deliverable 62: results summary report on anm simulations · results summary report on anm...

ADINE is a project co-funded by the European Commission

Project no: TREN/07/FP6EN/S07.73164/038533 /CONS

Project acronym: ADINE

Project title: Active Distribution Network

Deliverable 62:

Results Summary Report on ANM Simulations

Due date of deliverable: 30.9.2010

Actual submission date: 1.12.2010

Start date of project: 1.10.2007 Duration: 36 months

Organisation name of lead contractor for this deliverable: Tampere University of Technology

Revision [1.0]

Project co-funded by the European Commission within the Sixth Framework Programme (2002-2006)

Dissemination level

PU Public X

PP Restricted to other programme participants (including the Commission Services)

RE Restricted to a group specified by the consortium (including the Commission Services)

CO Confidential, only for members of the consortium (including the Commission Services)

ADINE Deliverable 62

2 (4)


TABLE OF CONTENTS:

1. INTRODUCTION .................................................................................................................................. 3


3 (4)


1. INTRODUCTION

This deliverable consists of eight brochures that present brief summaries of the interaction simulation studies

done in the ADINE project. All of these studies have already been presented in earlier deliverables of the

ADINE project.

The first brochure presents the ideas behind the network management concept in the ADINE project. This

forms the backbone for the studies performed in the ADINE project and shows how the studies are

interlinked.

The second brochure discusses the effect of DG on fast automatic reclosing and briefly presents the real-time

simulation studies dealing with this subject.

In the third brochure, real time simulations examining the contradiction between fault ride through

requirements and the performance of LOM protection are presented. These simulation studies were

performed by combining two different types of real time simulators (the dSPACE and the RTDS®) and a real

protection relay.

The following brochure presents studies were the non-detection zone for various loss-of mains (LOM)

protection functions were determined. This study was performed by using a real LOM protection relay and

the RTDS®. Choosing suitable LOM protection settings is a difficult task because one will have to settle for a

compromise between having high sensitivity and avoiding nuisance tripping of DG when active or passive

LOM detection methods are utilized.

Communication based LOM protection schemes, however, can provide high sensitivity and high resiliency

against nuisance tripping. The performance of a novel IEC-61850 communication based LOM protection

solution by ABB was studied in the ADINE project and a short summary of the results of the study is given

in the following brochure.

Hardware in loop simulations of STATCOM are introduced in sixth brochure.

Combined real-time simulation environment of RTDS and dSPACE were developed for the simulations of

wind turbine concepts. The seventh brochure presents these wind turbine concepts and simulation

environment.

The eight brochure presents a vision of the IT architecture of ANM to be used in DNO control centers in

future. The envisioned IT architecture is based on Service Oriented Architecture (SOA) and standardized

data transfer interfaces (IEC 61968) and information model (IEC 61970/61968). The main benefits gained

from this new approach are the reduction IT maintenance costs and better adaptability to changes occurring

in the business environment.


4 (4)


2. RESULTS SUMMARY REPORTS

ANM

Adine_fast_Autorecl

osing

LOM FRT Study

LOM NDZ

determination

IEC61850 based

LOM protection

STATCOM

Wind turbine concepts

ANM IT architecture

Active Network Management

The traditional passive network management or “fit&forget” principle in Distributed Generation (DG) connection needs to be changed into Active Net-work Management (ANM).

With proper management of active resources the overall system performance may be improved from presently used practices. DG provides a good potential as a controllable re-source for the active network. Other existing controllable re-sources are direct load control, reactive power compensation

and demand side manage-ment.

ANM method adds value by increasing the potential for re-newable energy, by improving efficient utilization of distribu-tion network assets and by supporting distribution network by ancillary services from cus-tomer-owned resources.

ANM coNcepT

The distribution network man-agement concept of ADINE project is based on existing systems like SCADA, Distri-

bution Management System (DMS), substation and distribu-tion automation and Advanced Metering Infrastructure (AMI). The ANM system operates on protection, decentralized con-trol and area control levels.

