Study Guides, Notes, & Quizzes

A Study on the Expert Voltage Control in Korean Power System

Description
1 A Study on The Expert Voltage Control in Korean Power System Won Kun Yu, Heung Jae Lee, Chan Ho Lim, Eun Jae Lee, Kyeong Seob Lee Tae Kyun Kim, Jeong Hoon Shin, Su Chul Nam  Abstract--Recent massive power outages in Europe and North America were caused by unbalances in reactive power, resulting in voltage collapse. In order to solve these problems voltage/reactive power control system have developed in many advanced countries for their own domestic power system. In terms of voltage stabilit
Published
of 4
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  1     Abstract  --Recent massive power outages in Europe and NorthAmerica were caused by unbalances in reactive power, resultingin voltage collapse. In order to solve these problemsvoltage/reactive power control system have developed in manyadvanced countries for their own domestic power system. Interms of voltage stability, as the KEPCO recognized the necessityof voltage/reactive power control system, we developed a hybridintelligent voltage/reactive power control system together withnumerical algorithm based on the sensitivity matrix. In thispaper, it is introduced that voltage control rules of expert systembase on empirical knowledge of power system operator tomanaging the bus voltage efficiently. And the search method of expert system has applied least-cost search for state-space that isapplied to quantization process. As a result, the hybrid intelligentvoltage/reactive power control system showed a goodperformance.   Index Te rms  ²  hybrid intelligent voltage/reactive powercontrol system, Least-cost search, quantization I.   INTRODUCE  ecent massive power outages in E urope and  N orthAmerica were caused by unbalances in reactive power,resulting in voltage collapse .   I n order to solve these problemvoltage/reactive power control system developed in manyadvanced countries for their own domestic power system [1]-[6] .   R  ecently, Korea power system has been operated moreclosely to stability limits in state heavy-load because of rapidgrowth in load-demand .   I n addition, reactive power demand isincreased due to long-distance transmission lines and increaseof reactive power losses .   U nfortunately, the power system voltage is maintained only by individual substations since it is not actually easy to secure T his work was supported in part by the the Power  IT   R  esearch Grant of theMinistry of  C ommerce, I ndustry and E nergy .  H . J . Lee is with the D epartment of  E lectrical E ngineering, Kwangwoon U niversity, Seoul Korea (e-mail: hjlee@kw . ac . kr) .   C. H . Lim is with School of  C ompute Multimedia E ngineering, Gyeongju U niversity, Gyeongju Korea (e-mail: chlim@gju . ac . kr) .  W . K  . Yu is with the D epartment of  E lectrical E ngineering, Kwangwoon U niversity, Seoul Korea (e-mail: ywk@kw . ac . kr) .   E. J . Lee is with the D epartment of  E lectrical E ngineering, Kwangwoon U niversity, Seoul Korea (e-mail: dmswo82@kw . ac . kr) .  K  . S . Lee is with the Juam E lectric & C ommunication C o . Ltd, SeoulKorea (e-mail: kyeong@kw . ac . kr) .   T. K  . Kim is with Power System Lab at Korea E lectric Power  R  esearch I nstitute (K  E P RI )/K  E P CO , D aejon, Korea (e-mail:tkkim@kepri . re . kr) .  J . H . Sin is with Power System Lab at Korea E lectric Power  R  esearch I nstitute (K  E P RI )/K  E P CO , D aejon, Korea (e-mail: jhshin@kepri . re . kr) .  S .   C.    N am is with Power System Lab at Korea E lectric Power  R  esearch I nstitute (K  E P RI )/K  E P CO , D aejon, Korea (e-mail: scnam@kepri . re . kr) .   the site for voltage compensation equipment and the regionalsystematic voltage control framework is not really prepared .  Accordingly, Korea recognized the necessity of voltage/reactive power control system in terms of voltagestability voltage and developed hybrid intelligentvoltage/reactive power control system for the domestic systemusing numerical algorithm based on the sensitivity matrix andintelligent technology[7] .   C urrently, the developed intelligentcontroller will validate the performance in the RTD S( R  eal T ime D igital Simulator) to apply domestic power system .   T hehybrid intelligent system is affected by the performance of expert system in main control system .   T ypically, the expertsystem consists of a knowledge-base and inference engine .  