0 ETS M.S. Thesis Mster of Siene in Sustinle Energy - Eletril Energy Systems Integrtion Strtegy for Fst-hrgers in Existing Power Grid Frederik Willim indner Kongens Lyngy 2016
DTU Eletril Engineering Deprtment of Eletril Engineering Tehnil University of Denmrk Supervisors: hresten Træholt Seyedmostf Hshemi Toghroljerdi Peter h ndersen Dniel Sndermnn Jensen Ørsteds plds uilding 25 2800 Kongens Lyngy, Denmrk Phone +45 4525 01 elektro@elektro.dtu.dk www.elektro.dtu.dk
strt Reent yers hs seen n enormous inrese in the fous on humn used limte hnge nd wys to void the ontinued temperture rise used y greenhouse gses. lrge prt of this prolem is the trnsition to n energy supply from sustinle soures. The trnsition to sustinle soiety inludes the trnsporttion setor. s of now, the solution seems to e eletri vehiles with tteries s energy storge. This thesis ims t nlyzing the Nordhvn power grid in the se where one or more 500kW+ fst-hrgers for E-uses re deployed. model for the relevnt prt of the Nordhvn grid is implemented in MTL Simulink. The hrgers will drw minimum of 500kW of power for short periods of time, mking lod nd voltge onstrint nlysis for the lol grid neessry. Simulink model for the hrger is developed nd implemented in the simultions long with lod profiles nd grid speifitions provided y DONG. The effets on the grid used y the fst-hrger is investigted in terms of urrent nd voltge level, oth during norml opertion nd fults. The effets of supporting the implementtion of fst-hrgers with ttery is investigted. Severl hrging ptterns, hrger lotions nd setups re explored to find the onstrints for suessful integrtion of fst-hrger.
ii
Resumé I løet f de seneste år hr verden st et øget fokus på de menneskeskte klimforndringer og ekæmpelsen f temperturstigninger som skyldes øgede mængder f drivhusgsser i tmossfæren. En stor del f udfordringen er omstillingen til æredygtige energiresourer. Omstillingen til et æredygtigt smfund omftter også trnsportsektoren, som i dg er seret på olie. Den mest lovende løsning ser i øjelikket ud til t være tteridrevne elektriske køretøjer. Målet med dette projekt er t nlysere Nordhvns elnetværket i det tilfælde t en 500kW hurtiglder til E-usser liver instlleret. En model f netværket er udviklet og indst i MTL Simulink for t undersøge prolemstillingen. Lderen vil som minimum påtrykke en elstning på 500kW i korte tidsperioder, dette nødvendiggør en nlyse f elstnings- og spændingsgrænser og niveuer. Der er udviklet en model for lderen smt elstninger på 10/0.4kV trnsformerne i netværket, derudover er en tterimodel også udviklet. Netværksmodellen er udviklet på en sis f speifiktioner leveret f DONG Energy. Det er undersøgt hvordn implementtionen f en hurtiglder påvirker strøm- og spændingsniveuer på to rdiler in elnettet. Der er undersøgt hvordn et tteri kn støtte under driften f hurtigldere. Sidst er der undersøgt hvd det vil kræve for et tteri for t vretge fuldstændig pek shving for forskellige ldemønstre.
iv
Prefe This report is 0 ETS point thesis nd requirement to omplete the Mster progrmme Sustinle Energy - Eletril Energy Systems nd therey otin the degree Mster of Siene in Engineering, Sustinle Energy t the Tehnil University of Denmrk. The thesis work ws rried out t the Deprtment of Eletril Engineering in ollortion with Dniel Sndermnn Jensen of DONG Energy nd is relted to the EnergyL Nordhvn reserh projet. The projet ws initited 5th of Mrh 2016 nd finished on the 5th of ugust 2016. The thesis dels with the onsequenes of instlling fst-hrger into n existing power grid, nd how it n e integrted so the power grid n ontinue opertion without expnding the pity of the grid. The intertion with ttery Energy Storge System is investigted s well s the effets of hving multiple fst-hrgers. The thesis onsists of five hpters. First, short introdution nd kground introduing the sujet to reder. Then Model desription nd prmeters hpter whih desries how the model is generted nd where the prmeters omes from. Next, results hpter, whih presents nd disusses the results of the simultions nd puts them into perspetive. t the end onlusion, summing up the whole projet. Kongens Lyngy, ugust 5, 2016 Frederik Willim indner
vi
knowledgements Severl people hve ontriuted with dvie, inspirtion nd support throughout this thesis journey. I owe ll of you my deepest grtitude. First nd foremost I would like to thnk my supervisors hresten Træholt, Seyedmostf Hshemi Toghroljerdi nd Peter h ndersen for the first lss supervision. gret thnks to you for letting me work with projet in my field of interest nd lwys tking the time to give me gret dvie when i needed it! It hs een n exiting hllenge to work on this projet. gret thnks to Dniel Sndermnn Jensen who offered me the opportunity to write my thesis in ollortion with DONG Energy nd ontriuted to seleting the min sujet nd mking it ll possile. For gret ptiene nd understnding, my most hertfelt thnks goes to Simone nd ugust, for enduring me during this time of high stress. Finlly, grteful thnks to my prents, my mother for sving the dy multiple times when ugust needed to e y-sitted nd for my fther for using hours upon hour on proof-reding nd mking suggetions to the report. Thnk you ll, I would never hve een le to pull through this if you hd not een there!