The intelligence of ANM is based on investments in con-trollability and ICT. Area control level may e.g. be used to coor-dinate individual resources and thereby increase the synergy benefits of network manage-ment.

Overview of the Active Network Management concept in the ADINE project. All hardware devices like protection relays, AVR of tap changer, AVR of DG units, STATCOM controller and power factor controller are working in decentralized way. On the top of the decentralized control system there exists also a centralized system for distribution network management.

Benefits of ANM are based e.g. on more efficient utilization of existing network assets. The example shows how much the size of DG may be increased when non-firm network capacity and ancillary services are applied in DG connection.

Contact person:Project coordinatorMs Mirva Seppä[email protected]

Contact person in Adine project:Mr Matti KärenlampiABB Ltd Distribution Automation [email protected]

Contact persons in Adine project: Mr Ralf [email protected] andMr Jarmo [email protected] www.areva-td.com

Contact person in Adine project: CEO Anders [email protected]

Contact person in Adine project: Prof. Olof [email protected]/english

Contact person in Adine project:Dr. Tech. Sami Repo, Department of Electrical Energy [email protected]

Adine project has been supported by the Centre of Expertise Program

ANM feATures

ANM concepts add new fea-tures for protection system and automatic control system levels. New protection system features are e.g. distance and differential protection schemes and commu-nication based LOM. The ANM concept includes at decentralized control system level local voltage, power quality and frequency con-trol, load shedding and produc-tion curtailment features. Many new features are also added for the area control level like coor-dinated voltage control, power flow management, fault location schemes, automatic network res-toration and island operation.

The aim of ANM is to add more flexibility for network management in order to utilize existing network assets more efficiently. The addition of flexibility comes from the utilization of active resources through grid code requirements or ancillary services provided by resources. Active resources are needed to integrate as a part of distribution system instead of just connecting. Active resources are typically utilized in extreme network conditions when network is not capable to transfer produced wind power.


Autoreclosing & LOM protection

Distributed generation (DG)capacity is growing rapidly.This is due to development ofDG technologies but alsostrongly driven by the EUpolicy. DG can have manybeneficial effects on theusage of distributionnetworks but there are also anumber of challengesinvolved. One of the majorconcerns is related to thefunctioning of fastautoreclosing (AR). AR has aremarkable significance forthe reliability of supply sincethe majority of faults on over-

head lines are temporary innature which means that theycan be cleared with the helpof AR. AR is based on de-energizing the faulted feederfor a short period of time(commonly from 0.2s to 2s).This is done by opening thefeeder circuit breaker (CB) fora brief moment and thenreclosing it. The challengingpart in this is the fact that allthe DG units on the feeder inquestion have to bedisconnected within the opentime of the AR. Thedisconnection of DG is taken

care by the LOM protectionwhich all DG units should beequipped with. Thesuccessfulness of AR isdependent on many factorssuch as the applied LOMdetection functions, the LOMprotection settings and thechosen open time of the AR.The effect of these factorswere examined in this study.Real IEDs could be utilized inthe study thanks to thepossibilities offered by the realtime simulation environmentwhich is shown below.

The left side of the figure illustrates themachine in loop type of functioning of the real time digital simulator (RTDS) based simulationenvironment. The right side of the figure shows the structure of the studied networkmodel. The aim of the study was to examine theperformance of LOM protection during fast autoreclosing with different protection settings and power balances on the feeder.

Contact person:Project coordinator Ms Mirva Seppälä

Contact person in Adine project: Mr Matti Kärenlampi ABB Ltd Distribution Automation

www.abb.com

Contact persons in Adine project: Mr Ralf Jessler [email protected] and Mr Jarmo Aho [email protected] www.areva-td.com



Contact person in Adine project:Dr. Tech. Sami Repo, Department of Electrical Energy Engineering



