Performance of Multiple Passive Islanding Detection ... - ESRGroups

Aimie Nadia Ab Salam1, Hasmaini Mohamad1, Nofri Yenita Dahlan1, Safdar Raza2

J. Electrical Systems 13-3 (2017): 568-578 Regular paper Performance of Multiple Passive Islanding Detection Technique for Synchronous Type of DG

JES Journal of Electrical Systems

Placing DG closer to the load in the system can improve the grid reliability and power quality, however it can lead to some technical issues such as islanding which requires attention from utility and Independent Power Producer (IPP). One of the important issues in power system distribution is islanding detection for protecting the DG when islanding operation occurred. The detection technique is employed so that the islanding event can be detected within the time frame. The main reason of detecting islanding situation at distribution level is to monitor the DG output or system parameters thus can determine the occurrence of islanding situation from the change of these parameters. Therefore, this paper presents the performance of multiple parameters of passive islanding detection technique in distribution system connected with synchronous type of DG. Sixteen passive parameters including rate of change of frequency (df/dt), rate of change of voltage (dv/dt), rate of change of active power (dp/dt), and rate of change of reactive power (dq/dt) are evaluated. The sensitivity analysis of these parameters are determined via simulation in terms of variation of loads, fault events and load switching on various events that are islanding and non-islanding events. The analysis indicates that each parameter displays a different pattern of sensitivity towards islanding event.

Keywords: Distributed generation; mini hydro generation; islanding detection; passive parameters. Article history: Received 19 March 2016, Accepted 16 June 2017

1. Introduction The presence of distributed generation (DG) can improve power quality, avoids capacity upgrades at power transmission and distribution, enhances voltage profile and also reduces the power losses in distribution system [1]. Despite of that, the connection of DG in distribution system would lead to the occurrence of islanding that could present a harmful effect to the DG. Islanding is a condition where DG still energizes a part of the system although grid is disconnected due to the loss of main. It is categorized into two: intentional and unintentional islanding. Intentional or so called planned islanding is designed for the DG to continuous supply power to the load particularly during maintenance services. Meanwhile, unintentional islanding is established when the utility losses control of the voltage and the frequency during islanding condition. To minimize the potential damage cause by islanding, IEEE Std 1547-2003 has stipulated that the DG should immediately disconnected from the system within 2 seconds following loss of main [2]. Thus, islanding detection technique is critically required to quickly detect the formation of islanding and subsequently stop the operation of DG to avoid the negative effect to the network, DG itself, power quality and equipment. Generally, islanding detection can be classified into two i.e. remote and local techniques. Local technique includes passive, active and hybrid techniques. Remote technique basically *

Corresponding author: Aimie Nadia Ab Salam, Hasmaini Mohamad, Nofri Yenita Dahlan, Safdar Raza, E-mail: [email protected], [email protected], [email protected], [email protected] 1 Faculty of Elecrical Engineering, Universiti Teknologi Mara, 40450 Shah Alam, Selangor, Malaysia 2 Department of Electrical Engineering, Jalan Universiti, 50603 Kuala Lumpur, Wilayah Persekutuan Kuala Lumpur, Malaysia

