Online Monitoring of Oscillation Modes For Small-Signal Security Assessment
Online Monitoring of Oscillation Modes For Small-Signal Security Assessment
Online Monitoring of Oscillation Modes For Small-Signal Security Assessment
This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
1
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
Abstract — This paper presents the main concepts of small- dynamic behavior (SSA) of these systems are analyzed [1-4].
signal security assessment of power systems, focusing on the online The SSA is performed through different methods and
monitoring of small-signal stability. In this work, a computational methodologies that can be easily found in the academic
tool for online monitoring of oscillation modes (OLMO) in power
literature. Some works are briefly described following.
systems is proposed. The objective of the methodology consists of
determining small-signal stability margins, security levels, and In [5], the authors deal with the monitoring of interarea
corrective measures that can improve the damping factor of oscillation mode by the utilization of phasor measurement units
oscillation modes present in such systems. These corrective (PMUs). The monitoring of interarea oscillation in power
measures can be associated with possible redispatches of power systems with renewable sources using Prony method is
plants. In this context, the Closest Security Boundary for evaluated in [6]. Oscillation modes of the BIPS are monitored
Generation Redispatch using Eigenvalue Sensitivities, known as
using ambient data in [7].
CSBGRES Method, could be applied to determine an optimized
redispatch for the power plants of the system to ensure a desired In [8], a real-time monitoring of interarea oscillations using
minimum damping factor, previously chosen, for oscillation PMUs is presented. Oscillation monitoring of synchronous
modes. The proposed tool can be applied for controlling power generators using parallel-Prony is addressed in [9]. In [10],
plant dispatches on Smart Grids, focusing on the damping factor [11], the authors use Hilbert-Huang Transform (HHT) for
of oscillation modes. A small test system is used to evaluate the monitoring low frequency oscillations in power systems. Fast
proposed computational feature and the results were presented to
frequency-domain decomposition (FFDD) is used in [12] for
demonstrate its importance to power system security. Results of
the utilization of OLMO on the Brazilian Interconnected Power ambient oscillation monitoring.
System (BIPS) are presented to prove the viability of using this In [13], the authors presented distributed optimization
method in large-scale power systems. algorithms for oscillation monitoring in power system using
interregional phasor measurement units and phasor data
Index Terms — Small-signal stability, Power systems security,
Online monitoring, Stability margins, Corrective measures,
concentrator (PMU/PDC). In [14], the authors combine PMUs
CSBGRES application, Optimal Power Flow, Power plant and Prony method for oscillation monitoring.
redispatches, Smart grid control, Smart grid monitoring. Different from the aforementioned research, this work is
focused on the Small-signal Security Assessment through
I. INTRODUCTION monitoring of oscillations modes, based on the mathematical
model of power system, instead of measurements.
Thiago Jose Masseran Antunes Parreiras is with Federal University of Rio Tiago Santana do Amaral is with Electrical Energy Research Center –
de Janeiro – UFRJ, Rio de Janeiro, RJ, Brazil, [email protected]. CEPEL, Rio de Janeiro, RJ, Brazil, [email protected].
Sergio Gomes Junior is with Electrical Energy Research Center – CEPEL, Marcelo Rosado is with Electrical Energy Research Center – CEPEL, Rio
Rio de Janeiro, RJ, Brazil and Fluminense Federal University – UFF, Niterói, de Janeiro, RJ, Brazil, [email protected].
RJ, Brazil, [email protected], [email protected]. Nicolas Abreu Rocha Leite Netto is with Electrical Energy Research Center
– CEPEL, Rio de Janeiro, RJ, Brazil, [email protected].
