MODELLING AND ANALYSIS OF THE ENHANCED TAPP SCHEME FOR DISTRIBUTION NETWORKS

Transcription

1 MODELLIN AND ANALYSIS OF THE ENHANCED TAPP SCHEME FOR DISTRIBUTION NETWORKS Maciej Fila Brunel University/EDF Energy, UK areth A. Taylor Brunel Institute of Power Systems Brunel University, UK Peter Lang EDF Energy, UK Abstract - At present the main aim for Distribution Network Operators (DNOs) is to develop flexible, reliable and efficient networks, in order to enable the connection of distributed generation. Active control of distribution networks is a feasible solution in order to reach that aim. One of the first and most commonly used active control devices is an on-load tap changer (OLTC) with its automatic voltage control (AVC) relay. Even though this technique for voltage control is well established, traditional AVC schemes can be unreliable particularly when the transformer arrangement is complex and conditions of the network variable. The main factors that undermine the performance of AVC schemes are; intermittent output of distributed generation, varying power factor, difference in primary voltage or nonidentical paralleled transformers. The Enhanced TAPP scheme can operate efficiently under the above conditions. The first objective of this paper is to present principles of the Enhanced TAPP scheme and mathematical models for AVC schemes in general. Then the functionality of this scheme will be demonstrated using software simulation for a range of distribution network case studies based upon realistic EDF Energy network scenarios. Results from the modelling and analysis of the Enhanced TAPP scheme and conclusions are also finally presented. Keywords: advanced voltage control, parallel transformer, TAPP scheme, distribution network, voltage control relay 1 INTRODUCTION Distribution Network Operators (DNOs) are obliged to maintain voltage profile across the networks within certain limits. These restrictions are due to system constraints such as insulation stresses but mainly due to statutory voltage limits defined in the Electricity Safety, Quality and Jonathan Hiscock Fundamentals Ltd Long Crendon, UK jhiscock@fundamentalsltd.co.uk Malcolm R. Irving Brunel Institute of Power Systems Brunel University, UK malcolm.irving@brunel.ac.uk Continuity Regulations 2002 (ESQC). Each customer in the UK connected to LV network needs to be supplied at 400/230 V with the tolerance +10/-6% whereas HV customers (11kV and 6.6kV) with the tolerance +/-6%. To satisfy these requirements on-load tap changers (OLTC) are used to produce appropriate output voltage under dynamic load conditions and various supply voltages as well as to reduce circulating current when transformers run in parallel. To simplify, automate and optimise performance of the OLTC s, automatic voltage control (AVC) schemes are used. Simple AVC arrangement is shown in the figure 1. The AVCs monitor voltage level V VT and the current I CT on the secondary side of the transformer. With the provision of these measurements AVC relay adjusts tap position of the transformer in order to maintain nominal target voltage, provide load drop compensation () boost and reduce circulating current when two or more transformers are paralleled. Figure 1: AVC arrangement. As the distribution networks become more complex and accommodate growing amount of distributed generation (D), conventional AVC schemes become inefficient. To cope with the voltage problems associated with the increasing penetration of the distributed generation as well as the rising load demand, distribution network 16th PSCC, lasgow, Scotland, July 14-18, 2008 Page 1

2 operators need more reliable and effective voltage control devices. The primary objective of this paper is to investigate performance of the innovative AVC relay SuperTAPP n+ and identify appropriate voltage strategies for the distribution networks with D under varying load conditions. This paper continues and completes the introductory research study presented by the same authors in [1], by extending the computer model and analysis of the Transformer Automatic Paralleling Package (TAPP) scheme and its successor the Enhanced TAPP scheme with distributed generation and load exclusion functionality. The paper structure is as follows. Section 2 describes common AVC schemes used in distribution networks and their characteristics. Section 3 presents a detailed model of the Enhanced TAPP scheme and its principles as well as distributed generation estimation technique and the load exclusion functionality. Section 4 analyses voltage profile of the network under various load and D output conditions and AVC performance. The results obtained from simulation are presented. Finally, conclusions are stated in Section 5. 