电气外文翻译智能电网纯手打

1、毕业设计（外文翻译材料）学 院：电气与电子工程学院 专 业：电气工程及其自动化 学生姓名：指导教师：2015年3月外文翻译（原文）En d-to-E nd Communi cati on Architecture for Smart GridsAbstractSmart grids heavily depe nd on com muni cati on in order to coord in ate the gen erati on, distributi on, and con sumptio n of en ergy -eve n more so if distributed power

2、 pla nts based on ren ewable en ergies are take n into acco unt. Give n the variety of com muni cati on part ners, a heteroge neous n etwork in frastructure con sisti ng of IP-based and suitable field-level networks is the most appropriate solution. This paper investigates such a two-tier infrastruc

3、ture and possible field-level networks with particular attention to metering and supervisory control and data acquisiti on applicati ons. For the problem of n etwork in tegrati on, a comb in ati on of gateway and tunn eli ng soluti ons is proposed which allows a semitransparent end-to-end connection

4、 between application servers and field no des. The feasibility of the approach and implementation details are discussed at the example of powerline com muni cati on and IP-based n etworks inv estigated in the Europea n research project on real-time energy management via powerlines and intern et. Nev

5、ertheless, it is show n that the com muni cati on architecture is versatile eno ugh to serve as a gen eric soluti on for smart grids.In dexTerms Communi cati onn etwork, meteri ng,powerli necom muni cati on, protocol tunn eli ng, smart grid.I. INTRODUCTIONIN AN increasingly deregulated and distribut

6、ed energy market , com muni cati on betwee n the points of en ergy gen erati on , distributi on, and consumption becomes an essential constituent of efficient grid control. The traditional approach of controlling power production by means of a few large power pla nts is less viable if a critical mas

7、s of small, distributed power pla nts have to be take n into acco unt, eve n more so whe n ren ewable sources like solar and wind en ergy are con sidered on a larger scale. These latter depend on the availability of natural resources that cannot be con trolled. If ren ewable en ergies are to play a

8、bigger roleas inten ded by the Europea n Un io n and other policy makerapid and un expected cha nges in power gen erati on become problematic. Wind power, for example, can vary between zero and rated power during15 min, which can yield variati ons in the range of several gigawatts depe nding on the

9、in stalled capacity. Such a situati on can be alleviated if many small, distributed power pla nts can be comb ined to a virtual power pla nt. This, however, n ecessitates appropriate con trol strategies on a local scale and suitable means of com muni cati on to excha nge in formati on and coord in a

10、te the acti ons of the participa nts.On the other side of the power distribution grid, the load peaks gen erated by con sumers duri ng the day are a critical issue because they dema nd the existe nee of quick and available power reserves. One goal of energy management should therefore be the utmost

11、reduction of such peaks. This can be achieved in various ways, such as real-time pricing where the en ergy prices depe nd on the actual dema nd. Today, this is only realized for large industries, which get their power directly from the wholesale market. Smaller in dustries and households are lagg in

12、g beh ind.The other possibility to reduce load peaks is dema nd side man ageme nt, where con sumers regulate their con sumpti on on a relatively fine-grained appliance basis by appropriate scheduling of run times or by exploit ing (mostly thermal) en ergy storage capabilities of in dividual applicat

13、ion domains such as air conditioning systems, heating, refrigerators, hot water generator systems, or -on a larger scale-heat stations. Combined with load-dependent pricing strategies, such peak-avoidi ng en ergy man ageme nt will also be finan cially attractive for the customer.All these con cepts

14、rely on a sig ni fica nt amount of com muni cati on in order to function properly. Specifically, energy consumption must be measured, en ergy provisi on must be con trolled, and the distributi on grid must be monitored and adjusted according to the power flows. On the tech ni cal side, this requires

15、 reliable com muni cati on betwee n en ergy meters, substation controllers, power plant control rooms, and data acquisition centers for planning and billing. On the organizational side, the com muni cati on part ners are the customers, utility compa nies provid ing the en ergy, and grid operators. T

16、he com muni cati on relatio nships are not n ecessarily peer-to-peer(this could be releva nt for low-level dema nd side man ageme nt).Rather,a multiclie nt/multiserver sce nario is realistic, where a group of possibly un related server operators retrieves and changes data on clients distributed in t

17、he field.Communi cati on architectures that can meet the n eeds of smart grids are not straightforward, and the selectio n of suitable com muni cati on tech no logies is subject to many tech ni cal, legal, and strategic restricti ons. It appears, however, that a two-tier architecture con sisti ng of

18、 an IP-based backb one tech no logy and an adequatefield-level n etwork can meet the n eeds fairly well. The purpose of this paper is to inv estigate this further. Secti on II will discuss the specific n eeds for com muni cati on infrastructures imposed by advaneed grid and energy management con cep

19、ts; whereas, Secti ons III and IV an alyze possible soluti ons for the general architecture and field-level networks, respectively. The solution proposed in Secti ons VA/III focuses on powerli ne com muni cati on on the lower level and was developed within the scope of the EU-funded research project