The studies were divided intoseven cases. In each of these casesthe following were varied:- Power balance (18 variations)- Fault locations (4) - Types of the fault (2) Each case thus consisted of 144simulation runs. The protectionsettings were changed betweenthe cases as shown in the tablebelow. It is noteworthy that theprotection settings chosen in thefirst case were taken from a real setof recommendations. As the tablebelow shows, 50% of the fast ARs,however, failed when these LOM

The upper table presents the simulation results from the AR study. The bar diagram in the lower leftcorner illustrates how the AR problems are distributed in relation pre fault power balance on the feeder.It can be seen from the diagram that the problems are concentrated to the balanced power (NDZ) area.

protection settings were used. Therate of change of frequencyfunction (ROCOF) was added to theLOM relay in case 2. This improvedthe situation significantly as thetable shows. The protection settingswere further tightened In thefollowing cases which eventuallyled to a successfulness rate of 100%(case 7). The settings in case 7 were,however, so strict that they wouldvery likely cause unwanted trippingof the DG unit. Too strict protectionsettings are therefore not realisticand, moreover, they are alsoconflicting with the FRTrequirements.


FRT capability & LOM protection

The behavior of windturbines (WT) during voltagedips has received a lot ofattention lately. This isbecause wind power alreadyconstitutes a considerableshare of the generationpalette in certain regions. It isthus no longer irrelevant howWTs behave duringdisturbances in such regions.System operators have,therefore, issuedspecifications defining thesystem supportive behaviorthat WTs need have duringvoltage dips. These are called

as fault ride through (FRT)requirements. The upper leftcorner of the figure belowshows the conceptual shapeof FRT curve. It is, however,not enough that the WT hasFRT capability but also theloss of mains protection(LOM) has to be set to allowthe FRT. Otherwise the LOMprotection will alwaysdisconnect the wind turbinewhenever a voltage dipoccurs. This means the LOMprotection settings need tobe loosened so that the FRT isallowed. This degrades the

sensitivity of LOM protectionand thus increases the size ofthe non-detection zone (NDZ)of LOM protection. This isillustrated in the down leftcorner of the figure below.This study aimed to examinethe contradiction between theFRT requirements and LOMprotection. The simulationmodel used in the studies isshown on the right side of thefigure below. The studies wereperformed in a simulationenvironment combining tworeal time simulators (RTDS &dSPACE) and a real LOM IED.

The upper left corner represents the conceptual shape of FRT curves. The figure below this illustrates how the size of the NDZ of LOMprotection increases when the LOM protection is set to allow the FRT. The right side of the figure represents the networkmodel that wasused for carrying out the simulation studies examining the conflict between FRT requirements and LOM protection.



www.abb.com







A full power converter connectedwind turbine model was run by thedSPACE, whereas, the networkmodel was run by the RTDS. Theused IED was set to control thecircuit breaker at the connectionpoint of the WT. The studies weredivided into two cases. In the firstcase the LOM IED was set loosely inorder to allow the FRT, whereas, inthe second case the LOM IED wasset according to a Finnishrecommendation for DG unitswhich did not enable FRT. Thefigure below represents variousgraphs related to the FRT of the WTin the first simulation case. A 1.5slasting short circuit which caused

These graphs represent the behavior of themodeledWT during a 1.5s lasting deep voltage dip. TheWTwas able to ride through the fault and the LOM protection allowed it in this case.

the voltage to dip to 30% of itsnominal value was inflicted to thesupplying grid. The topmost graphon the left shows how the WT seesthe voltage dip. This reduces thepower supply capability of the gridside converter of the WT. Thispower imbalance causes the DC-linkvoltage to increase. In order toreduce this, the excess power isdissipated in the breaking chopper.The FRT concept of the modeledWT, however, also used activepower reduction in parallel with thebreaking chopper. The resultsshowed that the WT was able toride through the fault with relativelysimple control actions.


