A Recursive DFT Based Technique for Accurate

Estimation of Grid Voltage Frequency

Md. Shamim Reza, Student Member, IEEE, Mihai Ciobotaru, Member, IEEE, and Vassilios G. Agelidis, Senior Member, IEEE
Australian Energy Research Institute and School of Electrical Engineering and Telecommunications
The University of New South Wales, Kensington, Sydney, NSW 2052, Australia
E-mail: m.reza@student.unsw.edu.au, mihai.ciobotaru@unsw.edu.au, vassilios.agelidis@unsw.edu.au

Abstract-This paper proposes an estimation technique of the

single-phase grid voltage fundamental frequency under distorted
conditions. The proposed frequency estimation technique relies
on a band-pass filter and an algorithm based on three
consecutive samples. The band-pass filter relying on the recursive
discrete Fourier transform (DFT) and inverse recursive DFT is
used to extract the instantaneous grid voltage fundamental
component. The algorithm based on three consecutive samples is
then used to estimate the grid voltage fundamental frequency.
The estimated frequency is used to synchronize the window size
of the band-pass filter with the period of the grid voltage
waveform. The proposed frequency estimation technique is
relatively simple, computationally efficient and can also estimate
the fundamental frequency accurately from a grid voltage
waveform distorted by DC offset and harmonics. Simulation
results are presented to verify the performance of the proposed
technique.
Index Terms: Discrete Fourier transform (DFT), frequency
estimation, recursive DFT, and single-phase voltage system.

I. INTRODUCTION
The grid voltage waveform becomes distorted mainly by
harmonics due to the widespread use of the nonlinear loads
[1]. The grid voltage waveform may also contain DC offset
due to the analog-to-digital conversion for fixed point DSP
application, the grid faults or the saturation phenomenon in
current transformer [2, 3]. Moreover, the grid voltage
waveform is time-varying in nature due to the continuous
change of the load conditions, system configurations and rapid
proliferation of the distributed energy sources.
The grid voltage frequency is a fundamental operational
parameter for effective power control, load shedding, load
restoration, safety, stability and efficiency of the electrical grid
[4, 5]. The grid voltage fundamental frequency is time-varying
in nature mainly due to the mismatch of the power generation
and consumption and hence reflects the dynamic energy
balance between the load demand and generated power [4, 5].
The variation of the fundamental frequency is much more
likely to occur, if the loads are supplied by a generator isolated
from the grid [4]. The fundamental frequency decreases when
large loads are connected or when a large generation source
goes offline [4]. On the other hand, the grid frequency
increases when the generation exceeds the consumption i.e.
due to the disconnection of large block of loads or when a
large generation source comes online. Under frequency or
over frequency relays are used to automatically shed blocks of

