CIGRÉ-697 2015 CIGRÉ Canada Conference 21, rue d Artois, F-75008 PARIS http : //www.cigre.org Winnipeg, Manitoba, August 31-September 2, 2015 GIC Analysis using PSS E K.V. PATIL Siemens Power Technologies International Schenectady, New York, USA SUMMARY Geomagnetic disturbances (GMDs) headed toward earth can produce severe fluctuations in the earth s magnetic field. Such GMDs induce potential gradients on the surface of the earth causing the earth surface potential (ESP) to rise. This leads to induced voltages in power system transmission lines. Such induced voltages can cause induced currents (called geomagnetically-induced currents or GICs) to flow into the electric power system if there is a closed path for the current to flow. This path is typically provided by grounding connections of substation transformers and shunts. GICs are often described as being quasi-direct current (dc), although the frequency variation of GICs is governed by the time variation of the induced electric field [1, 2]. North American Electric Reliability Corporation (NERC) standard TPL-007-1 [11] requires that each Planning Coordinator and Transmission Planner shall maintain ac system models and dc GIC system models within its respective area for performing the studies needed to complete its GMD vulnerability assessment. This paper describes development of a dc GIC system model using a power system planning base case power flow data model [12] and guidelines provided in the NERC GMD Task Force GIC Application Guide [4]. A GMD vulnerability assessment requires calculating the GICs flowing in the power system network; these GIC flows are then used to determine transformer reactive power absorption [11]. The resulting transformer reactive power losses are represented as constant current loads in the base case network when performing GMD vulnerability assessment studies. This paper describes the use of PSS E [13] to evaluate power system performance during GMD events; power flow, contingency analysis and QV analysis are discussed. KEYWORDS Geomagnetic Disturbance Geomagnetically Induced Currents Transformer Half-Cycle Saturation Power System Modeling Power System Analysis PSS E. krishnat.patil@siemens.com
1 POWER SYSTEM MODELING FOR GIC ANALYSIS GICs are low frequency currents that flow into the electric power system if there is a closed path for the current to flow between the electric network and the earth. This path is typically provided by grounding connections of substation transformers and shunts. For the purpose of GIC analysis the ac power system is reduced to a dc resistive network with the following network assumptions [1, 4]: Transmission lines are modeled with their dc resistance in series with induced dc voltage [6]. Transmission line reactors and charging are ignored. Series compensated transmission lines block GIC flow and hence are ignored. The windings of two and three winding transformers that have ground paths are modeled with their dc resistance to ground. The series winding and common winding (if grounded) of auto transformers are modeled. Bus shunts (if grounded) are modeled with their dc resistance to ground. The grounding impedance of transformers and shunts is modeled with their dc resistance. Equivalent substation grounding dc resistance is considered. Generators are isolated at dc from the rest of the network, hence are excluded. Figure 1 illustrates a dc network representation of a typical ac power system used in GIC calculations. Notations: GIC induced voltages on a transmission lines: V1, V2, V3 and V4 Transmission Line dc resistances: Rline1, Rline2, Rline3, Rline4 Transformer Winding dc resistances: Rtrn1-w2, Rtrn2-w2 Shunt dc resistances: Rsh1, Rsh2, Rsh3 Substation ground dc resistances: Rgrd2, Rgrd3, Rgrd4 Substations: Sub1 (switching station), Sub2, Sub3, Sub4 Figure 1: DC representation of an AC Power System used in GIC Calculations The induced dc voltages (V1, V2, V3 and V4) in transmission lines are calculated [3, 4] as in equation (1). = + (1) where E N and L N are the Northward electric field and Northward distance, while, E E and L E are the Eastward electric field and Eastward distance. Typical power flow programs require ac network data. However, for GIC calculations, dc resistance data of transmission lines, transformer windings, bus shunts, and effective substation ground are required. Additionally, the configuration (vector group) of transformer windings and bus shunts is required to determine their ground connection. In order to calculate the induced dc voltage in transmission lines using equation (1), the geographical locations (latitude and longitude) of transmission lines are also required. A power system dc network as shown in Figure 1 is converted into its Norton equivalent and solved to determine the GICs flowing in the network. 2
