Comparison of Possible Future Passive Islanding Detection Methods

Transcription

1 International Review of Electrical Engineering (I.R.E.E.), Vol. 8, N. 3 ISSN May June 2013 Comparison of Possible Future Passive Islanding Detection Methods H. Laaksonen 1 Abstract Due to growing distributed generation penetration level at all voltage levels in the future Smart Grids there is an increased need for high performance islanding detection (antiislanding protection) methods. A large non-detection zone near a power balance situation and unwanted distributed generation tripping due to other network events have been the major drawbacks of traditional, passive local islanding detection methods. Usually these traditional methods have also been dependent on the type of the distributed generation unit. In this paper the performance of three different possible future, local measurements based, passive islanding detection methods are compared. Performance comparison of these islanding detection methods is done in terms of size of non-detection zone, detection speed and possible detection speed dependency from power unbalance before islanding as well as possibility to mal-operate during frequency fluctuations and faults in the utility grid. Copyright 2013 Praise Worthy Prize S.r.l. - All rights reserved. Keywords: Distributed Generation, Islanding Detection, Loss-of-mains protection, Antiislanding, Islanding, Smart Grids I. Introduction Active control and management of distribution networks, including control of distributed energy resources (DER), will be in key role in future Smart Grids to fulfill increasing energy efficiency and reliability requirements [1]. With active network management the capacity utilization of lines can be improved, large voltage deviations can be avoided, system losses and interruptions can be minimized. Active network management and adaptive protection requires more information (measurements) from different points of distribution network as well as utilization of fast and cost-efficient communication technologies and further development of standardization (IEC related standards) [2, 3]. Information about distribution network status for control and monitoring purposes will be obtained in the future increasingly from sensors across the network through high-speed wireless 4G networks and optical fibers which can also be integrated in power cables [4]. One key functionality required from future distributed generation (DG) interconnection IEDs (intelligent electrical devices) in Smart Grids is reliable detection of both unintentional and intentional islanding [5]. Techniques proposed for islanding detection have been generally divided into two categories: communication based and local detection based (active and passive) methods. Proposed local detection methods have traditionally been dependent from DG type. Only communication based islanding detection schemes can be generally applied for every type of DG units. Two essential benefits of communication based islanding detection are lack of non-detection zone (NDZ) near power balance situation and lack of unwanted DG trips (nuisance tripping) due to other network events (e.g. utility grid fault, parallel feeder fault, capacitor connection). These have been the major challenges with traditional, passive local islanding detection methods like for example frequency (f), df/dt, vector shift (VS) or voltage (U) based methods. However, the communication-based islanding detection requires high-speed operation. This high-speed operation can be achieved e.g. by utilizing the IEC GOOSE message-based tripping signals [6] from MV feeder IEDs (communication e.g. through optical fibre or with wireless technologies such as LTE or WiMAX) to DG interconnection IEDs. The main challenges of the communication-based transfer trip islanding detection schemes are the availability and cost of high-speed communication as well as the flexibility to network topology changes. The flexibility of the communicationbased islanding detection schemes for network topology changes could be improved by pre-configuring the different IEC GOOSE signal-based transfer trip schemes and activating them after topology changes e.g. by command from grid automation controller. Also more flexible centralized, communication-based islanding detection schemes have been proposed in [7] and [8]. Active local methods have been questionable, because they introduce disturbances into the distribution network, which may become a serious problem when the number of DG units increases. In references [9] [25] different Copyright 2013 Praise Worthy Prize S.r.l. - All rights reserved

2 passive local islanding detection methods have been proposed. Traditional passive islanding detection methods work based on the assumption that in almost all circumstances islanding will result in a measurable variation in voltage, frequency and/or power. Passive local islanding detection methods are based on monitoring one or more system parameters locally and they make their trip decision without directly interacting with system operation. The traditional passive methods cannot guarantee a totally selective operation with other network events which may cause nuisance tripping of DGs like capacitor switching or connection of parallel transformer at HV/MV substation. Also different hybrid islanding detection methods which try to minimize the drawbacks of traditional passive and active methods have been proposed e.g. in [26] and [27]. In this paper the performance of three different possible future, local measurements based passive, islanding detection methods will be compared. Performance of these islanding methods is compared in terms of size of NDZ, detection speed and possible dependency of detection speed from power unbalance before islanding as well as possibility to mal-operate during frequency fluctuations and faults in utility grid. In Section II some requirements for future islanding detection method will be viewed. Section III of the paper presents shortly the compared islanding detection method and in Section IV the simulation results will be shown. In Section V some discussion and conclusions will be stated. II. Requirements for Future Islanding Detection Methods In the future, if the amount of DG units in distribution networks increases as expected, also the risk of power