Optimal PMU Placement on Network Branches for Intentional Islanding to Prevent Blackouts

Optimal PMU Placement on Network Branches for Intentional Islanding to Prevent Blackouts Mohd Rihan 1, Mukhtar Ahmad 2, M. Salim Beg 3, Anas Anees 4 1,2,4 Electrical Engineering Department, AMU, Aligarh, India; 3 Electronics Engineering Department, AMU, Aligarh Abstract Power grids, throughout the world, are operated as large interconnected networks to ensure reliability of supply in an economic way. The interconnected nature of power network makes it vulnerable and sometimes even a small disturbance in the grid may propagate and result in a blackout. Investigations into major blackouts of the decade in the world have identified inadequate monitoring of the dynamic phenomena occurring in the power grid as the primary reason behind blackouts. To overcome this drawback, it was recommended in various investigation reports that Phasor Measurement Units (PMUs) should be deployed throughout the grid. The investigation reports also recommended that even if a disturbance occurs, the system should be split into a number of independent islands to prevent propagation of fault and avoid blackout in the network. Earlier work on PMU assisted islanding assumed infinite channel capacity of PMUs available which is unrealistic. The primary objective of the present work is to show that the channel capacity of PMU should be an integral consideration while devising an intentional islanding methodology. The islanding algorithm is applied to IEEE 14 bus network which is installed with a minimum number of limited channel PMUs to make the system observable. The algorithm is able to split the IEEE 14 bus network in two independent, self sufficient islands. Keywords Islanding, PMU placement, Power grid I. INTRODUCTION Power grids, throughout the world, are operated as large interconnected networks to ensure reliability of supply in an economic way. In India also the power grid is a huge integrated network having five regional grids; Northern, Southern, Eastern, Western and North-Eastern [1]. Despite offering various advantages, the interconnected nature of power network makes it vulnerable and sometimes even a small disturbance in the grid may propagate and result in a blackout. Blackout means the complete shutdown of power networks. Such blackouts occur when a minor disturbance cascades into a sequence of low probability events [2]. Although blackouts are rare but when they occur, the effect is catastrophic. For example on July 31 st India faced one of the world s worst power grid blackouts. Three regional grids connecting more than 20 states collapsed triggering a major power crisis across the country. The blackout affected more than 60 crore people and virtually brought the entire railway system to a complete halt. This large scale blackout was preceded by a blackout in the northern grid on just the very previous day. These back to back blackouts caused real hardships to the public and resulted in loss of billions of rupees [3]. An expert committee was constituted by the government of India to analyze the causes of these disturbances and to suggest measures to avoid recurrence of such events in future. One of the most important causes identified by the committee was the inadequate monitoring of the dynamic phenomena occurring in the power grid [4]. To overcome this drawback, the committee recommended that Phasor Measurement Units (PMUs) should be deployed throughout the grid. This recommendation was in line with the recommendations of task force [5] and expert committee [6] which investigated the other major blackouts of the decade which occured in North American and European power grids. PMUs provide a dynamic snapshot of the system which helps in better monitoring and control of the power grid. Apart from better real time monitoring, PMU data will be helpful in faster and better state estimation and better utilization of transmission line capacity [7]. Moreover, the enquiry committee recommended that even if a disturbance occurs, the system should be split into a number of independent islands to prevent propagation of fault and avoid blackout in the network. An important requirement of controlled or intentional islanding is the availability of real-time system wide measurements. These measurements may be provided by the PMUs deployed in the grid [8]. Controlled islanding of the power grid based on PMU data will also help in faster restoration of the system to normal operation. An algorithm for intentional islanding of a power grid was presented in [9, 10]. The algorithm presents islanding scheme for a power network with all islands observable and maintaining power balance. An important drawback of the algorithm was that it did not work for all the networks e. g. the IEEE 14 bus network could not be split into islands. 73