NDZ of LOM protection

The functioning of loss ofmains (LOM) protection isone of the main concernsrelated to the integration ofDistributed generation (DG).LOM protection shouldensure that no DG units areleft unintentionally feedingan islanded network area.Unintentional islanding is notprohibited because:- DG can energize lines thatshould be de-energized(safety hazard for utility staff)- DG can ruin the functioningof autoreclosing and causeout of phase reclosings

- Power quality in the islandmay be dangerously poor forcustomer devicesIt has, however, beenobserved that LOMprotection can fail to detectislanding when the active-and reactive powerimbalance in the powerisland is not sufficiently large.This area where LOMprotection fails to detectislanding can be defined inactive- reactive powercoordinate system. It is calledas the non-detection zone(NDZ).

This study aimed, however, todetermine the shapes of theNDZs for various LOMprotection functions. Thesestudies were performed usinga real time digital simulator(RTDS) and a real LOM IED. Thesimulation model used in thisstudy is shown in the left sideof the figure below, whereasthe right side of the figurerepresents the determinedNDZ for a frequency andvoltage based LOMprotection. As the figureshows, this NDZ is not exactlya rectangle as often presented.

The left side of the figure represents the networkmodel that was used for determining the non-detection zones (NDZs). The figure onthe right represents the NDZ determined by simulations for voltage and frequency based LOM protection when the loads in thenetwork weremodeled as constant power loads.



www.abb.com







The figure below tries to explainthe reasons for the shape of theabove presented NDZ. First of all,the under- (UV) and overvoltage(OV) boundaries are bent becausethe more active power the DG unitproduces, the more the voltage atthe point of common coupling(PCC) of the DG unit rises. Thisvoltage rise either speeds up orslows down the UV / OV protectiondepending on which of the fourquarters in the figure is underexamination. For instance whenlooking at the 4th quarter in thefigure, the more we move towardsleft, the more the generatorproduces active power and thus

The NDZ for frequency- and voltage based LOM protection does not exactly have a rectangular shape asoften presented. This figure tries to explain the reasons for the bending of this NDZ.

the more the voltage at the PCCrises. This, of course, speeds up theoperation time of OV protectionsince in the 4th quarter there is alsosurplus in reactive power. Thevertical boundaries, on the otherhand, are bent due to the voltagedependency of the loads. Forinstance, when moving downwardsin the 4th quarter the reactivepower surplus increases andthereby also the voltage at the PCCrises. This increases theconsumption of the voltagedependent loads which causes theactive power imbalance in the 4thquarter to reduce. This naturallyslows down frequency protection.


Loss of Mains Protection Issues

Unintentional islanding is notallowed in distributionnetworks due to a number ofreasons:- Safety hazards for repaircrews- Potential risks ofcomponents being damaged- Unintentional islanding cancause autoreclosing failures

Because of these reasons, it isobligatory that all DG unitsare equipped with a LOMrelay which ensures thatunintentional islanding doesnot occur.

The most utilized LOMdetection methods fail todetect islanding when theproduction matches closelywith the consumption in theislanded zone. This blind areais called the non detectionzone (NDZ). The size of theNDZ can be reduced bytightening the LOM relaysettings but it may causeunwanted tripping. Strictersettings are also problematicin the sense that DG shouldsupport the power systemduring voltage dips (Fault ridethrough (FRT) requirements).

Many protection studiesincluding real protectionrelays have been carried outwith the help of a real timedigital simulator (RTDS) in theADINE project:- The functioning ofautoreclosing in a networkincluding DG- The contradiction betweenFRT and LOM protection- The determination of theNDZs of various LOMprotection functions- Testing of a specificcommunication based LOMprotection method by ABB

The left figure illustrates how fulfilling the FRT requirements extends the size of the NDZ due to the necessity of loosening undervoltageprotection setting of the LOM relay. The right figure, which shows the form of the NDZ of a real LOM protection relay, was determinedbased on amyriad number of simulations which were carried out with the help of a real time digital simulator (RTDS)



www.abb.com







ENHANCED LOM PROTECTIONWITH FAST COMMUNICATIONThe Adine project hasdemonstrated substantial benefitsin generator protection using fastcommunication between IEDs. Thecommunication uses bothstandard IEC 61850 GOOSEmessages and user definablesignals using Binary Signal Transfer(BST) that RED 615 IED offers.

