Carlo Muscas University of Cagliari, Italy A P + M Phasor Measurement Unit Workshop Synchrophasor estimation processes for Phasor Measurement Units: algorithms and metrological characterization December 9, 2014 EPFL, Lausanne, Switzerland 1
Outline Introduction Proposed PMU algorithm Simulations Experimental tests Conclusions 2
Outline Introduction Proposed PMU algorithm Simulations Experimental tests Conclusions 3
Introduction Synchrophasor standard IEEE C37.118.1-2011 defines required limits for PMU performance under steady-state and dynamic conditions (IEEE C37.118.1a-2014 revises or suspends some limits) ü P-class: protection applications, fast response ü M-class: measurement applications, good accuracy The user must choose a performance class that matches the requirements of each application 4
Introduction Goal of the PMU design P+M class PMU at reporting rate 50 fps ü Protection and measurement applications ü Fast response and good accuracy (tradeoff required) ü No selection needed by the user (one size fits all...) Measurement performed on a single-phase basis 5
Introduction Comparison of algorithms for synchrophasor estimation In Impact of the Model on the Accuracy of Synchrophasor Measurement P. Castello, M. Lixia, C. Muscas, P. A. Pegoraro IEEE Transactions on Instrumentation and Measurement, August 2012 TFT-WLS (Taylor Fourier Transform based on Weighted Least Squares) outperforms the other methods in all conditions except under step tests Proposal : adaptive version of the TFT-WLS, which detects fast signal changes and choses the best phasor estimation to enhance the performance under dynamic conditions. 6
Introduction Dynamic phasor s(t)=a (t )cos(2π f 0 t+φ(t )) a t t t=0 Sampled signal 7
Introduction Method TFT-WLS (De La O Serna et al.) Taylor expansion of the phasor (order K) Signal samples as a linear combination of the phasor derivatives The columns of B are the Taylor Fourier basis vectors 8
Introduction Method TFT-WLS (De La O Serna et al.) WLS solution Amplitude and phase of the synchrophasor Frequency ROCOF 9
Outline Introduction Proposed PMU algorithm Simulations Experimental tests Conclusions 10
Proposed PMU algorithm Block scheme Sample by sample implementation (10 ksample/s in the tests) 11
Proposed PMU algorithm Channel for steady-state signals 1st order TFT Kaiser window, 5.8 cycles (@ 50 Hz) with factor 3.7 12
Proposed PMU algorithm Channel for steady-state signals ROCOF evaluated as frequency derivative Boxcar filter to improve performance 13
Proposed PMU algorithm Channel for steady-state signals Latency > 40 ms window centered 20 ms before reference time Realignment based on computed phasor derivatives 14
Proposed PMU algorithm Channel for dynamic signals 2nd order TFT Kaiser window equal to 3.8 cycles (@ 50 Hz) with factor 5.55 15
Proposed PMU algorithm Frequency feedback Precomputed matrices Discrete frequencies (1 Hz step) in the range 45-55 Hz 16
Proposed PMU algorithm PMU algorithm: fast changes detector Based on evaluation of amplitude and frequency derivatives Thresholds: 0.1 p.u./s and 0.6 Hz/s 17
Outline Introduction Proposed PMU algorithm Simulations Experimental tests Conclusions 18
Simulations Uncertainty sources in PMUs 19
Simulations Uncertainty sources in PMUs 20
Simulations The synchophasor standard IEEE C37.118.1 requirements steady-state dynamic step change Synchrophasor Max TVE % Max TVE % Response time Delay time Max over/undershoot Frequency Max FE Max FE Response time ROCOF Max RFE Max RFE Response time In the following, the results obtained in the worst conditions between P-class and M-class (either stricter requirement or wider test range, or combination of both) will be shown. 21
Simulations Steady state, modulations and frequency ramp Synchrophasors TVE% 3,0% 2,5% Simulations IEEE C37.118.1 limit 2,0% 1,5% 1,0% 0,5% 0,0% Off-nominal Harmonics Out-of-Band Modulations Frequency Ramp 22
Simulations Steady state, modulations and frequency ramp Frequency FE [mhz] 70,0 60,0 Simulations IEEE C37.118.1 limit 50,0 40,0 30,0 20,0 10,0 0,0 Off-nominal Harmonics Out-of-Band Modulations Frequency Ramp 23
