Passive and Active Methods of Islanding Detection for Grid Connected PV Distributed Generators

Transcription

1 From the SelectedWorks of Almoataz Youssef Abdelaziz Summer June, Passive and Active Methods of Islanding Detection for Grid Connected PV Distributed Generators Almoataz Youssef Abdelaziz Available at:

2 International Journal of Applied Engineering Research, ISSN 97-6 Vol. No. () Passive and Active Methods of Islanding Detection for grid Connected PV Distributed Generators A. Y. Abdelaziz, Sanjay Kumar, Biplab Bhattacharyya, M. Ezzat, W. Sameh Electrical Power and Machines Department,,, Faculty of Engineering, Ain Shams University, Cairo, Egypt Faculty of Electrical Engineering Department, Indian School of Mines, Dhanbad, India, Research Scholar Department of Electrical Engineering, Indian School of Mines, Dhanbad. India, - Abstract. This study proposes a comparison between two islanding detection methods (IDMs) for gridinterconnected distributed generators (DG). The first method is a passive technique based on monitoring the frequency of the system and the rate of change of frequency (ROCOF) at the point of common coupling (PCC).The second method is an active technique based on the injection of a current waveform distortion to the original reference current of the inverter, to force a frequency drift in case of islanding operation. The test results show that there is a difference between the two techniques. Islanding detection methods are investigated, simulated and evaluated in MATLAB / SIMULINK. Key-Words: - Islanding detection, Distributed generator, Rate of change of frequency, Single phase system, PV I. INTRODUCTION The advancement in new technology like photo voltaic, fuel cell, wind turbine, and new innovation in power electronics, customer demands for better reliability and power quality are forcing the power industry to shift for distributed generations (DG). Hence distributed generation has recently gained a lot of momentum in the power industry due to environmental concerns and market deregulations []. Islanding occurs when the distribution generator system becomes electrically isolated from the remainder of the power system, and the local system continues to be energized by DG connected to it. Islanding situations can damage the grid itself also can damage the equipment connected to the grid and can even affect the security of the maintenance personnel that service the grid. Therefore, a distributed generator should be identified and disconnected in seconds after the loss of the main supply according to IEEE standard. To achieve such a goal, each distributed generator must be equipped with an islanding detection device, which is also called anti islanding devices. All islanding detection methods have benefits and drawbacks. Islanding detection techniques can be divided into two techniques remote and local. Remote islanding detection techniques are based on communication between utilities and DGs. Although these techniques may have better reliability than local techniques, they are expensive to implement and hence uneconomical. Transfer trip scheme which basic idea to monitor the status of all the circuit breakers that could island a distribution system. Supervisory Control and Data Acquisition (SCADA) systems can be used for that [] also power line signaling scheme [], [] can be used to determine when the distribution system is islanded. Therefore, local techniques are used to detect islanding situation and they can further be divided into passive and active techniques. Passive methods measure the system parameters continuously like voltage, harmonic distortion, frequency, etc. Based on the system characteristics, one of these parameters or more may greatly change when the system is islanded. Differentiation between an islanding situation and a grid connected condition can be done by setting a proper threshold values. Rate of change of output power of DG [], rate of change of frequency over power [6] rate of change of frequency [7], the vector surge relay [8], change of source impedance [9], voltage unbalance [], rate of change of phase angle difference (ROCPAD) [], frequency monitoring with reconfiguration of frequency relay [] and harmonic distortion [], [] are some examples of passive islanding detection techniques. Passive methods are easy to implement, simple and do not introduce any change to the power quality of the system. The main problem with the passive techniques is that those techniques have a large non-detection zones as it is difficult to detect islanding situation when the load and generation from the DG in the islanded system closely match. So, a special care should be taken while setting the boundary limits for these parameters. If the boundary limit is too low, then it could result in nuisance tripping of DG and if the boundary limit is set too high, islanding may not be detected. Active methods directly introduce small perturbations or deliberate changes into the system. These small perturbations will result in a significant change in system parameters when the islanding situation takes place, and the change will be negligible when the DG is connected to the grid. The positive feedback for power loop method [], the correlation factor method [6], slip mode frequency shift algorithm (SMS) [7], active frequency drift (AFD) [8], automation phase shift method (APS) [9], active frequency drift with pulsation chopping fraction (AFDPCF) [], Active Frequency Drift with Positive Feedback (AFDPF) [], impedance measurement method [], negative sequence current injection [], are some examples of active islanding detection techniques. The main problems with these techniques are introducing few perturbations in the system which may affect the power quality of the grid and detection time is slow as a result of the time needed to analyze the response of the system. Also, these perturbations are injected at defined intervals and if the

