1. Introduction. universities.

Transcription

1 Joural of Electrical Egieerig PHASOR MEASUREMENT UNIT BASED DATA ACQUISITION AND TRANSFER SYSTEM FOR THE SOLUTION OF THE ECONOMIC DISPATCH PROBLEM USING THE REAL TIME DIGITAL SIMULATOR Sethil Krishamurthy Carl Kriger Gaesa Deivakkau Ceter for Substatio Automatio ad Eergy Maagemet Systems (CSAEMS) Cape Peisula Uiversity of Techology P.O.Box 1906, Bellville, South Africa ad Abstract: The electric power utilities play a vital role i the geeratio, trasmissio ad distributio of the electrical power to the ed users. The power utilities geerally face two major issues firstly, power systems are expected to operate close to the maximum capacity, ad secodly, there is a eed for accurate moitorig ad cotrol of the power system etwork usig the moder techological advaces together with their associated cofiguratio tools. These two issues are itercoected as better moitorig allows for better cotrol of the power system. The developmet of the ew stadard-based power system techologies cotribute to cocept of buildig of a smarter grid. The challege is that this process requires the developmet of ew cotrol ad operatio architectures ad methods for data acquisitio, data trasfer, ad cotrol computatio. These methods require data for the dyamic state of the etire power system i real-time, which allows for the itroductio of sychrophasor-based moitorig ad cotrol of the power system. This paper describes the research work for itegratio of the ewer existig power system techologies to build fully automated systems for the real-time solutio of power system eergy maagemet problem, icorporatig data measuremet ad acquisitio, data trasfer ad distributio through a commuicatio etwork, ad data storage ad retrieval i oe complete system. The paper further details the developed methods, algorithms, procedures, software ad hardware tools for implemetatio of a lab-scale prototype of the power system ad the acquisitio ad trasfer of the data to the cotrol ceter i order to allow for the solutio of the optimal power dispatch problem i real-time usig real-time data. The developed prototype systems are tested for the fivebus power system model, ad the system operatio test results are preseted. The model is developed withi the software ad the output sigals are fed to the software-based. The data from the GTNET- is trasferred to the Phasor Data Cocetrator (C) usig the Etheret. The developed data acquisitio, data trasfer, data retrieval ad data storage system algorithms ad software programs ca be expaded for use i the power grid eergy maagemet system for the ecoomic dispatch solutio i regioal or atioal cotrol ceters, smart grid applicatios, educatioal courses uiversities. ad postgraduate research at Key words: Smart grid, Sychrophasor, Phasor Data Cocetrator, Real Time Digital Simulator, Data Acquisitio system, Power System moitorig ad Cotrol, Ecoomic Dispatch problem ad Lagrages method. 1. Itroductio The curret techological developmets to address the challeges of the itroductio of a smarter grid iclude, sesig ad measuremet, itegrated commuicatio system, improved cotrol strategies, iterfaces ad kowledge-based decisio-makig systems i [1] ad [2]. The eed for further developmet of the above techologies requires a itegrative approach ad kowledge ot oly about the physical power system, but also about the commuicatio architecture ad the cotrol of the coected power system compoets. Additioally, the requiremets for the operatio of the smart grid are that it acts as a autoomous, fully automated ad optimized system. The challeges brought about by these eeds ad requiremets iclude which existig methods, techologies, hardware ad software tools are applicable that ca further be developed, or which ew oes require ivestigatio ad implemetatio. This meas that there is a eed for the developmet of ew techiques for itercoectivity of systems ad approaches betwee the power system elemets, ad a detailed uderstadig of the ew approaches, methods, algorithms, hardware ad software ecessary to achieve full itegratio of the operatio of the smart grid. The eergy maagemet systems are oe of the mai parts of the smart grid as they cotrol ad reduce the ucertaity ito the daily operatio of the power system, utilizig various methods ad tools for moitorig, visualizatio, state estimatio, real-time stability assessmet, restoratio, optimisatio of power flow, ecoomic dispatch, real-time cotrol, etc. Amog those cotrol strategies, this paper provide the solutio 1