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
2
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
small-signal analysis is the offline studies using different function, as will be detailed in section IV.
operating points, prepared to cover typical and critical operation The damping factor, also called damping ratio, is a measure
conditions of power systems. of the damping of an oscillation represented by a complex pole
−𝜎
Contingency analysis may be included for each operating 𝜆 = 𝜎 + 𝑗𝜔. It is given by 𝜉𝑑 = 2 2 and this measure is
√𝜎 +𝜔
point. Note that contingency analysis in SSA consists in
valid for any pole frequency.
calculating the oscillation modes in a new operating point
Small-signal analysis consists of calculating oscillation
obtained from the steady state of the line outage into the base
modes and the verification of their damping factors, in order to
case. The transient of the fault and switching is nonlinear and
determine modes with small or negative damping factors, which
SSA is not used for this analysis.
represent stability issues.
As will be seen in the paper, it is possible to include an off-
Generally, SSA can be performed through three different
line monitoring of oscillations to complement the conventional
methods. The first is the Damping Nomogram Method (DNM),
offline analysis of regular operating conditions. In this case,
the second is the Root-Locus Method (RLM) and the third is
more severe and realistic conditions may be covered, improving
the Online Monitoring of Oscillation Modes (OLMO), which is
security and approaching simulation from the actual behavior
proposed and detailed in this paper.
of power system.
The Damping Nomogram Method is based on the analysis of
Other application of the methodology includes the operator
several operating points, aiming to determine the worst damped
training. In this case, the feature of the proposed monitoring
oscillation mode for each one of these situations. The operating
oscillations may be implemented into a software for training
points can be classified based on the values of the damping
operators to act at situations of oscillation problems.
factors and small-signal security regions can be obtained.
This paper is organized as follows: Section II presents a
The Root-Locus Method is based on the root-locus
review of SSA and its main concepts and characteristics,
calculation through power plants dispatches variation or
focusing on the Online Monitoring of Oscillation Modes
interchanging power of interest, aiming to determine the
(OLMO); Section III presents a review of the Closest Security
oscillation modes displacement in the complex plane related to
Boundary for Generation Redispatch using Eigenvalue
the variation considered for the dispatches.
Sensitivities, known as CSBGRES method [15], which can be
The OLMO is based on the calculation of oscillation modes,
used to determine a redispatch for power plants, ensuring a
which can be obtained through the mathematical model of the
desired damping factor for an oscillation mode of interest;
power system. The frequencies and damping factors of these
Section IV describes the OLMO and its computational tool
oscillation modes can be monitored over time, which may be
implemented in PacDyn software [16] for monitoring
observed graphically by the operators of the power system.
oscillation modes in real power systems; Section V presents the
The main objective of the OLMO is to study and monitor the
OLMO results obtained for the Test System; Section VI
power system to determine possible oscillation damping
presents the OLMO results obtained for the Brazilian
problems, trying to identify operation points where the system
Interconnected Power System (BIPS); and Section VII
would be presenting undesired dynamic behaviors.
concludes the paper.
During the execution of OLMO, it is possible to identify or
preview the decline of oscillation damping factors and try to act
II. SSA METHODOLOGIES AND OSCILLATION MONITORING
in the system, in order to improve its dynamic behavior,
Power systems security assessment studies are based on the increasing these damping factors until a desired damping factor
analysis of a desired power system considering many scenarios value (for example, 5%). Examples of OLMO of a secure
and critical contingencies, aiming to determine the most critical system, where the damping factor of the monitored mode is
operating points, where a fault may leave the system to a always over the desired minimum value (5%), and of an
collapse and, consequently, generate interruptions in electrical insecure system, where the damping factor of the monitored
energy supply [1-3,17-20]. mode decreases to undesired values, are shown respectively in
The security assessment allows the determination of power Fig. 1 and Fig. 2.
system security limits, which can be represented by stability The OLMO could also present results of forecast data, where
margins [20], security indexes [19], or security regions that the dynamic behavior of the system could be determined in real-
delimitate security boundaries for system operation [1-2]. time for some future tens of minutes, for example.