2 REVIEW OF EXISTIN AVC SCHEMES In order to meet the engineering recommendation for the security of supply and for the higher reliability of supply, it is common practice in distribution networks to parallel transformers on one site or across the network. Under such system configurations another aim of AVC schemes is to keep transformers at a desired tap position in order to minimise circulating current and prevent transformers from run away tapping. Negative Reactance Compounding, True Circulating Current, Master-Follower and TAPP are standard voltage control schemes for parallel transformers. There are several factors affecting performance of the AVCs such as varying power factor, presence of the D (intermittent output), difference in primary voltage or dissimilar transformer impedances. Following section presents advantages and weaknesses of the most common AVC schemes under varying load condition and network configuration. 2.1 Negative Reactance Compounding One of the most common AVC schemes in the distribution networks is Negative Reactance Compounding (NRC) technique. NRC was introduced to distribution networks in the 1960s due to deficiencies of the Master-Follower scheme. It uses settings with the negative value of reactance. The relationship between settings and NRC setting can be defined as follows: Z = R + jx (1) Z = R jx (2) NRC The principles of NRC scheme are presented in figure 2. In this strategy the transformer current is used to create voltage drop I T Z NRC and modify measured voltage from V VT to V AVC. A circulating current flows between transformers when they are on different tap positions. This circulating current rotates the transformer current and consequently phasor of the voltage drop I T Z NRC. When the transformer is on the higher tap position, the effective measured voltage V AVC is higher than the measured voltage V VT and the relay tends to tap down. When transformer is on the lower tap position, the voltage V VT seen by AVC is reduced and as a result tap position is increased. This action is performed until circulating current is minimised and the target voltage achieved. I T1 Figure 2: NRC principle. In this scheme connections between transformers are not required as each AVC is able to act independently. This feature makes NRC scheme suitable for the paralleling transformers across the network and paralleling non-identical transformers. However this scheme has also its disadvantages. The main one being that this scheme is the most accurate at unity power factor, and the error in the performance increases with the power factor deviation. Another disadvantage is that a compromise between strength of the compounding and the susceptibility of the scheme to produce voltage errors for varying power factor must be found. The third is degradation of the performance due to the change in the V AVCT1 V VT I T1Z NRC Transformer at a higher tap position I T2 V AVCT2 I T2 Z NRC V VT Transformer at a lower tap position 16th PSCC, lasgow, Scotland, July 14-18, 2008 Page 2

3 polarity of the X setting. To keep the same boost, the value of R must be increased [2]. 2.2 Master-Follower Another AVC scheme used by DNOs is master-follower. One of the parallel transformers in the scheme monitors the voltage at the bus-bar and adjusts tap position to provide desirable voltage level. When master transformer finishes the action all other transformers in the scheme replicate it. The scheme might be used along with and while settings are adjusted properly, the masterfollower scheme operates correctly under varying power factor, reverse power flow and with presence of distributed generation. However, because connection between relays is required, it is impractical to use this scheme to parallel transformers across a network. Complex switching is needed when one of the transformers is taken out. Additionally, circulating current will flow between paralleled transformers using this scheme unless the transformers are identical, such that they have the same impedance, number of taps and incoming voltage [1]. 2.3 True Circulating Current The true circulating current scheme can be used in order to control voltage at the bus-bar, eliminate circulating current between transformers and prevent transformers from run away tapping. In this scheme interconnection between controllers is used to deduce transformers currents in order to ( IT1 IT 2 ) calculate circulating current I CIRC = (3) 2 For the transformer on the higher tap, I CIRC has a positive sign, whereas on the lower tap it is negative. I CIRC is used to create a voltage bias which adjusts the relay target voltage such that circulating current is minimised (e.g. transformer on the higher tap position taps down). Similar to the master-follower scheme, true circulating current might be used with and it operates correctly under varying power factor, reverse power flow and with presence of distributed generation. Likewise, this scheme has disadvantages such as: - difficulty with paralleling transformers across the network, - necessary connection rearrangement while taking one of the transformers within the scheme out of service, - transformers must be similar (impedance, incoming voltage, connections), - Inaccurate in the presence of embedded generator output. 