Knowledge-base stored data to solve problems and voltagecontrol rule, the inference engine is used to resolve a given problem by connecting the rule and the fact .   I n particular, theinference of an expert system is result of searching in thestate-space, so we need an effective search method to solve agiven problem . So, this paper developed new voltage controlrules and applied least-cost search for state-space that isapplied to quantization process . As a result it was performancethat the hybrid intelligent voltage/reactive power controlsystem improved .   II. S TRUCTURE   OF   H YB RID   I  NTE LL I G ENT   V O L T AG E / R  E A CTI V E   P O W ER    C ONTRO L S YS TE M   T he following F ig . 1 is a structure of developed intelligentvoltage/reactive power control system .   T he intelligent power system controller stored various data in dynamic databasefrom power system, inference engine performed voltagecontrol problem process by using voltage control rules of knowledge-base and numerical algorithm based on thesensitivity matrix .   F inally, the control signal was sent to power system .    A. Knowledge-base T he knowledge in a specific problem domain is classified by truth and rule and then stored in the database and rule base .   T hen again, D atabase is divided by static database anddynamic database, and stored as immutable truth in a specificdomain or hypothetic truth derived from the inference process .   A Study on T he E xpert Voltage C ontrol in Korean Power System Won Kun Yu, Heung Jae Lee, C han Ho Lim, E un Jae Lee, Kyeong Seob Lee T ae Kyun Kim, Jeong Hoon Shin, Su C hul  N am   R     2 F ig . 1 . Structure of Hybrid I ntelligent Voltage/ R  eactive Power  C ontrolSystem T his paper developed a database of the following in order toresolve the problem efficiently .   1) Database    U  pper and lower limit of each bus voltage    U  pper and lower limit of the voltage regulation    U  pper and lower limit quantity of compensation devices    Generator 0 .95 ~ 1 . 0 5 [p . u]  Priority of compensation devices  Quantization level of generator terminal voltage 2) Rule base    I f the bus voltage exceed upper and lower limit of each busvoltage the system operate the controller     I f abnormal voltage occurred, firstly controller constitutesensitivity tree .      C ontroller select the compensation device of largestsensitivity    I f selected reactive power compensation device's capacityis lack, controller select the second highest compensationdevice    C ontroller operate the specified priority of compensationdevices    I f abnormal voltage occurred in several bus, controller operated based on the greatest abnormal bus voltage .      I f a bus voltage don¶t adjust within voltage regulation byusing compensation device of first ranking, compensationdevice of next ranking is committed .      R  eactive power compensation amount determined Linear Prediction method .      T he controller solved power flow to identify control effect III. S T A TE - SPA CE M ODE L OF   T H E P RO P O S ED V O L T AG E   CONTRO L P RO BL E MS   T his paper applied the weighted evaluation function such asexpression (1) to improve the performance of developed thehybrid intelligent voltage/reactive power control system .   I t is possible to minimize the weighted evaluation function such asexpression (1) when all compensation devices use discretecontrol quantity .    ¡  )min( E FK P K  FEP ³³³³³³((( §§§ iQ K T  giV vioV  (1) vi oV  is newly occurred abnormal bus voltage in normal busBasically, the transformer  T ab and the switched shunt useddiscrete quantity . But, compensation device such as generator terminal voltage used continuous quantity .   T herefore,continuous quantity such as generator terminal voltage must be converted to a discrete quantity through the quantization process .   T he following F ig . 2 is in partly expanded state-spacemodel of discrete effect quantity in power system composed of each 2 devices (generator, transformer, switched shunt) .   F ig . 2 .   E xpanded state-space model  A. Least-cost search After a given problem define the representation model, tosolve the problem we need a several strategy and one of thekey strategies is search .   T he searches will be defined by thetrial process to assess possible solution paths and reach frominitial state to final state .   I t may be divided into two categories: blind searches and heuristic searches .   