balance situations and hence the risk of possible operation in the NDZ of the traditional passive islanding detection methods will increase. Although it is likely that in the future grid codes intended island operation possibility will be more commonly allowed to realize the possible supply reliability benefits of DER, there still is a need to reliably detect the islanding situation to make correct operations, e.g. change the setting group of the DG interconnection IED or change the control principles and parameters of the DG unit. To enable and guarantee the stability after the transition to island operation, islanding and the change of DG unit control parameters or principles must be performed very fast, for instance, in less than milliseconds (ms). Therefore, also a very rapid islanding detection is required. In new grid codes, traditional islanding detection parameters such as f, U and df/dt will be increasingly used to define the DG unit fault-ride-through (FRT) requirements. Therefore, the use of these parameters for a reliable and selective, e.g. with auto-reclosing schemes, islanding detection will become even more difficult than today. Communication based transfer trip and centralized solutions for islanding detection will probably become more common in the future. In ENTSO-E grid code RfG [28] it has been also stated that islanding detection should not be based only on network operator s switchgear position signals. In addition, if communication fails or is not high-speed enough also reliable islanding detection scheme based on local detection is still required in future, but it has to overcome the major drawbacks of traditional islanding detection methods. Also if for example traditional frequency based method is still used for islanding detection in the future, then one possibility to avoid nuisance tripping and coordinate with grid code FRT requirements is to try differentiating local faults in MV networks from external disturbances coming from upstream voltage levels (e.g. 110 kv). For instance in order to use narrow frequency band for islanding detection in DG interconnection IED it must be activated based either on external signal (communication) from network operator etc. or based on local positive, negative and zero sequence voltages measurements of IED which indicate that there is fault at the corresponding MV network. However, there will still be situations when this method cannot detect islanding e.g. due to opening of a MV feeder circuit breaker for maintenance work and also the frequency based islanding detection method still has NDZ near power balance situations. III. Compared Islanding Detection Methods III.1. Method 1 The first method is based on detection of change in voltage positive sequence phase angle U pos_angle (Fig. 1). Fig. 1. Principle of the proposed voltage positive sequence phase angle based algorithm for islanding detection (method 1)

3 Negative sequence and zero sequence voltage components are usually utilized to detect unbalance and/or fault situations in networks. Therefore, to minimize the unwanted effect of other network events (i.e. other than islanding) to the islanding detection and possible nuisance tripping, only values based on positive sequence voltage have been used for islanding detection with method 1. Islanding detection based on change in islanding detection parameter U pos_angle is started only after needed (relatively large) change in the used value has been detected (Fig. 1) when compared to 300 ms earlier value. III.2. Method 2 The second, multi-criteria based, method relies on simultaneous detection of change in voltage total harmonic distortion U THD_15 of all phase voltages (15 harmonics) and voltage unbalance VU calculated from phase-to-phase voltages as described in Fig. 2. In [29] it was stated that this multi-criteria based islanding detection algorithm is not dependent from DG unit type, it is able to detect very fast and selectively islanding situations in a perfect power balance without NDZ and no nuisance tripping is likely to occur due to other network events. The use of these parameters together makes the islanding detection more reliable by confirming the islanding detection of another parameter and by improving the selectivity, for example in case of connection of the capacitor bank, single-phase load or in case of unsymmetrical voltage dips or unsymmetrical faults on other MV feeders. III.3. Method 3 The third method is based on discrete wavelet transform (DWT). DWT is a signal-processing tool which can be applied when time-varying harmonics must be evaluated and, as in the case of the detection of the islanding condition, time localization is required. In recent years, the passive and hybrid islanding detection methods based on the utilization of the wavelet transformation have been suggested in e.g. in references [30] [35]. In terms of NDZ and nuisance tripping risk it was though that DWT based islanding detection method could achieve similar performance as method 2. In simulations many different mother wavelets with different orders were studied and Daubechie s of order 8 (db8) was found to be one of the most promising ones. Therefore, in this paper DWT based islanding detection method was based on change in db8 coefficient level D4 i.e. Db8(D4) ave_4khz calculated from phase-to-phase voltages with 4 khz sampling frequency (Fig. 3). This means that frequency band for level D4 was in this case Hz. Fig. 3. Principle of the proposed phase-to-phase voltages and DWT based algorithm for islanding detection with sampling frequency 4 khz (method 3) Fig. 2. Principle of the multi-criteria-based islanding detection algorithm (method 2) [29] Islanding detection based on change in islanding detection parameters VU& U THD_15 is started only after needed (relatively large) change in the used values has been detected (Fig. 2) when compared to 30 ms earlier value. Islanding detection based on change in islanding detection parameter Db8(D4) ave_4khz is started only after needed (relatively large) change in the used value has been detected (Fig. 3) when compared to 60 ms earlier value. IV. Simulation Results In this section simulation results from two different cases are presented. In case 1 there were two directly connected rotating type of DG units and in case 2 two converter connected DG units.