The small-signal security assessment is directly related to Examples of OLMO of secure and insecure systems, with
small-signal stability analysis. In this analysis, the system forecast data, can be observed in Fig. 3 and Fig. 4.
variables are automatically initialized, and equations are By evaluating the results presented by the OLMO, it is
linearized around the initial values, using the first-order Taylor possible to detect small-signal stability or security issues in the
expansion for the non-linear terms. The oscillation modes are monitored power system.
represented by the eigenvalues of the resulting system matrix of With the OLMO results, it is also possible to consider
Differential and Algebraic Equation (DAE) system. corrective measures that can improve the performance of the
In large scale power systems, partial eigensolution power system, increasing the damping factor of the monitored
algorithms are used. The poles of interest can be calculated oscillations modes. These corrective measures could be related
using dominant methods by an adequate choice of a transfer to new dispatches for the power plants or new tunning for
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
3
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
controllers used in system equipment. previewing a collapse in the damping factor of the monitored
mode, but, with the measure, this damping factor was increased
to an acceptable value, respecting the security limits [4].
The CSBGRES method [15] could be used to indicate to the III. CSBGRES METHOD REVIEW
operators possible redispatches for some power plants, aiming
to improve system dynamic behavior. In [15] the authors presented the development of an extension
An illustrative corrective measure during the execution of an of the Hopf bifurcation method described in [22] to the
OLMO can be observed in Fig. 5, where the forecast data was application of this method to determine the minimum variation
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
4
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
5
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
This feature starts from a base case, with the dynamic model of the operating points being sent to the software during the
of the power system, where the oscillation modes of interest are monitoring. Note that SAGE has a service to update power flow
obtained. Then, the monitoring is started and PacDyn keeps data, which is not represented in Figure 6, but the monitoring of
observing the power flow data of this system obtained this updating data process, performed by PacDyn, is indicated
periodically from an Energy Management System (EMS). at the decision block of this figure.
The power flow data are generated from a conventional power Using the operating point information and the updated
flow program inside EMS using data of the calculated static oscillation modes, graphical visualization of the variation of the
state estimation. The results are the converged magnitudes and frequencies and the damping factors of the mode over time can
phases of the voltage in each bus of the system. be created, in a timeline result, as shown in Section II.
When power flow information is updated, PacDyn software The structure utilized in the integration between PacDyn and
uses one of the available partial eigensolvers (DPSE [23] or SAGE is presented in Fig. 7, where all the information shared
SDPA [24]) to calculate the oscillation modes for this new by both software is shown.
operating condition of the system. The integration between SAGE and PacDyn is based on the
The algorithm of the OLMO tool can be observed in Fig. 6, utilization of communication protocols and message queues,
which summarizes the processes involved in the oscillation where the power flow data is sent from SAGE system to
mode monitoring feature. PacDyn and the results obtained in the eigenvalue computations
Power flow data is updated and sent to PacDyn by the SAGE are sent back, from PacDyn to SAGE.
[25], which is a Brazilian EMS/SCADA system, commonly Basically, PacDyn receives the initial power flow data, of the
used by companies in the electrical energy sector in Brazil. base case, and the dynamic model of system equipment and
SAGE uses measurements and state estimators to update the obtains the initial values of the oscillation modes of interest.
power flow data of the system and then send it to PacDyn to The OLMO tool is executed and PacDyn starts to monitor the
compute updated oscillation modes. power flow data of the system.
At the same time, the SAGE system is collecting data from
the measures of the power system and uses the state estimator
to obtain all power flow data for the current date and time,
which is supplied to PacDyn.
PacDyn sees the update in power flow data and calculates the
new values for the oscillation modes of interest. With the results
of the frequencies and damping factors of the monitored modes,
graphics of the variation of these variables over time are
created, resulting in timeline information.
These graphical results can be observed in a dedicated
computer, which is executing PacDyn, and can be seen as well
in the SAGE environment by the system operators.
Results of the OLMO tool applied to a Test System will be
presented in the sequence. After that, the results of the
monitoring tool applied to the Brazilian Interconnected Power
System will be presented, proving that the computational tool
can be used in the analysis of large-scale power systems.