2 Joural of Electrical Egieerig of the ecoomic dispatch problem i real-time. At preset the Eergy Maagemet System (EMS) operatio is performed by existig SCADA systems, which supervise, cotrol ad maage the geeratio, trasmissio ad distributio systems. The operatio of the SCADA systems does ot satisfy today s requiremets for fast iformatio i order to produce ad implemet real-time cotrol i the power system. This situatio is due to the fact that the SCADA system is desiged ad implemeted usig the priciples of the older techology,[3] for measuremet, commuicatio ad cotrol as follows: The 3-phase curret ad voltage sigals from the CTs (Curret Trasformers) ad PTs (Power Trasformers) are sampled ad set to the eergy maagemet system oly represeted by their magitude ad the lie active ad reactive powers. The measuremet widow is large, which produces data that is ot as accurate with the system chages. The assumptio that durig the widow period the load parameters will chage ad the system will remai costat ad is ot valid durig the measuremet widow. The SCADA system does ot allow measuremet of the power system voltage ad curret agles ad does ot cosider sychroized wide area measuremet. Aalysis of the system status is based o a multiple power flow approach, which creates a computatioal burde that is ot fast eough for realtime implemetatio. The existig situatio is that critical data is required to moitor the system durig trasiets ad disturbaces which require fast sychroized data i order to capture the system dyamics. The, based o these data, creatig a realtime situatioal awareess, it is possible to implemet fast real-time cotrol. Olie moitorig of the curretly large itercoected power systems is importat to visualize the etwork i real-time, which has become possible with the advet of the sychrophasor techology. The phasor measuremet uit provides real-time measuremet data such as bus voltages ad brach curret phasors from remote locatios to a phasor data cocetrator at the cotrol cetre that is sychroized ad time-stamped i [1] ad [2]. The satellite-based Global Positig System (GPS) is used to sychroize all the s, which are located at differet buses with oe pulse per secod (1 pps) trigger, havig a ucertaity i the order of aosecods. The samplig clock of the is sychroized to the Uiversal Time Coordiator (UTC) sigals to eable sychroized phase agle measuremet from the istataeous values of voltage ad curret. This paper uses a software-based, which is simulated withi the software iterface of the RTDS. The data streams are trasmitted to the software-based Phasor Data Cocetrators (Cs),[4] which are exteral to the Simulator [5-6] ad is accomplished usig the firmware optio for the GTNET card, which provides sychrophasor output data streams accordig to the IEEE C stadard, [7-11]. The RTDS GTNET model ca represet ad provide for up to 8 outputs with symmetrical compoet iformatio such as three-phase voltage ad currets usig the User Datagram Protocol (UDP) or the Trasmissio Cotrol Protocol (TCP) coectios. The reportig rate of each ca be set idividually betwee 1 ad 60 frames per secod. The eed for high resolutio sychroized data is addressed by the Phasor Measuremet Uits (s), which are sophisticated digital recordig devices capable of exportig GPSsychroized data with a high samplig rate [11]. This fact supported the approach to take decisios for power system operatio based o steady-state data from a local area to be shifted to payig attetio to the system dyamic behavior i a global or wide area. The use of the sychrophasor data requires a proper, fast, reliable, ad safe commuicatio etwork betwee the s ad the cotrol ceter that supports both the existig fuctioalities ad the future operatioal requiremets to be developed. Usig sychroized data at the cotrol ceter requires restructurig of the available methods ad software algorithms for solutio (calculatio) of the eergy maagemet problems ad the trasfer of the obtaied cotrol actios of the elemets of the power systems. O the basis of the above, it ca be cocluded that there is a eed for iterdiscipliary research to develop a close itegrated framework of the latest achievemets i the field of measuremet, commuicatio, optimisatio ad cotrol of power systems. This paper cocetrates o research work i this field followig the priciples of the measuremet based Smart Grid system ad aimig to produce real-time workig prototypes implemetig a real-time solutio of the power dispatch optimisatio problem for give power etworks. The case studies for the implemetatio for such a system have bee developed i the laboratory of the Cetre for Substatio Automatio ad Eergy Maagemet Systems (CSAEMS), at the Cape Peisula Uiversity of Techology, to demostrate to the power utilities how such a measuremet, commuicatio ad cotrol optimisatio framework ca operate from oe side, ad to support the iterdiscipliary educatio ad research for graduate ad postgraduate cotrol ad power system egieerig studets i the field of the ew power system moitorig, protectio, automatio ad cotrol techologies, from the other side. At preset the educatioal programs at uiversities are very arrow specialised programs, ad the egieers produced here are ot fully equipped to meet the challeges preseted by the ew smarter grid. A few of the papers reviewed solve the problem for dispatch of power for a real system ad usig real-time data ad optimisatio i [12-14]. The required chages to the power productio are also ot commuicated i real-time to the power 2

