MONITORING OF LOW FREQUENCY
OSCILLATIONS IN POWER SYSTEM
Presented by:-
1. Javed Mustakim Borbora (E-25/16)
2. Syed Rounak Nishar (E-23/16)
3. Pranab Jyoti Kalita (E-21/16)
4. Arshad Asfak Hussain (E-24/16)
Under the guidance of:-
Dr. Shekha Rai
Assistant Professor
DEPARTMENT OF ELECTRICAL ENGINEERING JORHAT
ENGINEERING COLLEGE
CONTENTS
Introduction
Problem Formulation
Proposed Methodology
Simulation Process
Analysis of Result
Conclusion and Future Work
References
INTRODUCTION
Power systems are viewed as complex and non linear system due to increase of
loads which leads to a number of instability problems.
When instability occur in synchronous machines due to a small disturbance is
known as small signal instability.
Under this, lack of damping torque results in Low Frequency Oscillations(LFO).
Low frequency oscillations which are generator rotor angle oscillations having a
frequency between 0.1 -2.0 Hz.
The root cause of these oscillations are the unbalance between power demand and
available power at a period of time.
CONTD.
In earlier LFO are non observable as generators are closely connected to
loads.
Nowadays due to large demand, a huge power is transmitted through long
transmission lines which results in increasing oscillations.
The LFO can be classified as local and inter-area mode-
Local modes - Oscillations occurred only to the small part of the power system.
Inter area modes - It generally occurs in weak interconnected power systems
through long tie lines.
PROBLEM FORMULATION
A complete understanding of low frequency oscillations would help in finding
effective remedial measures and ways to control them.
Detection of critical modes is important to suppress the oscillations, which
would otherwise lead to system collapse or blackout.
Several techniques such as FFT, ERA, Matrix Pencil are fail to provide the actual
modal parameters in real time.
Prony method can directly estimate the signal’s frequency, phase, amplitude and
damping. The use of online based Prony algorithm can estimate of the desired
critical modes with noise also.
PROPOSED METHODOLOGY
We have used a Prony based approach is used to estimate the low-frequency
oscillatory modes in power system.
Prony method extracts valuable information from a uniformly sampled signal
and builds a series of damped complex exponentials or sinusoids. This allows
for the estimation of frequency, amplitude, phase and damping components of
a signal.
The efficiency of the proposed method is demonstrated in the presence of
different signal-to-noise ratio (SNR) of test signal to detect natural frequencies
and attenuation factor.
MATHEMATICAL DEDUCTION
The Prony method approximates the signal s(n) as a set of exponential
components as shown below
,n (1)
Where (n) is the noise added.
For real signal there are two complex conjugate exponents
Here, and are the amplitude, attenuation factor and frequency respectively.
The work of proposed model as shown in (2)
A=
=− (2)
A.g = h (3)
g = h (4)
Where, A, and g represents the Hankel matrix, pseudo inverse of A and the vector
of backward prediction coefficient respectively. For this purpose, an index
described in has been used.
Q(i)= (5)
where Q(i) and are the monotonously increasing function and the singular value
of the i-th term respectively. The coefficients obtained from (4) is then used to
determine the as shown in (6)
1+ (6)
Now since = , j = 1, 2, …….,L, the power system modes can be easily identified.
As from here can be calculate and which will give and the attenuation and
frequency of the signal.
SIMULATION PROCESS
For Local and Inter area mode a test signal is passed through Prony
program code where at first, noises are added to the signal through noise
generator function.
After the formation of Hankel matrix the Prony function is simulated for
10000 times for both the modes to finding out the modes of the signal
which are frequency and damping.
From the simulated damping matrix, we are finding the mean of data and
also variance of data is calculated to check deviation from actual value of
damping. These are also represented through figures. Similarly for
frequency mean, variance and figures are taken.