20、 on real-time energy management via powerlines and Internet (REMPLI). Section IX presentssome results from a field test, and Secti on X provides con clusi ons.II. REQUIREMENTS FOR THE COMMUNICATION SY STEMCommuni cati on n etworks suitable for en ergy man ageme nt applicati ons eve n in a loose sens

21、e-n eed to provide dist inct qualities and services which are closely related to applicati on requireme nts and disti nguish them from other n etworks.1) High reliability and availability are standard requirements for n early every com muni cati on system. Nodes should be reachable un der all circum

22、sta nces. While this is n ormally n ot a problem in a wired n etwork, it may be challenging for wireless or powerline infrastructures because com muni cati on cha nn els can cha nge duri ng operati on. In the particular case of powerline systems, such a change may be introduced by distributio n n et

23、work man ageme nt which bala nces the power con sumptio n load on the power grid, particularly on the medium-voltage (MV) level. Switching actions are initiated via various supervisory control and data acquisiti on (SCADA) and con troll ing systems (or eve n manu ally) using specific com muni cati o

24、n protocols that may not be modified. Therefore, there is no straightforward way to simply inform the com muni cati on system management about topology changes that are about to occur. Rather, the com muni cati on system itself must be desig ned for robust ness. If distributi on n etwork man ageme n

25、t switches a sec on dary tran sformer station from primary station A to primary station B, a request and response to and from the node may have to go via primary station A, the con firmati on already via primary statio n B.2) Automatic managementof redundanciesis closely related to the previous requ

26、irement. As some applications are time critical, real-time properties of the network have to be maintained even during topology changes. As stated before, such changes must not be regarded as exceptional situations due to error conditions but occur in normal operati on.3) High coverage and dista nce

27、s. Evide ntly, the no des to be conn ected by the com muni cati on n etwork are distributed in a wide area. Network concepts based on telecommunication systems or powerlines have the potential to fulfill this requirement.4) Large nu mber of com muni cati on no des. If we assume that only one energy

28、meter per customer is connected, a primary station can supply up to tens of thousa nds of no des, particularly in areas of large apartme nt block concen trati on. Eve n though the comma nds and data packets are usually short, total data volume to be transferred in the network is substa ntial, and co

29、m muni cati on overheads can become an issue.5) Appropriate com muni catio n delay and system resp on sive ness s .The Quality-of-Service (QoS) management needs to take care of different data classes such as metering, control, or alarm data. Even if the predo minant com muni cati on relati on ship i

30、s clie nt/server (i.e., an applicati on server polls the meter data or issues con trol comma nds), it may be n ecessary to foresee someth ing like a fast eve nt cha nnel to tran smit, e.g., alarms from the meters to the con trol room.6) Communication security. Data related to energy distribution are

31、 con sidered critical, in particular, whe n they are releva nt for billi ng purposes or grid con trol. Secure com muni cati on is therefore importa nt. Surveys among utilities showed that in tegrity (no malicious modificati on) and authenticity (origin and accessrights are guaranteed)are the most im

32、portant security goals for energy distribution networks, whereas the con fide ntiality aspect is not con sidered to be an issue.7) Ease of deployment and maintenance. For any distributed com muni cati on system, mecha ni sms must be foresee n which facilitate not only the initial installation but pa

33、rticularly the maintenance of the infrastructure during the operation. Features like error mode analysis and error localization, easy update of firm- and software and remote configuration are essential.III. NETWORK TOPOLOGIES AND COMMUNICATION ARCHITECTURESA com muni cati on system for a distributio

34、 n n etwork will preferably have a gen eric two-tier architecture as depicted in Fig. 1.Metering, billing, SCADA, third-party servicesIP-based netv/orkField devices (meters, control equipment)Ownership pcssibihlisUtility corrpanyGrid operatorThird-party service providerintranetPublic network provide

35、rGrid operatorNtslwurk rovidwrGrid operatorNelwork providerUtility worn阳nyGrid operatorThird-party service providerFig. L Principle two-level communication architecture for energy distributioa networks.The mostly centralized high-level equipment such as back-office application servers used for meter

36、ing and billing purposes but also add-on services, possibly offered by third parties, is exclusively used in an IP-based en vir onment like a compa ny intranet or a wide area n etwork, ofte n based on fiber optical n etworks. The distributed com muni cati on en tities in the lower level of the com m

37、uni cati on system such as en ergy meters, switch gears, or con trol equipme nt n eed to be conn ected by an appropriate field-level n etwork. The in terc onn ecti on of the two n etwork doma ins is achieved by n etwork eleme nts we denote in the followi ng as Access Points (APs). This term must not

38、 be mixed up with the APs used in wireless n etworks; it just describes the fact that these n etwork eleme nts permit the higher n etwork level to access data and services of the lower level.TechnologyProsConsDedicated lineIndependent Peimanent connection Large bandwidthExtremely high installation c

39、ostsPOTS ISDN GSMEasy to handle Low modern costs 64 Kbit/s (ISDN) 9.6 Kbit/s (GSM)Operator-dependent services No influence in case of problems Reachability No permanent connection Limited bandwidth (POTS, GSM)GPRSEasy to handleLow modem costs Permanent connection Cost-effective tariffs 56-114 Kbit/s

