Monitoring of Active Distribution Networks in Steady State and Transient Conditions by means of accurate synchrophasors measurements Mario Paolone École Polytechnique Fédérale se Lausanne - EPFL
Dec 27, 2015
Monitoring of Active Distribution Networks in Steady State and Transient Conditions by means
of accurate synchrophasors measurements
Mario PaoloneÉcole Polytechnique Fédérale se Lausanne - EPFL
Outline
Introduction PMU requirements for power distribution
network applications Proposed algorithm for the synchrophasor
estimation PMU prototype based on the NI Compact Rio Application examples and experimental activities Conclusions
Introduction
Evolution of distribution networks passive active major changes in their operational procedures; main involved aspect is the network monitoring; massive use of advanced and smarter monitoring tools
that would result into faster and reliable real-time state estimation;
possible use of distributed measurement of synchrophasors based on the use of phasor measurement units (PMUs).
In transmission networks WAMS (Wide Area Monitoring Systems) are typically based on the measurements of bus voltages synchrophasors realized by means of Remote Terminal Units typically synchronized by means of the Universal Time Code – Global Positioning System.
Peculiar characteristics of distribution networks lower p.u.l. inductances with non-negligible R; low power flows values; high harmonic distortion levels.
Improved accuracy of synchrophasors measurements
PMU requirements for power distribution network applications
60 Hz
Δθ*
Ee
E
2 2
2 2
( ) - ( ) -r r i i
r i
X n X X n XTVE
X X
(**) IEEE Standard for Synchrophasors for Power Systems, IEEE Std. C37.118, 2005
Total Vector Error (TVE) (**):magnitude of the difference between the theoretically true phasor and the estimated one, in per unit of the true phasor magnitude.
PMU requirements for power distribution network applications
synchrophasor #1
111 ,, QPI
1E 2E synchrophasor #2
222 ,, QPIjX Estimated phasor errors:
RMS: ΔEphase: Δθ
1 2,E E
2
2
2E
E
p e
q e
phase angle difference between phasors : δ Δδ=2∙Δθ
1 2,E E
PMU requirements for power distribution network applications
synchrophasor #1
111 ,, QPI
1E 2E synchrophasor #2
222 ,, QPIjX
1 2,E E
PMU requirements for power distribution network applications
Structure of the developed algorithm for the synchrophasor estimation
I. Sampling of the waveforms (voltage/current), within a GPS-PPS tagged window T (e.g. 80 ms, i.e. 4 cycles at 50 Hz), starting in correspondence of the GPS-PPS wave-front.
II. Identification of the fundamental frequency tone within a specific frequency window Δf (i.e. f0 ± Δf).
III. Reconstruction in time-domain of the identified fundamental frequency tone and improved estimation of the dynamic phasor amplitude, phase and frequency.
Proposed algorithm for the synchrophasor estimation
Input signal:
0 0,1
cos , , n
h h t tt Î Th
s t s s h t s DC s t Gaussian noise
-1-
1
sin- ,
sin
Nn iN
h N h Nh
G k f S D k f f T being D eN N
Discrete Fourier Transform
Problems related to the identification of the fundamental frequency tone:a. spectral leakage effects caused by the finite length of time window T;
b. identification of the correct frequency value that may fall between two subsequent frequency values provided by the DFT.
Proposed algorithm for the synchrophasor estimation
Problem a: simple solution by applying a proper windowing
Problem b:
1
( 1) ( -1)1- , ( ) ( ) -
2 2
nN N
H h N h N Nh
D DG k f S H k f f T H D
0 , 0 1, f m bin f bin m
HP: since the number of samples N per time window T is very large (fsampling=100 kHz >> f0=50-60 Hz); the sine function in the denominator of the Dirichlet kernel DN(ϑ) can be approximated by its argument:
Also, the following approximation applies:-1
--1
Ni
Ne iN
Proposed algorithm for the synchrophasor estimation
sinN N
These assumption led to a linear expression of GH(kΔf) that allows to express Δbin as:
2a bbin
a b
Where a e b are the highest and the second highest tone magnitudes in the discrete spectrum GH. Once Δbin is known, the complex amplitude of the fundamental frequency tone S1 at frequency f0 is given by:
1
2 11
sini bin
H
bin binS e bin G m f
bin
The knowledge of f0 and S1 allows to reconstruct the fundamental frequency tone in the time domain in order to improve its phase estimation.