3 Joural of Electrical Egieerig producers. The reaso for these problems are that commuicatio systems for data acquisitio, data trasfer, data storage, ad data retrievig i real-time have ot bee developed yet. The aim of this paper is the developmet of lab-scale variats of the power system supportig the real-time data acquisitio ad data trasfer usig the available hardware ad software to demostrate the possibilities for the realtime solutio of the optimal power dispatch problem usig real-time data. A lab-scale variat of the system for data acquisitio ad data trasfer usig the GTNET card Phasor Measuremet Uit () sychrophasor data is developed to ivestigate differet possibilities for capturig the power system data from the Real-Time Digital Simulator (RTDS) simulatio model ad to trasfer the data to a local or remote cotrol ceter usig the software-based C i [4]. The ivestigatios are doe for the five-bus power system model i [12], which are costructed withi the software, which is the software suite of the RTDS. This paper is structured with the itroductio as preseted above; the implemetatio of a lab scale data acquisitio system is give i sectio 2; the problem formulatio ad algorithm for the solutio of the dispatch problem is preseted i sectios 3 ad 4 respectively. The simulatio results ad discussio of these results are preseted i sectios 5 ad 6 respectively, with the coclusio preseted i sectio Implemetatio of a lab scale data acquisitio system usig the GTNET card Phasor Measuremet Uits (s) The fudametal cocepts goverig the sychrophasor techology-based data acquisitio ad data trasfer of power system ecoomic dispatch problem solutios is explaied i this sectio. The fuctios of the RTDS -based system are show i Fig. 1 below where the power system model is first simulated withi the software (orage box). The data from this model is firstly acquired (gree box) from the simulatio model, ad the trasferred to a remote ed computer which is regarded as the power system cotrol ceter, cotaiig a database (red box) ad also the optimisatio algorithm solutio (yellow box). Fig.1: RTDS GTNET card -based lab-scale systems fuctios The block diagram i the figure above is expaded i the flow diagram i Fig. 2 below. The simulatio model outputs are trasferred to the RTDS GTNET card where the physical sigal (Measuremet) is trasferred via Etheret (Data trasfer) ad coformig to the IEEE-C stadard. The coectio Tester software at the remote-ed PC, verifies the itegrity ad quality of the trasferred data sigals from the virtual s. The OpeC software represets a virtual Phasor Data Cocetrator which combies sigals from multiple Phasor Measuremet Uits (s). The data from multiple s is stored i a database (Data storage) o the same PC. The data from the database is read by the MATLAB algorithm that uses the Lagrage s method to solve the ecoomic dispatch problem i real-time (Data optimisatio problem solutio). The block diagram show i Fig. 3 which is a expaded view of both Figures 1 ad 2, is that of the developed RTDS GTNET card -based laboratory scale system to solve the ecoomic dispatch problem i real-time, where the RTDS geerates the voltages ad currets withi the software. RTDS Power system model simulatio Power system simulatios Data from the simulatio Measuremet RTDS GTNET Card Etheret coectio Tester Data trasfer Remote PC Data Acquisitio OpeC software database -PC Remote PC Data read by MATLAB to solve the ED problem Data storage ad optimisatio problem solutio Lagrages algorithm to solve Ecoomic Dispatch problem i MATLAB Fig.2: Flow diagram of the lab-scale data acquisitio ad data trasfer system usig the RTDS GTNET techology to acquire ad trasfer data ad to solve ED problem 3

4 Joural of Electrical Egieerig Ope C database GTNET Moitorig ad Cotrol statio Lagrages algorithm to solve Ecoomic Dispatch problem i Matlab Cluster of computer RTDS Fig. 3: Block diagram for the lab-scale data acquisitio system based o the RTDS GTNET card to solve the ecoomic dispatch problem These sigals are the fed to the software-based s (GTNET-s). The measured data from the GTNET is trasferred to the virtual Phasor Data Cocetrator (C)-OpeC [4], usig the Etheret commuicatio protocol. The GTNET card used is cofigured withi the software ad is sychroized to the Global Positioig System (GPS) usig a SEL-2407 satellite clock with a atea [15]. The RTDS GTNET card allows for the use of a maximum of 8 idepedet s to provide symmetrical compoet iformatio related to the three-phase sets of voltages ad currets, coformig to the IEEE-C stadard, [8-9]. A sigle ca provide a total of 12 phasors (three-phase voltage ad curret magitudes ad phase agles), the measured frequecy, the rate of chage of frequecy as well as four aalogue values ad a 1-to-16-bit digital value, [56]. The described RTDS output sigals with the desired parameters are geerated from withi the RTDS usig a GTNET card. From the GTNET card fixed i Rack-1 of the RTDS at the CSAEMS lab, voltage ad curret sigals are set to the remote ed via Etheret. At the remote ed, the OpeC software [4] is used to capture the RTDS sigals i real-time. These sigals are archived i the database. Every five miutes, the data is retrieved from the database by the MATLAB [16] optimisatio algorithm. The Lagrage s algorithm [13] is used to solve the Ecoomic Dispatch (ED) problem usig the real-time data which is retrieved from the database. The flow chart of the algorithm for the real-time solutio of the problem is give below i Fig. 4. Node voltage ad brach currets i Phasor Measuremet Uit Measured phasor sigals Voltages Via, Vib, Vic Currets Iia, Iib, Iic, Phaseagles ia1, ia2, ib1, ib2, ic1, ic2, Data trasferred usig TCP/IP Etheret protocol IEEE C stadard Motiored phasor sigals Voltages Via, Vib, Vic Currets Iia, Iib, Iic, Phaseagles ia1, ia2, ib1, ib2, ic1, ic2, coectio tester ad ope C data gatherig database storage Stored phasor sigals Voltages Via, Vib, Vic Currets Iia, Iib, Iic, Phaseagles ia1, ia2, ib1, ib2, ic1, ic2, iscalculated Lagrage s algorithm to solve usigthe storedphasor sigals m the ecoomic dispatch problem P Di i 1 Optimal P P1, P2,...P Operator Cotrol of geerators Fig. 4: Flow chart of the real-time solutio of the dispatch problem usig the GTNET card -based data acquisitio ad trasfer system 4

