Line Impedance Estimation Using SCADA Data

Transcription

1 Line Impedance Estimation Using SCADA Data Presenter: Ramiro Da Corte - Power System Engineer Prepared by: James Shen - Principal Engineer, AESO Nov. 5, 214

2 Background AESO is responsible for grid reliability using real-time power flow analysis. Lines and transformers parameters are important for the power flow results. Impedances are from TFO to the AESO EMS system. Not like transformers, it is difficult to test and validate a line impedance. The parameters of a line are calculated based on construction information. It will be helpful to actually verify the calculated parameters. All existing line impedance estimation are phasor based. This project collaborates with U of A to use EMS SCADA measurements for estimating the impedances. 2

3 Methodology of Impedance Estimation U of A researchers discovered a way to estimate a line impedance using only the SCADA measurements at both ends of a line, plus line length data SCADA measurements: Voltage magnitude (Vrms) without angle Active power (P) Reactive power (Q) Line impedance results: R, X, G and B To reduce the noise impact from SCADA measurement, multiple points of time are used to average the calculated impedance results 3

4 Project Data Five lines have been selected for the impedance estimation Two days SCADA measurements with 5 sec. interval were used Winter peak date in 213 Summer peak date in 214 SCADA data at two ends of the line are Voltage magnitude (V rms ) Active power (P) Reactive power (Q) Line lengths are also required 4

5 Line Impedance Model G, the shunt conductance representing corona loss, can also be estimated using the U of A algorithm 5

6 Impedance Estimation R = (P s + P r (V s 2 + V r2 ) x G/2) / (I rms ) 2 X = (Q s + Q r + (V s 2 + V r2 ) x B/2) / (I rms ) 2 Where G is shunt conductance and B is shunt susceptance Above two equations are for R and X estimation Considering SCADA measurements random errors for P, Q and V, the estimation results can vary. I rms is the denominator, which means the impact of error decreases with increased I rms. Therefore, larger load current can be used to filter results. 6

7 kv MVar MW Estimation Results Line 1 data: summer and winter MW, MVar and kv Sending end-summer Receiving end-summer Sending end-winter Receiving end-winter

8 p.u. p.u. p.u. p.u. Line 1 Estimated Results Original estimated results 15 summer R-est X-est R-ref X-ref 15 winter R-est X-est R-ref X-ref Sorted by Current 15 1 summer R-est X-est R-ref X-ref winter R-est X-est R-ref X-ref I (ka) rms I (ka) rms 8

9 Final Estimated Impedance Given Impedance Summer Est. Winter Est. Line 1 - R ± ±.23 Line 1 X ± ±.13 Line 2 R ± ±.8 Line 2 - X ± ±.9 Line 3 - R.78 /.59 ±.9 Line 3 X 4.8 / 3.55 ±.8 Line 4 R.78 /.74 ±.7 Line 4 - X 4.8 / 3.85 ±.8 Line 5 R 4. / 3.84 ±.5 Line 5 X 9.61 / 1.56 ±

10 Current [ka] Current [ka] Voltage [kv] Voltage [kv] Impedance Estimation by Fault Recorder The impedance estimation by SCADA also has been crosschecked by impedance estimation using measurements from Fault Recorder. The fault recorder data contain waveforms so phasor information can be extracted to calculate line impedance. Waveform of Phase C Voltage at 27s Waveform of Phase C Voltage at 235s t [s] Waveform of Phase C Current at 27s t [s] Waveform of Phase C Current at 235s t [s] t [s] 1

11 Estimation Comparison - SCADA and Fault Recorder Line 925L R X Value used for load flow case SCADA (Summer peak) 2.14 ± ±.38 SCADA (Winter peak) 1.74 ± ±.31 SCADA (Fault day) 1.63 ± ±.3 Fault Recorder Line 93L R X Value used for load flow case SCADA (Summer peak) 1.48 ± ±.8 SCADA (Winter peak) 1.17 ± ±.1 SCADA (Fault day) 1.45 ± ±.21 Fault Recorder

12 Application of Line Impedance Estimation In the AESO, EMS system has real-time State Estimation to check if the power flow will have a valid solution. Big line impedance error will cause big delta between SCADA measurements and State Estimation solution. State Estimation can screen out suspected parameter or SCADA measurements errors. Impedance estimation using SCADA data can verify suspected parameter errors from model, to improve the accuracy of power flow solution. Potentially this impedance estimation can be built as an application in real-time EMS system for State Estimation tuning. 12

13 Collaborative Experiences The AESO EMS is in a project to integrate power system model between EMS and Planning by mapping facilities, including facility impedances checking. APIC seminar at the AESO in 213 provided trigger for this project, when presenter from U of A talked about the innovated way to estimate impedance using SCADA data. At a short meeting by both sides, it was agreed that this initiative is a good match. The AESO then reviewed U of A s estimation methodology, and decided to co-operate with U of A, by providing project source data and reviewing the results. In future, the AESO will consider to use this program from U of A to check suspected line impedance in EMS model. U of A will develop the software for use by the AESO (and other APIC companies). U of A plans to investigate the change of line impedance with temperature and the variation G-loss with time. The findings might be useful for condition monitoring of lines. 13

14 Summary In the study cases, estimated line parameters are very close to the given value. Winter peak data is more suitable for line parameter estimation, as its load is larger. This conclusion is consistent with the finding that the larger load indicates better estimation. SCADA measurements used for line parameters estimation is crosschecked by fault recorder data and it is found that they are close. Therefore proposed algorithm using SCADA data for line impedance estimation is valid. SCADA measurement errors in the estimation is acceptable. This estimation approach will help to verify suspected lines impedances in data model. 14

15 Thank you