Professional Documents
Culture Documents
fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
I. INTRODUCTION
The grid voltage fundamental frequency in AC electricity
networks is a key parameter that constantly is required to be
monitored and controlled. This requirement is necessary in the
context of modern technological developments such as
microgrids in order to maintain the active power balance [15]. The frequency control of a microgrid is important during
the transitions between grid-connected and islanding modes,
where high deviations of frequency can occur [6, 7]. The
IEEE standard 1547 defines that the frequency of a microgrid
cannot differ from the grid frequency more than 0.1% during
the reconnection with the grid [6, 7].
Any digital signal processing (DSP) technique used for
grid voltage frequency estimation should be fast, accurate and
robust in the presence of grid disturbances such as voltage
transients, DC offset and harmonics. The technique should
also be simple and computationally efficient which will
contribute to an economical realization of the smart grid
vision [8] for phasor measurement units [9, 10], smart meters
[11, 12], load shedding and restoration functions [13, 14], and
power quality analysis [15].
Many DSP techniques for estimating the grid voltage
fundamental frequency have been reported in the technical
The authors are with the Australian Energy Research Institute, The
University of New South Wales, Sydney, NSW 2052, Australia, and also
with the School of Electrical Engineering and Telecommunications, The
University of New South Wales, Sydney, NSW 2052, Australia (e-mail:
m.reza@unsw.edu.au;
mihai.ciobotaru@unsw.edu.au;
vassilios.agelidis@unsw.edu.au).
1 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
Vk n
v i e
2 ki
N
(1)
i n N 1
Vk n 1
v i e
2 ki
N
(2)
i n N
2 kn
N
(3)
vk n
2 kn
N
(4)
j
1
Vk n 1 e
N
2 kn
N
1
v n v n N
N
(5)
2 k
N
j
1
Vk n 1 e
N
2 kn
N
(6)
2 k
N
1
v n v n N
N
(7)
H k ( z)
vk z
v z
1 zN
N 1 e j 2 k / N z 1
2 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
(8)
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
(1 b ) z
bz
L 1
vk n
2
N
2 kn
2 kn
Re Vk n cos
Im Vk n sin N
N
2
N
Magnitude (abs)
Phase (deg)
0
-90
-180
0
2 n
2 n
Re V1 n cos
Im V1 n sin N
N
50
100
150
200
250
300
Frequency (Hz)
350
400
450
500
Fig. 1. Bode plot of the transfer function H1(z), as given by (8), where k=1.
Step Response
1
Input
Output
0.5
-0.5
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
Time (s)
(9)
0.5
90
1.5
(b)
Voltage Waveform
(p.u.)
0
180
(a)
Magnitude
(p.u.)
The Bode plot of H1(z) (for k=1) at the centre frequency equal
to the grid voltage nominal fundamental frequency (50 Hz) is
shown in Fig. 1. As it can be noticed, H1(z) can reject odd and
even harmonics including the DC offset. Moreover, H1(z)
provides unity amplitude and does not introduce any phase
lag/lead at the fundamental frequency. During the frequency
dynamics, the time-varying harmonics can also be removed
adaptively from the grid voltage fundamental component by
updating the value of N=fs/f, where fs is the sampling
frequency. However, the value of N can be an integer or noninteger under dynamic conditions. H1(z) can reject harmonics
effectively when the value of N is integer. For a non-integer
value of N i.e. L<N<L+1, a rounded function, such as ceil or
floor, can be used to obtain an integer value, where L is a
positive integer, floor(N)=L and ceil(N)=L+1. However, H1(z)
cannot reject harmonics effectively due to the rounding error
produced by the functions floor or ceil. Nevertheless, a linear
interpolation can be performed between the samples v(n-L)
and v(n-L-1) to obtain the value of v(n-N) in order to improve
the performance of H1(z). The linear interpolation operation
can be obtained by (9), where b=N-L [7].
Input
Output
1
0.5
0
-0.5
-1
-1.5
0.98
0.99
1.01
1.02
1.03
1.04
1.05
1.06
Time (s)
Fig. 2. Response of the BPF (for v1) based on the RDFT and IRDFT under (a)
step input. (b) +50% fundamental voltage amplitude step and harmonics, as
given in Table I.