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
6
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
The results presented in this Section were obtained by using The results obtained for the aforementioned contingencies
the test environment of SAGE system, also known as TOPSIM, were calculated as follows: the transmission line of interest was
developed for the training of power system operators. This tripped out, a power flow calculation was made without this
computational feature is responsible for simulating the actual line, then this operation condition was used in PacDyn in order
behavior of power systems through load flow solutions. to calculate the oscillation modes for this situation.
Through Fig. 10, one can note that in some operating points,
an oscillation mode of the system became poorly damped, with
a damping factor lower than 5%, which is the minimum desired
Fig. 9. Load curve of Test System.
value for damping factor in the monitoring.
The value of 5% is not a consensus of limit criterion in
Before starting the OLMO tool, the base case of Test System academic literature, but it is commonly used in practical studies
should be analyzed, aiming to identify the oscillation modes of of small-signal stability, mainly when considering interarea
interest, such as the lowest damped ones. oscillation modes. In this work, the aforementioned value of 5%
Two oscillation modes were identified and used for was chosen as the small-signal security limit.
monitoring purposes. They were described as follows: In those operating points, where the system was insecure,
• The mode with a frequency of 7 rad/s represents an SAGE system sends a warning to the system operator,
oscillation between power plants US10 and US11 of the indicating the bad dynamic behavior forecast for the system
Test System; (based on the damping factor found with a value under 5%).
• The mode with a frequency of 5 rad/s represents an Every time an operating point with an undesired damping
oscillation between power plant US01 and the rest of the factor is identified, PacDyn runs the CSBGRES method aiming
system. to obtain a minimum redispatch for the power plants of the
The Fig. 10 presents the results obtained for damping factors system, to increase the damping factor of the monitored
and frequencies of both oscillation modes during the OLMO oscillation modes over to, at least, 5%, improving the dynamic
tool test and validation. behavior of the system, from the point of view of small-signal
Besides the base case of the system, two important stability. The results of online monitoring of oscillation modes,
contingencies were considered in the online monitoring: the trip considering the redispatch suggestions obtained through the
of the transmission line between buses 15 and 16; the trip of the CSBGRES method, are presented in Fig. 11.
transmission line between buses 28 and 29.
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
7
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
Note that damping factors changed to 5%, approximately, in • The mode with a frequency of 2.8767 rad/s (0.46 Hz),
Fig. 11 at the operating points where they were below the limit with a damping factor of 3.14%;
in Fig. 10. The frequencies on the bottom of Fig. 10 and Fig, 11 The reasons for selecting these poles are their small damping
did not significantly change with the solution of CSBGRES, factors and their global influence, involving oscillations in a
since the mitigation of poorly damped oscillations has a small large number of power plants.
impact on the oscillation mode frequencies. The normalized rotor speed mode shapes of both oscillation
modes of interest are shown in Fig. 12. The phasors in mode-
shapes are complex numbers that represent the amplitude and
the phase of the oscillation modes for each generator speed.
Fig. 12. Rotor speed mode shapes for both monitored modes.
Fig. 11. Frequencies and damping factors of monitored oscillation modes, Observing results for the 0.63 Hz pole, one can notice that the
considering the results of the CSBGRES method.
machines in North and Northeast regions oscillates coherently
and with larger amplitudes against machines of South and
This tool can be combined with the CSBGRES method, which Southeast. Therefore, the oscillation mode with a frequency of
can calculate an optimized redispatch for the power plants of 3.9667 rad/s (0.63 Hz) represents an oscillation between the
the monitored system, to increase the damping factor of modes North-Northeast and the South-Southeast regions of Brazil.
of interest, improving the stability behavior of the system. Considering, now, results for the 0.46 Hz pole, one can notice
This combined application of the OLMO tool with the that the oscillation mode with a frequency of 2.8767 rad/s
CSBGRES method contributes to improving the robustness and (0.46 Hz) represents an oscillation between the Madeira River
reliability of the power system that is being monitored. Complex (Jirau and Santo Antônio power plants) and Samuel
against, mainly, North-Northeast regions.