5 Joural of Electrical Egieerig The paper cosiders oly the real-time solutio of the optimisatio problem based o real-time acquired ad trasferred data ad does ot solve the problem for the cotrol of the geerators. The GTNET card provides a user-defied actio to select ay umber of s from a maximum of eight available s per GTNET card. This paper uses four s based o the IEEE C stadard [8] to trasfer the three-phase curret ad voltage sigals from the power system model simulatio measuremets at the remote ed. Each cofiguratio has user-defied sigal frame rates at either 50 or 60 Hz. The istrumet trasformer s (PTs ad CTs) turs ratios are defied i the GTNET card AC source. Fig. 5 below shows the four s with the user-defied iput sigal ames for the three-phase voltage ad curret sigals. 3. Formulatio of the dispatch problem The objective fuctio of the ecoomic dispatch problem is to miimize the fuel cost or the emissio criterio fuctio of the thermal power plat as give i Equatios (1) ad (2) respectively. Miimize FC Fi ( Pi ) ( ai Pi 2 bi Pi ci ) $ / hr (1) Miimize EC Ei ( Pi ) ( di Pi 2 ei Pi f i ) kg / hr (2) Where: FC is the total Fuel Cost EC is the total emissio value Fi(Pi) is the fuel cost of the ith geerator Ei (Pi) is the emissio value of the ith geerator Pi is the real power geeratio of uit i ai, bi, ci are the fuel cost coefficiets of geeratig for uit i di, ei, fi are the emissio value coefficiets of geeratig for uit i is the umber of geeratig uits Uder the costraits determied by the power balace, trasmissio loss, ad geerator limits, as give below. 1) Power balace costrait P P i G PL [ MW ] (3) Where: Fig. 5: GTNET AC Source The default turs ratio settig used i the PTs ad CTs are 2000:1 ad 600:1 turs respectively. Oce the coectio cofiguratio is tested usig the coectio tester, the correspodig cofiguratio ad coectio files for each are saved o the remote Persoal-Computer (PC) where the OpeC software is ruig. This sectio preseted the implemetatio of the lab-scale system. The ext sectio presets the formulatio of the dispatch problem. PG is the total power geeratio of the system is the total demad of the system ad PL is the total trasmissio loss of the system 2) The trasmissio loss costrait ca be expressed as PL Pi Bi j Pj B0i Pi B00 [MW ] j 1 (4) Where Bij, B0i, Boo are the trasmissio loss coefficiets 3) Geerator operatioal costraits 5

6 Joural of Electrical Egieerig Pi,mi Pi Pi,max, i 1, (5) Where (7) is as follows: Pi,mi is the miimum value of the real power allowed to be produced by the geerator i Pi,max is the maximum value of the real power allowed to be produced by the geerator i. L 2ai Pi bi 2 Bij Pj B0i 1 0, i 1, (9) j 1 Pi Two optimisatio dispatch problems are formulated o the basis of the criteria (1) or (2) ad the costraits give i Equatios (3) to (5) as follows: Fid the values of the active power Pi, i 1, i such a way that the criteria Equatios (1) or (2) are miimized uder the costraits (3) to (5). The algorithm for solutio of both optimisatio dispatch problems are the same. They are described i the ext sectio usig the criterio give i Equatio (1). This sectio preseted the formulatio of the ecoomic dispatch problem. The followig sectio describes the algorithm for the solutio of the ecoomic dispatch problem. 4. Algorithm for solutio of the dispatch problem A method for the Lagrage s solutio of the ecoomic dispatch optimisatio problem Equatio (1), subject to the costraits give i the Equatios (3), (4) ad (5) is developed o the basis of the fuctio of Lagrage (L) by itroductio of Lagrage s multiplier λ, [13]. 2 (ai Pi bi Pi ci ) L Pi Bi j Pj B0i Pi B00 j 1 (6) Pi The optimisatio problem i Equatio (1), subject to the costraits give i Equatios (3), (4) ad (5) is trasferred to the problem for miimizatio of L accordig to Pi, i 1, ad maximizatio of L accordig to λ, uder the costraits give i Equatio (5). The coditios for optimality are derived for the solutio of the problem preseted by Equatios (5) ad (6) as follows: Accordig to Pi, L 0, i 1, Pi L 0 Accordig to, Equatio (5.9) ca be writte i a matrix vector form as follows: b1 a1 1 B01 B1 P B11 B12 1 b (10) P a2 1 B 2 B22 B2 1 1 B02 21 : 2 b a1 P B 1 B 2 B 1 B0 1 If the value of the Lagrage s multiplier λ is kow, Equatio (10) ca be solved accordig to the ukow vector P, usig the MATLAB commad: P E\D (11) Where b1 a1 1 B01 B1 B11 B12 b ad a2 E 1 1 B02 D B21 B22 B2 b a1 B B 2 B 1 1 B0 1 The value of λ is ukow ad has to be calculated usig the ecessary coditio for optimality as defied by Equatio (11): L P B i j 1 ij Pj B 0i Pi B00 P 0 i (12) Equatio (12) is ot a fuctio of λ, but represets the gradiet of L accordig to λ. At the optimal solutio this gradiet has to be equal to zero. A aalytical solutio for λ is ot possible ad a gradiet procedure for calculatio of λ is developed as follows: (7) k 1 k k, 0 (8) Where is determied by the equatio (12), k is the idex of the gradiet procedure, ad is the step of the gradiet procedure that starts with some give (13) k The derivatio of the coditio preseted i Equatio 6