2.4 TAPP scheme: TAPP scheme is based on the negative reactance compounding principle [3]. However in this scheme two separate circuits are used, one for the purpose of and one for the purpose of compounding, in order to eliminate the need of trade-off between strength of the compounding and the susceptibility to produce voltage errors for varying power factor. Additionally this scheme effectively reduces circulating current, which can flow between parallel transformers, using numerical techniques based on the target power factor. Circulating current in the TAPP method is evaluated by comparing the measured transformer load current (I TR ) with the target power factor (pf targ ) as is shown in figure 3. One disadvantage of the TAPP scheme is the incurred voltage error as the load power factor deviates from the set power factor. This is because circulating current is considered a part of the load current. This drawback is eliminated in Enhanced TAPP scheme [3, 4]. α β I TR I CIRC Figure 3: Principles of TAPP scheme. pf target All the above AVC schemes have a common drawback associated with the presence of D. This is because the voltage and current measurements fed to the relay are taken locally as shown in figure 1. Therefore, the AVC schemes are unable to act appropriately when a voltage increase occurs, at a remote point of the feeder as a consequence of the presence of D. 3 ENHANCED TAPP SCHEME AND SuperTAPP n+ FUNCTIONALITY V VT There are several proposed solutions to improve voltage profile in distribution networks with D. Solutions such as network reinforcements, line reconducting, building a dedicated line are currently available and used in distribution networks. Active voltage control with remote voltage 16th PSCC, lasgow, Scotland, July 14-18, 2008 Page 3

4 sensing units (i.e. enavc), line voltage regulation, scheduling of distributed generation and several other techniques are still in development and trial. The SuperTAPP n+ relay offers firstly, very effective AVC performance based on the Enhanced TAPP algorithm, secondly the innovative technique for voltage control in the distribution network with D. The key benefit of this scheme is that all measurements are taken locally and there is no need for remote communication with the generators. To ensure effective voltage control at the substation, to which D is connected, the AVC relay must be resistant to the intermittent character of the D and varying power factor of the network. Enhanced TAPP scheme fulfils these criteria as demonstrated in the [1]. This scheme is the combination of the TAPP and Circulating Current methods bringing together their advantages stated in Section 2. Its principle is presented in figure 4. αt1 αt2 β Figure 4: Enhanced TAPP principle. I T2 IT1 - I CIRC As shown in figure 4, the circulating current is calculated using the measured transformer load current (I T1 for AVC T1 and I T2 for AVC T2 respectively) and summed load current (I T1 + I T2 ). The voltage drop on the compounding setting (I CIRC Z T ) is used to increase voltage seen by AVC of the transformer on the higher tap position and decrease voltage seen by AVC of the transformer on the lower tap position. Such action is performed until the voltage level at the bus-bar is within the bandwidth of the target voltage and the circulating current is minimised. Figure 5 shows a system arrangement where two parallel transformers are controlled by SuperTAPP n+ relays. The innovative technique employed in the SuperTAPP n+ relay is the ability to estimate output of the generator which is connected at remote point on the feeder. This is achieved by the additional current measurement I on the feeder with D and ratio E ST which represents the load share between feeders with embedded generation to those that do not have generator [5, 6]. ICIRC VAVCT2 - ICIRCZT V VT V AVCT1 I CIRCZ T I T1+I T2 E ST load on feeders with generators I = = load on feeders without generators I 1 2 = I I I This ratio is calculated prior to the connection of generator or when output of the generation is zero, consequently I =0 and I =I 1. Representing sum transformer currents as follows: n = ITn N = 1 I = I + I T1 T 2 (4) the generator current I can be determined as follows: I = ( EST ( I I )) I (5) Knowing the current output of the generator I, the voltage rise at the point of connection can be evaluated and appropriate generator compensation bias (buck) can be applied to the AVC. The generator compensation bias is calculated in reference to the voltage rise at the maximum generator current I MAX as shown in equation 6. V I = VMAX % (6) IMAX This value corresponds to necessary voltage reduction at the substation in order to bring voltage level at the point of connection of D within statutory