978-1-4799-0224-8/13/$31.00 2013 IEEE

loads or some generation, respectively, for restoring the

frequency to its operating range. An accurate and fast
estimation of frequency is required for proper operation of the
under frequency and over frequency relays. The reliable
estimation of the fundamental frequency is also required for
the grid-connected power converters [6], monitoring and
power quality analysis [7], to mention just a few. Thus, a
suitable computationally efficient digital signal processing
(DSP) technique is required which would be able to deal with
a distorted grid voltage waveform to provide a fast and
accurate estimation of fundamental frequency.
In the technical literature, zero crossing detection (ZCD) [8],
phase-locked loop (PLL) [9], frequency-locked loop (FLL)
[10], discrete Fourier transform (DFT) [11, 12], least-square
(LS) [13, 14], least-mean-square (LMS) [15], Kalman filter
(KF) [16], differentiation filter [17, 18] and finite-impulseresponse (FIR) filter [4, 19-23] are some of the used DSP
techniques for the estimation of grid voltage frequency.
The ZCD technique is simple to implement but may provide
inaccurate results under the noise condition [8]. The challenge
of the single-phase PLL is that it requires a virtual orthogonal
voltage system to be generated [9, 24], unlike the three-phase
systems where the orthogonal voltage system is provided by
the Clarke transformation. Moreover, the presence of DC
offset and harmonics introduces oscillations into the PLL
estimated parameters [2, 3, 25]. Therefore, the PLL requires
an optimal tuning of the controller parameters to obtain a
trade-off between good dynamics and estimation accuracy.
Another drawback is that the PLL presents a large
overshoot/undershoot in the frequency estimation during the
phase angle jumps under grid faults [26]. The performance of
the FLL based on the quadrature signal generator (QSG) using
a second order generalized integrator (SOGI) is sensitive to
the presence of the DC offset and it also requires a
compromise between good dynamics and estimation accuracy.
The DFT, as reported in [11], is not suitable for high number
of harmonics. On the other hand, during the time-varying
frequency cases, the DFT requires large moving window of
fixed size to avoid the interference caused by the low order
harmonics , as reported in [12]. The LS technique may suffer
from matrix singularities and also provides inaccurate results
from a small size window [27]. The LMS technique requires
orthogonal waveforms and may also diverge due to the
inaccurate tuning of the controller parameter [15]. The KF is a
computationally demanding technique due to the evaluation of

6420

transcendental functions [16]. Moreover, the performance of

the KF depends on the proper tuning of the parameters [16].
The adaptive technique, as reported in [17, 18], estimates
frequency from the instantaneous phase angle but requires
window based differentiation filter. On the other hand, the
frequency estimation technique relying on the three
consecutive samples requires the implementation of the
computationally complex FIR filter in order to reject the
harmonics and noise effect [4, 19-23].
The objective of this paper is to propose an estimation
technique of the single-phase grid voltage fundamental
frequency under distorted conditions. The proposed technique
relies on an adaptive band-pass filter (BPF) and an algorithm
based on three consecutive samples. The BPF is implemented
using recursive DFT and inverse recursive DFT. The three
consecutive samples based algorithm is used to estimate the
fundamental frequency from the instantaneous grid voltage
fundamental component. The proposed technique is relatively
simple, computationally efficient and can also reject the
negative effects caused by the presence of the DC offset and
harmonics.
The rest of the paper is organized as follows: the proposed
fundamental frequency estimation technique is presented in
section II. Section III contains the simulation performance of
the proposed technique. Finally, conclusions of the paper are
summarized in section IV.

v n

v n v0

h 1,2,...

vh n

v h n Ah n sin h n nTs h

v0, M, =2f, f and Ts are the DC offset, order of harmonics,

fundamental angular frequency, fundamental frequency and
sampling
time
period,
respectively,
and
Ah(n),
h(n)=h(n)nTs+h and h are the amplitude, instantaneous
and initial phase angle of the h (h=1,2,,M) frequency
component, respectively.
The proposed DSP technique, as shown in Fig. 1, is used to
track the fundamental frequency of the grid voltage waveform
given by (1). As it can be seen, the adaptive digital BPF based
on the recursive DFT and inverse recursive DFT is used to
extract the amplitude normalized instantaneous grid voltage
fundamental component v1u(n), where the superscript u
indicates that the instantaneous fundamental component has
unity amplitude [28]. The fundamental frequency is tracked by
using the algorithm based on the three consecutive samples of
the amplitude normalized instantaneous grid voltage
fundamental component. The actual fundamental frequency
information required by the adaptive BPF is also updated by
the estimated frequency. It can be noticed that the number of

f n

f s / .

samples (N) in one fundamental time period is calculated as

follows: N =f s /f , where fs=1/Ts and f are the sampling and
estimated fundamental frequencies, respectively.
A. Digital Band-Pass Filter
In this subsection, the implementation of the digital BPF
based on the recursive DFT and inverse recursive DFT is
briefly discussed [28]. The DFT of the grid voltage waveform,
as given by (1), at the (n-1)th sampling instant can be
expressed by [28]
V k n 1

n 1

v i e

2 k i 1

(2)

in N

where k=0,1,2,,N-1 and j is the complex operator. The value

of N in (2) is chosen as the number of samples in one
fundamental time period, hence the frequency resolution (f)
of the DFT is fundamental frequency i.e. f=f. Similar to (2),
the DFT of (1) at the nth sampling instant can be expressed by
Vk n

v i e

2 k i 1

(3)

i n N 1

By combining (2) and (3), the following recursive relation is

obtained.