RESULTS- Fast tripping of generator whenthe fault is in the generator feederor on the substation- Securing FRT when the fault is onthe other feeders

The demonstration networkmodel having two feeders and 4 IEDs utilizing GOOSE and BSTmessages

DEMONSTRATED CASESFault in location Fault1 causes avoltage dip. IED of CB1 trips andsends block as a goose message toIED of CB2. IED of CB2 sends themessage via BST to IED of CB3. IEDof CB3 sends the block message toLOM IED as a goose message.

DG unit was not disconnected

In case of fault 2 overcurrentprotection in IED of CB2 trips andsends a command using BST to IEDof CB3. This IED sends trip messageto LOM IED.

DG unit was disconnectedwithout unnecessary delays.

Static synchronous compensator

Short-term voltage distur-bances may inflict consid-erable harm in the form of damaged equipment, lost production and reduced productivity. An effective solution for power qual-ity improving is STATCOM, Static synchronous com-pensator. STATCOM is ca-pable of mitigating voltage sags/swells, imbalance, flicker, harmonics and im-proving the system capa-bility to ride through fault events.

A typical solution for the con-tinuous control of reactive power is static VAr compen-sator (SVC), a combination of thyristor-controlled reactor (TCR) and thyristor-switched capacitors (TSC). The reactive

power output of SVC is adjust-able in the order milliseconds but dependent on the system voltage which limits the ability to mitigate the network voltage instabilities. The new breed of VAr compensators is based on self-commutated AC/DC power converters. Their performance is superior to SVCs, being ca-pable of operating at rated power nearly independent of the system voltage and control-ling the reactive power output in the order of microseconds. A representative of converter based VAr compensators is STATCOM, Static synchronous compensator.

STATCOM

STATCOM is a self-commu-tated AC/DC power converter

connected in parallel with the power system through coupling reactors. The power electronic switches of the converter are controlled to produce approxi-mately sinusoidal output AC-voltages, being in phase with the mains voltages, from the DC-source. Depending on the magnitude of the AC-voltages produced the VARs are either generated or absorbed.

STATCOM provides a superior solution for VAr control, voltage regulation, flicker compensation, and fault-ride-through improvement. Also grid current harmonic filtering is possible if sufficiently high switching frequency can be used. Typical applications include flicker compensation of large industrial loads

Reactive current output range of Static VAr Compensator (SVC) and Static Synchronous Compensator (STATCOM).

Principle of hardware-in-the-loop simulation of STATCOM








such as arc furnaces and VAr control of wind farms. Benefits of STATCOM are improved power quality and network stability, increased transmission capacity, and improved fault-ride through capability and grid code compliance of renewable generation.

Hardware-in-the-loop simulation of STATCOM

Hardware-in-the-loop (HIL) simu-lation is a technique used in the development and testing of em-bedded systems. In Adine-project HIL simulation is used in the test-ing of STATCOM controller. The tests are based on the Real-Time Digital Simulator (RTDS) which is a multiprocessor power system simulator capable of HIL simula-

tions with latency times below 2 µs. The models of the pow-er system and the STATCOM studied are implemented in the RTDS. The STATCOM control-ler interfaces the RTDS simula-tion through analog and digital I/O’s of the RTDS hardware. The analog voltage and current mea-surements required in the STAT-COM control are output from the simulation and passed to the controller that performs the con-trol routines. In order to achieve the desired STATCOM current and voltage the controller adjusts the gating signals that control the power electronic switches of the STATCOM in the RTDS simula-tion. With the help of HIL simula-tions the controller operation can be tested and verified prior to real-life demonstration.

Wind turbine concepts

The penetration of wind gen-eration has increased in many areas to a significant level. In such areas, modern wind tur-bines are required to be able to endure deep voltage dips. Otherwise, major problems to the power systems stabil-ity would occur. In addition to staying connected during the fault, modern wind turbines should be able to support the grid voltage during the volt-age dip by injecting reactive power.

