hybrid networkarchitecture is still far below what they areexpected to be. Therefore, there is a need forthe development of novel communication proto-cols for hybrid network architectures and thus,many open research issues, such as coordinatednetwork management, security, mobility support,integration of heterogeneous networks, need tobe resolved.

• Cost vs. benefit analysis: While providing commu-nication requirements of automation applica-tions, the hybrid network architecture shouldenable rapid implementation and possible modi-fications of the electric utility network. In thisregard, the cost of the network should also beconsidered in order to make it feasible subjectto budget constraints of the electric utilities.Hence, a detailed cost vs. benefit analysis is

required to evaluate the performance of thehybrid network architecture.

• Wireless communications technologies: Wirelesscommunications technologies (WiMAX, wirelesssensor networks, wireless mesh networks) shouldbe developed for deployment in electric systemautomation applications (see Sections 4 and 5).WiMAX is expected to become commercial in2007–2008 and brings several advantages, suchas mobility support and large coverage area.Wireless sensor networks and wireless mesh net-works are under development and offer electricutilities low installation costs, increased reliabil-ity and self-configuration.

• Power line communications technologies: Powerline communications (PLC) technologies shouldbe developed for deployment in electric systemautomation applications. PLC has becomeimportant in recent years due to developmentsin technology, which enable PLC’s potential usefor medium and high speed communications overmedium (15/35 kV) and low (120/240 V) voltagepower lines. However, there are still several tech-nical problems and regulatory issues that areunresolved (see Section 3.1). Moreover, a com-prehensive theoretical and practical approachfor PLC is still missing and there are only fewgeneral results on the ultimate performance thatcan be achieved over the power line channel. Asa result, commercially deployable, high speed,long distance PLC still requires further researchefforts. International standards are also neededfor building power system applications andcustomer services on top of PLC technologies.

Acknowledgements

The authors would like to thank Tom Weaver,James Bales, Doug Fitchett, Eric Rehberg, RayHayes (AEP); Brian Deaver (Baltimore Gas & Elec-tric); Dan Landerman (Cooper Power Systems);Brad Black, Shawn Ervin, Jeff Daugherty (DukePower); Jerry Bernstein (Entergy), Wayne Zessin,Mark Browning (Exelon); Pat Patterson, MartinGordon (NRECA); Mark Gray, Bill Robey(PEPCO); David White (South Carolina Electric& Gas); Brian Dockstader (Southern CaliforniaEdison); Larry Smith, Bob Cheney, Mac Fry, BobReynolds (Southern Company), Joe Rostron(Southern States); Frank Daniel (TXU) for theirvaluable comments that improved the quality of this


paper. This work was supported by NEETRAC un-der Project #04-157.

References

[1] Y. Abe et al., Development of high speed power linecommunication modem, SEI Technical Review 58 (June)(2004) 28–33.

[2] J. Adam et al., EHT control systems and wireless commu-nications: the wave of the future, in: IEEE IndustryApplications Society Petroleum and Chemical IndustryConference, September 2001, pp. 169–178.

[3] I.F. Akyildiz, X. Wang, W. Wang, Wireless mesh networks:a survey, Computer Networks Journal (March) (2005).

[4] I.F. Akyildiz, W. Su, Y. Sankarasubramaniam, E. Cayirci,Wireless sensor networks: a survey, Computer Networks 38(4) (2002) 393–422.

[5] I.H. Cavdar, A solution to remote detection of illegalelectricity usage via power line communications, IEEETransactions on Power Delivery 19 (October) (2004) 1663–1667.

[6] F. Cleveland, R. Ehlers, Guidelines for implementing sub-station automation using UCA-SA (Utility CommunicationsArchitecture-Substation Automation), Electric PowerResearch Institute Technical Report 1002071, 2003.

[7] R. Cohen, On the establishment of an access VPN inbroadband access networks, IEEE CommunicationsMagazine 41 (February) (2003) 156–163.

[8] E. Ekici, I.F. Akyildiz, M.D. Bender, A multicast routingalgorithm for LEO satellite IP networks, IEEE/ACMTransactions on Networking 10 (2) (2002) 183–192.

[9] B.R. Elbert, Satellite Communications Applications Hand-book, Artech House Publishers, 2004.

[10] G.N. Ericsson, Communication requirements—basis forinvestment in a utility wide-area network, IEEE Transac-tions on Power Delivery 19 (January) (2004) 92–95.

[11] G.N. Ericsson, Classification of power systems communica-tions needs and requirements: experiences from case studiesat Swedish National Grid, IEEE Transactions on PowerDelivery 17 (April) (2002) 345–347.

[12] S. Galli, A. Scaglione, K. Dosterl, Broadband is power:Internet access through the power line network, IEEECommunications Magazine 41 (May) (2003) 82–83.

[13] A.L. Garcia, I. Widjaja, Communication Networks: Fun-damental Concepts and Key architectures, McGraw-Hill,2004.

[14] F. Goodman et al., Technical and system requirements foradvanced distribution automation, Electric Power ResearchInstitute Technical Report 1010915, June 2004.