7 Joural of Electrical Egieerig iitial value of the 0 variable. Lagrage s Whe 0, durig the iteratios the optimal solutio optimal solutio for the eergy that has to be produced by the geerators as a solutio of Equatio (10). for the Lagrage s variable is obtaied. It determies the The flowchart of the algorithm for the optimisatio solutio is show i Fig.6. The obtaied solutios for Pi, i 1, have to belog to the costrait domai determied by Equatio (5). That is why for every idex k of the gradiet procedure, the obtaied solutio is to fit to the costrait domai followig the procedure: Read the system data λ0, ε, a, b, c, Pmi, Pmax,, α, IterMax, B00,Bo1,B Calculate E(λ0) ad D(λ0) usig equatio (10) k if Pi Pi,mi P i,mi, k k Pi Pi, Pi,max, k if Pi.mi Pi Pi,max k if Pi Pi,max (14) Calculate P(0)= E(λ0)\ D(λ0) usig equatio (11) Check costraits Pmi, Pmax for every elemet usig equatio (14) The coditio for the ed of the iteratios is: k, or k m (15) Where 0 is a small umber ad m is the give λ(k+1)=λ(k)+ α Δ λ(k) maximum umber of iteratios. Yes The algorithm of the method is developed i [13], ad is preseted as follows: 1) The iitial value of the Lagrage s multiplier is guessed:, ad the value of the coditio for optimality is give. 0 2) I Equatio (10) matrix E 0 ad D 3) Equatio (11) is solved ad P determied. 0 0 are formed. 0 0 E \ D is 4) The obtaied vector P 0 is fit to the costrait domai (14). 5) 0 is calculated usig Equatio (12) where 0 P is substituted. 6) The coditio (15) is checked. If it is fulfilled the calculatios stop, if ot, improved value of 1 is calculated usig Equatio (13). 7) Calculatios Pi Pi 1 of the Calculate Δλ usig equatio (12) Usig equatio (13) improved values of are doe as i Equatio (14) ad so o. Iteratios cotiue util the coditio (15) is satisfied or the maximum umber of iteratios is reached. 8) The optimal solutio is used to calculate the total fuel cost ad emissio usig Equatio (1) ad (2) respectively. Δλ ε No Yes k IterMax Stop No Calculate the cost fuctio usig equatio (1) Fig.6: The flow chart of the ecoomic dispatch problem solutio algorithm usig the Lagrage s method, [13] This sectio preseted the algorithm for the solutio of the dispatch problem. The followig sectio describes the model ad presets the simulatio results. 5. Power system model, simulatio ad assigmet of the load demad powers to the GTNET card ad simulatio results The five-bus Idia power system model was desiged ad implemeted withi the software as show i Fig. 9 below. The five-bus power system has three costat source model geerators (gree boxes) ad four loads compoets (yellow boxes) ad their power demads are [20, 45, 40, 60] MWs. Currets eterig ito the four loads are moitored by coectig four circuit breakers (blue boxes) to the loads as is show. The voltage sigals of the loads are observed by moitorig the bus voltages at the poits where the load termials are coected. 7

8 Joural of Electrical Egieerig The buses are itercoected with pi-type trasmissio lies (orage boxes). These voltage ad curret sigals of the load compoets are assiged to the GTNET card withi the software. The voltage ad curret phasor sigals of the first two loads are assiged to 1 ad 2, ad the third ad fourth loads are assiged to the 3 ad 4 respectively. The moitored voltage ad curret sigals are time-stamped by the GPS clock at the host ed ad are trasferred to the remote-ed PC usig the Etheret commuicatio IEEE C stadard. The Rutime widow for 1 ad 2 oly, is show i Fig.7 where the three-phase voltage ad curret magitudes, the phase agles, ad the real power values of the idividual loads (power demads) are show. Fig. 7: Moitorig the voltage ad curret magitudes ad the phase agles of the 1 ad 2 sigals i the software (Rutime Widow The trasferred sigals are moitored i real-time at the remote ed usig the Ope Phasor Data Cocetrator (C) software. Fig.8 shows the GTNET voltage ad curret phasor sigals together with their respective time stamps beig moitored i the OpeC. TIME STAMP Fig.8: Moitorig the GTNET card 1 ad 2 sigals i the OpeC 8

9 Joural of Electrical Egieerig Fig.9: Five-bus power system etwork modelled i the with additioal compoets to implemet data acquisitio ad data trasfer usig the GTNET card sigals The OpeC data is stored i the database situated at the remote-ed computer. The data is read by the optimisatio dispatch algorithm i the MATLAB at the remote ed computer every 5 miutes. The SQL query is used to select the particular sigal from the database. The developed MATLAB software routie extracts the GTNET sigals from the database ito the MATLAB. Usig these captured data sigals the ecoomic dispatch problem based o the Lagrage s algorithm is solved oce every five miutes. Table 1 shows the moitored three-phase voltage ad curret sigals both i the ad i the s for the iitial loadig coditio of the power demad, =[ ] MW s respectively of the five-bus power system model. Table 1: sigal moitored i the ad for the iitial loadig coditio voltage ad curret sigals #1 (1) #2 (2) #3 (3) #4 (4) Va va Vb vb Vc vc Ia Ia Ib Ib Ic Ic It is observed that sigals moitored i the ad the s are very similar ad hece it is assumed that the laboratory set-up developed is capable of trasferrig the RTDS power system sigals from the local to the remote ed without ay loss of data. The same procedure is repeated for 9