1.1 Calculation of GIC Induced DC Voltage in Transmission Line The induced dc voltage in transmission lines determined by equation (1) depends on geomagnetic field and earth conductivity modeling. Three different approaches are considered. i. A Uniform geomagnetic field of assumed magnitude and direction and ignoring earth conductivity variations [3] ii. The Benchmark geomagnetic disturbance event scaled to account for local geomagnetic latitude and local earth conductivity structure [5] iii. A Non-uniform geomagnetic field and 1-D earth conductivity models [4]. The variation of earth conductivity with depth is represented using 1-D layered, laterally uniform earth models. This method uses the complex image method [7]. 1.2 Transformer Reactive Power Losses due to GIC Flow The transformers are subjected to half-cycle saturation due to the flow of GICs. This results in increased reactive power (Mvar) losses in these equipments. The GIC-saturated transformers appear as constant reactive current sinks to the rest of the network [1, 2]. The reactive power losses in transformers due to the flow of GICs are determined by considering K factors, transformer core design and winding voltage levels [9]. Using the effective GICs [4, 8] flowing in transformers, the reactive power losses are calculated as [9]: 3 = (2) where I eff is the effective GIC and K factor is a Mvar/ampere scaling factor. This scaling factor could be a generic value based on the transformer type, or could be a specific value obtained for the specific transformer. It is important to note that having accurate K factors from transformer tests and specifications will determine the accuracy of these studies. The GIC analysis results presented in the next section use a generic K factor provided in [9]. A different approach for calculating transformer reactive power consumption due to GIC flow is described in [10]. The base power flow network modified by adding constant current loads representing effects of GICs flowing in transformer windings is then used for performing various system planning studies. 2 GIC ANALYSIS PSS E IMPLEMENTATION Figure 2 shows GIC analysis implementation in PSS E. The salient features of this implementation are listed below, and correspond to numbered sections of the figure: 1) Specify and simulate various GMD event scenarios 2) In case of non-uniform GMD event modeling, specify desired electrojet characteristics 3) 1-D Earth Conductivity Models for USA and Canada defined in [4, 11] are implemented as standard earth models 4) Specify required GIC network model data using the GIC data file. Also, any geographic specific 1-D earth conductivity model, if available, can be specified in this file. 5) (a) Specify network subsystem of interest, and (b) Specify intertie levels of neighboring networks to consider 6) The default GIC to Mvar loss scaling factors are shown in Figure 3. Also, a specific K factor for any transformer can be provided in the GIC data file. Specific modeling information on power system equipment such as two and three winding transformers, auto transformers, phase angle regulators, and underground cables can be provided via a GIC data file. As the bulk electric power system is spread over wide physiographic regions, PSS E s
GIC model allows representation of different GMD-induced voltages as well as different earth conductivity values among these regions via the GIC data file. Figure 2: GIC Analysis PSS E Implementation Figure 3: Transformer GIC Mvar Loss Scaling Factors PSS E Defaults 4
3 POWER SYSTEM GIC ANALYSIS The effects of GMD on power system performance and operation are evaluated by adding the transformer reactive power losses due to GIC flow to the base power system network as constant reactive current loads. An example network consisting of 42 buses, 23 transformers, 56 branches, and 20 substations is considered for the analysis presented here. For this example network, the power flow analysis, contingency analysis and QV analysis are discussed, accounting for the effects of GMD. For the purpose of this analysis, the dc resistance values for transmission lines and transformers are calculated from ac resistance values. The geographical locations of the substations are assumed. 3.1 Power Flow Analysis For the example network, the power flow is solved with and without GMD effects. Two GMD events are applied to the example network: (a) the Benchmark event [5] with 8 V/km electric field magnitude, and (b) a Uniform electric field of 7 V/km. The earth conductivity model assumed is Shield [5, 11]. The electric field direction is varied to find the orientation that would result in maximum Mvar losses. Figure 4 shows the network bus voltage maps for the example network under the Uniform field and the Benchmark field. These figures display the lowest voltage found in the power flow solution for buses terminated in that substation and for the electric field orientation that would result in maximum Mvar losses. The transmission line colors indicate voltage levels. The bus voltage colors indicate bus voltage range; with green indicating 0.95<=pu<=1.05. The substations are indicated by numbers, and are located on the map by longitude and latitude. The size of the circle at the substation indicates the deviation of bus voltage from 1 pu. As seen in the network map, all bus voltages are within normal operating limits for the Benchmark event. However, for the Uniform electric field event, the bus voltages are much lower than normal operating limits; even as low as 0.76 pu. The size of the circles easily illustrates the locations of the network with significant voltage dip problems, and therefore shows the locations where additional capacitive reactive power support may be required. (It is noted that for the example network the power flow does not converge with an 8 V/km at 96 o Uniform electric field event; hence 7 V/km was used.) Figure 4: Network Bus Voltages for the Uniform and Benchmark GMD Events Figure 5 shows transformer reactive power losses due to GIC flow for the electric field orientation that results in maximum Mvar losses. As expected and seen here, the reactive power losses for the Uniform field event are much higher than the comparable-strength Benchmark event. 5