Moreover the proposed algorithm was based on the unrealistic assumption of an infinite channel capacity of available PMUs. But in practice the PMU has a limited channel capacity, in fact some of the studies have shown that it is neither economical nor necessary to use PMUs with a channel limit of more than 3 or 4 [11, 12]. Here it may be noted that in the present work a PMU channel implies how many branch currents the PMU can measure in addition to the bus voltage at which it is installed. The actual number of PMU channels may be more as it has to measure three phase current and voltage. The primary objective of the present work is to show that the channel capacity of PMU should be an integral consideration while devising an intentional islanding methodology. To make the study general, it is assumed that the PMUs available can monitor current through only one branch in addition to the voltage of the bus at which the PMU is installed. These PMUs are termed as branch PMUs as the PMU belongs to a branch and may be installed on any of the two buses at the ends of that branch, unlike a particular bus in case of multiple channels PMU. Some of the large utilities in the world have deployed a special type of PMU having only a single channel. The branch PMU offers significant benefits like uniform distribution in the network, adaptable to deployment in multiple stages, higher reliability and these units may be daisy-chained if more branches are to be monitored at a bus [13]. In the present work, the islanding algorithm is applied to a network which is installed with a minimum number of single channel PMUs to make the system observable. In section II of the paper, optimal PMU placement scheme is devised for IEEE 14 bus system. Then with these PMUs installed in the system, the islanding algorithm is applied to the system as discussed in section III. Conclusions drawn from the present work are provided in section IV. But fortunately it is still possible to derive the benefits of PMU installation, if the units are placed in such a way that the system is observable. Various algorithms/methods have been reported in literature for placement of a minimum number of PMUs to make the system observable. These methods are based on topological or numerical formulation and take into account a wide range of constraints related to PMU placement. Some of these constraints are [14]: 1) Effect of zero injection buses 2) Effect of conventional measurements 3) Single or multiple PMU loss 4) Single or multiple branch outage 5) Effect of PMU channel capacity Consider the IEEE 14 bus system shown in fig.1. The objective is to find an optimal location set for branch PMUs which makes the system observable. For the purpose of identifying the branch locations for PMU placement, the branches in the standard 14 bus system have been numbered arbitrarily. The number assigned to each branch has been indicated on top of it. For a power network having number of branches, the optimal PMU placement problem can be defined as [13]: Where (1) Such that (2) is the number of branches in the system is the cost of installation of a PMU on branch j is a vector and has binary values defined as: { (3) II. OPTIMAL PMU PLACEMENT Despite the numerous advantages associated with deployment of PMUs in the power grid, it may not be feasible to install a PMU at every bus of the system. Several constraints are imposed due to high cost of PMU as well as technical and economic requirements of the associated communication circuitry. is called the bus to branch connectivity matrix defined as: { (4) b is a vector having all its elements equal to 1. 74

Figure1. IEEE 14 Bus System Solving the optimal placement problem using Binary Integer Linear Programming (BILP), the branches for optimal location of PMUs for the IEEE 14 bus system are given in table I and their placement is shown in fig. 2. TABLE I Optimal PMU locations for IEEE 14 bus system Branch No. 2, 3, 8, 12, 14, 18, 20 Here it may be noted that the number of PMUs obtained is the minimum number required to maintain the system observability under normal operation conditions. However if the contingencies like failure of one or more PMUs or outage of a line or more than one lines are taken into account then the number of PMUs required will increase considerably. Moreover if the available conventional meassurements or presence of zero injection buses is considered then the number of PMUs required to maintain observability will reduce. However since the objective of present work is to show the effect of PMU channel capacity on islanding scheme, these considerations have been neglected. Figure2. IEEE 14 Bus System with Minimum PMUs III. ISLANDING IN THE PRESENCE OF PMUS Intentional or controlled islanding is an effective technique which may be utilized to segregate the system into several smaller subsystems. This will prevent the propagation of a fault from the faulty or weak area to more stable parts of the system. In order to maintain integrity of operation of the individual islands some constraints should be satisfied. For making each island a self sufficient power network, it is assumed that each island should have at least one generating unit, one load and one PMU. Moreover maximum possible islands should be defined such that the power balance may be ensured in each of the islands. Another constraint to be satisfied is that all the islands should be individually observable. Keeping in view the above constraints, the maximum possible islands are defined as Where { { } (5) { 75

{ The next step is to find island vector which is defined as: Where, are central buses of the system. A central bus is a bus which contains a generator, or a PMU or a load or combination of these elements [11]. Network buses vector B is defined as. (10) (9) is an upper triangular matrix and is defined as (14) In next step, the connected buses that do not belong to other islands are allocated to the boundary lines of previous step`s islands. Hence, for the system under consideration (15) Allocation is continued till all the buses are assigned to island vector (16) Once Island vector is formed, such that the buses connected to two ends of transformers are assigned to the same islands, the buses can be renamed. (17) Next step is to check the observability of individual islands. The observability criterion requires that for an island to be observable: (11) Where in (11) is minimum of the 3 terms in equation (5). For example if total number of generators is less than total number of PMUs or loads then are taken as the number of generators. For 14 bus system, a minimum of seven single channel PMUs are required to make the system observable. These PMUs are required to be placed at buses 1, 2, 4, 6, 7, 10, 13. Therefore from equation (5), the number of maximum possible islands may be determined as:,, (12) It means that a maximum of two islands may be created. Hence 2. Therefore elements will be generators (13) (18) Where k is the number of buses in the island For 14 bus system having minimum number of branch PMUs, the Island 1 connectivity matrix A 1 determined as: (20) may be Now the oservability matrix for island 1may be obtained as: Hence, the island vector at first step from (9) is calculated as: 76