10 Joural of Electrical Egieerig differet loadig coditios i order to test the data trasfer capability of the developed lab-scale system. The differet loadig coditios used are: = [ ] MW is the iitial loadig coditio. The sigal for the iitial loadig is give i Table 1. The slider (1) i the ru-time widow is used to icrease load oe from 20 to 25 MW. The the power demad becomes = [ ] MW ad it is called the first loadig coditio. Table 2 displays the sigals moitored i both the ad s for the first loadig coditio. Table: 2: sigals moitored i ad for the first loadig coditio voltage ad curret sigals #1 (1) #2 (2) #3 (3) #4 (4) Va va Vb vb Vc vc Ia Ia Ib Ib Ic Ic The slider (2) i the ru-time widow is used to icrease the load two from 45 to 50 MW. The the power demad becomes = [ ] MW ad it is called the secod loadig coditio. Table 3 shows the sigals moitored i both the ad the s for the secod loadig coditio. Table 3: sigals moitored i the ad for the secod loadig coditio voltage ad curret sigals #1 (1) #2 (2) #3 (3) #4 (4) Va va Vb vb Vc vc Ia The slider (3) i the ru-time widow is used to icrease the load three from 40 to 45 MW. The the power demad becomes = [ ] MW ad Ia Ib Ib Ic Ic it is called the third loadig coditio. Table 4 displays the sigals moitored i both the ad the s for the third loadig coditio. 10

11 Joural of Electrical Egieerig Table 4: sigals moitored i the ad for the third loadig coditio voltage ad curret sigals #1 (1) #2 (2) #3 (3) #4 (4) Va va Vb vb Vc vc Ia The slider (4) i the ru-time widow is used to icrease load four from 60 to 65 MW. The the power demad becomes = [ ] MW ad it is called Ib Ib Ic Ic Ia the fourth loadig coditio. Table 5 shows the sigals moitored i both the ad the s for the fourth loadig coditio. Table 5: sigals moitored i the ad for the fourth loadig coditio Voltage ad curret sigals #1 (1) #2 (2) #3 (3) #4 (4) Va va Vb vb Vc vc Ia Ia Ib Ib Ic Ic The moitored voltage, curret, ad phase agle sigals are used to calculate the real power of the loads usig Equatio (16). The voltage ad curret data are stored i the database; they are the read by the MATLAB-based optimisatio algorithm where the real power of the load compoets is calculated usig the data phasor sigals acquired at the remote ed usig Equatio (16). m V I cos (16) ia ia ia1 ia 2 Vib I ib cos ib1 ib 2 Vic I ic cos ic1 ic 2 Via,Vib,Vic are bus voltage magitudes i volt for the ith load, i 1, m Iia, Iib, Iic are brach currets i Amps for the ith load, i 1, m ia1, ia 2 are phase agle betwee the voltage ad curret i phase A for the ith load, i 1, m ib1, ib 2 are phase agle betwee the voltage ad curret i phase B for the ith load, i 1, m Where ic1, ic 2 are phase agle betwee the voltage ad is the total power demad i Watts curret i phase C for the ith load, i 1, m 11

12 Joural of Electrical Egieerig A compariso of the measured load power demads with the power demads calculated usig the trasferred GTNET card sigals ad the OpeC software is give i Table 6. It is observed that the sigal from the, coformig to the IEEE C stadard is successfully trasferred to the remote ed ( database) usig the GTNET card -based system ad the OpeC software, without ay data loss. It ca therefore be cocluded that the implemeted laboratory setup for the power system data acquisitio ad data trasfer is operatig accordig to the requiremets ad the data measuremets have bee validated. This method of data acquisitio ad data trasfer ca be used by the cotrol ceters i the moitorig ad cotrol of the power grid i order to solve the eergy optimisatio dispatch problem i realtime. Table 6: Compariso of the measured load power demads i the with the power demads calculated usig the GTNET card sigals ad OpeC software #1 (1) MW #2 (2) MW #3 (3) MW The Ecoomic dispatch solutio usig the RTDS GTNET card data acquisitio ad data trasfer system is give i Table 7. The total power demad varies from 165 #4 (4) MW to 185 MW with a icremet of 5 MWs. It is observed that the fuel cost icreases with the icrease i the power demad. Table 7: Ecoomic dispatch problem solutio usig the GTNET card sigals Lagrage s algorithm Geerator real powers i MW Fuel cost i Rs/hr i M W MATLAB (Optimal) (Simulatio) MATLAB (Optimal) (Simulatio) MATLAB (Optimal) (Simulatio) P1 P P3 PL i MW Number of iteratios (optimal) Computatio time i sec (Optimal) The Lagrage s method ca be used to provide optimal set poits for the geerator power productio. The load flow solutio withi the software uses the Newto-Raphso umerical method. By solvig the power flow problem the real power of the three geerators are kow ad are used to calculate the MATLAB (Lagrage s) (Newto Raphso) fuel cost (o the far right) of the power system which is show i Figure 10 for the case of =165 MW. Also from Table 7 it ca be see that the fuel cost is lower usig the Lagrage s method whe compared to the results obtaied by the simulatio. Figure 10: Ecoomic dispatch problem solutio i the software for a power demad of 165 MW 12