Two most common variable speed wind turbine types are the doubly-fed induction gen-erator (DFIG) wind turbine and wind turbine with full-power converter interface to the grid. However, the behavior of the concepts during the grid volt-age dip differ significantly. The DFIG suffers from high rotor voltages and currents. Hence, the converter connected to the rotor windings may have to be disconnected in order to pro-tect the converter. Then, the wind turbine is not controllable

anymore which complicates the reactive power injection for network voltage support.

The wind turbine with full-pow-er converter is capable to ride through voltage dips. Actually the generator itself does not recognize the voltage dip at all since frequency converter decouple the generator from the grid. In addition, full-power converter concept is also ca-pable to inject reactive power to the network for voltage sup-port purposes.

Doubly-fed induction generator wind turbine concept and full-power converter wind turbine concept.








Study environment

The study is done by using novel real-time simulation environment consisting of Real-time Digital Simulator (RTDS), dSPACE and protection relay. The combination of the powerful power system sim-ulator (RTDS) and efficient control system simulator (dSPACE) pro-vides excellent environment for wind turbine and network interac-tions studies. The simulation en-vironment used makes it possible to develop control strategies for wind turbines to qualify more de-manding grid codes in the future. Minimization of simulation time is additional benefit.

Project target

In this project, the fault ride through (FRT) capability of both wind turbine concepts present-ed above are assessed. In ad-dition, the different methods for achieving the FRT capability are compared and analyzed. The simulation results show that the full-power converter wind turbine concept has much more stable performance during network fault compared to DFIG concept. However, the fault ride through of DFIG concept can be enhanced by reasonable converter control during the fault.

IT architecture of ANM

The integration strategy of IT systems in Distribution Network Operator (DNO) control centers needs to be changed from point-to-point to Service Oriented Architecture (SOA).

The utilization of Active Net-work Management (ANM) in-creases the number of soft-ware applications needed in distribution network manage-ment. In DNO control centers this results in new information systems and data transfer in-terfaces. Without sophisticated

integration strategy the IT ar-chitecture can become overly complex which in turn leads to high IT maintenance costs.

INTEGRATION STRATEGY

Traditionally information sys-tems have been integrated with company and product spe-cific, often proprietary, point-to-point interfaces. In current state of distribution network development this type of inte-gration strategy is not anymore justified.

From now on the information systems should be integrated through Enterprise Service Bus (ESB) using SOA principle, standard interfaces and stan-dard information model. The most promising interface and information model standard is known as Common Information Model (CIM) which consists of IEC standards IEC 61970 and IEC 61968. At data transfer level the most important tech-nologies are XML and Web Services.

A change is needed in IT system integration strategy in distribution network companies. Traditionally used point-to-point integration strategy needs to be changed to strategy based on SOA, ESB, standard interfaces and standard information model. This new approach reduces total amount of interfaces, accelerates commissioning of new systems and clarifies communication between different parties.

Vision of the information system architecture of ANM. Applications communicate with each other through standard interfaces and ESB. Utilization of standard infor-mation model (e.g. CIM) reduces the need for data transformations between different information systems.








SYSTEM ARCHITECTURE

In the future the information sys-tem architecture is required to provide better openness, expand-ability, modularity, scalability, performance, usability and secu-rity than the systems used today. These goals can be achieved with utilization of interface and information model standards and SOA based integration strategy. Decentralization and partitioning of large information systems and decoupling of user interfaces, ap-plications and data sources from each other can furthermore help in reaching these objectives.

CONTROL CENTER ENVIRONMENT

From ANM point of view the most important information systems used today are Supervisory Con-trol and Data Acquisition (SCADA) and Distribution Management System (DMS). Currently these systems are used simultaneously with separate user interfaces. As the number of software applica-tions used in DNO control centers increase, the need for unified user interface becomes constantly greater. In the end this leads to an environment where the role of SCADA is to act only as a data broker between distribution net-work and higher level systems.

deliverable 62: results summary report on anm simulations · results summary report on anm...

Documents