VI. RESULTS OF BIPS In the OLMO simulation presented in this Section, 1182
The OLMO tool was tested in the Brazilian Interconnected operating points of BIPS were monitored, by using the
Power System (BIPS), by using real-time data from the integration between PacDyn software and SAGE system.
Brazilian National Operator Company (ONS) [26]. The base case was the scenario obtained for day 05/22/2016,
These data were related to the operation of BIPS at 0h08. OLMO tool was able to monitor the data of 1 week of
corresponding to days between 04/22/2016 and 04/29/2016. BIPS operation (168 hours). The average time of operating
The dynamic data of the model of the system were adapted from point acquisition is about 8.5 minutes (168 hours multiplied by
the database of ONS related to 2016. 60 minutes and divided by 1182 points).
Two interarea oscillation modes were obtained in the base The results obtained in the OLMO test for the aforementioned
case of the BIPS and used for monitoring purposes, to verify the oscillation modes of the BIPS can be seen in Fig. 13, which
performance of the OLMO tool in a large-scale power system: presents the frequencies of the modes, and in Fig. 14, which
• The mode with a frequency of 3.9667 rad/s (0.63 Hz), presents their damping factor.
with a damping factor of 5.86%;
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
8
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
One can notice that oscillation mode frequencies are less The zoom presented in Fig. 15 is utilized to show the
variable than the damping factor, over the monitoring time. highlighted last points of the monitoring. The diamond shaped
Some operating points, without contingencies, present slightly mark highlights the most recent calculated oscillation modes
negative damping factors for one of the monitored modes. and the circle shaped mark highlights the other four last
In practice, there were no instability occurrences in the BIPS, oscillation modes obtained in the OLMO tool.
during the time period of the monitoring. This issue is explained The oscillation mode displacement presented in Fig. 15 is
by some inaccurate data utilized in the model of the BIPS for caused by the variation of the loads of the BIPS over time and
simulations, which is continuously being improved to the redispatch of the power plants made by their speed
approximate the obtained monitoring oscillations to the governor, automatic generation control (AGC) and operator’s
observed oscillations in reality. actions.
Fig. 15 presents a different visualization of OLMO tool In order to show the correlation of the pole dispersion along
results, which is the monitoring of oscillation modes seen the monitoring and the power dispatch of power plants, Fig. 16
through their displacements in the complex plane. The line in shows the active power of the main power plants in Brazil:
red indicates the 5% damping factor limit. One can notice that Itaipu (14 GW), Complex of Madeira (6.9 GW), Tucurui
the poles go higher and lower the 5% limit during monitoring. (8.4 GW), Complex of Paulo Afonso (4.2 GW), Xingo
(3.2 GW) and Ilha Solteira (3.4 GW). In particular, Itaipu in
South and Tucurui in North have a large amount of generation
compared to the other power plants.
Fig. 14. Damping factors of both monitored modes. Fig. 16. Generation of the main power plants in Brazil during the monitoring.
Fig. 15. Displacement of monitored modes in complex plane (and zoom). Fig. 17. Generation and load variation in the BIPS, during the monitoring.
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
9
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
VII. COMPARISON WITH THE CONVENTIONAL ANALYSIS The OLMO tool rises as a new computational tool, bringing
For better evidence of the advantages of the proposed important information on the system stability for the power
oscillation monitoring method, a comparison is made with the system operator. Sending warning when undesired damping
conventional analysis of small-signal stability. factors are obtained during monitoring, this new feature can
In this methodology, different power flow data are prepared alert the operator that the system may be going to insecure and
representing typical and critical operating points. undesirable operating conditions.
For simplicity, three typical scenarios of operating conditions With this information, the operator could act on the power
of the Brazilian Interconnected Power System were used in this system, aiming to mitigate a possible stability problem, leading
comparison: heavy load of October of 2015 (I), heavy load of the system to a better operating scenario.
February of 2016 (II) and medium load of October of 2016 (III).