(1)

where

Fundamental
Frequency
Estimation

Fig. 1. Block diagram of the proposed single-phase grid voltage fundamental

frequency estimation technique.

II. PROPOSED FREQUENCY ESTIMATION TECHNIQUE

The DC offset and harmonically distorted single-phase grid
voltage waveform, v(n), at the nth sampling instant can be
expressed by

Adaptive Digital v1u n

Band-Pass
Filter

Vk n Vk n 1 v n v n N e

2 k n 1
N

(4)

Expression (4) is called the recursive DFT and can be used to

estimate the spectral content present in (1). On the other hand,
the time domain signal at the single frequency kf can be
extracted by taking the inverse transform of (4) i.e. the inverse
recursive DFT. Therefore, based on the inverse recursive DFT,
the time-domain signal at the frequency kf can be expressed
by [28]
vk n G k Vk n e

2 k n 1
N

(5)

where
1 / N for k 0, N /2
Gk
2 / N otherwise

The digital BPF at the center frequency kf can be

implemented by cascading the recursive DFT and the inverse
recursive DFT, as given by (4) and (5), respectively, in series.
The implementation of the digital BPF at the center frequency
f (where k=1) based on the recursive DFT and the inverse
recursive DFT is shown in Fig. 2. As it can be seen, the BPF is
simple to implement and requires only few mathematical

6421

1
N

1 zN

1 zN

j 2 / N 1
j 2 / N 1
z
1 e
z
1 e

(6)

The detail derivation of (6) can be found in [28]. The Bode

plot of (6) is shown in Fig. 3. As it can be seen, the BPF can
reject all the even and odd harmonics including the DC offset.
Moreover, the BPF does not introduce any lag/lead phaseangle at the fundamental frequency. Therefore, during the
time-varying fundamental frequency cases, the DC offset and
harmonics of the input grid voltage waveform can be
completely eliminated, if the value of N corresponding to the
actual fundamental frequency is updated. If the value of N is
not integer i.e. L<N<L+1, where L is a positive integer, then
linear interpolation is performed between the samples v(n-L)
and v(n-L-1) to obtain v(n-N). Thus, based on the adaptive
tuning, the output of the presented digital BPF {v1(n)}, as
shown in Fig. 2, at the nth sampling instant can be expressed
by
v1 n A1 n sin n nTs 1

(7)

Re V1 n Im V1 n

v1 n
2
N

Re V1 n + Im V1 n

v1 n

2/N

Im V1 n

N
Fig. 2. Digital band-pass filter based on the recursive DFT and inverse
recursive DFT to extract the instantaneous grid voltage fundamental
frequency component {v1(n)}.

0
-20
-40
-60
-80
-100
180
90
0
-90
-180
10

10

10

10

Frequency (Hz)

Fig. 3. Bode plot of the transfer function H(z) i.e. the frequency response of
the BPF based on the recursive DFT and inverse recursive DFT shown in Fig.
2, where f=50 Hz, fs=10 kHz and N=200.

cos{2 (n -1) / N }

(8)

v n

v1u n

. 2 . 2

zN

Therefore, the amplitude normalized instantaneous

fundamental voltage component can be obtained by
v1u ( n )=

Re V1 n

sin{2 (n -1) / N }

It can be seen from Fig. 2 that the fundamental voltage

amplitude can also be estimated from Re{V1(n)} and
Im{V1(n)}, respectively, and is given by
2
A1 n
N

z 1

v n

Magnitude (dB)