viii
ontents strt Resumé Prefe knowledgements ontents i iii v vii ix 1 Introdution 1 1.1 Reders guide................................. 1 1.2 kground.................................. 1 1. E-uses.................................... 2 1.4 Sope..................................... 1.5 Projet nd pproh............................. 4 1.6 Projet limittions.............................. 5 1.7 EnergyL Nordhvn............................. 7 2 Grid model nd prmeters 9 2.1 Model desription............................... 9 2.2 ttery Energy Storge System model................... 11 2. Lod omposition............................... 12 2.4 Fult situtions................................ 14 2.5 Pek lod mitigtion y lodshift...................... 15 2.6 hrging pttern derivtion......................... 15 2.7 hoie of hrger dimensions......................... 18 2.8 Summry of modelling ssumptions..................... 19 se study 21.1 Senrios.................................... 21.2 Summry of senrios............................. 25 4 Results 27 4.1 seline senrios............................... 27 4.2 Simultion using two 500kW units nd ESS............... 4
x ontents 4. ttery support during feeder fult..................... 4 4.4 Lod shifting during pek hour....................... 6 4.5 Pekshving estimtion............................ 42 5 onlusion 49 5.1 Future Work.................................. 51 ppendix : Model Overview 54 iliogrphy 57
HPTER 1 Introdution 1.1 Reders guide This hpter explins the motivtion of this Msters thesis nd the effet of integrting fst-hrger into n existing power grid long with the potentil nd possiilities offered y hving grid sle ttery Energy Storge System (ESS) providing support in fult nd pek demnd situtions. The sope, limittions nd outline of the thesis is presented. It is ssumed tht the reder possesses knowledge of eletril energy systems in ddition to n understnding of power system opertion, eletril omponents nd the network dministrtor regultion nd governne surrounding the opertion of power grid. 1.2 kground Reent yers hs seen n enormous inrese in the fous on humn used limte hnge nd wys to void the ontinued temperture rise used y greenhouse gses. lrge prt of this prolem is the trnsition to n energy supply from sustinle soures. In this ontext, Denmrk hs n mitious gol of eing ompletely fossil fuel free y 2050[1]. This trnsition does nturlly inlude the trnsporttion setor, where n lterntive to oil s primry fuel hs to e implemented. s of now, the solution seems to e eletri vehiles with tteries s energy storge. Of the ttery types ville, Lithium-ion hs eome the most dominnt due to its reltively high energy density, power output nd eptle durility. Li-ion tteries hve seen quik mturtion over few yers s the demnd for the tehnology hs inresed rpidly. urrent tteries offer signifintly etter performne, oth in solute terms nd in performne per investment[2]. This hs opened the opportunity to eletrify the puli us network in ities to derese point pollution, noise level nd help omplete the trnsition to fossil fuel free soiety. Multiple ities ross the world re lredy running projets where some us lines re servied y E-uses[]. The ontinued trnsition to n eletrifition of our soiety poses new tehnil hllenges for oth Trnsmission System Opertor (TSO) nd for the Distriution System Opertor (DSO) s the infrstruture hs to e upgrded to e le to hndle the new system with deentrlised plnts, Distriuted Energy Resoures (DERs) nd inresed
2 1 Introdution demnd prtilly used y the eletrifition of the heting nd trnsporttion setors. E-uses dds different hllenge thn personl EVs or PHEVs s they hve very inflexile demnd nd, in some onfigurtions, lk the ility of providing lning servies for the grid during idle hours. Furthermore, their inflexile driving pttern puts dditionl strin on the Medium Voltge (MV) distriution grid s. 1. E-uses The gol of eing fossilfuel free is not rehed without inluding the trnsporttion setor, oth privte nd puli. Sine the S-trin nd metro net re lredy fully eletri, the regionl trins re plnned to e eletrified in the oming yers, the next step is to eletrify the us network. Eletrified uses (E-us) offers some tehnil dvntges over diesel uses, mong these re regenertive rking where power is regined to the ttery from using the motor s genertor to rke[4], no point pollution, options for different fuels depending on wht method is used to generte the eletriity. 1..1 hrging methods Severl methods re ville to implement this, the three mjor types eing trolleyuses, ttery powered uses with lrge pity nd lstly ttery powered uses with smller tteries. Eh offers its own set of dvntges nd disdvntges. Trolleyuses trolleyus does not rely on on-ord energy storge s it is onstntly onneted to the eletril grid y overhed wires. This elimintes the need for hrging sttions, ut stritly limits its movements to the route of the overhed wires. The dependene on wires mkes this rigid solution with low rerouteing possiilities if rod work, idents et. requires it. E-us with opportunity hrging This type of us rries smller ttery or superpitor, whih needs to e rehrged fter shorter distne eg. 0km[5]. The dvntges of this design re the lower weight, whih results in higher effiieny in terms of km/kwh nd higher reliility[5]. Disdvntges re the need for frequent rehrging, lower energy density if using superpitors nd higher investment ost for hrging equipment s it hs to e more powerful nd instlled t more lotions.
1.4 Sope Night-/Slow hrging Similrly to onventionl EV, this type of us relies on lrge ttery whih llows for longer trips etween hrging periods. The lrge tteries will e hrging for longer periods, espeilly during night time, whih mkes it possile to use them s network utility nd for the owner to profit on the trding of power when the uses re idle nd therefore ville to provide system servies on the sme level s dedited ESS in terms of voltge regultion, frequeny support et. 1.4 Sope The fous of this thesis is to investigte the effets fst-hrger my hve on MV distriution grid in residentil re. The MV grid is hosen due to DONG Energy, the lrgest DSO in Denmrk [6], eing ollortor in the thesis. DONG hs provided dt for two of the 10 kv rdil lines in their grid in Nordhvn. The ngle of investigtion is from the DSO point of view, mening the onsumption on the 0.4kV side of the 10/0.4kV trnsformers re ggregted nd following the sme 24-hour lod profile with sling orresponding to the numer of registered ustomers per 10/0.4kV trnsformer. The onnetion to the 60kV grid is regrded s n infinite soure onneted y 60/10kV tp hnging trnsformer. This mens tht the grid frequeny does not ply ny role, oth in terms of the dded onsumption ffeting the grid frequeny nd fst-hrger ehvior during frequeny normlities. s prt of the full sle energy development projet, EnergyL Nordhvn[7], it seeks to identify the hllenges tht my rise when instlling fst-hrgers in n existing grid nd wht ountermesures n e tken to void overloding the grid or feeder in oth urrent nd future lod ptterns. The primry ftors used for evlution of the grids response to fst-hrger re line urrent nd voltge level. The thesis seeks to first disover wht prt of the grid the fst-hrger should e instlled t to use the worst performne in terms of urrent levels in feeders nd the voltge level t the lod frthest from the 60/10kV trnsformer on the respetive feeder. The onnetion point found in the forementioned senrios will e used s point of investigtion when determining the possile risks of diretly integrting fst-hrger into the power grid. This is investigted with respet to the grid still eing le to omply with norml opertion riteri for voltge level nd feeder lod perentge. nother onstrint it must omply with is to not overlod supporting feeder during fult on the feeder rrying the fst-hrger. lterntives to grid upgrdes y instlling lrger feeder les is explored, s future power grids will likely rely on energy storge servies in order to ope with the flututions of DERs. It is intuitive to mke use of these tehnologies, leding to the