Such as in the proposed OLMO tool, the partial eigensolver REFERENCES
SDPA was used for the computation of the oscillation modes of [1] ALVES, F. R. M., HENRIQUES R. M., PASSOS FILHO, J. A.,
interest. Table I shows the frequencies and damping factors of GOMES JUNIOR, S., BORGES, C. L. T., RANGEL, R. D., FALCAO,
D. M., AVELEDA, A. A., TARANTO, G. N., ASSIS, T. M. L.,
the oscillation modes obtained for these three conditions. “Integrated tool for static and danymic security assessment of large scale
power systems”, XII Simpósio de Especialistas em Planejamento da
TABLE I Operação e Expansão Elétrica - SEPOPE, Rio de Janeiro, May, 2012. (in
portuguese)
OSCILLATION MODES FOR CONVENTIONAL ANALYSIS
[2] HENRIQUES, R. M., ALVES, F. R. M., PASSOS FILHO, J. A.,
Mode 0001 Mode 0002 GOMES JUNIOR, S., RANGEL, R. D., BORGES, C. L. T., FALCAO,
Scenario Freq Damp. Freq Damp. D. M., TARANTO, G. N., ASSIS, T. M. L., “Improvement on electrical
(rad/s) (%) (rad/s) (%) system security assessment through detailed representation of equipment
and oscillation damping monitoring”, XXII Seminário Nacional de
I 4.6150 13.25 3.3350 2.06 Produção e Transmissão de Energia Elétrica - SNPTEE, Brasília,
II 4.4316 16.91 2.7107 0.46 October, 2013. (in portuguese)
III 4.5214 14.44 3.2321 5.77 [3] MORISON, K., HAMADANIZADEH, H., WANG, L., “Dynamic security
assessment tools”, IEEE Power Engineering Society Summer Meeting,
v. 1, p. 282-286, July, 1999.
From Table I, Mode 0002, which represents oscillations [4] ASSIS, T. M. L., “Calculation of dynamic transmission capability in power
between Madeira River Complex and North-Northeast regions, systems through integrated tools and intelligent”, D. Sc. Thesis,
presented poorly damped behavior, having similar results to the COPPE/UFRJ, Rio de Janeiro, RJ, Brazil, 2007. (in portuguese)
[5] KAKIMOTO, N. SUGUMI, M., MAKINO, T., TOMIYAMA, K.,
OLMO, depicted in Fig. 14. However, instability or poorly “Monitoring of interarea oscillation mode by synchronized phasor
damped behavior were not detected for Mode 0001, which measurement”, IEEE Transactions on Power Systems, v. 21, n. 1, pp. 260
represents oscillations between North-Northeast and South- - 268, January, 2006.
Southeast regions, such as appeared in Fig. 14. [6] TERZIJA, V., CAI, D., FITCH, J., “Monitoring of inter-area oscillations in
Therefore, one may conclude that the oscillation problems power systems with renewable energy resources using Prony method”,
CIRED 2009 - The 20th International Conference and Exhibition on
identified in the proposed monitoring method may not be Electricity Distribution - Part 2, pp. 1-14, December, 2009.
correctly identified and quantified in the conventional small- [7] PRIOSTE, F. B., SILVA, A. S., DECKER, I. C., “Monitoring oscillations
signal stability analysis, bringing more risk to system security. modes of the Brazilian Interconnected Power System using ambient data”,
On the other hand, the proposed OLMO tool can trace pole IEEE Trondheim PowerTech, pp. 1-7, September, 2011.
[8] ALI, H. R., HOONCHAREON, N., “Real-time monitoring of inter-area
changes during an actual system operation. power oscillation using phasor measurement unit”, 2013 10th
International Conference on Electrical Engineering/Electronics,
VIII. CONCLUSION Computer, Telecommunications and Information Technology, pp. 1-6,
July, 2013.
This paper presented a computational methodology for [9] SARAF, P., VENAYAGAMOORTHY, G. K., LUITEL, B., “Online
monitoring oscillation modes of power systems with possibility oscillation monitoring of synchronous generators using parallel-Prony
of improving poorly damped oscillations by redispatch. analysis”, Innovative Smart Grid Technologies, pp. 1-5, May, 2014.