4 IV.1. Case 1 Two Directly Connected Rotating DG Units In following, PSCAD simulation results for case 1a are shown. The used simulation model and simulation sequence (cases 1a & 1b) are presented in Fig. 4 and Table I respectively. In case 1a, with small power unbalance, there were only synchronous generator (P = 800 kw, Q = -100 kvar) and induction generator (P = 788 kw, Q = -123 kvar) based DG units and power unbalance before islanding was P = +21 kw, Q = +31 kvar. Respectively in case 1b with larger power unbalance before islanding, the active (P) and reactive power (Q) of synchronous generator was P = 850 kw, Q = 150 kvar) and induction generator (P = 788 kw, Q = -123 kvar) based DG units and power unbalance before islanding was P = +67 kw, Q = +291 kvar. 1) Case 1a Small Power Unbalance before Islanding In Fig. 5 frequency and voltage behavior during simulation sequence in case 1a (Table I) are shown. Correspondingly in Fig. 6, 7 and 8 simulation results for different islanding detection methods 1, 2 and 3 are presented. Fig. 4. Studied MV network in cases 1a & 1b TABLE I SIMULATION SEQUENCE IN CASES 1A & 1B (FIG. 4) Time (s) Event type Connection of large 1-phase load in phase C 3.5 Phase shift (-12 degrees) due to topology change in HV network 4.0 Capacitor switching at HV/MV substation, 0.8 MVAr 1-phase-to-earth (C-G) fault at the end of the same MV feeder (fault resistance Rf = 50 Ω) at t= s connection of parallel resistance at the centralized compensation unit at HV/MV substation Frequency fluctuation due to system wide disturbance phase 60 % voltage dip in HV network Step change of centralized compensation (momentary) 2-phase-to-earth (B-C-G) fault at the end of the same MV feeder (Rf = 1500 Ω) at t= s connection of parallel resistance at the centralized compensation unit at HV/MV substation phase (C-A) fault at the end of the adjacent MV feeder (Rf = 5 Ω) 20.5 Islanding, change of diesel generator control mode after islanding with 200 ms time delay 22.5 Re-connection back to utility grid Fig. 5. a) Frequency and b) voltage behavior during simulation sequence in case 1a Fig. 6. Voltage positive sequence phase angle (method 1) behavior during simulation sequence in case 1a

5 Fig. 7. Voltage THD and unbalance (method 2) behavior during simulation sequence in case 1a Fig. 10. Voltage positive sequence phase angle (method 1) behavior after islanding in case 1b Fig. 8. DWT coefficient level db8(d4) (method 3) behavior during simulation sequence in case 1a 1) Case 1b Larger Power Unbalance before Islanding In Fig. 9 frequency and voltage behavior after islanding in case 1b are shown. Correspondingly in Fig. 10 simulation results for islanding detection method 1 are presented. Fig. 11. Voltage THD and unbalance (method 2) behavior during simulation sequence in case 1b Fig. 12. DWT coefficient level db8(d4) (method 3) behavior during simulation sequence in case 1b Fig. 9. a) Frequency and b) voltage behavior after islanding in case 1b IV.2. Case 2 Two Converter Connected DG Units In this section PSCAD simulation results for case 2 are presented. The used simulation model and simulation sequence (case 2) are presented in Fig. 13 and Table II respectively. In case 2 (Fig. 13) there were only converter connected DG unit in MV network (P = 1226 kw, Q = -100 kvar) and converter connected energy storage unit at MV/LV distribution substation (P = 55 kw, Q = 12 kvar). Power unbalance before transition to island operation was in case 2 only P = +3 kw, Q = +1 kvar.