H ( z)

cos{2 (n -1) / N }

Phase (deg)

operations and one trigonometric function (other trigonometric

function can be estimated by cos{2(n-1)/N}=[1-sin2{2(n1)/N}] or vice versa). As it can also be seen, Re{V1(n)} and
Im{V1(n)} are the real and imaginary parts of the recursive
DFT of the fundamental frequency component. Similar to Fig.
2, other BPF for different values of k>1 can be implemented to
extract the instantaneous single harmonic component of
interest.
The frequency response of the BPF can be obtained from the
z-transform of (4) and (5), respectively. Therefore, the ztransform of the transfer functions {v1(n)/v(n), k=1}, as shown
in Fig. 2, can be expressed by [28]

sin{2 (n -1) / N }

=sin ( n ) nTs + 1

(9)
The estimation of v1u(n) by using the recursive DFT and
inverse recursive DFT based BPF is shown in Fig. 4. As it can
be seen, the amplitude factor 2/N is removed which does not
have any effect on the estimation of v1u(n).
The BPF based on the recursive DFT and inverse recursive
DFT suffers from the accumulation errors in the estimated
amplitude and phase angle due to the input voltage dynamics
[29]. However, in the proposed technique, the amplitude
accumulation error is rejected due to the use of the amplitude
normalized instantaneous fundamental voltage waveform. On

Fig. 4. Digital band-pass filter based on the recursive DFT and inverse
recursive DFT to extract the amplitude normalized instantaneous grid voltage
fundamental frequency component {v1u(n)}.

the other hand, the effect of phase angle accumulation error is

neglected based on the assumption that the accumulated phase
angle error is constant within three consecutive samples. The
constant accumulated phase angle error does not have any
effect in the frequency estimation based on three consecutive
samples.
B. Fundamental Frequency Estimation
Based on the constant fundamental parameters within three

6422

v1u n

z 1

z 1
.

consecutive samples of (7), the following relation can be

expressed which is defined as the Teager Energy Operator
(TEO) [30-35].

sin 1 .

LPF

f n

1/(2Ts )

A12 n sin n Ts v12 n 1 v1 n v1 n 2 (10)

Fig. 5. Fundamental frequency estimation by using (12) relying on the three

consecutive samples of {v1u(n)}.

It is reported in the technical literature that the TEO concept

can be used to estimate the fundamental envelope of the grid
voltage waveform [30-35]. However, the TEO concept is used
in this paper to estimate the fundamental frequency. Based on
the amplitude normalized instantaneous fundamental voltage
waveform v1u(n), as obtained by (9), expression (10) can be
written as

5% of 3rd, 4% of 5th, 3% of 7th, 2% of 9th and 1% of 11th

harmonics, thus leading to a total harmonic distortion (THD)
of 7.42%.

sin n Ts v1u n 1

v1u n v1u n 2

(11)

Therefore, the fundamental frequency can be estimated from

(11) by using only three consecutive samples of the amplitude
normalized instantaneous fundamental voltage waveform and
is expressed by

v1u n v1u n 2

2 Ts

v n 1

The implementation of (12) is shown in Fig. 5. As it can be

seen, equation (12) requires three consecutive samples and can
provide fast estimation of frequency. However, as there is an
interdependent loop in Fig. 1, the faster tracking of frequency
as compared to the estimation of instantaneous fundamental
component will affect the stability of the proposed technique.
Therefore, a delay is used for the loop providing the estimated
frequency and hence a first-order infinite-impulse-response
(IIR) based low-pass filter (LPF) is used, as it can be seen in
Fig. 5 [36]. The time constant of the LPF determines the
dynamics of the estimated fundamental frequency. The
estimation of the fundamental frequency requires few
mathematical operations and one inverse sine function, and is
also simple to implement, as can be seen in Fig. 5.