4 1 Introdution investigtion of the possile intertion etween fst-hrger nd ttery Energy Storge System (ESS) The ESS is introdued to llevite the grid from the lrge lod spikes ssoited with the ursts of power the fst-hrger delivers. To mke more generl piture of the effets of fst-hrgers, senrios with two units will e investigted to explore more severe ses of fult opertion nd higher pek demnd. Finlly n investigtion of the pity requirements for ESS to keep the lod of feeder elow norml opertion onstrints will e rried out to give n impression of the possiility for this type of flexile support s n lterntive to onventionl grid upgrdes. Sine grid sle ESS is lredy plnned to e instlled in the Nordhvn grid, the speifitions used in the Simulink grid model resemles the prmeters for the tul Nordhvn ESS. The ttery is modelled s progrmmle lod with stte of hrge determining the mount of power ville. The prmeters for the ESS re sed off the tul ESS in terms of in-/output power nd ttery pity. It is ssumed tht the ESS is ompletely dedited to supporting the fst-hrger, so ttery degrdtion nd other servies the ESS might e le to utilize hs een omitted in this investigtion. To summrize the gols of this projet: dpt nd upgrde the existing model of the Nordhvn grid to inlude residentil lods with vrying demnd over 24-hour yle. Determine the power nd hrge length for fst-hrger tht n supply ity us onsumption. Develop ESS model tht n support the opertion of grid when one or more fst-hrgers is integrted. Determine the pity requirements for running multiple fst-hrgers within grid ode limits. 1.5 Projet nd pproh In order to investigte the effets of instlling 500kW fst-hrger in the existing power grid in Nordhvn, n existing model[8] of two rdil rnhes hs een dpted nd expnded to inlude ontrollle lods, rel physil prmeters for les nd ESS with power flow driven ontroller. The onsequenes of the fst-hrger, seen from the DSO perspetive, re mesured in the individul line power flows, urrents nd voltge level. Simultions re done to disover if the implementtion will violte ny of the urrent
1.6 Projet limittions 5 rules in the grid opertion ode, minly if 50 % line lod is exeeded during norml opertion nd if the mximum line pity is exeeded during fult senrios. Furthermore, the potentil synergy with grid size ESS hs een investigted in terms of pek shving nd ful sitution power support. suessful integrtion of fst-hrger in the power grid will meet the following riteri: Will e le to supply the power need for n E-us to finish hrging efore new rrives. Deliver the power without overloding the trnsformer it is onneted to. e le to run t ll times of the dy without exeeding the orresponding 10kV feeders 50% pity during norml opertion. Not overlod feeder during fult opertion where the remining feeder hs to rry two feeder lods. drg the voltge of the frthest lod elow the regultory threshold. Furthermore, suessful integrtion of fst-hrger in omintion with ESS will, in ddition to the ove, meet the following: Signifintly redue the urrent on remining feeder during fult. e le to disple lod during high demnd hours. 1.6 Projet limittions s the sope of this thesis is to investigte nd present the possile hllenges in regrds to line lods nd ongestions. The fst-hrger model is modelled s PQ lod nd its trnsient ehvior is not investigted. The lotion nd power of the hrger is ssumed to e of insignifint influene on the high voltge grid nd the frequeny is therefore omitted s ftor in the opertion riteri for the hrger. In relity, hrger of this size will likely e sujet to set of grid odes. ll lods re modelled s resistive lods, s detiled lod dt hs not een ville for the projet. The model used limits the lod nlysis to -phse onneted, symmetril omponents whih exludes investigtion of phse voltge unlnes. The tul plement of the physil ESS re not inluded in this model s the ESS used hs een pled t the sme us s the hrger to show the optiml synergy effet it my hve on the opertion of the fst-hrger. The ESS is ssumed to e dedited to the hrger, whih would not e the se in prtie s it would e required to provide grid servies in order to mke profit.
6 1 Introdution The PV instlled in Nordhvn is not inluded in the model nd its possily positive influene on the opertion is not investigted.
1.7 EnergyL Nordhvn 7 1.7 EnergyL Nordhvn EnergyL Nordhvn is full sle energy reserh nd urn development projet situted in openhgen. The gol of the projet is to develop nd implement new tehnologil solutions to the future energy infrstruture. The topis of reserh inlude: Utilistion of renewle energy, smrt grids, eletrified trnsporttion nd energy effiieny in uildings. The min gol is to unify ll these sutegories of the energy setor into n intelligent smrt ity. The projet is funded y the Dnish Energy Tehnology Development nd Demonstrtion Progrmme (EUDP) nd hs totl udget of 14 million DKK nd runtime period spnning 2014-2019[7]. Figure 1.1: Reserh re overview for projet EnergyL Nordhvn [7]. s seen in figure 1.1, the projet spns over mny topis, distriuted over severl work pkges within the portfolio. This thesis is prt of the work pkges WP6: Eletriity Infrstruture nd WP7: Trnsporttion Infrstruture whih fouses on EV integrtion nd eletril infrstruture. The WP leder for WP6 is DONG Energy, the DSO responsile for the entire Nordhvn power grid. The sope of WP6 is s follows: The WP investigtes flexile nd dynmi onsumption ptterns nd new grid designs. numer of innovtive dd-ons to the urrent grid lyout will e developed nd evluted for grid performne nd finnil fesiility[9].
8 1 Introdution nd the sope for WP7: There re severl eletri usses urrently in initil tril set-ups in severl regions of the world. In this Work Pkge, we will nlyze urrent eletri us projets nd their ssoited hrging tehnologies nd grid integrtive solutions. The findings from the nlysis will serve s referene point for Smrt Grid potentil for eletri uses[9] This projet ridges the two work pkges s it investigtes oth the possiility of hving E-uses present in Nordhvn nd the effets n integrtion of fst-hrgers for E-uses will hve on the eletril grid nd wht power grid dd-ons my do to mitigte the onsequenes of integrting sid fst-hrgers.