TABLE I
HARMONICS AS A PERCENTAGE OF FUNDAMENTAL COMPONENT BASED ON
SECOND LEVEL TEST CLASS IN IEC STANDARD 61000-4-13
Harmonic
THD
2nd
3rd
4th
5th
7th
3.0 %
8.0 %
1.5%
9.0 %
7.5 %
14.58%
N
2 n
Im V1 n Im V1 n 1 v n v n N sin
3 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
(12)
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
Re V1 n Im V1 n
(13)
v (n)=
v1 n
A1 n
=sin ( n) nTs +1
(14)
u
1
v1u n v1u n 2
(16)
v n 1
u
1
f n
v1u n v1u n 2 S0
(18)
2 Ts C0
f n f 0 f n
v n 1
sin n Ts
(15)
(19)
D.
Block Diagram Implementation of the RDFT-TEO
Technique
The block diagram implementation of the RDFT-TEO
technique for the single-phase grid voltage fundamental
frequency estimation is shown in Fig. 3. The BPF is
implemented based on (11), (12), (13) and (14). On the other
hand, the TEO is implemented based on (18). It can be noticed
from Fig. 3 that v1u ( n ) is obtained by using the frequency
adaptive BPF. The value of v1u ( n ) is then used to estimate the
fundamental frequency deviation by using the TEO. Only
three consecutive voltage samples are required for the TEO
and hence can provide fast estimation of frequency deviation.
The estimated fundamental frequency is also fed back to
obtain the number of voltage samples present in one
cos(2 n / N )
z 1
v n
. .
u
1
v n
z 1
f n 1 order f n
IIR LPF
st
1/2 Ts C0
z 1
.
z 1
(1 b) z L bz L 1
f0
S0
sin(2 n / N )
Band-Pass Filter
f s / .
Fig. 3. Block diagram implementation of the RDFT-TEO technique for the single-phase grid voltage fundamental frequency estimation.
4 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
f n
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
Fundamental
Frequency (Hz)
(a)
47
46
45
44
42
1.99
58
2.01
2.02
2.03
2.04
2.05
Fundamental
Frequency (Hz)
57
56
RDFT-TEO
Actual
55
54
53
52
51
50
1.99
2.01
2.02
2.03
2.04
2.05
0.005
0.01
Error (Hz)
(22)
0.004
0.008
0.003
0.006
0.002
0.004
0.001
0.002
0
42
44
46
48
50
52
54
56
TLPF TRDFT
Time (s)
Fig. 4. Simulated frequency steps estimation. (a) -7.5 Hz (50 Hz to 42.5 Hz).
(a) +7.5 Hz (50 Hz to 57.5 Hz).
(21)
48
43
(20)
49
TLPF 5 / cut
RDFT-TEO
Actual
50
(b)
0
58
(23)
It can be seen from Fig. 2 that the BPF and the TEO
require only few mathematical operations and two
trigonometric functions. A lookup table can be used for the
trigonometric functions for online implementation [52].
Therefore, the RDFT-TEO technique is relatively simple to
implement on a low cost digital signal processor for real-time
frequency estimation of the single-phase grid voltage
waveform.
III.
SIMULATION RESULTS
100%
(24)
IV.
EXPERIMENTAL RESULTS
5 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
vLN
~
Electrical
Grid
Programmable
AC Power Supply
Voltage
Sensor
dSPACE
1103
Personal
Computer
RDFT-OC
50 Hz
f 0.1 Hz/V
RDFT-TEO
RDFT-OC
50 Hz
f 1.0 Hz/V
RDFT-TEO
RDFT-OC
Fig. 8. Experimental case-2: frequency step (-7.5 Hz: 50 to 42.5 Hz) and
harmonics. Channel 1 contains the voltage waveform, and Channels 2 and 3
show the fundamental frequency estimated by the RDFT-TEO and RDFT-OC
techniques, respectively.