(a)

(12)

Grid Voltage
Waveform (p.u.)

u
1

0.5

-0.5

-1
1
50.02

(b)

f n

sin 1

Case-2: Steady-State with DC Offset and Harmonics

In this case, the grid voltage waveform is distorted by 5%

Fundamental
Frequency (Hz)

Case-1: Steady-State with Harmonics

The steady-state grid voltage waveform, as shown in Fig.
6(a), contains 7.42% THD. The estimation of fundamental
frequency at steady-state by using the proposed technique is
shown in Fig. 6(b). As it can be seen, the proposed technique
can provide accurate estimation of fundamental frequency
from a grid voltage waveform containing harmonics.

1.01

1.02

1.03

1.04

1.05

1.06

Actual

Proposed

50.01

50

49.99

49.98
1

1.01

1.02

1.03

Time (s)

1.04

1.05

1.06

Fig. 6. Case-1: Steady-state with harmonics. (a) Grid voltage waveform. (b)
Fundamental Frequency.

III. SIMULATION RESULTS

0.5

-0.5

-1
1
50.02

(b)
Fundamental
Frequency (Hz)

In this section, the performance of the proposed frequency

estimation technique is tested in MATLAB/Simulink under
several cases as follows:
i. Steady-state with harmonics (Case-1)
ii. Steady-state with DC offset and harmonics (Case-2)
iii. Frequency step and harmonics (Case-3)
iv. Frequency sweep and harmonics (Case-4)
v. Voltage flicker and harmonics (Case-5)
vi. Voltage sag and harmonics (Case-6)
The sampling and nominal grid fundamental frequencies are
chosen as 10 kHz and 50 Hz, respectively. As the window size
of the BPF based on recursive DFT and inverse recursive DFT
is one fundamental time period, hence the time constant of the
first-order IIR based LPF, as shown in Fig. 5, is chosen as
20ms (=1/50). The fundamental component of the grid voltage
waveforms presented in all the case studies are distorted by

(a)
Grid Voltage
Waveform (p.u.)

1.01

1.02

1.03

1.04

1.05

1.06

Actual

Proposed

50.01

50

49.99

49.98
1

1.01

1.02

1.03

Time (s)

1.04

1.05

1.06

Fig. 7. Case-2: Steady-state with DC offset and harmonics. (a) Grid voltage
waveform. (b) Fundamental Frequency.

6423

51.2

Actual

Proposed

(a)
Grid Voltage
Waveform (p.u.)

Fundamental
Frequency (Hz)

51
50.8
50.6
50.4

-0.5

-1
0.8
50.04

50.2

1.02

1.04

1.06

1.08

1.1

1.12

1.14

Time (s)

Fig. 8. Case 3: Frequency step and harmonics.

50.2

Actual

Proposed

(b)
Fundamental
Frequency (Hz)

50
49.8
0.98

0.5

1.2

1.4

1.6

1.8

2.2

Actual

Proposed

50.02

50

49.98

49.96
0.8

49.8

49.4

1.2

1.4

1.6

1.8

Time (s)

2.2

49

1.05

1.1

1.15

1.2

1.25

Time (s)

Fig. 9. Case 4: Frequency sweep and harmonics.

DC offset and 7.42% THD, as can be seen in Fig. 7(a). The

steady-state estimation of the fundamental frequency by using
the proposed technique is shown in Fig. 7(b). As it can be
noticed, the performance of the proposed frequency estimation
technique is not affected by the grid voltage waveform
distorted by DC offset and harmonics.
Case-3: Frequency Step and Harmonics
In this case, the performance of the proposed technique is
tested under +1 Hz fundamental frequency step and 7.42%
THD. The estimation of fundamental frequency step by using
the proposed technique is depicted in Fig. 8. As it can be
noticed, the proposed technique takes 76.6 ms as a settling
time and can track the frequency step accurately. After the
dynamics, the performance of the proposed technique is also
non-sensitive to the presence of the harmonics at off-nominal
fundamental frequency of 51 Hz, as can be seen from Fig. 8.
Case-4: Frequency Sweep and Harmonics
The grid voltage frequency varies slowly due to the large
inertia of the rotating shaft of the power generators. A -10
Hz/s fundamental frequency sweep is considered into the grid
voltage waveform containing 7.42% THD. The fundamental
frequency sweep estimation by the proposed technique is
shown in Fig. 9. As it can be noticed, the proposed technique
can track the fundamental frequency sweep accurately.
Case-5: Voltage Flicker and Harmonics
The grid voltage waveform, as shown in Fig. 10(a), contains
voltage flicker and 7.42% THD. The frequency and amplitude
of the triangular voltage flicker are 1 Hz and 5% of
fundamental amplitude, respectively. For this case, the