(21) Similarly For 14 bus system having single channel PMUs, the island 2 connectivity matrix A 2 will be: (22) To check the observability of buses in the island 2, its oservability matrix may be obtained as: (23) Since the matrices L 1 and L 2 for both the islands in 14 bus system contains all elements as 1, all the buses are observable and the system is divided into 2 islands as given in table 2 and shown in fig. 3. Island TABLE II The Two Islands of 14 Bus System Buses in the Island Island 1 1 2 5 6 11 12 13 Island 2 3 4 7 8 9 10 14 Therefore with a minimum number of single channel PMUs installed, the IEEE 14 bus system has been split in two independent, self sufficient islands. For infinite channel capacity, the 14 bus network could not be split into islands. Therefore it is obvious that the PMU channel capacity will affect the islanding scheme and it should be an integral consideration while devising system splitting strategies to prevent grid blackouts. Figure 3. IEEE 14 Bus System after Islanding IV. CONCLUSIONS This paper addressed the issue of utilizing the phasor measurement units in maintaining observability of islands during intentional islanding to prevent grid blackout. An important consideration for the manufacturers of PMUs is the number of a channel a PMU should have. The specific objective of the work was to show the effect of channel capacity of the available PMUs in islanding. Earlier work on PMU assisted islanding assumed an infinite channel capacity of the PMUs installed in the network. However practically PMUs are manufacture with limited channel capacity. This factor was considered in the present work and a successful islanding scheme has been developed for the system under consideration. The IEEE 14 bus system was shown to be successfully split in two self-sufficient independent islands. REFERENCES [1] Taylor, C W. 1999. Improving Grid Behaviour. J. IEEE Spectrum (June 1999) 40-45. [2] Amin, M. 2007. Preventing Blackouts. J. Scientific American (May 2007) 60-67. 77

[3] Romero J J. Blackouts illuminate India s power problems. J. IEEE Spectrum (October 2012) 11-12. [4] Report of the enquiry committee on grid disturbance in Northern Region on 30.07.2012 and Northern, Eastern and Northeastern Grids on 31.07.2012. [Online]. Available: www.powermin.nic.in/pdf/grid_enq_rep_16_8_12.pdf [5] US- Canada Power System Outage Task Force, Final Report on August 14, 2003 the Blackout in the United States and Canada, 2004.[Online]. Available: https://reports.energy.gov/blackoutfinal- Web.pdf [6] Anderson G, Donalek P, Farmer R, Hatziargyriou N, Kamwa I et al. Causes of the 2003 major grid blackouts in North America and Europe, and recommended means to improve system dynamic performance. 2005. J. IEEE Transactions on Power Systems (December 2005) 1922-1928. [7] Terzija, V. Valverde, G. Cai, D. Regulski, P. Madani, V. Fitch, J. Skok, S. Miroslav, M. B. Phadke A. 2011. Wide-AreaMonitoring, Protection, and Control of Future Electric Power Networks. J. Proceedings of the IEEE (January 2011) 80-93. [8] Senroy, N. Heydt, G. T. 2006. A conceptual framework for the controlled islanding of interconnected power systems. J. IEEE Transactions on Power Systems (May 2006) 1005-1006. [9] Nezam-Sarmadi,S.A., Nouri-Zadeh,, S., Ranjbar, A.M., Pishvaie M.R. 2010. An Islanding Algorithm to Restore a PMU Installed Power System. In Proceedings of Power and Energy Engineering Conference. [10] Nezam Sarmadi, S.A., Dobakhshari, A.S., Azizi, S., Ranjbar, A.M. 2011. A Sectionalizing Method in Power System Restoration Based on WAMS. J. IEEE Transactions on Smart Grid (March 2011) 190-197. [11] Korkali, M. and Abur A. 2009. Placement of PMUs with channel limits. In Proceedings of Power and Energy Society General Meeting. [12] Milijanic, Z., Djurovic, I., Vujosevic, I. 2012. Multiple channel PMU placement considering communication constraints. J. Energy Systems (September 2012) [Online]. Available: http://link.springer.com/article/10.1007%2fs12667-012-0069-6 [13] Emami, R. and Abur, A. 2010. Robust measurement design by placing synchronized phasor measurements on the network branches. J. IEEE Transactions on Power Systems (March 2010) 38-43. [14] Manousakis, N. M., Korres, G.N. 2012. Georgilakis, P.S. Taxonomy of PMU placement methodologies. J. IEEE Transactions on Power Systems (June 2012)1070-1077. 78