Directional Inverse Time Overcurrent Relay for Meshed Distribution Systems with Distributed Generation with Additional Continuous Relay Settings

Transcription

1 Directional nverse Time Overcurrent Relay for Meshed Distribution Systems with Distributed Generation with dditional Continuous Relay Settings Hebatallah Mohamed Sharaf, H. H. Zeineldin*,, Doaa Khalil brahim nd Essam ELDin bou ELZahab Faculty of Engineering, Cairo University, Egypt. *nstitute Centre for Energy, Masdar nstitute of Science and Technology, bu Dhabi, United rab Emirates. Keywords: nverse time overcurrent relays, Protection coordination, Distributed Generation. bstract Coordination between inverse time overcurrent relays within meshed systems are achieved by adusting two relay settings; pick up current and time multiplier settings. The operating time of these relays is also a function of two additional constants; one represents the constant for relay characteristics and the other one represents the inverse time type. For typical relays, each of these two constants has four definite values and choosing between them indicates the selected relay operating curve: either standard inverse, very inverse, extremely inverse or long time standby earth fault timecurrent relay characteristics. n this paper, a coordination strategy, that takes advantage of the available capabilities in microprocessor based relays, is proposed by considering the two relay characteristic constants as continuous variable settings that can be adusted in addition to the conventional pickup current and Time Dial Setting (TDS). The protection coordination problem is formulated as a nonlinear programming problem where the main obective is to minimize the overall time of operation of relays taking into account protection coordination constraints. The proposed approach is applied to the EEE 14 bus system and is compared with the conventional two setting relay. 1 ntroduction P ROTECTON relaying plays a vital role within the operation of any power system. Overcurrent protection is one of the basic protective relaying principles and inverse time overcurrent relays is considered as the backbone of the protection strategies in distribution networks where the overcurrent relay settings are chosen to achieve coordination, guaranteeing fast, selective and reliable relay operation to isolate the power system faulted section [1]. Digital microprocessor based overcurrent relays are currently widely used for safe and efficient protection with much more powerful capabilities than conventional electromechanical overcurrent relays [2]. These developments in the relays technology are essential to cope with the growing interest to develop the traditional electric power grids into Smart Grids where an important feature of this smart grid will be the increasing penetration of DG at distribution levels [3]. Generally, integration of DG has various impacts on distribution systems and one maor challenge is its effect on the protection system including the increase of short circuit levels, changing the distribution system nature into dynamic, bidirectional power flow nature in addition to the relay coordination failure [4].Directional Overcurrent Relays (DOCRs) become an attractive option for modern distribution systems. Such relays are coordinated optimally to minimize the overall time of operation of all relays [5]. To overcome miscoordination resulting from DG integration, new optimal relay settings need to be determined that take into account its presence. Minimizing the relay operation time becomes more critical in the presence of DG since DG can lose stability. Different optimization methods, including conventional and heuristic techniques, have been applied to determine the optimum time dial and pickup current settings of the relays that guarantee coordination and minimum total relay operating times [6,7]. Other proposed protection coordination techniques used different or modified groups of relay settings and characteristics [8, 9]. This paper investigates the impact of using two additional continuous settings for inverse time overcurrent where the timecurrent characteristics can be easily defined by the user. The coordination problem is formulated such that the relay coordination will be based on four settings: the conventional time dial setting (TDS) and pick up current (p) in addition to settings &B which control the timecurrent relation of the relay and consequently the four are optimally chosen. Choosing the operating timecurrent curve will depend on adusting the & B at the optimal value. The proposed study is implemented on directional inverse time overcurrent relays (DOCRs) for the protection of the power distribution system of the EEE 14 bus system equipped with synchronous based DG. Section of this paper explains the proposed coordination strategy. Section shows the optimization problem formulation. Section V presents the test system and the simulation setup. Section V gives the detailed results and in Section V conclusions are finally drawn.