40、Operator dependent services No influence in case of problems Reachability Additional security needed Limited bandwidthTETRAIndependent High availability High reachabilitySetup of dedicated infrastmeture Limited bandwidthWLANWPANIndependentPermanent connection Low component costsVery limited range卩 o

41、weNineIndependent High reachability Permanent connectionHigher technical effortWhile the basic com muni cati on architecture is rather straightforward, the actual network topologies can be very diverse and depend mostly on the field-level n etwork. With respect to the com muni catio n requireme nts,

42、 several options are possible which are summarized in Table I. Classical wired com muni cati on with dedicated data n etworks conn ect ing the field devices is one possibility .In such a sce nario, the APs would typically be located close to the field devices in the last transformer station (provide

43、d that it can be reached via the IP-based n etwork), and the cable would bridge the last mile. Dedicated wired data networks can be designed to fulfill all requirements, but the installation costs do not permit an intensive use. A less expensive option is wireless n etworks. Tech no logies for wirel

44、ess local area n etworks or pers onal area n etworks like IEEE 802.15.4 can, i n prin ciple, be used as a replaceme nt for wired links in the sce nario men ti oned above. Their in here nt problem is the limited range which would require the AP to be very close to the end devices (thus increasing the

45、 number of APs) or an additional network level (sometimes called neighborhood area network) to bridge the dista nee to more cen tralized APs.A more promisi ng wireless alter native is cellular n etworks as used in telecom muni cati ons, like global system for mobile com muni cati ons (GSM) or genera

46、l packet radio service (GPRS), or worldwide in teroperability for microwave access (WiMAX). They have high coverage and low installation costs for the end devices if the in frastructure already exists. The dow nside is the gen eral depe ndency on the network provider if public telecommunication netw

47、orks are used-regard ing both costs for the com muni cati on cha nn els and quality of service. In this scenario, the AP would be located close to the application servers in the IT in frastructure, e.g., the utility compa ny.The third possibility which is increasingly attracting interest is powerli

48、ne com muni cati on systems. They use the exist ing power cabli ng and n eed only moderate additi onal n etwork eleme nts but require a more complex tech no logy in order to overcome the rather poor com muni cati on cha nnel characteristics. Regard ing the n etwork topology, the AP can be located an

49、y where in the power grid, but a likely positi on is on the MV level in the primary substation.Tech no logical aspects are just one criteri on affect ing the selectio n of com muni cati on architectures. In deregulated markets ,ow nership of equipment, infrastructure, and services is distributed amo

50、ng different bus in ess en tities, which also affects the com muni cati on system in a smart grid. Fig. 1 shows possible ownership structures for the components of a com muni cati on n etwork. The end poi nts of a com muni cati on cha nn el, i.e., application servers and field devices, are usually i

51、n the hands of one compa ny. The com muni cati on cha nnel as such, however, may be provided by some third party .In the case of cellular n etworks, this could be a telecommunication provider. In the case of powerline networks, this could be the grid operator who may be independent from the utility.

52、 Furthermore, there might be add-on service providers n eed ing access to com muni cati on resources as well. The system architecture should therefore allow for a clear separatio n of com muni catio n cha nn els.IV. ACHIEVING END-TO-END COMMUNICATIONAs outl ined in the in troducti on, the predo mina

53、nt com muni cati on relationship in smart grids is the multiclient/ multiserver scenario, where com muni cati on cha nn els are n eeded betwee n applicati ons (primarily SCADA and metering software) and corresponding field hardware (sensors, switch gear, energy meters) in an end-to-end fashion. In m

54、any installations, application server and field hardware, software, and protocols already exist as sta ndards that should be reused. Therefore, the network must be able to handle several applications and, respectively, com muni cati on protocols in parallel. Naturally, not all of these protocols can

55、 be directly considered during the design phase. Rather, the com muni cati on architecture must be ope n to allow for an easy in tegrati on of further protocols in the future.A com mon approach advocated by many authors today is to use TCP/IP (usually based on IPv6 to cope with the address space lim

56、itation problem) as a con siste nt n etwork layer in order to achieve en d-to-e nd com muni cati on. This is appeali ng in theory but requires sufficie nt performa nee in all parts of the n etwork because the n ature of TCP and some com mon applicati on protocols above it impose a nu mber of require

57、ments on the lower layers of the protocol stack. In particular, TCP implicitly assumesthat IP packets can be emitted at arbitrary points in time in a peer-to-peer fashion without any adherence to the properties of the un derly ing medium. This behavior is difficult to support in large-scale n arrowb

58、a nd powerli ne n etworks, where ofte n a master/slave com muni cati on style is adopted (this is also the case in the example considered here). Transmitting IP packets from the master to the slave is not overly problematic because the master can, in principle,initiate data tran smissi on at any tim

59、e. The reverse directi on (i.e., whe n the slave wants to initiate a data transfer) is, however, more difficult. With a large number of slaves, it may take very long until the slave is ultimately polled by the master for its data. The resulting timing characteristics of the network may be unacceptable for applications needing fast responses. Furthermore, most of the curre ntly used meteri ng and SCADA protocols are not based on user d