HPTER 2 Grid model nd prmeters This hpter fouses on desriing the model tht hs een uilt in MTL Simulink nd used to simulte the senrios onstruted to investigte the effets of integrting fst hrger in MV power grid. It moves on to define the prmeters for the senrios suh s the power of the fsthrger unit, ESS ontroller, hrging yle nd demnd distriution. 2.1 Model desription model of of the Nordhvn MV power grid developed in [8], hs een dpted nd expnded. The model onsists of 60/10kV tp hnging trnsformer tht supplies two prllel 10kV feeders, eh with set of 5-7, 10/0.4kV trnsformers onneted long its length. S1 Externl grid 60kV / 10kV 10kV / 0.4kV S2 S 10kV / 0.4kV Figure 2.1: Overview of the grid tht hs een used in this thesis. Figure 2.1 is sketh of the grid implemented in the Simulink model. Swithes hve een deployed t the eginning of eh feeder for the ske of demonstrting opertion in worst se fult situtions whih is if feeder is ompletely disonneted t its strt. swith t the end of line 1 is lso present in the event of fult t either line. Swith S 1 nd S re oth losed during norml opertion nd swith S 2 t the end is open.
10 2 Grid model nd prmeters When fult sitution is to e simulted either S 1 or S is opened nd swith S 2 is losed t the sme time to irumvent the fult nd seure power to ll lods. omplete overview of the developed Simulink lok model n e seen in. 2.1.1 Simulink nd SimPowerSystems MTL Simulink is used to simulte the power grid in this projet. Simulink uses lok interfe, where ll mthemtil funtions re pkged nd onneted in lok digrm. This llows for grphil overview mimiking the shemtis of rel power grid. Simulink ontins the lirry Simspe Power Systems whih hs defined loks for the most ommon elements in power systems suh s genertors, line setions, trnsformers nd RL lods. The model used uses the uilt-in loks for trnsformers nd π-line segments whih hs een modified with prmeters provided y DONG[10]. The model from [8] hs een extended with lods on eh 10/0.4kV trnsformer, n externlly ontrollle lod hs een developed for the simultions done in this thesis. 2.1.2 Lod loks The lod lok interfe provided in the model omes from [8]. It provides onnetion etween SimPowerSystems- nd stndrd Simulink-signls. It is done y mesuring the phse-phse voltges V nd V. these re used to rete voltge signl, V1, given s: V 1 = 1 (V 2 V ) (2.1) where 2 is the omplex onjugtion e j 2π. This results in the sme voltge s the phse to ground for V. Using the expression for pprent power, the voltge is omined with the ontrollle power signl to derive the orresponding urrent. S = (V I ) (2.2) 2 2 S = I (2.) V The resulting urrent is then split gin nd phse ngles re pplied to rerete lned -phse urrent signl. This method of modelling the lod y muxing the three phse system into single phse, mnipulting the urrent nd then demuxing into three phse gin, is only possile s long s the system is lned. This limits the model from eing used for voltge unlne nlysis in its urrent form. The lod is hnged through the simultion y lookup tle ontining the lod profile developed in this projet.
2.2 ttery Energy Storge System model 11 2.1. Fst-hrger lok The fst-hrger model hs een developed with fous on the effet the dded lod used y the fst-hrger lod step hs on the lol power grid. The fst-hrger is onsidered to hve idel power eletronis le to mintin unity power ftor nd with ner-instntneous rmp up pility. The fst-hrger is using the sme interfe s the lod loks nd will therefore intert with the grid through the sme equtions shown in (2.1) nd (2.). It is ontrolled with inry on/off pulse genertor using the pttern tht will e onstruted in setion 2.6, whih toggles the 500kW lod step. 2.1.4 le speifitions The indutive retne, X, is given s [Ω/km] in the dt provided y DONG[10]. Sine Simulink Simspe EletriPowerSystems uses [H/km] s prmeter for its π line segments, it must e onverted y using the reltion elow: X[Ω] = 2πf L (2.4) X 2πf = L (2.5) 2.2 ttery Energy Storge System model The ESS tehnology hs een suggested s n opportunity to provide utility servies on grid level suh s frequeny support nd voltge support[11]. ESS unit is lso plnned to e prt of the Nordhvn projet[12], this ESS will e dimensioned to e le to produe or sor 1MW of power with pity of 0.5MWh[12]. Using this s design point, the ESS hs een modelled nd integrted in the simultion to explore the opportunities for synergy it my provide with fst-hrger. The ESS is used when investigting fult sitution support nd pek shving y lod shifting. It is ontrolled y proportionl ontroller whih uses the mgnitude of the power flow t the eginning of the relevnt 10kV feeder s input nd n externlly speified set point whih is hnged depending on the desired ggressiveness of the ttery ontrol strtegy. The ESS ontrol eqution is reted s: P out = (P set P mesured ) G P ESS (2.6) Where P set is the speified opertion point for the ESS, P mesured is the mesured power flow, G is the gin of the ontroller, P ESS is the mximum power output nd P out is the ommnded output y the ESS, if this is negtive it will put the ESS into hrge mode nd it will sor power from the grid. It is ssumed tht the ESS is dedited to supporting fst-hrger nd it is pled on