6 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
50 Hz
f 1.0 Hz/V
50 Hz
RDFT-TEO
RDFT-OC
f 0.5 Hz/V
RDFT-TEO
RDFT-OC
Fig. 11. Experimental case-5: voltage sag (30%) and harmonics. Channel 1
contains the voltage waveform, and Channels 2 and 3 show the fundamental
frequency estimated by the RDFT-TEO and RDFT-OC techniques,
respectively.
50 Hz
50 Hz
f 0.1 Hz/V
RDFT-TEO
RDFT-OC
f 1.0 Hz/V
RDFT-TEO
RDFT-OC
Fig. 10. Experimental case-4: voltage flicker (5%) and harmonics. Channel
1 contains the voltage waveform, and Channels 2 and 3 show the fundamental
frequency estimated by the RDFT-TEO and RDFT-OC techniques,
respectively.
Fig. 12. Experimental case-6: phase jump (-30) and harmonics. Channel 1
contains the voltage waveform, and Channels 2 and 3 show the fundamental
frequency estimated by the RDFT-TEO and RDFT-OC techniques,
respectively.
7 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
v
v
f 1.0 Hz/V
50 Hz
RDFT-TEO
RDFT-OC
50 Hz
f 0.1 Hz/V
RDFT-TEO
RDFT-OC
Fig. 14. Experimental case-8: voltage sag (-10%), frequency step (-1Hz),
phase jump (-10) and harmonics. Channel 1 contains the voltage waveform,
and Channels 2 and 3 show the fundamental frequency estimated by the
RDFT-TEO and RDFT-OC techniques, respectively.
TABLE III
PERFORMANCE AND COMPUTATIONAL EFFORT COMPARISONS OF THE RDFTTEO AND RDFT-OC TECHNIQUES
RDFT-TEO
Simplicity
RDFTOC
Interdependent loop
Comparison Cases
Computational burden
Harmonics
DC offset
Frequency step
Frequency sweep
Voltage flicker
Voltage sag
Phase jump
Notches and spikes
symbol denotes which technique performs better with respect to the other
one and = symbol indicates both techniques show similar performance for a
case provided in the left column of Table III.
V. CONCLUSIONS
The performance evaluation of a power system
fundamental frequency estimation technique has been
reported in this paper. The technique relies on the Teager
energy operator and a band-pass filter. The three consecutive
samples based Teager energy operator is used to estimate the
fundamental frequency. The band-pass filter is implemented
based on the recursive DFT and inverse recursive DFT. The
technique is computationally efficient and can also estimate a
wide range of frequency variation accurately under the DC
offset and harmonics. The technique also requires less
computational effort, can provide faster estimation and is less
8 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
[20]
[21]
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
9 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/JESTPE.2015.2405094, IEEE Journal of Emerging and Selected Topics in Power Electronics
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
Vassilios G. Agelidis (S
(S'89M'91
91SM
SM'00)
00) was born
in Serres, Greece. He received the B.Eng. degree in
electrical engineering from the Democritus
University of Thrace, Thrace, Greece, in 1988; the
M.S. degree in applied science from Concordia
University, Montreal, QC, Canada, in 1992; and the
Ph.D. degree in electrical engi
engineering
neering from Curtin
University, Perth, Australia, in 1997.
He has worked with Curtin University (1993
(1993
1999); the University of Glasgow, Glasgow, U.K.
(20002004);
(2000
2004); Murdoch University, Perth, Australia
(2005
(20052006);
2006); and the University of Sydney, Sydney, Aust
Australia
ralia (2007
(20072010).
2010).
He is currently the Director of the Australian Energy Research Institute,
School of Electrical Engineering and Telecommunications, University of New
South Wales, Sydney.
Dr. Agelidis was the recipient of the Advanced Research Fellowship
from the U.K.s Engineering and Physical Sciences Research Council in
2004. He was the Vice President for Operations with the IEEE Power
Electronics Society (PELS) from 2006 to 2007. He was an AdCom Member
of IEEE PELS from 2007 to 2009 and the Technical C
Chair
hair of the 39th IEEE
Power Electronics Specialists Conference in 2008 held in Rhodes, Greece.
10 / 10
2168-6777 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.