Copyright © JES 2017 on-line : journal/esrgroups.org/jes

J. Electrical Systems 13-3 (2017): 568-578

based on the communication between the DG and utilities. The examples of remote technique are signal produced by disconnect (SPD) [3], supervisory control and data acquisition (SCADA) [4], and power line signaling scheme [5]. The advantage of this technique is a better reliability can be achieved as compared to local technique but the implementation of remote technique in small power system is more expensive. Local technique is really simple, applicable and cost effective therefore it is widely being used for islanding detection. Local technique is based on the measurements of network parameters at the DG site. Active technique is one of the local islanding detection techniques. It detect by adding a perturbation signal to the DG at regular interval of time. Slip mode frequency shift (SMS) [6], active frequency drift (AFD) [7], Sandia Voltage Shift (SVS) [8] are the examples of active technique. Although this technique has smaller non-detection zone than passive method, it degrade the power quality of the system and if significant enough it will degrade the stability of the system [6,9]. Passive technique is based on the measurement of the system parameters such as frequency [9,10], current, voltage imbalance, total harmonic distortion [10], phase jump detection method [11], and harmonic impedance estimation technique [12]. The passive technique does not affect the power quality of distribution network but it has large nondetection zone and has difficulty to detect islanding when load and generation in the islanded system is closely matched. The combination principle of active and passive techniques is called a hybrid technique. By combining active and passive technique, it is not only possible to get benefit from their advantages, but it is also possible to overcome their undesirable features[13]. Examples of hybrid technique are voltage unbalance and frequency set point [14], technique based on voltage and real power shift [15], hybrid SFS and Q-f islanding technique [16], and voltage fluctuation injection [17]. This technique has a small non-detection zone and minimized power degradation but it will take longer time in detecting islanding phenomena. The latest innovation/research in passive detection technique was by adopting computational intelligent approach such as Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF), Decision Tree (DT) as classifier [18,19]. Based on the research, it shows that each technique has higher accuracy than common passive technique in detecting islanding. However, the system will be quite complex when using multiple passive parameters for the detection, which only feasible when using classifier. In addition, selection of parameters used could affect the detection performance since each parameter has a different tendency to react towards each disturbance. This paper presents sixteen different passive parameters to investigate the behaviour of the test system by evaluating the performance of passive parameters based on variation of loads, line to ground faults, line to line faults, line to line to ground faults, three phase faults and load switching at each bus. The rest of the paper is structured accordingly. Material and method that explain the modelling of test system, load model and passive parameter model is presented in section II. In Section III, the results and discussion are analysed. It includes events of islanding and non-islanding and also the case studies. Lastly, the conclusion is drawn in section IV. In Section III, the results and discussion are analysed. It includes events of islanding and non-islanding and also the case studies. Lastly, the conclusion is drawn in section IV. 569

Aimie Nadia Ab Salam: Performance of Multipe Passive Islanding Detection Technique ...

2. Material and Method 2.1. Test System Modelling Simulation study is performed on a test system of 11kV Malaysia distribution network using PSCAD simulation tools for evaluating the passive detection technique. This network as shown in Figure 1 consists of 7 buses, 9 loads and a mini hydro unit (DG) rated 2 MVA at the end of distribution line. There are 2 units of 33/11 kV step-down transformer rated 20 MVA at utility side and a transformer rated 2 MVA is used to step up the voltage to 11 kV from the DG unit. For the distributed generator, the excitation control is equipped to maintain the voltage within an acceptable range. . In order to control the mechanical torque for the water flow, the mini-hydro generator is also driven by a hydraulic turbine. The turbine model used in this work is the Non-Elastic Water Column without Surge Tank, while the excitation system model is IEEE type AC1A standard [20]. 2.2. Load Model The sensitivity of passive parameter highly depends on the type of load used in the test system. A static type of load is used to consider the important effect based on stability of the system. Model of load used in this study is an algebraic function of voltage and also frequency, however, real and reactive power is considered separately. Equations of active and reactive static model are shown in equations (1) and (2).

Fig. 1 Test System The range for exponent A is normally between 0.5 to 1.8, while exponent B is between 1.5 and 6. Besides, both exponents are equal to the slope of dp/dv and dq/dv at V = V0. Value of Kpf is set between 0 to 3 and range of Kqf is -2 to 0. The load model in this study set the exponent A and B to 1.0 and 2.0 respectively. The value of Kpf and Kqf are 1.0 and 1.0. Therefore, the load model is a voltage and frequency dependent type by using values mentioned above [21].