limits. Additional advantage of this method is improved performance of. In order to eliminate the error from the performance caused by D, the generator current is removed from the sum transformer currents I. LOAD = ( I I ) = ( I I ) ( 1 E ) (7) I + The voltage boost calculation is based on the true load current I LOAD not on the total transformer current, as it is in standard AVC relay. The generator compensation bias and the appropriate voltage bias, along with the circulation current voltage bias and target voltage are used in the AVC relay to calculate effective target voltage to control voltage profile in the system with embedded generation. Figure 5: SuperTAPP n+ relay arrangement. ST 16th PSCC, lasgow, Scotland, July 14-18, 2008 Page 4

5 4 SOFTWARE SIMULATIONS MODELLIN AND RESULTS 4.1 AVC modelling and implementation into OCEPS Operation and Control of Electrical Power Systems (OCEPS) software includes load flow package. It is based on the Newton-Raphson algorithm and uses a partitioned-matrix approach to the Jacobian equations. This algorithm has been tested on a variety of systems and proved to be a very robust tool for load flow studies in distribution and transmission networks [7]. To represent the tap position of the OLTC and its changes, the model of the transformer with variable turns-ratio T is implemented in the load flow software. The AVC monitors and compares actual voltage and target voltage magnitudes. If the difference between these two values exceeds the band width settings, then AVC calculates incremental turns-ratio change T. The calculation for T depends on which AVC scheme is used. The following equation represents how T is evaluated for the SuperTAPP n+ relay: T = VVT VTAR + V + V + V (8) CIRC where: V - measured voltage at bus-bur VT V - relay voltage setting TAR - the generator compensation bias V V - the bias V - the circulating current bias CIRC The new turns-ratio is calculated in order to produce desirable voltage output of the transformer as follows: T NEW OLD = T + T. (9) To reflect the effect of the changes in the turnsratio on the currents and voltages in the network the admittance matrix needs to be recalculated as follows: I I P S Y = T NEW Y NEW T Y V P NEW 2 ( T ) Y V S (10) The system shown in figure 5 was used to investigate and analyse functionality of the SuperTAPP n+. The system consists of two 33/11 kv parallel transformers, both rated 20 MVA with reactance 12% as well as two feeders and the generator connected at the remote point on feeder 1. The CT is located on the feeder 1 and the load ratio of feeder 1 to feeder 2 is set E ST = Case 1 - SuperTAPP n+ performance with In this case the performance of the SuperTAPP n+ was investigated. The distribution network under maximum load condition experienced unacceptable voltage drop on the feeder 2. To improve voltage profile the 2% boost at the 11 kv bus-bar was required. This strategy was effective until D was added. When the generator was producing 3 MW the effectiveness was significantly degraded and voltage magnitude at Bus 5 was below lower statutory limit at pu. This voltage profile of the system is presented in figure 6 by the dashed line. Voltage [pu] Bus Bus Distance Bus 3 Bus Bus SuperTAPPn+ TAPP Figure 6: SuperTAPP n+ performance with. In order to amend this undesirable state the output current of the generator was deducted from transformers load currents and boost based on the true load current was applied. This SuperTAPP n+ functionality enable AVC to keep voltage profile of the distribution network presented in figure 6 within statutory limits. 4.3 Case 2 - SuperTAPP n+ performance with enerator Bias and In the second case the system was under low load condition. The generator at its maximum power output of 4 MW caused voltage rise at the point of connection above upper limit. The AVC scheme was unable to detect this situation on the network with voltage magnitude at Bus 3 of dashed line in the figure 7. Voltage [pu] Bus Bus Distance Bus 3 Bus Figure 7: SuperTAPP n+ performance with enerator Bias and. Bus SuperTAPPn+ TAPP 16th PSCC, lasgow, Scotland, July 14-18, 2008 Page 5

6 When SuperTAPP n+ was used with the 2% generator compensation bias setting at the maximum output of the generator the AVC was able to reduce voltage level at 11 kv bus-bar in order to bring voltage at the Bus 3 within the limit. The performance of this AVC scheme is presented in figure 7 by the continuous line. 4.4 Study System Figure 8 shows the one-line diagram of the 132/11 kv local distribution network. The substation is equipped with two 30 MVA transformers with OLTCs. These transformers can apply voltage variations in the range of +/- 10% in 32 steps and are controlled by AVC relay. Five units of D with the total rated capacity of 5 MW are connected to the distribution network. All units are represented as the one generator connected at the Bus 3. The generator is requested