2 2 Coordination Strategy Using Four Continuous Settings The typical inverse time current characteristic of a directional overcurrent relay is formulated in the following Equation (1) in accordance with EC 60255: t TDS (1) B ( ) 1 where is the constant for relay characteristic, B is the constant representing inverse time type, TDS is the relay time dial setting and p is the pickup current setting. Typically, and B can have one of the four fixed standard values shown in Table 1. Relay Characteristic Type B Standard nverse Very nverse Extremely nverse 80 2 Long Time Standby Earth Fault Table 1: Different types of inverse characteristics curves. Different TDS values allow working within a range of curves for each relay characteristic type. Relay operating time depends on the chosen TDS and p. The proposed approach takes advantage of the flexibility available within digital inverse time overcurrent relays by allowing operating within a wider range of characteristics not limited to the four standard ones. Considering the and B as continuous variable settings of different values, in addition to the conventional TDS and p, allows working within different timecurrent characteristics. n Fig.1, with = 4 and B=0.04, the relay will operate faster than the case where the very inverse type is utilized considering the same value of TDS which is set to 1.The operating time is 2.5 seconds for a short circuit to pick current ratio of4. On the contrary, a relay set with =745 and B=0.02(with TDS=1) will result in a slower operating time of more than 6 seconds. With the additional two settings ( and B), a wide variety of curves can be achieved which results in higher flexibility while achieving relay coordination and thus reducing the minimum total relay operating times and this will be clarified more in the next section. 3 Formulation of the protection coordination problem The protection coordination problem can be optimized such that the obective is to minimize the coordination times of all the relays while maintaining the conditions of protection coordination. The obective function is taken to be the sum, T, of the coordination times of all the relays which needs to be minimized as follows: p Operating Time (seconds) T N M i 1 1 =745 B=0.02 ( t ) (2) pi t bi Standard nverse =4 B=0.02 Very nverse =13.5 B=1 =4 B= Current (Multiples of pickup current) Fig.1. Examples of inverse time overcurrent relay characteristics with different &B constants with TDS=1 Where t p is the operating time for the primary protection relay and t b is the operating time for the backup protection relay. Two approaches will be investigated while optimizing the and B settings. Firstly, all the system relays are considered to have the same timecurrent characteristics and in such case all the system relays will have the same optimized and B settings (same optimized user defined characteristics) that will be referred to in the following sections as optimized global & B settings where relay will have an operating time ( t ) equal to: TDS sci B ( ) 1 t (3) p The second approach optimizes and B for each relay such that each one could have its optimal timecurrent characteristics and that will be referred to in the following sections as optimized individual & B constants where the time to be optimized will be: TDS sci B ( ) 1 t (4) p Where in Equations (3) and (4), i is the fault location identifier, with the total number of fault locations investigated being N, and is the relay identifier, with the total number of relays being M. The following constraints must be satisfied for both studies while solving the optimized coordination problem:

3 t bi t CT (5) pi Where CT is the coordination time interval which indicates the minimum time between the primary and the backup relay for a fault next to the main relay. CT can range usually between s and it is chosen to be 0.3s in this paper. The relay settings (p, TDS,, B) will have a maximum and minimum values as given in the following constraints: p min TDS B (6) p p max min TDS TDS max (7) min max (8) min B B max (9) The minimum and maximum pick up current (p) will depend on the system s rated load currents and system s short circuit levels. TDS could take a value between and 3. For the &B constants; they have been chosen within the analysis to have the minimum value of 4 and 0.02 and a maximum value of 13.5and 1 respectively which represent the standardized values of the standard and the very inverse timecurrent relay characteristics. 4 System and Simulation Setup The proposed coordination strategies considering optimized global and individual and B are applied to the distribution portion of the EEE 14bus system shown in Fig. 2.The detailed system parameters can be found in [10].The system is fed through two 60 MV 132kV/33 kv transformers connected at buses 1 and 2. Two synchronous based DGs are connected at buses 3 and 5. Each DG is rated at 5MV and operates at unity power factor. Nodes are added at midway points of all lines (F8) representing fault locations at which three phase short circuit analysis will be carried out [11]. The proposed strategies has been tested using different case studies including the EEE 14 bus system without the addition of DG and with the addition of DG and considering different DG sizes and locations. The protection coordination problem is formulated as a nonlinear programming problem and is solved using the fmincon function in the Matlab Optimization Toolbox which relies on the gradientbased method that is designed to work on problems where the obective and constraint functions are both continuous and have continuous first derivatives [12]. 5 Results and nalysis n this section, the optimal relay settings and the optimal relay operating time for the assigned faults for the test meshed distribution system are presented. The simulations include the case where protection coordination is solved considering =4 and B=0.02 and then investigating the effect of adding the ( and B) constants as continuous adustable settings for an inverse overcurrent relay firstly considering one optimized ( and B) value for all system relays and then considering different individual relay characteristics. Simulations including different case studies: without DG installation and with DG installations at different DG locations and with different ratings are carried out. Table 2 shows the relay settings (TDS and p ) for the system shown in Fig. 4. ll the settings lie between the minimum and maximum limits and these settings result in the primary and backup relay operating times shown in Table 3 with a total relay operating time equal to 19.57s. ll the primary/backup relays pairs maintained a coordination time interval equal to or above 0.3s. The minimum primary relay operating time in this analysis was chosen to be seconds to avoid nuisance tripping due to transient conditions and in consequence the fastest backup relay operating time will be s (keeping the CT). From Table 3, it can be seen that no primary or backup relay operating time hits the minimum. Relay has the fastest operating time within the system s primary protection relays clearing the fault at node in s and its backup has the fastest backup time which equals s. Relay TDS(s) p(p.u) Relay TDS(s) p(p.u) Table 2: Optimal relay settings (TDS and p ) considering =4 and B= MV 5 F F MV 3 60 MV 60 MV Fig.2. The Distribution Portion of the EEE 14 bus system

4 Fault Location F8 F10 Operating times of relays in seconds (p = primary, b = backup) p b1 b Table 3: Optimal primary and backup relay operating times considering =4 and B=0.02 with DGs. Due to the meshed nature of the system will have to be also cleared through the primary/backup pairs, and which have TDS and p equal (23, ), (0.05, ) and (0.05, 039) respectively as shown in Table 2 and operating times equals , and s respectively as shown in Table 3. s mentioned earlier in Section 3, firstly the and B are chosen to be optimally chosen with the same value for all the relays in the system global & B. The optimal values of the global & B for the EEE 14 bus system under system under test are found to be 859 and 0.5 respectively. Table 4 shows the new TDS and p settings based on the new timecurrent characteristics of the directional inverse overcurrent relays of the system. Relay TDS(s) p(p.u) Relay TDS(s) p(p.u) Table 4: Optimal relay settings using optimized global & B constants for test system including DGs. The total relay operating time is reduced compared to using the standard inverse characteristics to be s instead of 19.57s. The new relay operating times are shown in Table 5 and it can be seen that 11 relays out of 16 (the total number of system relays) recorded the minimum primary protection operating time of s and 9 backup relays recorded also the minimum backup operating time of s. The highest primary relay operating time is the operating time of protecting against a fault at node which equals s. By examining the relays protecting against a fault at node, the operating time of the primary protection relay and the backup relay hits the minimum limits; and sec respectively. The operating time of the backup relay is reduced to be 0.73 sec. Despite that the operating time of the other primary protection relay doesn t hit the minimum but it has reduced remarkably (473 sec with a reduction more than 50% when using =4 and B=0.02). The operating time of the backup relays and are also reduced to and 473 sec, respectively. The previous results showed that operating times of the inverse time overcurrent relays within the system can be reduced when working on the same user defined timecurrent characteristics that are based on the optimized global &B constants. Further enhancement can be achieved by allowing each relay to have its optimal and B constants. n such case, each relay will have four settings to be optimized; TDS, p in addition and B that will identify its operating curve. Table 6 presents the optimal values for all fur variables and the results of the relays operating times for the different faults is shown in Table 7.The total relay operating times in the case of optimizing the and B constants for each relay experienced a small reduction in comparison to the previous case to be s. Fault Location F8 Operating times of relays in seconds (p = primary, b = backup) p b1 b