12 2 Grid model nd prmeters the sme us s the hrger. This is done in the ttempt to mke generl evlution of the enefits of hving ESS s support when integrting one or more fst-hrgers in power grid. 2. Lod omposition The Nordhvn grid onsists primrily of residentil lods. onsequently, the lod profile onsists of morning nd evening peks with medium onsumption during the dy nd low onsumption during the night. The lod type is regrded s resistive nd with no rotting mhines. 2..1 ggregted onsumption profile The model hs een dpted from stte without ny lod eing onneted to the 10/0.4kV trnsformers. These hs, s mentioned erlier een modelled in this thesis s ontrollle PQ lods. The lod shpe hs een derived from dt out the numer of onsumers onneted to eh trnsformer[10] in omintion with n verged lod profile onstruted from dt from Energinet.dk. Tle 2.1: Numer of onsumers on eh 10/0.4kV trfo used for sling lod on eh lod us in the model. Trfo no. onsumers Trfo no. onsumers 5295 5 52555 1 51968 408 52097 252 5191 51 55247 21 50970 15 5227 5 54106 1 5147 60 55296 0 The onsumer numers re normlized reltive to the highest numer of onsumers on trnsformer nd pplied to the model s rough sling of the lod distriution. Due to zero or ner-zero onsumer onnetions on some trnsformers, the lower numers hs een djusted to resemle more uniform distriution ross the feeders. Tle 2.2: Trnsformer lod sling, sorted y distne to 60/10kV trnsformer. Line ID Normlized onsumption sle Pink 0.89 1.00 0.89 0.46 0.20 0.20 0.20 lue 0.81 0.69 0.24 0.21 0.2 0.20 Sine the lod of network is vrying through the dy, lod urve hs to e pproximted. This is done on 24-hour yle whih repets if the simultion is longer. The
onsumption ftor 2. Lod omposition 1 lod profile is onstruted y sorting nd verging ll hours over yer into set of onsumption vlues overing one dy using the following eqution: P vg (h) = 65 d=1 P d (h) 65 (2.7) Where P vg (h) is the verged demnd for hour h of the dy nd P d (h) is the dy of the yer for the given hour h. The dt re gthered from Energinet.dk s dthu[1] nd overs the net onsumption for DK1+DK2 for the yer my 2015 - my 2016. The sling otined from the urve is pplied to the model through Look Up tle whih onsists of 24 points with interpoltion etween. To keep the prmeter ount of the investigtion low, sesonl hnges in demnd hve not een implemented. 1 onsumption sle 0.8 0.6 0.4 0.2 0 1 2 4 5 6 7 8 9 10 11 12 1 14 15 16 17 18 19 20 21 22 2 24 Time of dy Figure 2.2: verged onsumption used s 24-hour lod profile. The two slings re then omined nd the finl onsumption sle n e expressed s: P trfo,n (t) = P vg (t) P trfosle,n P sle (2.8) Where P trfo,n (t) is the lod t trnsformer n t time t, P vg (t) is the verged lod ftor t time t, P trfosle,n is the lod sling for trnsformer n nd P sle is the lod t pek onsumption time t the heviest loded trnsformer. Lod mesurement dt for the individul grid uses hs not een ville for this work nd n estimtion of lod sling hs een used insted. The tul lod t present time is derived on verge lod per onsumer in Denmrk nd numer of onsumers per trnsformer. The lod sle used in the simultions is signifintly higher thn the expeted urrent lod nd re sed off the expettion of more eletrified privte setor in the future due to het pumps nd EVs eoming more prevlent.
14 2 Grid model nd prmeters 2.4 Fult situtions mjor prt of the work presented in this thesis is the effet filure of one of the feeders hs on the opertion of the grid. fult n rise from multiple resons, le eing ut, reker triggering, short iruit et. The fult itself is not nlysed s this is outside the purpose of this work. Normlly when fult ours there will e period of time where the segment ehind the fult is ompletely disonneted efore the fult proedures re tivted nd the power is rerouted through the neighouring feeder(s). The simultion of fult sitution is hndled s instntneous, mening when fult is introdued it is hndled s if reker is tripped nd the swith t end of the lines is losed in the sme time step. 2.4.1 N-1 riterion The dimensioning rule for power grids re s minimum to e le to operte if the iggest unit in the system trips nd disonnets. This rule is ditted y the Europen TSO ENTSOE [14] nd y extension, the Dnish TSO Energinet.dk [1]. The rule rries over to the design riteri for MV grids where DONG hs speified the limits s 2/ of the mximum pity. This is due to DONG s grid design hoie where two neighoring feeders re set to over for one if fult ours, splitting the lod in two insted of one. Sine the model used for this thesis only ontins two feeders, onstrint of 50% lod during norml opertion hs een used s evlution prmeter.
2.5 Pek lod mitigtion y lodshift 15 2.5 Pek lod mitigtion y lodshift Pek lod mitigtion or pek shving is n option to defer or postpone trditionl grid upgrdes where the method is to reinfore the grid y instlling dditionl/lrger trnsmission lines. The onept hs een visulised in figure 2. Figure 2.: Pek shving exmple [15]. The ESS desried erlier llows for n lterntive solution where the grid is mde flexile in terms of deentrlized regultion units, in this se the ESS, supports the lod flow in the lol re insted of power plnt hving to push the power through the whole grid. In this se, the purpose of the ESS is to ttempt to keep the lod of the feeders elow the threshold of 50% during norml opertion. The ESS is to dishrge during peks nd to exploit the exess line pity to hrge during off pek hours. Eqution 2.8 is used to govern the ehvior of the ESS, with the Stte Of hrge (SO) nd mximum Power s limits to its opertion. It is of ourse not possile to utilize ll exess pity, whih is leds to the prolem of ttery size whih is investigted further in the lod shift simultion se. 2.6 hrging pttern derivtion The hrging pttern used for the simultion is sed off the ssumption tht the openhgen us route will servie the route tht is to e eletrified. This route termintes t Nordhvn sttion, mking it n ovious ndidte for n E-us fst-hrger lotion for the Nordhvn projet. The route is depited in figure 2.4.