3 International Journal of Applied Engineering Research, ISSN 97-6 Vol. No. () islanding occurs during this interval, then it has to wait until the next perturbation to be injected before detection, which will elongate detection time. To overcome these disadvantages, researches have focused on hybrid detection systems, relying on more than one parameter, or combining active and passive methods []. In this study, we assumed that a grid connected PV system consists of parallel local load RLC, and a transformer (trans.), and a breaker which will simulate the islanding, and the utility source as an example of DG system shown in Fig.. Fig.. (b) Equivalent circuits of an islanded DG The reactive power equation in normal operation is expressed in (). QL = QG + QU () Where QU = Fig.. Configuration of the grid connected PV system with local RLC load II. ROCOF DETECTION METHOD BASIC PRINCIPLE The ROCOF detection method is based on the transient caused by the imbalance in the reactive power in the islanded system then the frequency starts to vary during the islanding operation. This system behavior could be used to detect islanding condition. Therefore, measuring the ROCOF could show if the DG is operating in parallel with the grid or it is in islanding condition. Fig. (a) shows the equivalent circuit of a DG connected to the grid in normal operation, where QL is the reactive load demand; QU is the reactive power supplied by the power grid; QG is the reactive power generated by the PV; f is the system frequency; and V is the terminal voltage of local load. Fig. (b) shows the equivalent circuit of the DG while it is in islanding situation, where QL is the reactive load demand during islanding; V is the terminal voltage of local load during islanding; f is the system frequency during islanding; R, L, C is the local load parameters. Fig.. (a) Equivalent circuits of a DG parallel with the grid Q G = - fcv During islanding, load reactive power will be only generated from the DG, so there would be a reactive power imbalance Q expresses in (). Q = QL - QL () = = () As V V, () can be approximately expressed as: Q = = = () Q = () So, the ROCOF can be solved from () as: (6) As shown in (6), the reactive power imbalance ΔQ causes some transients in the islanded system, the frequency drifts down or up. It will trigger the frequency relay due to the deviation in the system frequency. But, if the reactive power imbalance ΔQ is small, then the frequency will change slowly, and may not trigger the frequency relay so the value of ROCOF could be used to detect islanding condition. III. PROPOSED DETECTION TECHNIQUES In this paper two IDMs will be discussed, the first IDM will be a passive technique based on monitoring the ROCOF of the system and take the islanding decision based on the threshold values which will be preset. The second technique is an active technique called active frequency drift (AFD) based on inserting small perturbations into the inverter current as shown in Fig., these small perturbations will cause the frequency drift in islanding situation. The inverter current waveform could be expressed in (7).

4 International Journal of Applied Engineering Research, ISSN 97-6 Vol. No. () i(t)=. - - I sin(ωt) ωt <π/ I sin(ωt) - K I π/ ωt <π I sin(ωt) π ωt <π/ I sin(ωt) +K I π/ ωt <π K orignal current waveform injected current waveform inverter current waveform Fig.. Inverter current with perturbations (7) IV. SIMULATION AND TEST RESULTS The simulation mainly use MATLAB/Simulink software to build a single phase grid connected PV distributed generation system which shown in Fig Fig. 8. Add Fig. 9. Islanding occurred Fig.. Sub Fig.. Islanding occurred. 8.7 Fig Add. Fig.. Islanding occurred Fig.. Islanding occurred Fig Fig. 6. Islanding occurred Fig. 7 Fig. 6 shows the frequency at the PCC at normal operation and islanding occurred at time.sec the higher ROCOF as show in Fig Fig. 6. Frequency at PCC Fig.. single phase grid connected PV generation system In this example, the voltage of the PV is set to be Volt, 6Hz. The system is mainly composed of PV panel, IGBT inverter and its control, single phase transformer, filter, load which is formed from parallel RLC and the utility (grid). The PV array is designed to supply a load 8KW and operates at unity power factor. Various events (islanding, capacitor switching, and load change) had been simulated as shown in TABLE to show the effectiveness of the proposed techniques and the differences between them. The output voltage waveform at the PCC is shown in Fig.. The ROCOF threshold value was set. Hz/sec. The results are shown in figures from Fig. 6 to Fig Fig. 7. ROCOF at PCC Fig. 8 shows the frequency at the PCC at normal operation and a load consists of KW added at time.sec then islanding occurred at time.sec the higher ROCOF as show in Fig Add KW.8 Fig.. Inverter voltage at PCC Table. Studied Cases Case # T P(KW) Q ind(kvar) Q Cap(KVar) Fig. (#) Fig. 6. Islanding occurred Fig Fig. 8. Frequency at PCC