T he blind searchesdivided to breadth-first search and depth-first search . But, thismethod is not good in case that state-space is large due to it donot contain intelligent decision .   T he heuristic searches is amethod to continuously search a solution after getting rid of the solution path that seems inappropriate by judgment such asheuristic knowledge or cost function .   T his method couldreduce state-space but might be occurred that do not solve .   T his paper used least-cost search to minimize weightedevaluation function such as expression (1) .   T he search processfollowing: 1) Step 1: About the bus where the abnormal bus voltage, a v1 nodethat the largest effect quantity (sensitivity value v   control  3 quantity) was selected by the system . And the system expandsv1 node as three quantized effect quantity .   U sing threequantized effect quantity, the system calculated liner  prediction and weighted evaluation function about all the busvoltage .   E valuation function quantities of expanded node are11, 9 ,10 as seen F ig . 2 .   T he system selected v1_2 node that thesmallest of evaluation function value and progress a selectionof compensation device of step 2 .   2) Step 2: T he system performs liner prediction by using effectquantity of v2 . As a result, the system selected transformer tapt1 because abnormal bus voltage occurred in normal bus . Andthrough the same process of v1 node, the system selected t1_3node that the smallest of evaluation function quantity .   F inally,if abnormal bus voltage is dissolved, the system selectedcompensation device of step 3 .   C onversely, if abnormal busvoltage is solved, the system finished the search process .   I V .   C AS E   S TUD Y   I n the case study, we found the abnormal bus voltage incase of a bus fault using PSS/ E in Jeju I sland .   I n order toverify intelligent system's operating characteristic, upper andlower limits of abnormal bus voltage are specified as 0 .9 77[p . u] ~ 1 . 023 [p . u] and upper and lower limits of voltageregulation are specified as 0 .9 77 [p . u] ~ 1 . 023 [p . u] . Also, priority of compensation device selected firstly the generator terminal voltage because of voltage control .   F or the transmission line outage joining Seongsan(200) busand Jochen(3 5 0) bus amongst all (n-1) transmission lineoutages, the voltage violations happened at Seongsan (200) bus as seen in T able    T he system recognizes that abnormal bus voltage occurred,and output a control action signal as the following T able  inorder to adjust Seongsan(200) bus voltage within voltageregulation .   T ABL E   I.   V O L T AG E P ROFI L E   OF B U S IN   C AS E   OF A TR  A  N SM I SS ION L INE   OUT AG E J OININ G S EON GSA  N   (200) B U S A  ND J OC H EN (3 5 0) B U S  B U S  NU MB ER   B U S  N AM E  V O L T AG E [p . u . ]120  N orth Jeju T P1 . 004 5  121  N orth Jeju C S1 . 004 5  122  N orth Jeju T S1 . 004 5  130 E ast Jeju1 . 003 5  140  N ew Jeju0 .99 8 9  1 5 0 Hanrim CC  1 . 0083160 Andeok 0 .999 7170 South Jeju T P1 . 0004180  N ew Seogwi0 .9 8371 9 0 Hanra0 .9 832 200 Seongsan 0 .9 6 55  210 Pyoseon0 .9 864220 Sanji1 . 0032330 Hanrim1 . 006 5  3 5 0 Jochen0 .99 48 T ABL E   II.   CONTRO L A CTION   -   DO    NOT APPL IED Q U A  NTIZ A TION  B U S  NU MB ER    CO MP EN SA TION   DE V ICE S CONTRO L Q U A  NTIT Y[p . u . ]20122  N orth Jeju T PGenerator terminal voltge+ 0 . 0 5  20173South Jeju T P#3Generator terminal voltge+ 0 . 0 5  20174South Jeju T P#4Generator terminal voltge+ 0 . 0 5   T ABL E    .   CONTRO L A CTION   - APPL IED Q U A  NTIZ A TION  B U S  NU MB ER    CO MP EN SA TION   DE V ICE S CONTRO L Q U A  NTIT Y[p . u . ]20122  N orth Jeju T PGenerator terminal voltge+ 0 . 0 5  20123  N orth Jeju C SGenerator terminal voltge+ 0 . 0333180 Shunt capacitor  5 [MVA R  ] - 1 bank  T he following F ig . 4 shows voltage profile of control action before and after  .  Looking at F ig . 4, the system adjusted abnormal bus voltagewithin voltage regulation and does not occurred another abnormal bus voltage .   I n other words, it could be identifiedthat intelligent voltage control action in satisfying the proposed weighted evaluation function F ig . 4 . Voltage Profile V .   C ONC L U S ION   T he proposed system developed voltage control rules of knowledge-base base on empirical knowledge of power system operator and applied the least-cost search satisfying the proposed weighted evaluation function using the continuouscontrol quantity that applied the quantization process .  