16 2 Grid model nd prmeters Figure 2.4: us route used for estimting onsumption[16]. Length of roundtrip is pproximtely 1 km ording to Google Mps[16]. Estimted onsumption rtes re sed on OWI report[17] where three onsumption levels re presented s low, medium nd high onsumption rte. The influening ftors re, mong others, lod weight, verge speed nd driving pttern. Using the vlues from [17] yields the energy onsumption seen in tle 2.. Tle 2.: onsumption estimtion for route with l roundtrip levels of. se kwh/km kwh/trip Low 1.28 16.64 Mid 2.00 26.00 High 2.75 5.75 = 1km using different The rnge of onsumption, s n e seen from 2., hs wide gp from highest to lowest, dding some unertinty to the estimtion of hrger size nd hrging time. To show some of the options ville, severl possile hrger power levels hs een investigted. These re hosen on the kground of the rnge presented y Solris us nd oh in []. lulting the hrging time for the three onsumption level re done for three fsthrger power levels, 0.90 effiieny ftor for the ttery hrging loss is used, the expression used is: t hrge = e us l roundtrip P hrger η hrge (2.9) Where t hrge is the time it tkes to hrge, e us is the onsumption of the us in [ ] kw h km, l roundtrip is the length of the us route, η hrge is the hrge yle effiieny nd P hrger is
hrger on/off 2.6 hrging pttern derivtion 17 the power delivered y the hrger in [kw]. This yields the hrging times seen in tle 2.4 Tle 2.4: hrging times t different levels of hrger power output nd onsumption rtes. hrging time [minutes] se kwh 150kW 00kW 500kW Low 18.49 7.4.7 2.2 Mid 28.89 11.5 5.8.5 High 9.72 15.9 7.9 4.8 With 6 minutes eing the shortest time etween deprtures from the hrging terminl, some of the hrging solutions re undesirle to implement in prtie s they would violte the time onstrint posed y the shedule. With one hrger ville, hrging times longer thn 6 minutes re uneptle. With 4.8 minutes eing the worst se senrio for the 500kW fst-hrger, n estimted hrging time rounded to five minutes will e used. The frequeny of rrivls of uses vries through 24 hour yle, with rush hour pek oiniding with residentil onsumption pek seen in figure 2.2. This retes higher risk of lod prolems on line with hrgers onneted. hrging frequeny through 24 hour yle with 5 minutes hrging period on off 0 2 4 6 8 10 12 14 16 18 20 22 24 Time of dy Figure 2.5: hrge frequeny used for simultion,with eh hrge yle set to 5 minutes. Figure 2.5 shows the hrging pttern derived from the time tle for us line. Tle 2.5 explins the rrivl frequeny of eh group seen in 2.5.
18 2 Grid model nd prmeters Tle 2.5: Frequeny groups for us line [18]. Setion Frequeny [minutes] 6-8 10 20-0 2.7 hoie of hrger dimensions The fst-hrger speifitions hs een hosen on the sis of us frequeny in tle 2.5, the onsumption nd orresponding hrge time in tle 2.4 nd the physil limits of the 10/0.4kV trnsformers in the grid. The fst-hrger lok used is then hosen to hve power of 500kW whih is sustined for five minutes for eh hrge yle. This hoie is to over the worst se hrging time for n E-us overing the 1km route. The power of the hrger orresponds to ville tehnology s seen in [] nd is le to keep within the 60kV limit of the urrent trnsformers instlled in the Nordhvn grid. The 500kW power does put high lod perentge on the trnsformer nd it might e worth onsidering n upgrde of the trnsformer to etter ope with the lod in prtie.
2.8 Summry of modelling ssumptions 19 2.8 Summry of modelling ssumptions The ssumptions mde in order to simulte nd evlute the senrios for n integrtion of fst-hrger in MV power grid re summrized here: Grid ssumptions le length on 0.4kV side is ssumed to e negligile 10/0.4kV trnsformer t hrger us is ssumed to e upgrded if lod exeeds pity Lod ssumptions Lod sle is mde on kground of Denmrk s whole Is lmost purely resistive hrger ssumptions Trnsient ehvior is not investigted Power eletronis re ssumed to e le to mintin unity powerftor Totl hrmoni distortion is ssumed to e negligile rmp up is ssumed to e ner instntneous ESS ssumptions ESS is dedited to supporting fst-hrger Is loted t sme us s fst-hrger Hs the sme properties s the plnned Nordhvn ESS in terms of pity nd mx power
20
HPTER se study This hpter desries the senrios used to determine the impts of the integrtion of fst-hrger in the power grid. The effets re investigted in terms of feeder lod nd voltge qulity. The investigtion is done to ensure tht the opertion of fst-hrgers does not rek ny grid odes or entirely overlod the system during fult on feeder. Synergy with ESS is eing investigted to see if it is possile to redue the stress on the grid used y fst-hrger y supporting during fult or y supplying power lolly when demnd is high. fter showing the effets of single fst-hrger, nother one is introdued in order to explore more hypothetil senrio where the grid proves too wek to operte with the fst-hrgers within the grid ode oundries. Finlly it will e explored if it is tehnilly possile to sty within norml opertion oundries with the help of ESS when two 500kW fst-hrgers re instlled..1 Senrios series of senrios hs een simulted to investigte the onsequenes of instlling fst-hrger in n existing residentil power grid nd to give piture the tions tht n e tken to void overloding the grid or violting norml opertion riteri desried in the previous hpter. The sope of the projet is to investigte the lod hnge on the grid during different opertion modes. The senrios re foused on norml opertion nd opertion during single fult on one feeder t time. The simultions hs een onduted in n order with inresing omplexity nd fult severity. First, the integrtion of single 500kW hrger is simulted t severl lotions to determine the worst se lotion to e used when simulting other senrios with different onfigurtions. Seond, the grid response to fult on one feeder is investigted with oth one nd two 500kW units. During 1000kW senrios, the pity of the trnsformer onneting the fst-hrgers hs een inresed from 60kV to 1.5MV to e le to hndle the extr lod. The forementioned senrios re used s se ses for omprison to senrios where
22 se study the ESS re then introdued to show the support it n provide to void reking the norml opertion riteri for the grid or to keep the grid from overloding in fult situtions..1.1 Influene of the lotion of fst-hrger in norml opertion simultion of three senrios hs een rried out to determine the differene of influene the distne from the fst-hrger lod us to the 60/10kV trnsformer hs during norml oper tion. oth swith S 1 nd S t the eginning the the feeders re losed during the length of the simultion nd the end swith,s 2, is open. The fst-hrger onfigurtion is single unit of 500kW. The senrio length is one hrge yle nd is t the pek onsumption time of dy t hour 18 to 19 s seen in figure 2.2. The senrios re ompred in terms of urrent flow in the feeders where the fst-hrger is loted nd y mesuring the voltge level t the lod us frthest from the 60/10kV trnsformer on the feeder ontining the fst-hrger. Figures.1,.2,. show the positions to e srutinized. S1 Externl grid 60kV / 10kV 10kV / 0.4kV S ESS S2 10kV / 0.4kV Figure.1: Overview of the grid with oth fst-hrger nd ESS onneted t lotion 1.