570


P = P0 x (

)A x (1 + Kpf x df)

(1)

Q = Q0 x (

)B x (1 + Kqf x df)

(2)

2.3. Passive Parameter Model Passive technique uses parameter measurement for instance frequency, active and reactive power, voltage and various derivatives to detect the occurrence of islanding. The detection can be based on a single parameter to a number of parameters. This study focused on a single passive parameter measurement at the DG as indicated in Fig. 1. Table I shows detail of each 16 passive parameters used in the simulation study. A direct measurement of voltage, frequency, active and reactive power is used to produce the respective derivatives by using the module in the PSCAD. The parameter that has been selected is based on evaluation sixteen different power system parameters and also on the basis of high sensitivity. Table 1: Sixteen Different Passive Parameters use in Test System No. of parameters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Passive Parameters Rate of change of active power (dp/dt) Rate of change of voltage (dv/dt) Rate of change of reactive power (dq/dt) Rate of change of frequency (df/dt) Rate of change of frequency over reactive power (df/dq) Rate of change of frequency over voltage (df/dv) Rate of change of active power over reactive power (dp/dt) Rate of change of active power over voltage (dp/dv) Rate of change of reactive power over frequency (dq/df) Rate of change of voltage over frequency (dv/df) Rate of change of voltage over reactive power (dv/dq) Rate of change of reactive power over voltage (dq/dv) Rate of change of active power over frequency (dp/df) Rate of change of frequency over active power (df/dp) Rate of change of reactive power over active power (dq/dp) Rate of change of voltage over active power (dv/dp)

3. Results and Discussion The analysis of the passive parameter sensitivity is carried out by simulating different islanding/non-islanding event on the test network. Detail of events and case studies are described in the next following section. 3.1. Islanding and Non-Islanding Events There are two main categories of event need to be premeditated in this study i.e islanding and non-islanding event. Islanding event/loss of mains happens when a part of distribution system connected with DG is disconnected from the main grid. and the DG continue to energize the islanded network [22]. There are many disturbances/fault can be categorized as non-islanding events. In this paper, variations of faults have been used including single phase to ground, double phase to 571


ground, three phases, and phase to phase fault. Fault with 70% occurrence in the power system is single phase to ground fault. It can occur in any of three phases. However, it is sufficient to analyze only one of the cases. In case of phase to phase fault, it occurs when two conductors are short circuited. Their chance of appearance is hardly 15% in the power system. Double phase to ground fault is fault breakdowns of insulation between two phases and earth occur. It is the most extreme type of fault but rarely happen in the power system. The chance of occurrence is 10%. Lastly, three phase faults which mainly occur due to a breakdown of insulation between all the three phases. Its appearance is rarely 2% to 3% in the power system. Load switching is also one of the non-islanding events. At certain duration of time, the grid breaker is opened at different system loading capacities to make sure the load is switched off at each bus. 3.2. Case Studies From the abovementioned events, ten cases of islanding and non-islanding events shown in Table II are simulated on the test system. All sixteen passive parameters are measured for every single case. Table 2: Case Studies of Test System Cases

Type of Cases

1

Loss of Main (LOM)

2

Line to ground fault

3

Line to line to ground fault

4

Line to line fault

5

Three phase fault

6

Load switching at bus 3

7


8


9


10


In order to vary the results, several different scenarios have been considered. Different levels of operating capacities are simulated on the test system based on power of utility grid (Pgrid). In this paper, only selected level has been chosen. Meanwhile, value of power at distributed generator (Pdg) will remain the same. The power losses are assumed to be negligible in the system. Table III below shows the operating capacities between Pdg and Pgrid at different levels. The value of Pdiff is calculated using the following equations. Pdiff = (Pgrid) / Pload

(3)

where, Pgrid = Pload – Pdg

572

(4)


Table 3: Different Operating Capacities No. 1 2 3 4 5

Pdg 1.578 MW 1.578 MW 1.578 MW 1.578 MW 1.578 MW

Pgrid 0.062 MW 0.112 MW 0.162 MW 0.212 MW 0.262 MW

Pload 1.64 MW 1.69 MW 1.74 MW 1.79 MW 1.84 MW

Pdiff 3.78% 6.627% 9.31% 11.844% 14.239%

All sixteen parameters are used in the simulation process while ignoring the initial transients until 0.13 seconds [23]. All parameters are tested for all islanding and nonislanding events. The values of operating capacities are from 0.062MW to 0.262MW by varying the total load in the system. The magnitude of all events of each parameter was tabulated in graph based on different case studies. The results of value based on Pgrid = 0.062MW are shown in Figures 2-11.