(a)
Grid Voltage
Waveform (p.u.)

49.2

48.8
0.95

Fig. 10. Case-5: Voltage flicker and harmonics. (a) Grid voltage waveform.
(b) Fundamental Frequency.

49.6

0.5

-0.5

-1
0.98
50.2

(b)
Fundamental
Frequency (Hz)

Fundamental
Frequency (Hz)

50

1.02

1.04

1.06

1.08

1.1

Proposed

1.12

1.14

Actual

50
49.8
49.6
49.4
49.2
0.98

1.02

1.04

1.06

Time (s)

1.08

1.1

1.12

1.14

Fig. 11. Case-6: Voltage sag and harmonics. (a) Grid voltage waveform. (b)
Fundamental Frequency.

fundamental frequency estimation is shown in Fig. 10(b). As it

can be seen, the performance of the proposed technique is
affected and provides ripples into the estimated frequency
during the voltage flicker.
Case-6: Voltage Sag and Harmonics
A grid voltage waveform containing 20% voltage sag and
7.42% THD is shown in Fig. 11(a). The estimation of the
fundamental frequency by using the proposed technique is
shown in Fig. 11(b). As it can be noticed, the proposed
technique provides undershoot during the voltage sag.
IV. CONCLUSIONS
A single-phase grid voltage fundamental frequency
estimation technique has been proposed in this paper. The
proposed frequency estimation technique relies on a band-pass
filter and three consecutive samples based algorithm. The
band-pass filter was implemented based on the recursive DFT
and inverse recursive DFT. The proposed estimation technique