6 Fig. 14. a) Frequency and b) voltage behavior during simulation sequence in case 2 Fig. 13. Studied MV network in case 2 TABLE II SIMULATION SEQUENCE IN CASE 2 (FIG. 13) Time (s) Event type Islanding Frequency fluctuations (in steps) due to system wide disturbances 2.3 Capacitor switching at HV/MV substation, 0.8 MVAr 2.6 Phase shift (10 degrees) due to topology change in HV network phase-to-earth (A-G) fault 150 ms at the end of adjacent MV feeder (Rf = 750 Ω) phase-to-earth (B-G) fault 150 ms at the beginning of adjacent MV feeder (Rf = 100 Ω) phase-to-earth (A-G) fault 150 ms in the middle of the same MV feeder (Rf = 1000 Ω) phase (A-B) fault 150 ms in the middle of adjacent MV feeder (Rf = 5 Ω) phase (A-B-C) fault 150 ms at the beginning of adjacent MV feeder (Rf = 1 Ω) phase 30 % voltage dip in HV network phase-to-earth (A-B-G) fault 150 ms at the end of the same MV feeder (Rf = 50 Ω) In Fig. 14 frequency and voltage behavior during simulation sequence in case 2 (Table II) are presented and in Fig. 15, 16 and 17 simulation results for different islanding detection methods 1, 2 and 3 are presented respectively. Fig. 15. Voltage positive sequence phase angle (method 1) behavior during simulation sequence in case 2 Fig. 16. Voltage THD and unbalance (method 2) behavior during simulation sequence in case 2 Fig. 17. DWT coefficient level db8(d4) (method 3) behavior during simulation sequence in case 2

7 V. Discussion and Conclusions Islanding detection operate times of compared islanding detection methods from studied cases with initially proposed settings are summarized and presented in Table III. Regarding to operate times of the compared methods in Table III, it should be noted that islanding detection is based on exceeding the start value of the change parameter i.e. U pos_angle (method 1), VU& U THD_15 (method 2) and Db8(D4) ave_4khz (method 3) and therefore operate time of the particular method depends very much from the chosen start value and time delay as well as from time difference used to detect change in the islanding detection parameter i.e. which was 300 ms with method 1, 30 ms with method 2 and 60 ms with method 3. From operate times shown in Table III (cases 1a and 1b) one can see how detection speed (operate time) of method 1 increases considerably when power unbalance before islanding is small (case 1a) when compared to case 1b with larger power unbalance. In corresponding cases 1a and 1b the operation times with methods 2 and 3 are quite similar (Table III) and not affected by the level of power unbalance. TABLE III ISLANDING DETECTION / OPERATE TIME IN STUDIED CASES WITH COMPARED ISLANDING DETECTION METHODS Islanding Time Operate Time in Threshold / Start Detection Method *) Value ***) Delay Case 1a, 1b and 2 [ms] **) [ms] Method 1 ± , 165 and 264 Method % (VU) & +0.5 % (UTHD_15) 50 89, 82 and 59 Method kv , 153 and 140 *) Islanding detection wakening (timer with chosen time delay**) is started) after change parameter exceeds start value***) when compared to 300 ms (method 1), 30 ms (method 2) or 60 ms (method 3) earlier value It can be also mentioned that, with method 3 and with chosen start value (Table III) in case 1b, if time delay would be very long (in this case over 1000 ms) the islanding detection would not operate due to fluctuation in measured parameter (Fig. 12). This fluctuation takes place due to oscillations in voltage and frequency after islanding in larger power unbalance (Fig. 9). This possibility for non-detecting islanding with method 3 could be totally avoided (i.e. also with time delay over 2000 ms) with lower start value setting (Table III) e.g kv in this case 1b (Fig. 12). However, if high-speed islanding detection in ms is considered this is not a problem. In general, all compared methods are sensitive to maloperate during large voltage dips due to faults if sensitive settings with short time delays are used to achieve fast detection speed. Therefore, to enable fast and reliable islanding detection, it is reasonable to use simultaneous fault (undervoltage) detection (blocking) as part of the islanding verification logic to prevent possible nuisance tripping (Fig. 2). This can be seen for example from simulation results of method 1 in case 2 (Fig. 15). However, from case 1 simulation results it could be also seen that voltage positive sequence phase angle (method 1, Fig. 6) like other voltage phase angle based islanding detection schemes proposed e.g. in [36], [37] are very sensitive to maloperate due to frequency fluctuations in the utility grid. Frequency fluctuations cannot be blocked in similar manner as undervoltage situations with method 1. Sensitivity of method 1 to operate during frequency fluctuations can be only reduced by usage of higher setting value, but then as a consequence the size of NDZ as well as detection speed of method 1will be increased. Instead, DWT coefficient level change based method 3 (Fig. 8) is not affected by frequency fluctuations like methods 1 and 2 (especially method 1). But as presented in Fig. 18, if frequency adaptive measurement is used with multi-criteria based method 2 [29], then the negative effect of frequency fluctuations between t = s can be removed. The main difference between methods 2 and 3 was found to be in the detection speed (Table III). With DWT based method 