.1 Senrios 2 S1 Externl grid 60kV / 10kV 10kV / 0.4kV S ESS S2 10kV / 0.4kV Figure.2: Overview of the grid with oth fst-hrger nd ESS onneted t lotion 2. S1 Externl grid 60kV / 10kV 10kV / 0.4kV S2 S 10kV / 0.4kV ESS Figure.: Overview of the grid with oth fst-hrger nd ESS onneted t lotion..1.2 Influene of the lotion of fst-hrger during feeder fult sitution worst se sitution fult is introdued on one feeder in this set of senrios. The swith on one of the feeders is opened during pek hour nd the other feeder hs to rry the lod through the end swith whih is losed t the sme time s the fult ours. For these simultions, pling the fst-hrger t the strt of the feeder is expeted to generte the lrgest prolem s this retes the longest distne from supply to frthest lod. This senrio is lso used to demonstrte the potentil for supporting the opertion of fst-hrger with ESS. These simultions re done with oth single 500kW fst-hrger nd with two units with the sme power. This is done to rete sis to ompre the improvements used y introduing ESS.
24 se study The performne of the grid is gin mesured in terms of urrent nd power flow in the feeders where the fst-hrger is loted nd y mesuring the voltge level t the lod us ontining the fst-hrger. The retive power is expeted to e of little signifine due to the resistive modelling of the lods, thus ll retive power needed will e from the feeders. Tle.1: Swith sttus during the three opertion situtions used for the se study. Norml opertion Feeder 1 fult Feeder 2 fult S 1 losed Open losed S 2 Open losed losed S losed losed Open Tle.1, gives n overview of the swith onfigurtion used in the senrios..1. Worst se fult sitution with ESS support The worst se plement of fst-hrger determined y the previous simultion is used here to show the improvement ESS n provide during fult. The ESS is loted on the sme lod us s the fst-hrger nd provides lol power injetion to relieve the surrounding grid. The simultion is done with either one nd two 500kW units onneted to the first trnsformer on feeder 1. 10-minute fult is then introdued t the sme time s ove to e le to ompre the grid ehvior. The ESS will injet power during the fult, leding to smller drop in voltge nd lower lod on the remining feeder..1.4 Pek shving y lod shifting with ESS Fult situtions our rrely ompred to demnd peks whih hppen dily, therefore it would e very ostly to instll grid sized ESS just to mitigte fult lods. nother purpose for ESS ould then e to ensure tht the norml opertion limits re not exeeded during pek hours. This is done y letting the ESS injet power proportionl to the mount ove the riteri limit on the respetive feeder. The dishrged energy is then repled during off pek hours nd thus the lod will e shifted. The lod shift pility is of ourse limited y the physil speifitions of the ESS in terms of mximum power nd energy pity. The simultion is rried out for oth one nd two 500kW units. The ESS is not expeted to e le to hndle ll of the exess demnd, the pity required is investigted fter.
.2 Summry of senrios 25.1.5 Estimtion of ESS pity required to fully over pek shving demnd The finl spet of this investigtion is n estimtion of the totl ttery pity needed to fully over the lod shift demnd shown in previous simultions. This is done y mesuring the urrent mgnitude t ll time steps where the feeder urrent exeeds 50% of mx nd then mesuring the gp when there is exess pity. The two resulting vetors re then plotted s umultive sums showing the totl power required s well s the ville power for hrging the ESS k up. The slope of the dishrge umultive sum revels where the ritil time period is, the steeper the urve, the more power is needed. Sutrting the hrge nd dishrge umultive sums will revel when there is defiit of hrging opportunities. To show the signifine of hrge time length, this investigtion is rried out for hrge time lengths of 5, 4 nd minutes..2 Summry of senrios Tle.2 shows n overview over the onduted simultions nd the si prmeters used. Tle.2: List of simultions nd se prmeters used to disover the impt of fsthrgers on the grid. hrger lotion P hrger Fult ESS Lotion omprison Feeder 1 500 kw No No Feeder 2 500 kw No No Feeder 1 end 500 kw No No N-1 omprison Feeder 1 500 kw Yes No Feeder 2 500 kw Yes No Feeder 1 end 500 kw Yes No N-1 with support Feeder 1 500 kw Yes Yes Feeder 1 1000 kw Yes Yes Lodshift pility Feeder 1 500 kw Yes Yes Feeder 1 1000 kw Yes Yes Lodshift ESS pity estimtion Feeder 1 1000 kw No No Feeder 1 refers to the pink/upper feeder in the model nd Feeder 2 to the lue/ottom
26 se study feeder. The ESS hs onstnt pity of 500kWh nd strts t 80% SO in ll simultions where it is present. This is done to reflet prtil opertion of ESS whih rrely rehes 100% SO.
HPTER 4 Results The senrios desried in the previous hpter hve een simulted using MTL Simulink nd the results hve een doumented in this hpter where it will e desried nd evluted in terms of lod flow, feeder urrents nd voltge levels. The otined results re presented in n order strting with n investigtion of the effet 500kW lod step hs on the power grid t three lotion. The response to 10 minute fult on one feeder is then shown t three lotions. sed on these results the worst se lotion is hosen for use in the following simultions. The fult sitution is then investigted gin ut with ESS onneted on the sme us s the fst-hrger, to show the improvements on grid strin tht my e provided y suh servie. Finlly, senrios re simulted with the ESS used for pek shving when the grid is operting in norml onditions. n investigtion of the required pity is done for vrious hrge time lengths. 4.1 seline senrios The se senrios shows the simplest onfigurtion of the grid, with single 500kW fst-hrger unit onneted t 10/0.4kV trnsformer on one of the feeders. The grid is simulted in three onfigurtions where the fst-hrger is loted t either one of the first trnsformers on the feeders, or t the end trnsformer of feeder 1. These senrios re eing used s enhmrk for the more dvned senrios where multiple fsthrger units nd ESS is onsidered. The simultions give sis for omprison t the pek onsumption hour s this will e the time where overloding will our if it hppens. 4.1.1 se senrio w/ hrgers in ll lotions The first simultions re done to determine the lod on the grid when fst-hrger of 500kW is introdued. The simultions re done over the spn of five minutes nd over lod step during the pek hour of the dy.