As a result, this paper improved the hybrid intelligentvoltage/reactive power control system through the inferenceengine and the knowledge-base improvement .  V I.   R  EFERENCE S    P  eriodicals: [1]   S . J .   C heng et al . , ³An E xpert System for Voltage and R  eactive Power  C ontrol of a Power System´, IEEE   T rans .  on PW R  S, Vol . 3 .    N o . 4,  N ov . 1 9 88, pp . 144 9 -14 55.  [2]   J . L . Sancha et al . , ³Spanish Practices in R  eactive Power Management and Voltage C ontrol´, IEEE   C olloquium on  4 I nternational Practices in R  eactive Power  C ontrol, April1 99 3, pp . 3/1-3/4[3]   A . Gomez E xposito et al . , ³Sensitivity-Based R  eactivePower  C ontrol for Voltage Profile I mprovement´, IEEE   T rans . on PW R  S, Vol . 8,  N o . 3, Aug . 1 99 3, pp .   9 37- 9 4 5  [4]   J . L . Martinez R  amos et al . , ³A Hybrid T ool to Assist the O  perator in R  eactive Power/Voltage C ontrol and O  ptimization´, IEEE   T rans . on PW R  S, Vol . 10,  N o . 2,May 1 995 , pp . 760-768[ 5 ]   T. L . Le et al . , ³  N etwork  E quivalents and E xpert SystemApplication for Voltage and VA R    C ontrol in Large-ScalePower Systems´, IEEE   T rans . on PW R  S, Vol . 12 .    N o . 4,  N ov . 1 99 7, pp . 1440-144 5    P  apers from Conference  P  roceedings (   P  ublished): [6]   C. Belhadj et al . , ³An I ntegrated Power System Global C ontroller  U sing E xpert System´, C anadian C onferenceon E lectrical and C omputer  E ngineering 1 99 6 . Vol . 1,May 1 99 6, pp . 446-44 9.  [7]   H . J . Lee et al . , ³Hybrid I ntelligent Voltage and R  eactivePower  C ontrol System F or Jeju Power System in Korea´8th WS E AS I nt .   C onf  . on P O W ER    SYS TE M .   pp118~123 V II. B IO G R  APH IE S   H eung-Jae Lee received the BS, MS and Ph .   D.  degrees from Seoul  N ational U niversity, in 1 9 83,1 9 86 and 1 99 0 respectively, all in electricalengineering . He was a visiting professor in the U niversity of Washington from 1 995 to 1 99 6 . Hismajor research interests are the expert systems, theneural networks and the fuzzy systems applicationto power systems including the computer application . He is a full professor in theKwangwoon U niversity .   Won-Kun Yu received his BS and MS degrees inelectrical engineering from Kwangwoon U niversity,in 2003 and 200 5 respectively .   C urrently, he is pursuing Ph .   D course at Kwangwoon U niversity .  His research interests include state estimation andvoltage control using expert system .  Chan- H o Lim received the BS, MS and PH .   D.  degrees from Kwangwoon U niversity, in 1 99 1,1 99 3, and 1 99 8 respectively, all in electricalengineering . His major research interests areexpert system, neural networks and fuzzy systemapplication to power systems including computer application . He is an associate professor inGyenongju U niversity .  Kyeong-Seob Lee received his BS, MS and Ph .   D.  degrees in electrical engineering from Kwangwoon U niversity, in 1 99 8, 2000 and 2006 respectively .   C urrently, he is manager of  R  & D and ProductionBusiness at Juam E lectric & C ommunication C o .  Ltd .   Eun-Jae Lee received his B . S degrees from Kwangwoon U niversity in Seoul Korea . He has been a U nified Master's and D octor's C ourse studentsince 2008 . His research interests include stateestimation, topology identification, and substationautomation Tea-Kyun Kim received the BS, MS and Ph .   D.  degrees from Hanyang U niversity, in 1 9 86, 1 9 8 9  and 1 99 3 respectively, all in electrical engineering .   C urrently, he is the R  esearch manager of Power System Lab at Korea E lectric Power  R  esearch I nstitute (K  E P RI )/K  E P CO , D aejon, Korea Jeong- H oon Shin received the BS, MS and Ph .   D.  degrees from Kyungpook  U niversity, in 1 99 3, 1 995  and 2006 respectively, all in electrical engineering .   C urrently, he is the Senior  R  esearch E ngineer of Power System Lab at Korea E lectric Power  R  esearch I nstitute (K  E P RI )/K  E P CO , D aejon, Korea Su-Shul Nam received his BS and MS degrees inelectrical engineering from Korea U niversity, in2001 and 2006 respectively .   C urrently, he is the R  esearch E ngineer of Power System Lab at Korea E lectric Power  R  esearch I nstitute (K  E P RI )/K  E P CO , D aejon, Korea 
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks