Published on

IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit

  • Be the first to comment

  • Be the first to like this

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. V.Veera Nagireddy, Dr. D.V. Ashok Kumar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 Vol. 2, Issue 5, September- October 2012, pp.986-990 Development Of Islanding Detection Technique For Utility Interactive Distributed Generation System V.Veera Nagireddy Dr. D.V. Ashok Kumar Assistant professor Professor EEE Department SDIT, NadyalABSTRACT Distributed generation (DG) on the protective relay that is located closest to the faultydistribution system provides many potential spot. As a result, a distributed generation tries tobenefits like peak shaving, fuel switching, supply its power to part of the distribution systemimproved power quality and reliability, increased that has been separated from the utilitys powerefficiency, and improved environmental system. In most cases, this distributed generationperformance. Impacts are steady state voltage assumes an overloaded condition, where its voltagerise, increase the fault level, power quality, and frequency are lowered and it is finally led toislanding. One of the problem is an islanding. stoppage. However, though this is a rare case, a So many islanding detection techniques generator (or a group of generators) connected to thisare available, each one having their own islanded system is provided with a capacity that isadvantages and drawbacks. A fuzzy rule-based large enough to feed power to all the loadspassive islanding detection technique is accommodated in the islanded system. When theimplemented in this project. The initial loads are fed power only from the distributedclassification model is developed using decision generations even after the power supply is suspendedtree (DT) which is a crisp algorithm. This from the power company, such a situation is called analgorithm is transformed into a fuzzy rule base by "islanded operation" or "islanding" [2]. They can bedeveloping fuzzy membership functions (MFs) broadly classified into two types locally built-in andfrom the DT classification boundaries. communication based detection schemes. Local detection methods detect islanding situations basedIntroduction on the information (such as voltage, frequency, Distributed generation (DG) generally refers harmonic. Local Detection methods further dividedto small-scale (typically 1 kW – 50 MW) electric into passive, active and hybrid techniques.power generators that produce electricity at a siteclose to customers or that are tied to an electric Technical challengesdistribution system. There are many reasons a Distributed generation (DG) is not withoutcustomer may choose to install a distributed problems. DG faces a series of integration challenges,generator. DG can be used to generate a customer‟s but one of the more significant overall problems isentire electricity supply; for peak shaving (generating that the electrical distribution and transmissiona portion of a customer‟s electricity onsite to reduce infrastructure has been designed in a configurationthe amount of electricity purchased during peak price where few high power generation stations that areperiods); for standby or emergency generation (as a often distant from their consumers, ”push” electricalbackup to wires owners power supply); as a green power onto the many smaller consumers.power source (using renewable technology); or forincreased reliability. A). Islanding Control some remote locations, DG can be less Depending on the amount of DG connectedcostly as it eliminates the need for expensive and the strength of the utility power system, theconstruction of distribution and/or transmission lines issues listed above can become substantial problems.. Of the challenges with DG, the problem of protection The distributed generation systems are against unplanned islanding is a significant one.operated in parallel with utility power systems, A single line diagram of a powerespecially with reverse power flow, the power quality distribution system is shown in Fig(a) . Theproblems become significant. Power quality substation is the sending end for several feedersproblems include frequency deviation, voltage where transmission voltage is stepped-down intofluctuation, harmonics and reliability of the power distribution voltage level. One of the feeders issystem. In addition, one of important problem is an shown in detail with many customer connectionislanding protection. A fault occurring in the power points. An islanding situation can occur indistribution system is generally cleared by the distribution system due to operation of an upstream breaker, fuse or an automatic sectionalizing switch in 986 | P a g e
  2. 2. V.Veera Nagireddy, Dr. D.V. Ashok Kumar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 Vol. 2, Issue 5, September- October 2012, pp.986-990response to fault or due to manual switching. The prime movers [3] Also, transients are created,islanding of the portion of network due to opening of which are potentially damaging to utility andre-closer switch „C‟ is shown in Fig. 1.1. The DG1 other customer equipment. Out of phasewill feed the power into the resultant island in this reclosing, if occurs at a voltage peak, willcase. The most common cause for a re-closer to open generate a very severe capacitive switchingis a fault in the downstream of the re-closer. An transient and in a lightly damped system, theislanding situation could also happen when fuse at crest over-voltage can approach three timesthe point „F‟ melts. In this case the inverter based rated voltage [4].DG3 will feed the local loads, forming a small  Various risks resulting from this include theislanded power system. Formation of islanding can degradation of the electric components as abe intentional or unintentional. consequence of voltage& frequency drifts.B). Intentional islanding E). Contributions The use of DG with proper control to supply  Implementing power distribution network withthe islanded portion of the network, during utility DG in MATLAB.outage conditions is called intentional islandingoperation. This improves the reliability of supply and  Development of passive islanding detection technique using fuzzy logic rule base.also brings many other benefits. Once the disturbanceis over the islanded portion of the network should be  An extensive simulation studies on islandingconnected back to the grid. The transition between detection based on above techniques usinggrid connected and islanding operation modes must seamless without any down time, in order toprovide uninterruptible power supply to critical and Another problem for power linesensitive loads. communication is the complexity of the network and the affected networks. A perfectly radial networkC). Benefits of intentional islanding operation with one connecting breaker is a simple example of The intentional islanding operation island signaling; however, more complex systemspotentially brings many benefits to the DG owner, with multiple utility feeders may find thatdistribution network operators (DNO) and customers. differentiation between upstream breakers difficult.Due to the increasing competition amongst energysuppliers to attract more and more customers, F).Fuzzy logic controllerintentional islanding of DG units are a valuable A fuzzy logic controller is a controloption. Intentional Islanding can improve the quality algorithm based on several linguistic control rulesof supply indices and reliability. The additional connected along them through a fuzzy implicationrevenue to DG owners can be achieved by the and a compositional rule of interference, togetherincreased power supplied during network outage. It with a defuzzification mechanism, that is to say, ahelps in increasing the customer‟s satisfaction due to mechanism that changes the action of fuzzy controlthe reduction in frequency and duration of in to one which is not fuzzy. Fuzzy logic controllers,interruptions from outages in the distribution as a focus for analysis nd design of control strategies,network. The intentional islanding operation can help are obtaining good results, increasing considerablythe society in general, by preventing civil their application in the last years. Fig 4.1 shows thedisobedience, assault, looting and disruption in configuration of fuzzy logic controller.essential services during utility power outageconditions G).DESCRIPTION OF SIMULATION MODEL Power distribution network with DGD). Issues In order to investigate the performance of Although there are some the proposed Technique during various contingenciesbenefits of islanding operation there are some a simulation modeldrawbacks as well. Some of them are as follows: was implemented. It is important that the  Line worker safety can be threatened by DG model reflects a real system in all vital parts. The sources feeding a system after primary sources behavior of the simulated system must be similar to have been opened and tagged out. what happens in a real situation. How this has been  The voltage and frequency may not be achieved is described in the following. In the maintained within a standard permissible level. preliminary study we have considered a system as Islanded system may be inadequately shown in the fig.a [11]. grounded by the DG interconnection.  Instantaneous reclosing could result in out of phase reclosing of DG. As a result of which large mechanical torques and currents are created that can damage the generators or 987 | P a g e
  3. 3. V.Veera Nagireddy, Dr. D.V. Ashok Kumar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 Vol. 2, Issue 5, September- October 2012, pp.986-990 𝚫f/ 𝚫t ≤ 2.18 Yes No 𝚫p/ 𝚫t ≤ 0.64 Class-1 Yes No 𝚫f ≤ 0.16 Class-1 Yes No Class-0 Class-1Fig. a : Block diagram of power distribution networkwith DG Fig. b: Decision tree chart [10].Transformers Class-1 means islanding and class-0 means non-Transformer T1 islanding.Rated MVA =25 f=60Hz Rated kV=69/13.8 Dyn1X1=0.1pu R1=0.00375pu Xm=500pu Rm=500pu I).Rule based fuzzy logic controllerTransformer T2 The most significant features 𝚫f/𝚫t, 𝚫p/𝚫t,Rated MVA =10 f=60Hz Rated kV=13.8/13.8 Ynd1 𝚫f are considered as X1, X2 and X3, respectively. TheX1=0.1pu R1=0.00375pu Xm=500pu Rm=500pu fuzzy MFs developed for variable X1 are A1 and A2, for X2 are B1, B2, and for X1 are C1, C2.Load data Per the above formulations, the rectangular MFs areLoad (L-1) = 10MW 3.5MVAR, load ( L-2) = 5MW derived as2.0MVAR A1=µ{X1, [2.18, 2.18, 34.0, 34.0]} A2=µ{X1, [-9.5, -9.5, 2.18, 2.18]}Transmission lines data B1=µ{X2, [0.64, 0.64, 19.0, 19.0]}Rated MVA =20 f=60Hz Rated kV=13.8 B2=µ{X2, [-0.5, -0.5, 0.64, 0.64]}X0L=0.0534ohms /Km R0L=0.0414ohms /Km C1=µ{X3, [0.16, 0.16, 0.6, 0.6]}X1L=0.0178ohms /Km R1L=0.01384ohms /Km C2=µ{X3, [-0.05, -0.05, 0.16, 0.16]}.X0CL=5.1nF /Km X1CL=17nF /Km line length=20Km The fuzzy MFs generated from the DT classification boundaries are rectangular in nature.Rated MVA =10 f=60Hz 54poles Yn Rated kV=13.8 But to further add fuzziness to the membershipInertia constant H =3.0sec. functions, the rectangular boundaries are skewed to a certain extent by heuristic tuning. The coordinates ofH).Decision tree the trapezoidal fuzzy MFs are decided after testing on Decision tree learning is a method several values around the initial values resulting fromcommonly used in data mining. The goal is to create DT [10]. Thus, the final fuzzy MFs area model that predicts the value of a target variablebased on several input variables. The basic ideainvolved in any multistage approach is to break up a Rule base R1: If X1 is A1 and X2 is B1 then Class-1complex decision into a union of several simpler R2: If X1 is A2 and X2 is B2 then Class-1decisions, hoping the final solution obtained this way R3: If X1 is A2 and X2 is B1 and X3 is C1 then Class-1would resemble the intended desired solution [17]. R4: If X1 is A2 and X2 is B1 and X3 is C2 then Class-0 Class-1 means islanding and class-0 means non-Decision tree chart islanding. Simulated Fig(b) Block diagram usingMATLAB/Simulink from that extract 𝚫f/𝚫t, 𝚫p/𝚫t, fthese terms and verified with below flow-chart .Thisflow-chart explain given network is islanding/non-islanding. 988 | P a g e
  4. 4. V.Veera Nagireddy, Dr. D.V. Ashok Kumar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 Vol. 2, Issue 5, September- October 2012, pp.986-990 Table1: Different test conditions results Event X1 X2 X3 Actualconditio no n E11 5.500 -0.20 0.234 Islanding E12 5.330 -0.30 0.020 Islanding E21 0.001 3.00 0.005 Non-islanding E23 0.001 5.40 0.005 Non-islanding E31 4.900 -0.01 0.005 Islanding E33 5.000 -0.10 0.005 Islanding E41 5.800 5.00 0.01 Islanding The table1 shows conventional results of our case study system. With fuzzy results are given below. The system study state reaches around 300ms. After 300ms system (fuzzy output results)output value reaches 0.5, it is islanding condition other ways non-Fig.c: Fuzzy inference system. islanding condition.J).SimulationResults 1) Tripping of the circuit breaker CB-1at 1sec Modeling of case study network (figure a) to simulate with PCC bus loads is P=0.5pu,using Mat lab was done. Simulations were carried out Q=0.175pu.considering the following critical conditions atdifferent PCC bus loads:  Condition-1: Tripping of the circuit breaker CB-1 to simulate the condition of islanding of the DG with the PCC bus loads.  