Proposed algorithm for the synchrophasor estimation
1 0 1 1 1
32 ' "
2f T T T
1T i t
1
1'
clock
T kf
1"
1
s i tT t
s i t s i t
Proposed algorithm for the synchrophasor estimation
PMU prototype basedon the NI Compact Rio
PXI-based system for the generation of reference signals used forthe experimental characterization
Application examplesand experimental activities
Time-Sync accuracy ±100 nswith 13 ns standard deviation
18-bit resolution inputs at 500 kS/s,analog input accuracy 980 μV over ±10 V input range (acciracy of 0.01%)
16-bit resolution, sampling frequency: 100 MS/s, carrier frequencies with 355 nHz resolution and PXI time synchronization skew < 20 ps
2 2
2 2
( ) ( )r r i i
r i
X n X X n XTVE
X X
Accuracy assessment with reference to steady-state signals (phase error)
-6 -4 -2 0 2 4 6
x 10-5
0.0001
0.05 0.1
0.25
0.5
0.75 0.9
0.95
0.99
0.999
0.9999
Phase error [rad]
Pro
ba
bili
ty
-6 -4 -2 0 2 4 6
x 10-5
0.0001
0.05 0.1
0.25
0.5
0.75
0.9 0.95
0.99
0.999
0.9999
Phase error [rad]
Pro
ba
bili
ty
Single tone signal(50 Hz)
Distorted signal(spectrum of std. EN 50160)
Application examplesand experimental activities
Steady-state signals (single-tone and distorted)
Application examplesand experimental activities
Note that the 3σ of the GPS card is ±100 ns 31.410-6 rad @ 50 Hz
Std.dev of the synchrophasor phase estimation with reference to 10 s frequency sweep ramp from 47 Hz to 53 Hz
Frequency-varying signals
Application examplesand experimental activities
Application examplesand experimental activities
Procedure:1. Calculation of signal waveforms (bus voltages) within
the EMTP-RV environment;
EMTP-RVnetwork model
National InstrumentsPXI Arb 6289
cRio PMUs
2. analog-generation of fault transient waveforms by means of the PXI reference signal generator system.
, 1m m ssKP P fsT
, ,m ssP f
1 2
13.8/20
Tr
+ SW
PQ 20kVRMSLL
2MW1MVAR
Load2
PI
+
Overhead_line
SM
13.8kV5MVA
SM
PQ 20kVRMSLL
2MW1MVAR
Load1Governor
BUS_line_startBUS_SM BUS_line_end
Accuracy assessment with reference to electromechanical transients
, 1m m ssKP P fsT
, ,m ssP f
1 2
13.8/20
Tr
+ SW
PQ 20kVRMSLL
2MW1MVAR
Load2
PI
+
Overhead_line
SM
13.8kV5MVA
SM
PQ 20kVRMSLL
2MW1MVAR
Load1Governor
BUS_line_startBUS_SM BUS_line_end
Application examplesand experimental activities
, 1m m ssKP P fsT
, ,m ssP f
1 2
13.8/20
Tr
+ SW
PQ 20kVRMSLL
2MW1MVAR
Load2
PI
+
Overhead_line
SM
13.8kV5MVA
SM
PQ 20kVRMSLL
2MW1MVAR
Load1Governor
BUS_line_startBUS_SM BUS_line_end
Voltagephasorsmeasured
by PMU
Application examplesand experimental activities
80 MW power plant:two aeroderivative gas turbine (GT) units and a steam turbine unit (ST) in combined cycle;
PP connected to a 132 kV substation feeding a urban medium voltage (MV) distribution network;
PP substation is linked, by means of a cable line, to the 132 kV substation that feeds 15 feeders of the local medium voltage (15 kV) distribution network and provides also the connection with the external transmission network throughout circuit breaker BR1.
80
0 m
ca
ble
lin
e
External transmission network
BR1
BR1-GT1
BR-ST
BR1-GT2
BR2-GT1 BR2-GT2
PMU3
PMU2
PMU1
Application examplesand experimental activities
PMU application example (real scale)
Distribution network voltage phasors angles differences during the islanding
Application examplesand experimental activities
80
0 m
ca
ble
lin
e
External transmission network
BR1
BR1-GT1
BR-ST
BR1-GT2
BR2-GT1 BR2-GT2
PMU3
PMU2
PMU1
Conclusions The use of PMUs in active distribution networks requires the
definition of improved performances of these devices compared to those available in the international standard (e.g. IEEE C37.118).
The developed PMU prototype allows to:
i) identify the fundamental frequency tone with accuracy levels adequate for distribution network applications;
ii) obtain accuracy levels not influenced by the harmonic distortion of the analysed signals and by their time-varying characteristics.