Fig. 2 Loss of Mains (Case 1)

Fig. 3 Line to Ground Fault (Case 2)

573


Fig. 4 Line to Line to Ground Fault (Case 3)

Fig. 5 Line to Line Fault (Case 4)

Fig. 6 Three Phase Fault (Case 5)

574


Fig. 7 Load Switching at Bus 3 (Case 6)



575



Fig. 11 Load Switching at Bus 7 (Case 10) As can be observed in the graphs plotted above, the magnitude are varies among sixteen parameters. The magnitude value of each parameter in loss of main case that shown in Fig. 2 are compared to the other nine cases to determine the most sensitive parameter to be selected. By simulating the other cases as in Fig. 3-11, the pattern shows significant changes during non-islanding events. For cases of faults variations in this study (Fig. 3-6), it is found that the magnitude parameters of df/dp, dq/dp and dv/dp have lower value compared to other parameters. Meanwhile, for load switching cases as in Fig. 7-11, parameters of df/dt, dq/dv, and df/dp have the least magnitude value. From the observation and consideration, the rate of change of frequency over active power (df/dp) is selected to be the most sensitive compared to the other fifteen parameters. It is due to all values of df/dp in non-islanding cases simulated are less than the value of df/dp in islanding case as shown in Fig. 12 below. Significant difference can be observed between islanding and non-islanding cases for the df/dp plot.

576


Fig. 12 All Cases of df/dp parameter It is quite difficult to differentiate between the islanding and non-islanding events for the others 15 parameters except for rate of change of frequency over active power (df/dp) due to inconsistent fluctuation of the magnitude value. Thus, this proved that only df/dp parameter can differentiate between islanding and non-islanding event accurately. 4. Conclusion In this paper, sensitivity behaviours of passive parameters based on islanding events and non-islanding events have been identified. The study involves a comprehensive analysis of 16 different parameters used for islanding detection. The parameter of df/dp is selected based on higher sensitivity and accuracy compared to other passive parameters. From the simulation results, parameter of df/dp is able to distinguish between islanding and also nonislanding events. Therefore, a new passive islanding detection technique that uses df/dp as discriminating factor can be proposed for future work. Acknowledgment The authors gratefully acknowledge the support of Institute of Research Management and Innovation (IRMI) and Ministry of Education, Malaysia (MOE) for the support and financial sponsor. This research is supported by University of Technology MARA (UiTM) Shah Alam with project code: 600-RMI/FGRS 5/3 (35/2015). References [1]

[2]

[3] [4] [5]

K. Prasartsuwan and P. Jirapong, "Analysis of islanding detection methods for grid-connected distributed generation in provincial electricity authority distribution systems," in Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), 2012 9th International Conference on, 2012, pp. 1-4. J. A. Laghari, H. Mokhlis, M. Karimi, A. H. A. Bakar, and A. Shahriari, "Artificial neural network based islanding detection technique for mini hydro type distributed generation," in Clean Energy and Technology (CEAT) 2014, 3rd IET International Conference on, 2014, pp. 1-6. D. V. Cesar Trujillo, Emilio Figueres and Gabriel Garcera (2010). Local and Remote Techniques for Islanding Detection in Distributed Generators, Distributed Generation. C. L. Yin. J, Diduch. C, , "Recent Developments in islanding detection for distributed power generation. In: Power Engineering," pp. p. 124-128, 2004. W. Wencong, J. Kliber, Z. Guibin, X. Wilsun, B. Howell, and T. Palladino, "A Power Line Signaling Based Scheme for Anti-islanding Protection of Distributed Generators: Part II: Field Test Results," in Power Engineering Society General Meeting, 2007. IEEE, 2007, pp. 1-1.