5 International Journal of Applied Engineering Research, ISSN 97-6 Vol. No. (). Add KW Fig. 9. ROCOF at PCC Fig. shows the frequency at the PCC at normal operation and load consists of KW disconnected at.sec then islanding occurred at time.sec the higher ROCOF as show in Fig Sub. KW Fig.. Frequency at PCC Fig.. ROCOF at PCC The previous case studies were done while the mismatch in active and reactive power is low ( P, Q ). So, passive technique couldn t detect islanding while active technique detected it as none detection zones (NDZ) in passive technique is larger than in active technique. But the next two cases done while there is a mismatch in active and reactive power. Fig. shows the frequency at the PCC at normal operation and islanding occurred at time.sec the frequency deviates from its nominal value in AFD technique than passive technique which will cause a higher ROCOF as show in Fig. but in this case the both techniques succeeded in islanding detection Add Cap..KVar.8. Sub. KW Fig.. ROCOF at PCC Fig. shows the frequency at the PCC at normal operation and a capacitor bank.kvar switched on at.sec then islanding occurred at time.sec the higher ROCOF as show in Fig Add Cap..KVar Fig.. Frequency at PCC Fig. ROCOF at PCC Fig. 6 shows the frequency at the PCC at normal operation and islanding occurred at time.sec the higher ROCOF as show in Fig. 7, but also the both techniques detect islanding as the mismatch in power is not small. Fig.. Frequency at PCC

6 International Journal of Applied Engineering Research, ISSN 97-6 Vol. No. () Fig. 6. Frequency at PCC Fig. 7. Frequency at PCC Table shows the response of the two techniques (passive and active) for cases -6. We could find that the active techniques have smaller NDZ than passive techniques but slower than them. Table : ROCOF response Case # Passive Active X X X Succeeded X X Failed 6 V. CONCLUSION This paper has showed a short overview of some passive and active techniques used in islanding protection systems and and also has showed short comparison between two techniques based on the ROCOF to detect islanding operation for PV grid connected generators. To verify the effectiveness of the proposed techniques, simulation using different kinds of typical loads and load variation were used in this study. The results showed that the active techniques have smaller NDZ than passive but slower and affect the power quality of the system. REFERENCES [] Nehrir, M.H., Vivek.: A Review of Issues Regarding the Use of Distributed Generators. Proceeding of the 7 th Annual North Power Symposium () pp. 99- [] Refern, M.A., Usta, O., Fielding, G.: Protection against loss of utility grid supply for a dispersed storage and generation unit. vol. 8, no., IEEE Trans. Power Del (99) pp [] Zhang, G., Kliber, J., Wang, W., Li, C., Wang, G., Xu, W.: A power line signaling based technique for anti-islanding protection of distributed generators Part I: Scheme and analysis. vol., no., IEEE Trans. Power Del (7) pp [] Aaker, K., Ropp, M., Sabhah, N., Haigh, J.: Using power line carrier communications to prevent islanding. in Proc. 8th IEEE Photovoltaic Specialist Conf () pp [] Barren, J.I., Redfernand, M.A., Usta, O.: A new microprocessor based islanding protection algorithm for dispersed storage and generation units. vol., no., IEEE Trans. Power Del (99) pp. 9 [6] Pai, F., Huang, S.: A detection algorithm for islanding-prevention of dispersed consumerowned storage and generating units. vol. 6, no., IEEE Trans. Energy Convers () pp. 6 [7] Warin, J., Allen, W.H.: Loss of mains protection, ERA Conf. Circuit Protection for industrial and Commercial Installation, London, U.K (99) pp..... [8] Huang, Z., Freitas, W. Xu, W.: A practical method for assessing the effectiveness of vector surge relays for distributed generation applications., vol., no., IEEE Trans. Power Delivery() pp. 7-6 [9] Hopewell, P.D., Jenkins, N., Cross, D.A.: Loss of mains detection for small generators., vol., no., in Proc. IEE Electric Power Applications (996) pp. [] Jang S.I, Kim, K.H.: A new islanding detection algorithm for distributed generations interconnected with utility networks.vol., in Proc. IEE Int. Conf. Developments in Power System Protection () pp. 7 7 [] Samui, A., Samantaray S.R.: Assessment of ROCPAD Relay for Islanding Detection in Distributed Generation, Vol.. IEEE Transactions on Smart Grid () pp [] Plamondon, M., Belhomme, R., Nakra, H., Desrosiers, G., Gagnon, C. Case study on the integration of a non-utility induction generator to the Hydro-Quebec distribution network. vol., no., IEEE Trans. Power Del (99) pp [] Jang S. Kim, K. Development of a logical rulebased islanding detection method for distributed