urrent [] urrent [] 28 4 Results Feeder 1 urrent, 500kW hrger with no ESS 120 100 80 60 40 20 11 112 111 110 Lotion 1 Lotion 2 Lotion 0 0 50 100 150 200 250 00 Time [s] Figure 4.1: urrent in feeder 1 during 500kW lod step t pek hour, the smll ox is zoom on the urrents t the eginning of the lod step. 100 80 60 Feeder 2 urrent, 500kW hrger with no ESS Lotion 1 Lotion 2 Lotion 40 20 0 0 50 100 150 200 250 00 Time [s] Figure 4.2: urrent in feeder 2 during 500kW lod step t pek hour, the feeder urrent is uffeted of lod steps on feeder 1. Figure 4.1 nd 4.2 show the urrent flow mesured t the eginning of eh feeder in the event of lod step of 500kW used y the fst-hrger. Sine the feeders do not hve the sme se lod due to differene in numer of trnsformers, the results hve een split nd shown on two figures, one for eh feeder. From figure 4.1 it n e seen tht the effet of hving the fst-hrger pled t the end of feeder of this length is insignifint in terms of urrent flow. Thus the resistive losses of moving the hrger to the k of the feeder is negligile nd is not signifint ftor in the identifition of worst se lotion senrio in norml opertion.
tive power [MW] tive power [MW] 4.1 seline senrios 29 1.4 1. 1.2 1.1 1 0.9 Feeder 1 tive power, 500kW hrger with no ESS 1.7 #106 1.6 1.5 Lotion 1 Lotion 2 Lotion 0.8 0 50 100 150 200 250 00 Time [s] Figure 4.: tive power flow in feeder 1 during 500kW lod step, the smll ox is zoomed re showing the eginning of the lod step. 1.1 1 0.9 0.8 0.7 0.6 Feeder 2 tive power, 500kW hrger with no ESS 1.08 #106 1.07 1.06 1.05 Lotion 1 Lotion 2 Lotion 0.5 0 50 100 150 200 250 00 Time [s] Figure 4.4: tive power flow in feeder 2 during 500kW lod step, the smll ox is zoomed re showing the eginning of the lod step. The hnge in urrent is trnslted into power hnge s seen in figures 4. nd 4.4. From the zoom on figure 4. the differene etween pling the fst-hrger t the eginning or t the end of feeder 1 is pproximtely 0.00MW or kw.
Retive power [kvr] Retive power [kvr] 0 4 Results 4 42.5 42 Feeder 1 retive power, 500kW hrger with no ESS Lotion 1 Lotion 2 Lotion 41.5 41 40.5 0 50 100 150 200 250 00 Time [s] Figure 4.5: Retive power flow in feeder 1 during 500kW lod step. Feeder 2 retive power, 500kW hrger with no ESS 25.8 25.6 Lotion 1 Lotion 2 Lotion 25.4 25.2 25 24.8 24.6 0 50 100 150 200 250 00 Time [s] Figure 4.6: Retive power flow in feeder 2 during 500kW lod step. The retive power flow omes exlusively from the loding of the feeders. s seen in figure 4.5, the differene etween the worst nd est lotion on feeder 1 is less thn one kvr. Thus the retive power is not the deiding prmeter for worst se lotion. The influene of the retive power would of ourse e greter if different power ftor hd een used for the lod models. This leds to the next evlution prmeter, whih is the voltge level. This hs een mesured on the end of the sme feeder s the fst-hrger is loted t, it is mesured during the sme lod step s the ove results.
Voltge [p.u.] 4.1 seline senrios 1 1 0.995 Voltge levels t end trnformers 500kW hrger with no ESS Lotion 1 Lotion 2 Lotion 0.99 0.985 0.98 0 50 100 150 200 250 00 Time [s] Figure 4.7: Voltge hnge during 500kW lod step, the voltge is mesured t the end of the respetive feeder rrying the fst-hrger. Figure 4.7 shows the voltges mesured t the end lod of the feeder rrying the fsthrger during the lod step. The voltge is shown in per unit to give etter reltion to the grid ode requirements for voltge qulity. Looking t the voltges llows one to distinguish the effet eh lotion hs on the grid response to the fst-hrger lod step. s it might e expeted, the longest pth hs the most severe effet on the lod voltge level. Lotion of the fst-hrger t the end of feeder 2 is not investigted s the end of feeder 1 will lwys hve worse onditions due to the numer of 10/0.4kV trnsformers is seven on feeder 1 ompred to five on feeder 2. The results show tht it is possile to integrte fst-hrger into the power grid with the urrent onfigurtion without exeeding ny oundries for norml opertion. From this simultion lone it mkes sense to hoose the end of feeder 1 s lotion for the fst-hrger in the following simultions. Sine the length of the feeder from trnsformer to fst-hrger hs shown its signifine, it is ovious tht further simultions must e onduted to show the most extreme length of feeder to fst-hrger us to determine the worst se lotion. This is done in the following simultions y looking t the fult opertion. 4.1.2 Feeder fult ll lotions The sme grid onfigurtions s efore is used gin for the investigtion of the ehvior during totl fult t the eginning of the feeder rrying the fst-hrger. The simultions lst for one hour t the pek onsumption hour of the dy. The fult is introdued fter 10 minutes of simultion nd lered gin fter nother 10 minutes. The mesurement point is lwys on the non-fult feeder in order to mesure urrent nd power during the fult. The fult is ssumed to e hndled instntneously nd S 2 t the end of the feeders is losed in the sme time step. It is importnt to note tht this would not e the se in
Voltge [p.u.] urrent [] 2 4 Results rel world opertion s ll proedures hve dely nd some even re mnully operted. 250 200 150 500kW hrger N-1 sitution urrent flow Line 1 Strt Line 2 Strt Line 1 End 100 50 0 18.00 18.0 19.00 Time of dy Figure 4.8: urrent hnge during lod step. Looking t figure 4.8 revels the urrent experiened y the remining feeder during fult sitution, the feeder hs to e le to rry oth lods without exeeding the physil limits of the le. The lk, dotted, horizontl line t the top of the plot signifies the 100% loding of the feeder. 1 0.99 0.98 500kW hrger N-1 sitution Voltge level Line 1 Strt Line 2 Strt Line 1 End 0.97 0.96 0.95 18.00 18.0 19.00 Time of dy Figure 4.9: Voltge hnge t frthest trnsformer during lod step. Moving on to the voltge level during fult. It n e seen from figure 4.9 tht the differene in voltge levels re more outspoken during fult. The severity of the voltge drop is worst when the fst-hrger is loted t the first trnsformer of feeder 1. This is onsistent with the ft tht feeder impedne is very influentil on the voltge level due to the omplex impedne of the les[19].