6424

is relatively simple, computationally efficient and can also

estimate the fundamental frequency accurately. Moreover, the
technique can reject the negative effects caused by the
presence of the DC offset and harmonics. The presented
simulation results confirmed the effective applications of the
proposed frequency estimation technique.
REFERENCES
[1] L. Asiminoaei, F. Blaabjerg, and S. Hansen, "Detection is key Harmonic detection methods for active power filter applications," IEEE
Ind. App. Mag., vol. 13, no. 2, pp. 22-33, July/Aug. 2007.
[2] M. Ciobotaru, R. Teodorescu, and V. G. Agelidis, "Offset rejection for
PLL based synchronization in grid-connected converters," in Proc. 23rd
Ann. IEEE Applied Power Elect. Conf. Expo. (APEC), 2008, pp. 16111617.
[3] M. Karimi-Ghartemani, S. A. Khajehoddin, P. K. Jain, A. Bakhshai, and
M. Mojiri, "Addressing DC component in PLL and notch filter
algorithms," IEEE Trans. Power Elect., vol. 27, no. 1, pp. 78-86, Jan.
2012.
[4] M. D. Kusljevic, "A simple recursive algorithm for simultaneous
magnitude and frequency estimation," IEEE Trans. Instrum. Measur.,
vol. 57, no. 6, pp. 1207-1214, June 2008.
[5] A. J. Roscoe, G. M. Burt, and J. R. McDonald, "Frequency and
fundamental signal measurement algorithms for distributed control and
protection applications," IET Gen. Trans. Dist., vol. 3, iss. 5, pp. 485495, 2009.
[6] F. Blaabjerg, R. Teodorescu, M. Liserre, and A. V. Timbus, "Overview
of control and grid synchronization for distributed power generation
systems," IEEE Trans. Ind. Elect., vol. 53, no. 5, pp. 1398-1409, Oct.
2006.
[7] I. Sadinezhad and V. G. Agelidis, "Frequency adaptive least-squaresKalman technique for real-time voltage envelope and flicker estimation,"
IEEE Trans. Ind. Elect., vol. 59, no. 8, pp. 3330-3341, Aug. 2012.
[8] V. Friedman, "A zero crossing algorithm for the estimation of the
frequency of a single sinusoid in white noise," IEEE Trans. Signal Proc.,
vol. 42, no. 6, pp. 1565-1569, June 1994.
[9] M. Ciobotaru, V. G. Agelidis, R. Teodorescu, and F. Blaabjerg,
"Accurate and less-disturbing active antiislanding method based on PLL
for grid-connected converters," IEEE Trans. Power Elect., vol. 25, no. 6,
pp. 1576-1584, June 2010.
[10] P. Rodriguez, A. Luna, M. Ciobotaru, R. Teodorescu, and F. Blaabjerg,
"Advanced grid synchronization system for power converters under
unbalanced and distorted operating conditions," in Proc. 32nd Ann. Conf.
IEEE Ind. Elect. (IECON), 2006, pp. 5173-5178.
[11] J.-Z. Yang and C.-W. Liu, "A precise calculation of power system
frequency," IEEE Trans. Power Del., vol. 16, no. 3, pp. 361-366, July
2001.
[12] M. S. Reza, M. Ciobotaru, and V. G. Agelidis, "Tracking of time-varying
grid voltage using DFT based second order generalized integrator
technique," in Proc. of IEEE Int. Conf. Power Sys. Tech. (POWERCON),
2012, pp. 1-6.
[13] R. Chudamani, K. Vasudevan, and C. S. Ramalingam, "Real-time
estimation of power system frequency using nonlinear least squares,"
IEEE Trans. Power Del., vol. 24, no. 3, pp. 1021-1028, July 2009.
[14] A. Abdollahi and F. Matinfar, "Frequency estimation: A least-squares
new approach," IEEE Trans. Power Del., vol. 26, no. 2, pp. 790-798,
Apr. 2011.
[15] A. K. Pradhan, A. Routray, and A. Basak, "Power system frequency
estimation using least mean square technique," IEEE Trans. Power Del.,
vol. 20, no. 3, pp. 1812-1816, July 2005.
[16] A. Routray, A. K. Pradhan, and K. P. Rao, "A novel Kalman filter for
frequency estimation of distorted signals in power systems," IEEE Trans.
Instrum. Meas., vol. 51, no. 3, pp. 469-479, June 2002.