Condition-2: Tripping of the circuit breaker CB-2 (isolating the PCC bus loads) to simulate disturbances on the DG.  Condition-3: Tripping of the circuit breaker Fig. d: E11 Crisp values plot. CB-3 to simulate the islanding of the DG The fig. d shows the crisp value plot, crisp value without the PCC- bus loads. reaches 0.5 after 300msec, islanding is occurred.  Condition-4: Three-phase fault on the 2) Tripping of the circuit breaker CB-1at1sec GEN_BUS with instantaneous (1 cycle) to simulate with PCC bus loads is P=0.3pu, fault-clearing time by the CB-1 which, in Q=0.105pu. turn, causes islanding of the DG.The PCC bus loads are1. PCC-bus loading P=0.5pu, Q=0.175pu.2. PCC-bus loading P=0.3pu, Q=0.105pu.3. PCC-bus loading P=0.625pu, Q=0.22PU.Table 1 provides the test results for differentconditions of inputs X1, X2, and X3for islandingdetection. In 1 table E11 means, E indicated event,first 1 indicated fault, second 1 indicated PCC bus Fig. e: E12 Crisp values plot.load condition. The fig. e shows the crisp value plot, crisp value reaches 0.5 after 300msec, islanding is occurred. 3) Tripping of the circuit breaker CB-2(1sec) to simulate with PCC bus loads is P=0.5pu, Q=0.175pu. 989 | P a g e
  5. 5. V.Veera Nagireddy, Dr. D.V. Ashok Kumar / International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 Vol. 2, Issue 5, September- October 2012, pp.986-990 [6] Ward Bower and Michael Ropp, “Evaluation of islanding detection methods forphotovoltaic utility-interactive power systems,” Report IEA PVPS Task 5 IEA PVPS T5-09: 2002, Sandia National Laboratories Photovoltaic Systems Research and Development, March 2002. [7] B. and O Kane, P. J. , “Identifying loss of mains in electricity distribution system, “ Patent Number: GB2317759, 1998. [8] J.W. “Loss of Mains Protection”, ERAFig. f: E21 Crisp values plot. Conference on Circuit Protection for The fig. 5.3 shows the crisp value plot, crisp Industrial and Commercial Installations,value not reaches 0.5 after 300msec, islanding is not 1990, London.occurred. [9] M., Begovic, M. and Rohatgi, "Analysis and performance assessment of the activeCONCLUSION frequency drift method of islanding This thesis describes and compares different prevention", IEEE Tran. Energy Conversion,islanding detection techniques. Fast and accurate vol. 14, no. 3, pp. 810-816, 1999.detection of islanding is one of the major challenges [10] “IEEE Standard for Interconnectingin today‟s power system with many distribution Distributed Resources with Electric Powersystems already having significant penetration of DG Systems,” Approved 12 June 2003,as there are few issues yet to be resolved with Reaffirmed 25 September 2008.islanding. Islanding detection is also important as [11] WalmirFreitas, M. Affonso and Zhenyuislanding operation of distributed system is seen a Huang, “Comparative Analysis Betweenviable option in the future to improve the reliability ROCOF andVector Surge Relays forand quality of the supply. A fuzzy rule-based passive DistributedGeneration Applications,”IEEEislanding detection is implemented in this project. Tran. power delivery, vol. 20, no. 2, aprilThe initial classification model is developed using 2005.decision tree (DT) which is a crisp algorithm. Thisalgorithm is transformed into a fuzzy rule base bydeveloping fuzzy membership functions (MFs) fromthe DT classification boundaries.The islanding detecting time obtained fromsimulation results is 100ms to 150ms, it‟s followedby IEEE Std. 1547-2003 [10].REFERENCES [1] K.S.Sidhu, “Non-conventional energy resources,” Punjab state electric board , PEC Campus, Chandigar,2007 [2] Philip P. Barker Robert W. de Mello, “Determining the Impact of Distributed Generation on Power Systems: Part1 - Radial Distribution Systems,” IEEE Power Technologies, Inc. 2000. [3] R. A. Walling, and N. W. Miller, “Distributed generation islanding implications on power system dynamic performance,” IEEE Power Engineering Society Summer Meeting, vol.1, pp. 92-96, 2002. [4] A. Greenwood, “Electrical Transients in Power Systems,” New York: Wiley, 1971, pp. 83. [5] Vivek Menon and M. HashemNehrir, “A Hybrid Islanding Detection Technique UsingVoltage Unbalance and Frequency Set Point,” IEEE transactions on power systems, vol. 22, no. 1, february 2007. 990 | P a g e