3 it was not possible to detect islanding as fast as with method 2, because time delay of D4 level rising after islanding was found to be ms (Fig. 8 and 13). In addition, from case 1 and 2 simulation results in Section III it could be seen that with methods 2 and 3 simultaneously detected undervoltage needs to be used as part of the islanding logic to prevent the false islanding detection due to removing 3-phase faults. Fig. 18. Effect of frequency adaptive measurement on voltage THD and unbalance (method 2) behavior during simulation sequence in case 1a (Fig. 7) In Fig. 19 dependencies between a) setting value (method 1) and b) intentional time delay (methods 2 and 3) with size of NDZ, islanding detection speed and sensitivity to frequency fluctuations are concluded. From Fig. 19a) it can be seen how higher setting value (e.g. to avoid operation due to frequency fluctuations) results also in larger NDZ and longer operation time (i.e. slower detection speed) with method 1. On the other hand, with methods 2 and 3 the size of NDZ and sensitivity to frequency fluctuations are not dependent on the setting

8 value (Fig. 19b). Nevertheless, minimum intentional time delay should be used with methods 2 and 3 to directly avoid islanding detection maloperation after 2- and 3- phase short circuit faults. phase angle based methods like method 1 and methods proposed in [36], [37] will not be sufficient and reliable enough. Islanding detection method 2 based on simultaneous detection of changes on U THD_15 of all phase voltages and voltage unbalance VU was found to be best, especially if frequency adaptive measurement of phaseto-phase voltages is used, from the three compared methods size of non-detection zone, detection speed and possible dependency of detection speed from power unbalance before islanding as well as possibility to maloperate during frequency fluctuations and faults in the utility grid. Almost equal performance could be achieved with DWT based method 3, but if very fast islanding detection is required (e.g. in less than ms) the usage of Db8(D4) with 4 khz sampling frequency may be not be adequate due to time delay in D4 level rising after islanding. Naturally with usage of higher sampling frequency, e.g khz, this time delay of DWT based method 3 could be reduced but simultaneously it would require more from the technology of IEDs than method 2. Acknowledgements This work was supported by Smart Grids and Energy Market (SGEM) research program of CLEEN Ltd, the Strategic centre for science, technology and innovation of the Finnish energy and environment cluster. Fig. 19. Dependencies between a) setting value (method 1) and b) intentional time delay (methods 2 and 3) with size of NDZ, islanding detection speed and sensitivity to frequency fluctuations In addition to FRT requirements, many new grid codes require utility grid voltage and frequency stability supporting functionalities increasingly also from smaller distribution network connected DG units. For example automatic voltage control by reactive power/voltage (Q/U) regulation and frequency control by active power/frequency (P/f) regulation can be required from DG units. Also frequency dependent demand response (DR) participation by automatic load shedding has been proposed to support utility grid. However, from islanding detection point of view these automatic voltage and frequency control functionalities required from DG units and DR are challenging, because in case of unintentional islanding these functionalities will also stabilize voltage and frequency in the islanded part of network. Therefore, in the future islanding detection based on traditional passive methods like f, U and df/dt as well as voltage References [1] H. Laaksonen, Technical Solutions for Low-Voltage Microgrid Concept, Ph.D. dissertation, Faculty of Technology, Dept. Elect. and Energy Eng., Univ. Vaasa, Vaasa, Finland, [Online]. Available: [2] H. Laaksonen, K. Kauhaniemi, Smart Protection Concept for LV Microgrid, International Review of Electrical Engineering (IREE), vol. 5, March-April 2010, pp [3] H. Laaksonen, Protection Principles for Future Microgrids, IEEE Transactions on Power Electronics, vol. 25, December 2010, pp [4] BDEW/ZVEI: Smart Grids in Germany: Fields of action for distribution system operators on the way to Smart Grids, June 2012, [Online] Available: art-grids-in-germany---fields-of-action-for-distribution-systemoperators-on-the-way-to-smart-grids.aspx [5] H. Laaksonen, F. Suomi, New Functionalities and Features of IEDs to Realize Active Control and Protection of Smart Grids, The 22 nd International Conference on Electricity Distribution ~CIRED 2013~, June 10-13, 2013, Stockholm, Sweden. [6] O. Rintamäki, K. Kauhaniemi, Applying modern communication technology to loss-of-mains protection, The 20 th International Conference on Electricity Distribution ~CIRED 2009~, June 8-11, 2009, Prague, Czech. [7] F. Coffele, P. Moore, C. Booth, A. Dyśko, G. Burt, Centralised Loss of Mains protection using IEC-61850, The 10 th IET International Conference on Developments in Power System Protection ~DPSP 2010~, March 29-April 1, 2010, Manchester, United Kingdom. [8] A. Timbus, A. Oudalov, C.N.M. Ho, Islanding detection in smart grids, The Energy Conversion Congress and Exposition, September 12-16, 2010, Atlanta, Georgia, USA.