5 F Table 5: Optimal primary and backup relays operating times using optimized global &B for the test system including DGs. Relay TDS(s) p(p.u) B Table 6: Optimal four relay settings using optimized individual & B constants for test system including DGs. s can be seen in Table 7, most of the relays kept their operating times unchanged or experienced a small reduction in their operating times except for three relays that had a slight increase, when compared to the values in Table 5. For the fault at node, the first group of relays (, and ) kept their operating times while in the second group and experienced a slight increase in their operating times to be 518 & 518 sec respectively while had a reduction to be sec. The other relay that had an increase in its operating time is acting as a backup protection relay for. The highest primary relay operating time experienced a reduction in operating time (885 sec). Fault Location F8 F10 Operating times of relays in seconds (p = primary, b = backup) p b1 b Table 7: Optimal primary and backup relays operating times using proposed coordination strategy and optimized individual &B for the test system including DGs. More simulations, including the system without DG addition and adding DGs at different locations and with different ratings, have been carried. Table 8 shows the total relay operating times (T) for some tested cases. The results show that optimizing and B can further reduce the total time for different DG penetrations.

6 DG Capacity and Location Two Settings Relay Global &B ndividual &B No DGs DGs rated 3 4 DGs rated 6 4 DGs rated 3 bus 7 DGs rated 6 bus 7 DGs rated 3 buses 4,7 DGs rated 6 buses 4, Table 8: Total relay operating time (T) in seconds for different DG sizes and locations. 6 Conclusion The paper proposes optimizing the constants that control the relay timecurrent characteristics ( & B) in addition to the conventional settings (TDS & p ). The relay coordination is formulated as a nonlinear optimization problem where the four variables are optimally chosen to achieve minimum total relay operating time. The proposed study is tested using the distribution portion of the EEE 14 bus system considering midpoint faults without the addition of DGs and with their addition at different locations and with different ratings. Two approaches were tested: either optimally choosing a Global & B constants such that all the system relays will follow the same characteristics or optimally choosing individual & B constants so each system relay will follow its own optimal characteristics. The simulation results show the effectiveness of using the four variables in reducing the total relay operating time through all the simulated case studies. Both approaches (global and individual &B constants) achieved a reduction with a small improvement for the individual &B approach over the global &B. power grids protection, Proceedings of 65 th nnual Conference for Protective Relay Engineers, (2012). [5] H.H. Zeineldin, Optimal coordination of microprocessor based directional overcurrent relays, Proceedings of Canadian Conference on Electrical and Computer Engineering, CCECE, (2008). [6] Waleed K..Nay, H.H. Zeineldin and Wei Lee Woon, Optimal protection coordination for microgrids with grid connected and islanded capability, EEE Transactions on ndustrial Electronics, vol.60,pp , (2013). [7] Tura mraee, Coordination of directional overcurrent relays using seeker algorithm, EEE Transactions on Power Delivery, vol.27, pp , (2012). [8] Chabanloo, R.M.; byaneh, H.; Kamangar, S.S.H; Razavi, F., Optimal combined overcurrent distance relay coordination incorporating intelligent overcurrent relay characteristic selection, EEE Transactions on Power Delivery, vol.26, pp , (2011). [9] Timo Keil and Johann Jager, dvanced coordination method for overcurrent protection relays using nonstandard tripping characteristics, EEE Transactions on Power Delivery, vol.23, pp.5257, (2008). [10] Univ. Washington, Power Systems Test Case rchive, Seattle, W. Mar.2006 [Online]. vailable: [11] lberto J. Urdaneta, Ramon Nadira, Luis G.Perez, Optimal coordination of directional overcurrent relays in interconnected power systems, EEE Transactions on Power Delivery, vol.3, pp , (1988). [12] References [1] Network Protection and utomation Guide, lstom [2] Y.Lee,.K. Ramasamy, F.Hafiz,.bidin, Numerical relay for overcurrent protection using TMS320F2812, Proceedings of the 9 th WSES international conference on Circuits, Systems, Electronics, Control &Signal Processing, (CSECS '10), (2010). [3] Mozina, C.J., mpact of smart grids and green power generation on distribution systems, EEE Transactions on ndustry pplications, vol.49, pp , (2013). [4] ntonova, G.; BB nc., Canada; Nardi, M.; Scott,.; Pesin, M., Distributed generation and its impact on