577


[6]

[7] [8] [9] [10]

[11] [12]

[13]

[14] [15]

[16] [17]

[18]

[19] [20]

[21]

[22] [23]

578

R. S. Kunte and G. Wenzhong, "Comparison and review of islanding detection techniques for distributed energy resources," in Power Symposium, 2008. NAPS '08. 40th North American, 2008, pp. 1-8. L. A. C. Lopes and S. Huili, "Performance assessment of active frequency drifting islanding detection methods," Energy Conversion, IEEE Transactions on, vol. 21, pp. 171-180, 2006. M. G. M. Abdolrasol and S. Mekhilef, "Robust hybrid anti-islanding method for inverter-based distributed generation," in TENCON 2010 - 2010 IEEE Region 10 Conference, 2010, pp. 13-18. J. N. Patel, "Islanding Detection Techniques for Distributed Energy Resources - Review," International Journal of Innovative Research in Science, Engineering and Technology, vol. 5, April 2016. J. Sung-Il and K. Kwang-Ho, "An islanding detection method for distributed generations using voltage unbalance and total harmonic distortion of current," Power Delivery, IEEE Transactions on, vol. 19, pp. 745-752, 2004. B. Singam and L. Y. Hui, "Assessing SMS and PJD Schemes of Anti-Islanding with Varying Quality Factor," in Power and Energy Conference, 2006. PECon '06. IEEE International, 2006, pp. 196-201. M. Sumner, B. Palethorpe, D. W. P. Thomas, P. Zanchetta, and M. C. Di Piazza, "A technique for power supply harmonic impedance estimation using a controlled voltage disturbance," Power Electronics, IEEE Transactions on, vol. 17, pp. 207-215, 2002. M. Khodaparastan, H. Vahedi, F. Khazaeli, and H. Oraee, "A Novel Hybrid Islanding Detection Method for Inverter-based DGs Using SFS and ROCOF," Power Delivery, IEEE Transactions on, vol. PP, pp. 1-1, 2015. V. Menon and M. H. Nehrir, "A Hybrid Islanding Detection Technique Using Voltage Unbalance and Frequency Set Point," Power Systems, IEEE Transactions on, vol. 22, pp. 442-448, 2007. P. Mahat, C. Zhe, and B. Bak-Jensen, "A Hybrid Islanding Detection Technique Using Average Rate of Voltage Change and Real Power Shift," Power Delivery, IEEE Transactions on, vol. 24, pp. 764-771, 2009. W. Chang, "Hybrid islanding detection method for distributed synchronous generators," presented at the The 2010 international power electronic conference, 2010. H. Vahedi, R. Noroozian, A. Jalilvand, and G. B. Gharehpetian, "Hybrid SFS and Q-f Islanding Detection Method for inverter-based DG," in Power and Energy (PECon), 2010 IEEE International Conference on, 2010, pp. 672-676. M. R. Alam, K. M. Muttaqi, and A. Bouzerdoum, "An Approach for Assessing the Effectiveness of Multiple-Feature-Based SVM Method for Islanding Detection of Distributed Generation," IEEE Transactions on Industry Applications, vol. 50, pp. 2844-2852, 2014. G. S. Mehrdad Heidari, Morteza Razaz, "An Intelligence-based islanding detection method using DWT and ANN," Turkish Journal of Electrical Engineering & Computer Sciences, pp. 381 - 394, 2015. M. Karimi, H. Mokhlis, A. H. A. Bakar,J. A. Laghari, A. Shahriari, M. M. Aman, "Combination of Adaptive and Intelligent Load Shedding Techniques for Distributed Network," IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia:6-7 June 2012. M. Karimi, H. Mokhlis, H. Mohamad, and A. H. A. Bakar, "Under-Frequency Load Shedding Scheme for Islanded Distribution Network Connected With Mini Hydro," International Journal of Electrical Power & Energy Systems, vol. 43, issue 1, pp. 127-138, November 2012. S. Kar and S. R. Samantaray, "Data-mining-based intelligent anti-islanding protection relay for distributed generations," IET Generation, Transmission & Distribution, vol. 8, pp. 629-639, 2014. S. Raza, H. Mokhlis, H. Arof, J. A. Laghari, and H. Mohamad, "A Sensitivity Analysis of Different Power System Parameters on Islanding Detection," IEEE Transactions on Sustainable Energy, vol. 7, pp. 461-470, 2016.