[17] R. Zivanovic, "An adaptive differentiation filter for tracking

instantaneous frequency in power systems," IEEE Trans. Power Del.,
vol. 22, no. 2, pp. 765-771, Apr. 2007.
[18] M. S. Reza, M. Ciobotaru, and V. G. Agelidis, "Robust estimation of
real-time single-phase grid voltage frequency under distorted grid
conditions," in Proc. of IEEE ECCE Asia Down Under, 2013, pp. 948954.
[19] A. Lopez, J.-C. Montano, M. Castilla, J. Gutierrez, M. D. Borras, and J.
C. Bravo, "Power system frequency measurement under nonstationary
situations," IEEE Trans. Power Del., vol. 23, no. 2, pp. 562-567, Apr.
2008.
[20] M. D. Kusljevic, "A simple recursive algorithm for frequency
estimation," IEEE Trans. Instrum. Measur., vol. 53, no. 2, pp. 335-340,
Apr. 2004.
[21] M. D. Kusljevic, J. J. Tomic, and L. D. Jovanovic, "Frequency estimation
of three-phase power system using weighted-least-square algorithm and
adaptive FIR filtering," IEEE Trans. Instrum. Measur., vol. 59, no. 2, pp.
322-329, Feb. 2010.
[22] M. D. Kusljevic, "Simultaneous frequency and harmonic magnitude
estimation using decoupled modules and multirate Sampling," IEEE
Trans. Instrum. Measur., vol. 59, no. 4, pp. 954-962, Apr. 2010.
[23] M. S. Reza, M. Ciobotaru, and V. G. Agelidis, "Real-time estimation of
single-phase grid voltage frequency using a modulating function based
technique," in Proc. of IEEE ECCE Asia Down Under, 2013, pp. 664669.
[24] M. Ciobotaru, R. Teodorescu, and F. Blaabjerg, "A new single-phase
PLL structure based on second order generalized integrator," in Proc.
37th IEEE Power Elect. Spec. Conf. (PESC), 2006, pp. 1-6.
[25] M. Karimi-Ghartemani and M. R. Iravani, "Measurement of
harmonics/inter-harmonics of time-varying frequencies," IEEE Trans.
Power Del., vol. 20, no. 1, pp. 23-31, Jan. 2005.
[26] M. K. Ghartemani, S. A. Khajehoddin, P. K. Jain, and A. Bakhshai,
"Problems of startup and phase jumps in PLL systems," IEEE Trans.
Power Elect., vol. 27, no. 4, pp. 1830-1838, Apr. 2012.
[27] M. S. Reza, M. Ciobotaru, and V. G. Agelidis, "Accurate estimation of
grid voltage parameters using singular value decomposition technique,"
in Proc. of 3rd IEEE Symp. Power Elect. Distrib. Gen. Sys. (PEDG),
2012, pp. 794-800.
[28] B. P. McGrath, D. G. Holmes, and J. J. H. Galloway, "Power converter
line synchronization using a discrete Fourier transform (DFT) based on a
variable sample rate," IEEE Trans. Power Elect., vol. 20, no. 4, pp. 877884, July, 2005.
[29] H. A. Darwish and M. Fikri, "Practical considerations for recursive DFT
implementation in numerical relays," IEEE Trans. Power Del., vol. 22,
no. 1, pp. 42-49, Jan. 2007.
[30] J. F. Kaiser, "On a simple algorithm to calculate the `energy' of a signal,"
in Proc. Int. Conf. Acoustics, Speech, and Sig. Proc. (ICASSP-90), 1990,
pp. 381-384.
[31] T. K. Abdel-Galil, E. F. El-Saadany, and M. M. A. Salama, "Online
tracking of voltage flicker utilizing energy operator and Hilbert
transform," IEEE Trans. Power Del., vol. 19, no. 2, pp. 861-867, Apr.
2004.
[32] M. A. Eldery, E. F. El-Saadany, and M. M. A. Salama, "A novel energy
operator algorithm for voltage envelope tracking," IEEE Trans. Power
Sys., vol. 20, no. 1, pp. 510-512, Feb. 2005.
[33] M. A. Eldery, E. F. El-Saadany, and M. M. A. Salama, "A simple energy
operator computational method for voltage flicker assessment," IEEE
Trans. Power Del., vol. 21, no. 3, pp. 1743-1750, July 2006.
[34] I. Kamwa, A. K. Pradhan, and G. Joos, "Robust detection and analysis of
power system oscillations using the Teager-Kaiser energy operator,"
IEEE Trans. Power Sys., vol. 26, no. 1, pp. 323-333, Feb. 2011.
[35] A. Subasi, A. S. Yilmaz, and K. Tufan, "Detection of generated and
measured transient power quality events using Teager energy operator,"
Elsevier Energy Conv. amd Management, vol. 52, pp. 1959-1967, 2011.
[36] W. Yi Fei and L. Yun Wei, "Three-phase cascaded delayed signal
cancellation PLL for fast selective harmonic detection," IEEE Trans. Ind.
Elect., vol. 60, no. 4, pp. 1452-1463, Apr. 2013.

6425

Powered by TCPDF (www.tcpdf.org)