13 Joural of Electrical Egieerig The compariso of the solutio of the dispatch problem usig the developed two lab-scale systems is give i Table 9. The table cotais a compariso for power demad, load power, the umber of iteratios, computatio time, ad the software platform used to perform the calculatio. The highlighted boxes idicate the compariso for the Fuel cost i Rupees per hour, where firstly the calculatio is performed i MATLAB with the sigal take from the GTNET card output (blue box), ad secodly the calculatio i MATLAB is performed with the output take from the frot pael aalog output PB5 processor card (red box). It is observed that the values for the fuel cost are very similar although the physical sigals are take from two differet outputs from withi the RTDS simulatio model. Table 9: Compariso of the ecoomic dispatch problem solutios usig the sigals take from the RTDS GTNET card ad the frot pael aalog output RTDS PB5 card sigals Sychrophasor Techology usig GTNET card sigals i MW PL i MW Fuel cost i Rs/hr RTDS PB5 card frot pael aalog output sigals MATLAB i MW Lagrage s method Number of Computatio iteratios time i sec This sectio preseted the power system model, the simulatio thereof, the assigmet of the load demad powers to the GTNET card ad the simulatio results. The followig sectio describes a brief discussio o the data acquisitio ad trasfer systems usig the GTNET card sigals. 6. Discussio o the operatig characteristics of the data acquisitio ad data trasfer systems usig the GTNET card sigals This paper describes the applicatio of data acquisitio ad data trasfer usig the RTDS GTNET card sigals i order to solve dispatch optimisatio problems at the remote ed cotrol cetre PC. The five-bus Idia utility power system model is used to validate the data trasfer capability. Four differet loadig coditios are used to validate the data trasfer capability of the developed lab-scale systems. The followig coclusios ca be made: The GTNET card supports the IEEE C stadard to trasfer the large, complex power system etwork data from the local to the remote ed i real-time. The data trasfer capability of the sychrophasor techology usig the RTDS GTNET card is very fast ad accurate. The loss of data durig the process of commuicatio ad sigal trasformatios is very small. PL i MW Number of iteratios Computatio time i sec Fuel cost i Rs/hr MATLAB The compariso of the ecoomic dispatch problem solutio at both the local ad remote ed is doe. It has bee prove that the ecoomic dispatch solutio usig the Lagrage s method provides accurate values for real power that is to be produced by the geerators, ad the fuel costs are reduced usig the outputs from the GTNET. 7. Coclusio This paper ivestigates the capabilities of the developed data acquisitio ad data trasfer systems based o the RTDS GTNET card -based sigals to provide real-time data for the solutio of a dispatch optimisatio problem. The ivestigatios are preseted for the five-bus Idia power system model usig differet power demads. It is prove that the sychrophasor techology is a ideal solutio for data acquisitio ad data trasfer betwee the host ad the remote PCs without ay data loss. The optimized real power of the geerators ca be trasmitted back to the RTDS models i as set-poits for the geerators i the simulatio model. The GTNET card -based data acquisitio ad data trasfer systems creates favourable coditios for the optimisatio algorithm to produce a better solutio. The developed lab-scale systems will play a vital role i the future developmets of a smart grid eergy maagemet system. The paper provides the results of 13