Figure 5: Transformer Reactive Power Losses for the Uniform and Benchmark GMD Events Figure 6 shows GICs flowing in substation ground and transmission lines for a Benchmark GMD event of 8V/km at 95 o. The substation ground GIC flows show the substations that have considerable GICs and where GIC mitigation measures may be needed. Figure 6: Substation and Transmission Line GIC flows for the Benchmark GMD Event 3.2 Contingency Analysis For the example network, the contingency events comprised of single or parallel transmission line outages are simulated. The GIC flows and transformer reactive power losses are calculated for the Benchmark Event efield magnitude. For each contingency, efield orientations are varied in order to determine the GMD direction that results in the maximum transformer reactive power loss; this direction is shown in the lower portion of Figure 7. The associated transformer loss in Mvars is shown in the upper portion of Figure 7. Note that for some contingency scenarios, it is possible not to achieve a converged power flow solution when GMD loads are added to the base case. 6
Figure 7: Contingency Analysis with the Benchmark GMD Event 3.3 QV Analysis Figure 8 shows the QV analysis at one of the network buses in a substation where the voltage dip is largest. For any given voltage level, reactive power requirements at the bus are less under base case conditions or the Benchmark event than under the Uniform field condition. Therefore, more reactive power support is required for the Uniform Field event compared to the Benchmark Event. Figure 8: QV Analysis with the Uniform and Benchmark Event Fields 7
4 CONCLUSION This paper describes recommended electric power system modeling to calculate Geomagnetically Induced Currents flowing in the network and implementation of this modeling in PSS E. The paper illustrates the techniques for conducting power system planning studies such as power flow, contingency, and QV analysis considering the Benchmark or Uniform Field GMD events. The conversion of GIC flows in a transformer to reactive power losses is crucial to ascertain and quantify GMD effects on power system performance. It is important to note that having accurate K factors from transformer tests and specifications will determine the accuracy of these studies. Also the earth conductivity models play a deterministic role in evaluating the GIC flow in the network. For conservative power system performance evaluation under GMD events, it is necessary to determine the efield orientation that would result in maximum reactive power losses in transformers. PSS E provides a means for simulating various GMD events, using standard and user-defined earth conductivity models and performing power system studies to determine the effects of Geomagnetic Disturbance Events on power system operation. BIBLIOGRAPHY [1] IEEE Transmission and Distribution Committee Working Group on Geomagnetic Disturbances and Power System Effects Report, Geomagnetic Disturbances Effects on Power Systems, IEEE Transactions on Power Delivery, Vol. 8, No. 3, July 1993, pages 1206-1216. [2] North American Electric Reliability Corporation (NERC), 2012 Special Reliability Assessment Interim Report, Effects of Geomagnetic Disturbances on the Bulk Power System. [3] R. Horton, D. Boteler, T.J. Overbye, R. Pirjola, and R.C. Dugan, A Test Case for the Calculation of Geomagnetically Induced Currents, IEEE Transactions on Power Delivery, Vol. 27, No. 4, October 2012, pages 2368-2373. [4] North American Electric Reliability Corporation (NERC), Computing Geomagnetically - Induced Current in the Bulk-Power System Application Guide, 2013. [5] North American Electric Reliability Corporation (NERC), Benchmark Geomagnetic Disturbance Event Description, 2013. [6] D. H. Boteler and R. J. Pirjola, "Modeling Geomagnetically Induced Currents produced by Realistic and Uniform Electric Fields", IEEE Transactions on Power Delivery, Vol. 13, No. 4, October 1998, pages 1303-1308. [7] D.H. Boteler and R.J. Pirjola, The Complex Image Method for calculating the Magnetic and Electric Fields produced at the Surface of the Earth by the Auroral Electrojet, Geophys. J. Int., Vol. 132, 1998, pages 31-40. [8] K.V. Patil, Modeling and Evaluation of Geomagnetic Storms in the Electric Power System, Paper # C4-306, CIGRE 2014 [9] X. Dong, Y. Liu, J. G. Kappenman, Comparative Analysis of Exciting Current Harmonics and Reactive Power Consumption from GIC Saturated Transformers, Proceedings IEEE, 2001, pages 318-322. [10] R.A. Walling and A.H. Khan, Characteristics of Transformer Exciting-Current during Geomagnetic Disturbances, IEEE Transactions on Power Delivery, Vol. 6, No. 4, October 1991, pages 1707-1714. [11] North American Electric Reliability Corporation (NERC), TPL-007-1: Transmission System Planned Performance for Geomagnetic Disturbance Events [12] PSS E Power Flow Data Model [13] Siemens Industry, Inc., Siemens PTI, Power System Simulator for Engineering (PSS E) 8