9 [9] C. B. Cooper, Standby generation problems and prospective gains from parallel running, The International Power System Protection Conference, September, 1989, Singapore. [10] M. A. Redfern, O. Usta, G. Fielding, Protection against loss of utility grid supply for a dispersed storage and generation unit, IEEE Transactions on Power Delivery, vol. 8, July 1993, pp [11] O. Usta, M. A. Redfern, J. I. Barrett, Protection of dispersed storage and generation units against islanding, The 7 th Mediterranean Electrotechnical Conference, April 12-14, 1994, Antalya, Turkey. [12] M. A. Redfern, J. I. Barrett, O. Usta, A new loss of grid protection based on power measurements, The 6 th International Conference on Developments in Power System Protection ~DPSP 1997~, March 25-27, 1997, Johannesburg, South Africa. [13] S. K. Salman, D. J. King, G. Weller, New loss of mains detection algorithm for embedded generation using rate of change of voltage and changes in power factors, The 7 th International Conference on Developments in Power System Protection ~DPSP 2001~, April 9-12, 2001, Amsterdam, Netherlands. [14] S. K. Salman, Detection of embedded generator islanding condition using elliptical trajectory technique, The 6 th International Conference on Developments in Power System Protection ~DPSP 1997~, March 25-27, 1997, Johannesburg, South Africa. [15] F. Pai, S. Huang, A detection algorithm of islanding prevention of dispersed consumer-owned storage and generating units, IEEE Transactions on Energy Conversion, vol. 16, December 2001, pp [16] X. Ding, P. A. Crossley, D. J. Morrow, Islanding Detection for Distributed Generation, Journal of Electrical Engineering & Technology, vol. 2, March 2007, pp [17] C. F. Ten, P. A. Crossley, Evaluation of Rocof Relay Performances on Networks with Distributed Generation, The 9 th International Conference on Developments in Power System Protection ~DPSP 2008~, March 17-20, 2008, Glasgow, Scotland, United Kingdom. [18] S. I. Jang, K. H. Kim, A new islanding detection algorithm for distributed generations interconnected with utility networks, The 8 th International Conference on Developments in Power System Protection ~DPSP 2004~, April 5-8, 2004, Amsterdam, Netherlands. [19] S. I. Jang, K. H. Kim, An islanding detection method for distributed generations using voltage unbalance and total harmonic distortion of current, IEEE Transactions on Power Delivery, vol. 19, April 2004, pp [20] P. Mahat, Z. Chen, B. Bak-Jensen, Review of Islanding Detection Methods for Distributed Generation, The Third International Conference on Electric Utility Deregulation and Restructuring and Power Technologies ~DRPT 2008~, April 6-9, 2008, Nanjing, China. [21] V. Menon, M. H. Nehrir, A hybrid islanding detection technique using voltage unbalance and frequency set point, IEEE Transactions on Power Systems, vol. 22, February 2007, pp [22] S. I. Jang, K. H. Kim, Development of a logical rule-based islanding detection method for distributed resources, The IEEE Power Engineering Society Winter Meeting, January 27-31, 2002, New York, NY, USA. [23] M. Hanif, M. Basu, K. Gaughan, A Discussion of Anti-islanding Protection Schemes Incorporated in a Inverter Based DG, The 10 th International Conference on Environment and Electrical Engineering ~EEEIC 2011~, May 8-11, 2011, Rome, Italy. [24] H. Kabayashi, K. Takigawa, E. Hashimato, Method for preventing islanding phenomenon on utility grid with a number of small scale PV systems, The Second IEEE Photovoltaic Specialists Conference, October 7-11, 1991, Las