7 International Journal of Applied Engineering Research, ISSN 97-6 Vol. No. () resources, vol., in Proc. IEEE Power Engineering Society Winter Meeting () pp [] Kabayashi, H., Takigawa, K. Hashimato, E.: Method for preventing islanding phenomenon on utility grid with a number of small scale PV systems. vol., in Proc. nd IEEE Photovoltaic Specialists Conf (99) pp [] Du, P., Nelson, J.K Ye, Z.: Active anti islanding schemes for synchronous-machine-based distributed generators, vol., no., IEE Proc. Generation Transmission Distribution () pp [6] Chang, W.Y: A correlation factor based islanding detection method for distributed synchronous generators. Vol. 7, Issue, WSEAS Transactions on Systems (8) pp. 9-8 [7] Smith, G.A., Onions, P.A, Infield, D.G.: Predicting islanding operation of grid connected PV inverters. vol. 7, in Proc. IEE Electric Power Applications () pp. 6 [8] Rohatgi, A., Ropp, M.E, Begovic, M.: Analysis and performance assessment of the active frequency drift method of islanding prevention. vol., no., IEEE Trans. Energy Convers (999) pp [9] Yin, J., Chang L. Diduch, C.: A New Adaptive Logic Phase-Shift Algorithm for Anti-Islanding Protections in Inverter-Based DG System. IEEE 6 th Power Electronics Specialists Conference () pp [] Jung Y., Choi, J., Yu, B., Gwonjong Yu, :Optimal Design of Active Anti-islanding Method Using Digital PLL for Grid Connected Inverters. IEEE 7 th Power Electronics Specialists Conference ( 6) [] Gwon-jong Yu, Jeong Hoon So, Young-Seok Jung, Seung-Gi Jeong, Ki-Hyun Kim and Ki-ok Lee, :Boundary conditions of reactive power variation method and active frequency drift methodfor islanding detection of grid-connected photovoltaic inverters. Photovoltaic Specialists Conference () pp [] Fox B. Kane, P.O.: Loss of mains detection for embedded generationby system impedance monitoring. in Proc. 6th Int. Conf. Developmentsin Power System Protection (997) pp [] Yazdani, A. Karimi, H., Iravani, R. :Negative Sequence current injectionfor fast islanding detection of a distributed resource unit. vol., no., IEEE. Power Electron (8) pp [] Mahat, Pukar, Chen, Zhe, Jensen, Birgitte, Bak.: A Hybrid Islanding Detection Technique UsingAverage Rate of Voltage Change and Real Power Shift. Vol., no., IEEE Transactions On Power Delivery(9).