Jean-Pierre Braun obtained the B.E. degree from the Ecole d'ingénieurs de Genève, Switzerland, in 1980; the M.E.M. degree from the University of

Transcription

1 Jean-Pierre Braun obtained the B.E. degree from the Ecole d'ingénieurs de Genève, Switzerland, in 1980; the M.E.M. degree from the University of Technology Sydney, Australia, in 1993; and the M.Eng.Sc. from the University of New South Wales, Sydney in In 2007, he joined the Power and Energy Laboratory, Federal Office of Metrology (METAS), Bern-Wabern, Switzerland, as a Scientific Collaborator. He has over thirty years of experience in product development and has held positions such as Hardware Design Engineer, System Engineer and Group Manager. His current research interests include the field of power quality and signal processing.

2 The Calibration and Certification of PMUs Jean-Pierre Braun

3 Outline Introduction Why certification and calibration PMUs and their calibration Architecture of a PMU calibrator Design of the METAS PMU calibrator 8.Calibration of PMU calibrator 9.Signal processing of waveforms 10.Static performances of PMUs 11.Dynamic performances of PMUs 12.Conclusions 3

4 Why certification and calibration The 1995 IEEE standard for PMUs provided: no measurement methods no test methods Magnitude error vs Frequency Phase error vs Frequency Frequency Frequency Depablos et al., Comparative testing of synchronized phasor measurement units, PE Society General Meeting, USA, 2004 PMUs of different manufacture could not be compared at off-nominal frequency Need for better standard and compliance tests 4

5 Why certification and calibration First Traceable PMU Calibrator (NIST, USA) First System for Static Tests 2005 Standard Gerald Stenbackken 2006: Start of development work Improvement interoperability of PMUs through traceability Work paralleled the development of IEEE C standard Second System for Static & dynamic Tests 2011 Standard 5

6 PMUs and their calibration Synchrophasor Measurement of the phasor of the fundamental power frequency (voltage or current) with a common time reference (UTC) In addition to phase and magnitude, PMUs report: 6

7 PMUs and their calibration PMU Error PMUs are defined in the IEEE standard C (IEC ) First issue 1995 Basics aspects Second issue 2005 Introduces static tests Third issue 2011 Introduction of dynamic tests Fourth issue 2014 New limits of ROCOF 7

8 PMUs and their calibration Tests defined in the IEEE C standard Steady-State Compliance (Introduced in 2005) Signal frequency range Signal magnitude Voltage Signal magnitude Current Phase angle Harmonic distortion (single harmonic) Out-of-band distortion Dynamic Compliance (Introduced in 2011) Amplitude modulation Phase modulation Ramp of system frequency Step change in magnitude and phase Reporting latency 8

9 Architecture of a PMU Calibrator Reference PMU Precision Signal Generator Characterisation of Signal 9

10 Architecture of a PMU Calibrator Accuracy Maximal Error 50 Hz TVE Mag. Phase. Time ppm deg us PMU 1% Test device 0.25% Reference grade test device 0.05% Improving the TVE error by a factor 10 demands: Reduction of timing, magnitude and phase error 10x lower TVE 0.1% TVE 0.01% TVE 0.001% Magnitude error [ppm] Max Timing Error 3.18 μs Magnitude error [ppm] Max Timing Error 318 ns Magnitude error [ppm] Max Timing Error 31.8 ns Phase error [μrad] Phase error [μrad] Phase error [μrad] 10

11 Design of the METAS PMU calibrator Hardware platform: NI PXI Software: LabVIEW 12 PMU: Arbiter 1133A Amplifiers: Omicron CMS 156 Sampling Rate: 18 ksps (integer multiple of 50 Hz, 60 Hz & all C reporting rates) 11

12 Design of the METAS PMU calibrator PMU PMU Configuration I Shunts Amplifiers (V & I) UTC Synchronization Waveforms generation / acquisition 12

13 Design of the METAS PMU calibrator Time synchronisation GPS High Quality GPS Receiver Meinberg LANTIME M600 METAS LAN IEEE 1588 Switch Hirschmann MAR 1040 PPS GPS IEEE 1588 IRIG-B 10 MHz OCXO Master/Slave Clock 10 MHz 50/60 Hz Events NI PXI &

14 Design of the METAS PMU calibrator Data frame Continuous waveform generation / acquisition (No dead time) The UTC time of every sample (1/18 khz), data frame (0.1 s), test point (n 1 s) is known PPS Example: Signal frequency range test 45 to 55 Hz with 0.2 Hz increments 10 frames per second Nb. of test points: 51 Test point duration: 5 s Set up time: 1 s Hold time: 1s Duration of test: 357 s 14

15 Design of the METAS PMU calibrator Waveform generation Nominal conditions Harmonics 2 nd 3 rd Harmonics 50 th - None Magnitude step - 20% Amplitude modulation Phase modulation 15

16 Design of the METAS PMU calibrator Reacquisition of PMU signals PMU Voltage Active Voltage Divider PPS ADCs PMU Current Current Shunt Reacquisition sampling synchronous to generation 18 ksps 24 bits 16