14 Joural of Electrical Egieerig the developed data acquisitio ad data trasfer systems to solve the ecoomic dispatch problem i real-time usig the real-time data. Ackowledgemet The authors gratefully ackowledge the support ad for the use of the equipmet ad facilities withi the Ceter for Substatio Automatio ad Eergy Maagemet Systems at the Electrical Egieerig Departmet of the Cape Peisula Uiversity of Techology, South Africa to carry out this research work. This research work is fuded by the Natioal research foudatio (NRF) THRIP grat TP ad ESKOM TESP for CSAEMS developmet ad growth. Refereces [1] [2] J. De La Ree, V. Ceteo, J. S. Thorp, ad A. G. Phadke, Sychroized phasor measuremet applicatios i power systems, IEEE Tras. Smart Grid, vol. 1, o. 1, pp , [3] Mak, K.-H., Hollad, B. Migratig electrical power etwork SCADA systems to TCP/IP ad Etheret etworkig, Power Egieerig Joural, pp , [4] D. M. Laverty, R. J. Best, P. Broga, I. Al Khatib, L. Vafretti, ad D. J. Morrow, The Ope platform for ope-source phasor measuremets, IEEE Tras. Istrum. Meas., vol. 62, o. 4, pp , [5] D. Ouellette, M. Desjardie, R. Kuffel, Y. Zhag, ad E. Xu, Usig a real time digital simulator with phasor measuremet uit techology, Dev. Power, pp. 6 10, [6] R. Kuffel, J. Giesbrecht, T. Maguire, R. P. Wierckx, ad P. McLare, RTDS a fully digital power system simulator operatig i real time, i WESCANEX 95. Commuicatios, Power, ad [14] Krishamurthy, S., ad Tzoeva, R. Ivestigatio of the Methods for Sigle area ad Multi area Optimisatio of a Power System Dispatch Problem, Iteratioal Review of Electrical Egieerig (IREE), Praise Worthy Prize, Vol.7, No.1, Page: , [15] I.-K. Rhee, J. Lee, J. Kim, E. Serpedi, ad Y.C. Wu, Clock Sychroizatio i Wireless Sesor Networks: A Overview, Sesors, vol. 9, o. 1, pp , [16] S. J. Chapma, Matlab Programmig for Egieers, Maagemet, pp , [17] Giri, J., "Ehaced power grid operatios with a wide-area sychrophasor measuremet & commuicatio etwork," i Power ad Eergy Society Geeral Meetig, 2012 IEEE, vol., o., pp.1-3, July [18] Kezuovic, M., Heydt, G.T., DeMarco, C., Mout, T., "Is teamwork the smart solutio?," i Power ad Eergy Magazie, IEEE, vol.7, o.2, pp., March-April Phadke, A.G., Thorp,J.S., Sychroized phasor measuremets ad their applicatios, New York: Spriger, Computig. Coferece Proceedigs., IEEE, 1995, vol. 2, pp vol.2. [7] K. E. Marti, Phasor Measuremet Systems i the WECC, 2006 IEEE PES Power Syst. Cof. Expo., pp , [8] I. Power ad E. Society, C IEEE Stadard for Sychrophasor Measuremets for Power Systems, vol. 2011, o. December [9] K. E. Marti, Sychrophasor Measuremets Uder the IEEE Stadard C With Amedmet C a, IEEE Tras. Power Deliv., vol. 30, o. 3, pp , [10] Ke, M., Bruello, G., Adamiak, M., ad Atoova, G., IEEE Sychrophasor Stadard for Commuicatios PAC World magazie., December [11] Tholomier, D., Kag, H., Cvorovic, B., Phasor measuremet uits: Fuctioality ad applicatios, Power Systems Coferece, [12] Kothari, D.P., ad Dhilo, J.S., Power System Optimisatio, Secod Editio, PHI Learig Pvt. Ltd, Idia, ISBN: , pages 732, [13] Krishamurthy, S, Developmet of decompositio methods for solutio of a multiarea power dispatch optimisatio problem, DTech Thesis: CPUT, [19] Paul, T., Kimball, J.W., Zawodiok, M., Roth, T.P., Mcmilli, B. Ivariats as a uified kowledge model for Cyber-Physical Systems, 2011 IEEE Iteratioal Coferece o ServiceOrieted Computig ad Applicatios (SOCA),

15 Joural of Electrical Egieerig AUTHOR BIOGRAPHIES Dr. Sethil Krishamurthy received the BE ad ME i Power System Egieerig from Aamalai Uiversity, Idia ad Doctorate degree i Electrical Egieerig from Cape Peisula Uiversity of Techology, South Africa. He has bee a lecturer at the SJECT, Tazaia ad Lord Vekateswara ad E.S. College of Egieerig ad Techology, Idia. Sice 2011 he has bee workig as a Lecturer at the departmet of Electrical, Electroic ad Computer Egieerig, Cape Peisula Uiversity of Techology, South Africa. Mr. Gaesa Deivakkau received the Bachelor degree i Computer Sciece Egieerig from Aamalai Uiversity, Idia ad Master Degree i Electrical Egieerig from Cape Peisula Uiversity of Techology from South Africa i 2007 ad 2013 respectively. He worked as a Egieer i reputed Telecommuicatio compaies betwee 2008 ad His research focus areas are Power System Moitorig ad Cotrol, Sychrophasor Techology ad Database Maagemet Systems (DBMS). Dr S Krishamurthy is a member of the Niche area Real Time Distributed Systems (RTDS) ad of the Cetre for Substatio Automatio ad Eergy maagemet Systems supported by the South Africa Natioal Research Foudatio (NRF). He is a member of the Istitute of Electrical ad Electroic Egieers (IEEE), Istitutio of Egieers Idia (IEI), Istitutio of Egieers Tazaia (IET), ad South Africa Istitutio of Electrical Egieers (SAIEE). His research iterest is i the fields of Power Systems, Eergy Maagemet Systems, Parallel Computig, Computatioal Itelligece ad Substatio Automatio. Mr. Carl Kriger received the B.Eg degree from the Saxio Hogeschole Eschede i the Netherlads. He completed MTech i Electrical Egieerig at the Cape Peisula Uiversity of Techology. He is curretly the deputy leader of the ceter for Substatio Automatio ad Eergy Maagemet Systems withi the Departmet of Electrical, Electroics ad Computer Egieerig at the Cape Peisula Uiversity of Techology, Cape Tow, South Africa. His research iterest icludes cotrol systems, simulatio of Cotrol Systems, Eergy Maagemet Systems, Substatio Automatio ad Coditio Moitorig. 15