Vegas, NV, USA. [25] J. Saeed, Loi Lei Lai, DG Operation Detection Methods in Combination of Harmonics Protection Schemes, The IEEE PES Innovative Smart Grid Technologies EUROPE ~ISGT Europe 2011~, December 5-7, 2011, Manchester, UK. [26] P. Mahat, Zhe Chen, B. Bak-Jensen, A Hybrid Islanding Detection Technique Using Average Rate of Voltage Change and Real Power Shift, IEEE Transactions on Power Delivery, vol. 24, April 2009, pp [27] H. H. Zeineldin, J. L. Kirtley, A Simple Technique for Islanding Detection With Negligible Nondetection Zone, IEEE Transactions on Power Delivery, vol. 24, April 2009, pp [28] ENTSO-E, ENTSO-E Network Code for Requirements for Grid Connection Applicable to all Generators, 26 June 2012, [Online] Available: [29] H. Laaksonen, New Multi-criteria-based Algorithm for Islanding Detection in Smart Grids, The IEEE PES ISGT Europe 2012 ~ISGT Europe 2012~, October 14-17, 2012, Berlin, Germany. [30] C.T. Hsieh, J.M. Lin, S.J. Huang, Enhancement of islandingdetection of distributed generation systems via wavelet transformbased approaches, International Journal of Electrical Power & Energy Systems, vol. 30, December 2008, pp [31] A. Pigazo, M. Liserre, R. A. Mastromauro, V. M. Moreno, A. Dell Aquila, Wavelet-Based Islanding Detection in Grid- Connected PV Systems, IEEE Transactions on Industrial Electronics, vol. 56, November 2009, pp [32] M. Hanif, U.D Dwivedi, M. Basu, K. Gaughan, Wavelet Based Islanding Detection of DC-AC Inverter Interfaced DG Systems, The 45 th International Universities' Power Engineering Conference ~UPEC 2010~, August 31-September 3, 2010, Cardiff, Wales, United Kingdom. [33] N. Lidula, N. Perera and A. D. Rajapakse, "Investigation of a fast islanding detection methodology using transient signals," The Power & Energy Society General Meeting ~PESGM 2009~, July 26-30, 2009, Calgary, Canada. [34] S. R. Samantaray, T. M. Pujhari, B. D. Subudhi, A new approach to Islanding detection in Distributed Generations, The 3 rd International Conference on Power Systems, December 27-29, 2009, Kharagpur, India. [35] S. Shrivastava, S. Jain, R. K. Nema, Wavelet Entropy: Application in Islanding Detection, WSEAS Transactions on Power Systems, vol. 7, July 2012, pp [36] A. Dyśko, G. Burt, R. Bugdal, Novel Protection Methods for Active Distribution Networks with High Penetrations of Distributed Generation, Year II Report, DTI Centre for Distributed Generation and Sustainable Electrical Energy, June [37] C. An, G. Millar, G. J. Lloyd, A. Dyśko, G.M. Burt, F. Malone, Experience with accumulated phase angle drift measurement for islanding detection, The 11 th International Conference on Developments in Power Systems Protection ~DPSP 2012~, April 23-26, 2012, Birmingham, UK. Biography 1 ABB Oy, Medium Voltage Products, Vaasa, Finland. H. Laaksonen was born in Vaasa, Finland, on November 22, He received his MSc degree (2004) in Electrical Power Engineering from Tampere University of Technology and PhD degree (2011) in Electrical Engineering from University of Vaasa. His employment experience includes working as a research scientist in University of Vaasa, VTT Technical Research Centre of Finland and Institute of Power Engineering department at Tampere University of Technology. Currently he works with ABB Oy, Medium Voltage Products in Vaasa, Finland. His fields of interest are protection of Smart Grids, integration and active management of distributed energy resources in smart distribution networks and development of new functionalities and algorithms for future Smart Grid concepts (e.g. microgrids).