The OLMO tool was implemented with an integration [10] PENG, J. C. H., KIRTLEY, J. L., “An improved empirical mode
decomposition method for monitoring electromechanical oscillations”,
between PacDyn and the SAGE system, supplying information Innovative Smart Grid Technologies, pp. 1-5, May, 2014.
on the frequencies and damping factors of oscillation modes of [11] YANG, P., YUE, L., “The study on monitor of power system low
interest over time for the system operator, sending warnings frequency oscillation base on HHT method”, 014 China International
when an undesired damping factor is calculated. Conference on Electricity Distribution (CICED), pp. 305-307, December,
A first Test System was used to verify the implementation of 2014.
[12] KHALILINIA, H., ZHANG, L., VENKATASUBRAMANIAN, V., “Fast
the OLMO tool in a small-scale power system, evidencing the Frequency-Domain Decomposition for Ambient Oscillation Monitoring”,
importance of the new information that the new computational IEEE Transactions on Power Delivery, v. 30, n. 1, pp. 1631 - 1633,
feature can supply to operators of power systems. February, 2015.
This system was also utilized to show how the CSBGRES [13] NABAVI, S., ZHANG, J., CHAKRABORTTY, A., “Distributed
method can be used combined with the OLMO tool, increasing Optimization Algorithms for Wide-Area Oscillation Monitoring in Power
Systems Using Interregional PMU-PDC Architectures”, IEEE
system robustness and reliability. Transactions on Smart Grid, v. 6, n. 1, pp. 2529 - 2538, March, 2015.
The BIPS was utilized to demonstrate that the OLMO tool can [14] ALMUNIF, A., FAN, L.., “PMU Measurements for Oscillation
be used in large-scale power systems successfully, being able Monitoring: Connecting Prony Analysis with Observability”, IEEE
to monitor the oscillation modes of interest in the Brazilian Power and Energy Society General Meeting, pp. 1-5, January, 2020.
power system, such as some interarea modes.
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.
This article has been accepted for publication in IEEE Transactions on Power Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/TPWRS.2023.3249477
10
> REPLACE THIS LINE WITH YOUR MANUSCRIPT ID NUMBER (DOUBLE-CLICK HERE TO EDIT) <
[15] PARREIRAS, T. J. M. A., GOMES JUNIOR, S., TARANTO, G. N., CEPEL-since 1996, working on the development of SAGE network analysis
UHLEN, K., “Closest security boundary for improving oscillation applications. His research interests include network analysis algorithms,
damping through generation redispatch using eigenvalue sensitivities”, supervisory and control systems, operator training simulators and energy
Electric Power System Research, vol. 160, pp. 119-127, July, 2018. management systems.
[16] CEPEL, “PacDyn - Software for small-signal stability analysis and control
- Version 9.9.16”, User's manual, Rio de Janeiro, RJ, Brazil, 2021. Nicolas Abreu Rocha Leite Netto received the B.Sc. degree in Electrical
[17] KERIN, U., BALAURESCU, R., LAZAR, F., KREBS, R., BALASIU, F., Engineering from Federal University of Rio de Janeiro (2014) and M.Sc. in
“Dynamic security assessment in system operation and planning - First Electric Engineering from Federal University of Rio de Janeiro (2016).
experiences”, IEEE Power and Energy Society General Meeting, pp. 1-6, Currently he is an engineer researcher at Cepel, Rio de Janeiro, Brazil, acting
July, 2012. on the following subjects: high-performance computing; power systems
dynamics, control, stability, and reliability; situational awareness and human
[18] WANG, L., HOWELL, F., KUNDUR, P., CHUNG, C. Y., XU, W., “A
cognition; power flow algorithm and modeling.
tool for small-signal security assessment of power systems”, IEEE
International Conference on Power Industry Computer Applications, pp.
246-252, May, 2001.