to operate at unity power factor. The distribution network supplies residential and commercial customers with a load demand of maximum 50 MVA, minimum 15 MVA and overall power factor of Feeder 1 supplies Load 1 and has the generator connected along its length. Feeder 2 represents the feeder of the network with the lowest voltage profile. Load 3 relates to the rest of the load of the substation. historical data and set E ST =0.18. The was set to provide 2% boost at the maximum load, and the generator compensation bias was set at 2% at the 5 MW generator output. The simulation was performed under various load and generator output conditions. The results are presented in figure 9. Figure 9: Performance of the SuperTAPP n+ in the distribution network. The AVCs maintained the voltage profile of the system within statutory limits at the substation as well as at the remote points of the feeders. The voltage magnitude at the point of connection of the generator at Bus 3 was kept below 1.06 pu while voltage level along feeder 2 and at Bus 5 was kept above 0.94 pu. This voltage improvement prevented undesirable voltage rise at the Bus 3 and helped to avoid unnecessary tripping of the generator. 5 CONCLUSIONS AND FURTHER WORK Figure 8: One-line diagram of 132/11kV distribution network. Currently the generator is connected to the distribution network with an overvoltage protection relay. When the voltage at the point of connection increases to an unacceptable level, the generator is requested to reduce its output. The generator has experienced several tripping due to high voltage, especially under low load conditions. The OCEPS software was used to analyse SuperTAPP n+ relay functionality in the above system. Enhanced TAPP algorithm with the generator output calculation technique was used to control AVCs of the transformers. An additional CT was employed on the feeder with the generator in order to provide current measurement to the relays. The load ratio of the feeder with generator to the feeders which do not have generators was calculated based on the In this paper the functionality of the innovative AVC scheme was presented. The performance of the SuperTAPP n+ relay in realistic distribution network was demonstrated and analyzed. The study results of this paper confirm that this AVC technique can be an accurate method of voltage control at the substation bus-bar as well as at the remote points of the feeders in the distribution network with D. The main advantage of this scheme is that all measurements are taken locally. With the accurate settings, the proper voltage profile of heavily loaded network with significant penetration of D can be maintained. Also with the appropriate settings of the generator compensation bias the system where D causes unacceptable voltage rise at some point of the feeder can be controlled. The numerical results presented in this paper correspond to the specific system however the load flow software created to perform it can be used for variety of distribution network cases. The software can be used as a universal tool for 16th PSCC, lasgow, Scotland, July 14-18, 2008 Page 6

7 network planning engineers in order to investigate opportunities to accommodate additional D and for the voltage profile improvements using the SuperTAPP n+ relay. It is important to note that as long as one feeder attached to the substation under consideration remains without D then it can be used as a reference feeder the SuperTAPP n+ scheme. In such circumstances the scheme can then be applied generally in order to consider networks with D present in several feeders. Further work will demonstrate the application of the SuperTAPP n+ scheme in such scenarios based upon real network models with varying load conditions. REFERENCES [1] M. Fila,.A Taylor, J. Hiscock, Systematic modelling and analysis of TAPP voltage control schemes, 42 nd International Universities Power Engineering Conference, UPEC 2007, pp , Brighton, UK, 4-6 September 2007 [2] M. Thomson, Automatic-voltage-control relays and embedded generation, Power Engineering Journal, No. 14, pp , 2000 [3] V.P. Thorney, N.J. Hiscock, Improved voltage quality through advances in voltage control techniques, IEE Seventh International Conference on Developments in Power System Protection, pp , 2001 [4] "Technical Specification for Advanced Voltage Control Relay - SuperTAPP n+, Fundamentals Ltd, 2006 [5] J. N. Hiscock, D. J. oodfellow, A Voltage Control Scheme for High Voltage Power Transformers, UK Patent Application B , 2004 [6] J. N. Hiscock, D.J. oodfellow, Voltage Control of Electrical Networks with Embedded eneration, UK Patent Application B , 2004 [7] M.R.Irving, M.J.H.Sterling, Efficient Newton-Raphson algorithm for load flow solution in transmission and distribution networks, Proc IEE, C,134,5,1987, pp [8] Automatic Voltage Control EI , EDF Energy Networks, th PSCC, lasgow, Scotland, July 14-18, 2008 Page 7