17 Calibration of PMU Calibrator Timing Universal Time Interval Counter SR MHz Clock Start Stop Atomic Clocks PPS PPS UTC CH PXI

18 Calibration of PMU Calibrator Magnitude Uncertainty < 200 ppm 18

19 Calibration of PMU Calibrator Phase Alignment of 50 Hz to UTC Time alignment of 50 Hz square wave is known within a few ns with respect to UTC Both signals are sampled synchronously The delay through the filter is identical for both the sine wave and the fundamental of the square wave The relative angle represents the time misalignment 19

20 Signal Processing of Waveforms Principe «PMU Reference data» need to be extracted from reacquired waveforms Waveform Samples (3V + 3I) Phasors Estimation _ + PMU Data Error The test signals required by IEEE C are deterministic Due to system synchronisation time is known with great accuracy All frequencies present in test waveforms are locked to UTC 20

21 Signal Processing of Waveforms PMU tests with a single sinewave Waveform equations for steady-state tests (single phase shown) Signal frequency, magnitude & phase tests (single sine wave): yy tt = XX m cos(ωωωω + φφ 0 ) Estimation of unknown parameters through least-square fitting yy tt = aa cos ωω tt + bb sin ωω tt + cc aa bb cc = (AA T AA) 1 AA T yy(tt 0 ) yy(tt N 1 ) with AA = cos ωωtt 0 sin ωωtt 0 1 cos ωωtt 1 sin ωωtt 1 1 cos ωωtt 2 sin ωωtt 2 1 XX m = aa 2 + bb 2 φφ 0 = arctan bb, aa 21

22 Signal Processing of Waveforms PMU tests with modulated sinewaves Waveform equations for dynamic tests (single phase) Amplitude and phase modulation yy tt = XX m 1 + kk x cos(ωωωω) cos ωω 0 tt + kk a cos(ωωωω π) Ramp of system frequency yy tt = XX m cos ωω 0 tt + πrr f tt 2 Step change in phase & magnitude yy tt = XX m 1 + kk x ff 1 (tt) cos ωω 0 tt + kk a ff 1 (tt) Non linear equations => Iterative least-square fitting 22

23 Signal Processing of Waveforms PMU tests with modulated sinewaves (2) Steps 1. Linearisation yy = yy xx 0 + AA xx (First order Taylor approximation) =yy 00 with AA = (xx,tt 1 ) (xx,tt 1 ) (xx,tt 1 ) xx 1 xx=xx00 xx 2 xx=xx00 (xx,tt 2 ) (xx,tt 2 ) xx 1 xx=xx00 (xx,tt nn ) xx mm (xx,tt 2 ) xx 2 xx=xx00 xx mm (xx,tt nn ) (xx,tt nn ) xx 1 xx=xx00 xx 2 xx=xx00 2. Residues rr = yy yy mmmmmmmm = yy 00 + AA xx yy mmmmmmmm = AA xx yy mmmmmmmm yy 00 =yy oooo 3. Least square condition rr TT rr minimum xx mm xx=xx00 xx=xx00 xx=xx00 23

24 Signal Processing of Waveforms 0.02 Comparison of measured PMU data with reference values 0.01 Phase [deg] Meas - Ref [deg] calc meas fit Time [s] 24

25 Steady-State Compliance Tests Frequency range (45 HZ to 55 Hz) PMU Set up Nominal frequency: 50 Hz Window: Flat Top (6 cycles) Frame rate: 50 Frame/s 25

26 Steady-State Compliance Tests Amplitude range (10 % to 120 %) PMU Set up Nominal frequency: 50 Hz Window: Flat Top (6 cycles) Frame rate: 50 Frame/s 26

27 Steady-State Compliance Tests Harmonics (2 nd to 50 th ) PMU Set up Nominal frequency: 50 Hz Window: Flat Top (6 cycles) Frame rate: 50 Frame/s 27

28 Steady-State Compliance Tests Interharmonics (10 Hz to 100 Hz) PMU Set up Nominal frequency: 50 Hz Window: Flat Top (6 cycles) Frame rate: 50 Frame/s 28

29 Steady-State Compliance Tests Influence of Window Length Interharmonic test (Hann - 6 cycles) Interharmonic test (Hann - 4 cycles) Interharmonic test (Hann - 2 cycles) Interharmonic test (Hann - 1 cycles) 29

30 Dynamic Compliance Tests Phase modulation (0 Hz to 5 Hz) PMU Set up Nominal frequency: 50 Hz Window: Flat Top (6 cycles) Frame rate: 50 Frame/s 30

31 Dynamic Compliance Tests Step change in phase Rise time 31.1 ms Delay time 159 μs Overshoot 5.09 % 31

32 Conclusions PMU Calibrator is operational for certifications & calibrations Suitable for PMUs designed according to IEEE C Time required to calibrate a PMU: ~1 day Informal intercomparison with NIST taking place Latency tests to be completed Present work focuses on increased accuracy Future PMUs likely to be designed for distribution networks (TVE 0.0X%) Higher phase and magnitude accuracy Higher immunity to PQ disturbances PMU calibrators need improved accuracy Higher accuracies also required for the calibration of PMU calibrators Non-standard tests 32

33 Thanks for your interest