[15] V.C. Gungor, A forecasting-based monitoring and tomog-raphy framework for wireless sensor networks, in: Proc. ofIEEE ICC, Istanbul, Turkey, June 2006.

[16] Y. Hu, V.O.K. Li, Satellite-based Internet: a tutorial, IEEECommunications Magazine 39 (March) (2001) 154–162.

[17] P.H. Ibarguengoytia et al., Real time intelligent sensorvalidation, IEEE Transactions on Power Systems 16 (4)(2001) 770–775.

[18] IEEE P1675: Standard for Broadband over Power LineHardware. Available from: <http://grouper.ieee.org/groups/bop/>.

[19] IEC 61850: Standard for Substation Automation Systems.Available from: <http://www.61850.com/>.

[20] A. Jamalipour et al., Guest editorial broadband IP networksvia satellites Part I, IEEE Journal on Selected Areas inCommunications 22 (2) (2004) 213–217.

[21] J. Jun, P. Peddabachagari, M. Sichitiu, Theoretical maxi-mum throughput of IEEE 802.11 and its applications, in:IEEE International Symposium on Network Computing andApplications, April 2003, pp. 249–256.

[22] W. Liu, H. Widmer, P. Raffin, Broadband PLC accesssystems and field deployment in European power linenetworks, IEEE Communications Magazine 41 (May)(2003) 114–118.

[23] J. Lowrey, Automation systems work best when theycommunicate with each other, Rural Electric (April) (2003)38–41.

[24] A. Minosi et al., Intelligent, Low-power and low-costmeasurement system for energy consumption, in: Proc. ofIEEE VECIMS, July 2003, pp. 125–130.

[25] F.J. Molina, et al., Automated meter reading and SCADAapplication for wireless sensor network, in: Proc. ofADHOC-NOW, Canada, 2003, pp. 223–234.

[26] J.C. de Oliveira, C. Scoglio, I.F. Akyildiz, G. Uhl, Newpreemption policies for DiffServ-Aware traffic engineering tominimize rerouting in MPLS networks, IEEE/ACM Trans-actions on Networking 12 (August) (2004) 733–745.

[27] N. Pavlidou, A.J.H. Vinck, J. Yazdani, B. Honary, Powerline communications: state of the art and future trends, IEEECommunications Magazine 41 (April) (2003) 34–40.

[28] D. Pompili, C. Scoglio, V.C. Gungor, VFMAs, Virtual-flowmulti-path algorithms for MPLS, in: Proc. of IEEE ICC,Istanbul, Turkey, June 2006.

[29] J. Rabaey et al., Smart energy distribution and consumption:information technology as an enabling force, TechnicalReport of UC Berkeley, 2001.

[30] J. Roman et al., Regulation of distribution network busi-ness, IEEE Transactions on Power Delivery 14 (2) (1999)662–669.

[31] M. Stephenson et al., Exploiting emerging tools in shortrange wireless technologies, in: International Conference on3G Mobile Communication Technologies, June 2003, pp.348–353.

[32] N.K. Tan, Building VPNs: with IPSec and MPLS, McGraw-Hill Networking, 2003.

[33] A. Tisot, Rio Grande electric monitors remote energy assetsvia satellite, Utility Automation & Engineering T&DMagazine (July) (2004).

[34] T. Tommila, O. Venta, K. Koskinen, Next generationindustrial automation—needs and opportunities, Automa-tion Technology Review (2001).

[35] M.C. Vuran, V.C. Gungor, O.B. Akan, On the interdepen-dence of congestion and contention in wireless sensornetworks, in: Proc. of ICST SenMetrics, San Diego, CA,July 2005.

[36] Z. Xie et al., An information architecture for future powersystems and its reliability analysis, IEEE Transactions onPower Systems 17 (3) (2002) 857–863.

[37] WiMAX Forum. Available from: <http://www.wimaxforum.org>.

[38] G.E. Ziegler, Protection and substation automation, Electra206 (February) (2003) 14–23.

http://grouper.ieee.org/groups/bop/

http://grouper.ieee.org/groups/bop/

http://www.61850.com/

http://www.wimaxforum.org

http://www.wimaxforum.org

mput

Vehbi C. Gungor received his B.Sc. andM.Sc. degree in Electrical and Electron-ics Engineering from Middle East Tech-nical University, Ankara, Turkey, in2001 and 2003, respectively. He is cur-rently a Research Assistant in theBroadband and Wireless NetworkingLaboratory and pursuing his Ph.D.degree at the School of Electrical andComputer Engineering, Georgia Insti-tute of Technology, Atlanta, GA. His

current research interests include wireless sensor networks,wireless mesh networks, and WiMAX.

V.C. Gungor, F.C. Lambert / Co

Frank C. Lambert received his B.E.E andM.S.E.E. degree from Georgia Instituteof Technology, in 1973 and 1976,respectively. He is currently the Electri-cal Systems Program Manager at theNational Electric Energy Testing,Research and Applications Center,Atlanta, GA. His current researchinterests include power delivery equip-ment, automation and systems; powerquality; communications for electric

utility automation, and grid connected hybrid vehicles.

er Networks 50 (2006) 877–897 897

a survey on communication networks for electric system ...a survey on communication networks for...

Documents