RFE [Hz/s] Simulations Steady state, modulations and frequency ramp ROCOF 0,25 Simulations 2,5 0,20 IEEE C37.118.1 limit 2,0 0,15 1,5 0,10 1,0 0,05 0,5 0,00 Off-nominal Frequency Ramp 0,0 Modulations 24
Test Magnitude step change (±10%) Phase step change (±10 ) Simulations Step tests TVE response time Frequency response time ROCOF response time limit results limit results limit results 40 ms 19 ms 90 ms 57 ms 120 ms 75 ms 22 ms 69 ms 76 ms Over/under-shoot < 5 % (but close to the limit) Delay time is very low Latency: compliance to P-class compliance to M-class 25
Simulations PMU tests: summary of compliance P class Steady state Modulation/Ramp Step change Latency M class P class M class P class M class ON H ON H OOB MOD RAMP MOD RAMP a φ a φ P class (< M) Synch. Freq. ROCOF ON: off-nominal H: harmonic disturbance OOB: out-of-band MOD: modulation 26
Outline Introduction Proposed PMU algorithm Simulations Experimental tests Conclusions 27
Experimental tests Uncertainty sources in PMUs 28
GPS Receiver Experimental tests Test setup IRIG-B Test Signals Reference time OMICRON CMC 256plus Reduced set of tests with respect to simulations PPS National Instruments PXI NI PXI-6133 - acquisition (14 bit, 2,5 MSample/s) NI PXI-6682 - timing Three-phase implementation Positive sequence synchrophasor Measurement results Off-line performance evaluation 29
TVE % Experimental tests Syncrophasors 3,0% 2,5% 2,0% Simulations Experimental IEEE C37.118.1 limit 1,5% 1,0% 0,5% 0,0% Off-nominal Harmonics Out-of-Band Modulations Frequency Ramp 30
FE [mhz] Experimental tests Frequency 70,0 60,0 50,0 Simulations Experimental IEEE C37.118.1 limit 40,0 30,0 20,0 10,0 0,0 Off-nominal Harmonics Out-of-Band Modulations Frequency Ramp 31
Experimental tests ROCOF RFE [Hz/s] 0,25 Simulations 2,5 0,20 Experimental 2,0 IEEE C37.118.1 limit 0,15 1,5 0,10 1,0 0,05 0,5 0,00 Off-nominal Frequency Ramp 0,0 Modulations Very low impact on step changes tests 32
Experimental tests Compact RIO single-phase implementation Controller real-time NI crio 9024 Acquisition module NI 9215 (16 bit, 100 ksample/s) GPS synchronization module NI 9467 (PPS ± 100 ns) Some changes needed Feedback every 20 ms Different tuning of the processing parameters Data output compliant to IEEE C37.118.2 Preliminary results: compliance with P+M class 33
Outline Introduction Proposed PMU algorithm Simulations Experimental tests Conclusions 34
Conclusions ü Proposal: two TFT-WLS estimations with frequency feedback and automatic choice of the output based on rapid changes detection ü Performance: P+M class compliant synchrophasor, frequency and ROCOF measurement at reporting rate 50 fps ü Feasibility: PMU prototype implemented in a National Instruments CompactRIO platform 35
Conclusion ü Two TFT-WLS estimations with frequency feedback and automatic choice of the output based on rapid change detection ü P+M class compliant synchrophasor measurement at reporting rate 50 Hz ü P class compliance for frequency and ROCOF measurement ü Frequency and ROCOF measurements comply with M class except for out-of-band conditions ü Feasibility: proposed PMU solution was implemented in a CompactRIO platform of National Instruments, exploiting both the controller and the FPGA modules. Thank you! 36
List of publications ü P. Castello, M. Lixia, C. Muscas, P.A. Pegoraro: Impact of the Model on the Accuracy of Synchrophasor Measurement, IEEE Transactions on Instrumentation and Measurement, Vol. 61, No. 8, August 2012, pp. 2179-2188. ü P. Castello, M. Lixia, C. Muscas, P.A. Pegoraro: Adaptive Taylor-Fourier synchrophasor estimation for fast response to changing conditions, IEEE I2MTC 2012, Graz (Austria), May 13-16, 2012, pp. 294-299. ü P. Castello, J. Liu, C. Muscas, P.A. Pegoraro, F. Ponci, A. Monti: A Fast and Accurate PMU Algorithm for P+M Class Measurement of Synchrophasor and Frequency, IEEE Transactions on Instrumentation and Measurement, Vol. 63, No. 12, December 2014, pp. 2837-2845. ü P. Castello, C. Muscas, P.A. Pegoraro, S. Sulis, S. Toscani: Experimental Characterization of Dynamic Methods for Synchrophasor Measurements, IEEE International Workshop on Applied Measurements for Power Systems, IEEE AMPS 2014, Aachen (Germany), September 24-26, 2013. 37