[19] KERIN, U., TUAN, T. N., LERCH, E., BIZJAC, G., “Small signal security
index for contingency classification in dynamic security assessment”,
IEEE Trondheim PowerTech, pp. 1-6, June, 2011.
[20] VAAHEDI, E., MANSOUR, Y., TSE, E. K., “A general purpose method
for on-line dynamic security assessment”, IEEE Transactions on Power
Systems, v. 13, n. 1, pp. 243-249, February, 1998.
[21] HUANG, X., LIU, J., LIU, Y. XU, L., YANG, J., NI, Y., “A GA based
optimized PMU-location-decision algorithm considering WAMS
reliability”, IEEE International Conference on Power System Technology,
pp. 433-439, October, 2014.
[22] GOMES JUNIOR, S., MARTINS, N., PARREIRAS, T. J. M. A.,
“Computing the closest small-signal security boundary in the control
parameter space for large scale power systems”, Electric Power System
Research, v. 149, pp. 10-18, August, 2017.
[23] MARTINS, N., “The dominant pole spectrum eigensolver for power
system stability analysis”, IEEE Transactions on Power Systems, v. 12, n.
1, pp. 245 - 254, February, 1997.
[24] GOMES JUNIOR, S., MARTINS, N., PORTELA, C., “Sequential
computation of transfer function dominant poles of s-domain system
models”. IEEE Transactions on Power Systems, vol. 24, no. 2, pp. 776-
784, May, 2009
[25] CEPEL, “SAGE – Open Energy Management System”, Site:
www.sage.cepel.br. (in portuguese)
[26] ONS, “SIN – Electromechanical Transients Data Base”, Site:
www.ons.org.br. (in portuguese)
BIOGRAPHIES
Thiago Jose Masseran Antunes Parreiras received the B.Sc. degree (2009),
M.Sc. degree (2012) and D.Sc. degree (2017) in electrical engineering from the
Federal University of Rio de Janeiro (UFRJ). In 2011, he was an analyst at the
Brazilian Electrical Centre (ELETROBRAS), Rio de Janeiro, RJ, Brazil. From
2011 to 2014, he was an analyst at the Energy Research Company (EPE), Rio
de Janeiro, RJ, Brazil. From 2014 to 2021, he was a researcher at the Electrical
Energy Research Center (CEPEL), Rio de Janeiro, RJ, Brazil. Since 2021, he is
a professor at the Federal University of Rio de Janeiro (UFRJ), Rio de Janeiro,
RJ, Brazil. He has experience in the field of electrical engineering, with
emphasis on power systems, acting mainly on the following topics: power
system planning, equipment modelling, power system dynamic, linear analysis,
control systems and computer application.
Sergio Gomes Junior received the B.Sc. degree in electrical engineering from
Fluminense Federal University, Niteroi, Brazil, the M.Sc. and D.Sc. degrees in
electrical engineering from Federal University of Rio de Janeiro, Rio de Janeiro,
Brazil, in 1992, 1995 and 2002, respectively. Since 1994, he has worked at
CEPEL, Rio de Janeiro, Brazil, developing methods and computer tools for
power system analysis. Since 2010, he is a Professor at the Department of
Electrical Engineering of UFF, Niteroi, Brazil.
Tiago Santana do Amaral received the B.Sc. degree (2004), M.Sc. degree
(2007) and D.Sc. degree (2021) in electrical engineering from the Federal
University of Rio de Janeiro (UFRJ). Since 2006 he has been a researcher at
Cepel working in the research and development of computer programs for the
analysis of power systems. His main areas of interest are reliability, control and
computational modeling of power systems.
Marcelo Rosado da Costa received the B.Sc. degree (1993), and M.Sc. degree
(1998) in electrical engineering from the Federal University of Rio de Janeiro
(UFRJ). He has been a researcher at the Department of Systems Automation at
© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Chalmers University of Technology Sweden. Downloaded on March 10,2023 at 08:04:24 UTC from IEEE Xplore. Restrictions apply.