FLOWCHART OF THE ALGORITHM
Noise Hankel Matrix Prony
Test Signal Generation Form Function Call
Calculation of Mean
Estimating Damping
and standard
and Frequency
deviation
Analysis of Results
TABLE 1 FOR LOCAL MODE
TRUE VALUE: FREQUENCY = 1.8 Hz, ATTENUATION= - 0.1
SNR Damping Frequency
(dB)
Mean Variance Mean Variance
50 -0.1000 8.1616 e-09 1.8000 2.134 e-10
40 -0.1000 8.3701 e-08 1.8000 2.1201 e-09
30 -0.0998 7.9717 e-07 1.8000 2.1130 e-08
20 -0.0985 8.5116 e-06 1.8000 2.1192 e-07
Analysis of Results
TABLE 2 FOR INTER AREA MODE
TRUE VALUE: FREQUENCY = 0.5 Hz, ATTENUATION = -0.06
SNR Damping Frequency
(dB)
Mean Variance Mean Variance
50 -0.0600 6.0359 e-09 0.5000 1.6499 e-10
40 -0.0600 5.8193 e-08 0.5000 1.6160 e-09
30 -0.0598 5.9156 e-07 0.5000 1.5976 e-08
20 -0.0584 6.1614 e-06 0.5000 1.6454 e-07
ANALYSIS OF RESULTS
I. GRAPH FOR LOCAL MODE
A. DISTRIBUTION OF DAMPING AT SNR =50 DB
ANALYSIS OF RESULTS
I. GRAPH FOR LOCAL MODE
B. DISTRIBUTION OF DAMPING AT SNR =40 DB
ANALYSIS OF RESULTS
I. GRAPH FOR LOCAL MODE
C. DISTRIBUTION OF DAMPING AT SNR =30 DB
ANALYSIS OF RESULTS
I. GRAPH FOR LOCAL MODE
D. DISTRIBUTION OF DAMPING AT SNR =20 DB
ANALYSIS OF RESULTS
II. GRAPH FOR LOCAL MODE
A. DISTRIBUTION OF FREQUENCY AT SNR =50 DB
ANALYSIS OF RESULTS
II. GRAPH FOR LOCAL MODE
B. DISTRIBUTION OF FREQUENCY AT SNR =40 DB
ANALYSIS OF RESULTS
II. GRAPH FOR LOCAL MODE
C. DISTRIBUTION OF FREQUENCY AT SNR =30 DB
ANALYSIS OF RESULTS
II. GRAPH FOR LOCAL MODE
D. DISTRIBUTION OF FREQUENCY AT SNR =20 DB
ANALYSIS OF RESULTS
III. GRAPH FOR INTER AREA MODE
A. DISTRIBUTION OF DAMPING AT SNR =50 DB
ANALYSIS OF RESULTS
III. GRAPH FOR INTER AREA MODE
B. DISTRIBUTION OF DAMPING AT SNR =40 DB
ANALYSIS OF RESULTS
III. GRAPH FOR INTER AREA MODE
C. DISTRIBUTION OF DAMPING AT SNR =30 DB
ANALYSIS OF RESULTS
III. GRAPH FOR INTER AREA MODE
D. DISTRIBUTION OF DAMPING AT SNR =20 DB
ANALYSIS OF RESULTS
IV. GRAPH FOR INTER AREA MODE
A. DISTRIBUTION OF FREQUENCY AT SNR =50 DB
ANALYSIS OF RESULTS
IV. GRAPH FOR INTER AREA MODE
B. DISTRIBUTION OF FREQUENCY AT SNR =40 DB
ANALYSIS OF RESULTS
IV. GRAPH FOR INTER AREA MODE
C. DISTRIBUTION OF FREQUENCY AT SNR =30 DB
ANALYSIS OF RESULTS
IV. GRAPH FOR INTER AREA MODE
D. DISTRIBUTION OF FREQUENCY AT SNR =20 DB
ANALYSIS OF RESULTS
Table 1 and Table 2 shows mean and variance of attenuation factor and
frequency of the test signals corresponding to local mode and inter area mode
respectively.
The proposed method is able to estimate the model parameters accurately as
depicted in the Tables. It has seen that for higher SNRs such as 30dB and 20dB
the result of damping and frequency deviated from its true value, so variance
increases.
Hence, it has been concluded that as the SNR increases error will decrease.
CONCLUSION
The project is aimed to design a Prony based
estimator for Power System stability. The Prony
method provides an accurate estimate of modal
parameters as shown in the simulation results.
FUTURE WORK
Modification of existing Prony algorithm in
presence of high level noise and missing
data.
BIBLIOGRAPHY
[1] P. Kundur, Power System Stability and Control,. New York: McGrawHill, 1994.
[2] J. Xiao, X. Xie, Y. Han, and J. Wu, “Dynamic tracking of low-frequency
oscillations with improved Prony method in wide-area measurement system,” in
Proc. IEEE Power Eng. Soc. General Meeting, Jun 2004, pp. 1104–1109 Vol.1.
[3] P. Tripathy, S. C. Srivastava, and S. N. Singh., “A modified TLSESPRIT-based
method for low-frequency mode identification in power systems utilizing
synchrophasor measurements,” IEEE Trans. Power Syst., vol. 26, no. 2, pp. 719–
727, May 2011.
[4] M. He, V. Vittal, and J. Zhang, “Online dynamic security assessment with
missing pmu measurements: A data mining approach,” IEEE Trans. Power Syst.,
vol. 28, no. 2, pp. 1969–1977, May 2013.
QUESTIONS AND
SUGGESTIONS ?
THANK
YOU
