International Conference on Renewable Energies and Power Quality (ICREPQ’16)

Madrid (Spain), 4th to 6th May, 2016

exÇxãtuÄx XÇxÜzç tÇw cÉãxÜ dâtÄ|àç ]ÉâÜÇtÄ (RE&PQJ)
ISSN 2172-038 X, No.14 May 2016

Power Quality Analysis in an Industrial Electrical System by Probability Density

Function

G. B. Gibelli1, M. Oleskovicz2 and J. C. M. Vieira2

1
Department of Engineering of Energy – FAEN
Federal University of Grande Dourados - UFGD
Dourados - MS, Brazil
Phone number: 55 67 3410 2160, e-mail: [email protected]
2
Department of Electrical and Computer Engineering
University of São Paulo – USP/EESC
São Carlos – SP, Brazil
Phone number: 55 16 3373 8142, e-mail: [email protected], [email protected]

Abstract. This research aimed to analyze the Power Quality economic impact on energy companies and end
(PQ) in an Industrial Electrical System (IES) through event consumers (industrial) according to the different
classification by applying the Probability Density Function associated costs and resulting from problems occurred
(PDF). The aim is to obtain PQ indicators that enable an due to the lack of PQ.
informative action (warning sign) for the protection devices It is worth mentioning that there are, in the literature,
associated through the classification of disturbances conducted researches that apply strategies for the detection, location
by PDF. The methodology and results are based on the modeling and classification of disturbances in the PQ area. In [2],
and simulation of a real IES containing Three Phase Induction
Motors (TPIM), using the DIgSILENT PowerFactory software.
the authors present a review of techniques and
The modeling was performed to generate representative methodologies developed for PQ analysis in transmission
situations of IES operation featuring short duration voltage and distribution systems, seeking to better classify the
variations. The classification of the situations by the proposed observed power disturbances. According to the proposed
methodology could be initially checked by observing this revision, the Wavelet Transform (WT), Transform-S
disturbance. Satisfactory results attest to the potential of the (TS), Kalman Filter (KF), filter banks, and sometimes,
approach developed until now. the combination of two of them are techniques usually
applied for the estimation of PQ indicators. As for the
Key words classification of events, the trend is to use algorithms
based on Artificial Intelligence (AI) and Support Vector
Power Quality, Industrial Electrical System, Probability Machine (SVM), where the main problem for their
Density Function, Three Phase Induction Motor. application is the computational effort required and the
large amount of data that should be available and
1. Introduction analyzed.
The authors in [3] present techniques for automated
analysis of various types of data collected in an EPS,
Especially in the last 15 years, there has been a growing
emphasizing the need for an intelligent analyzer that is
concern among experts and researchers because of the
able to classify, identify the types and causes, and locate
large amount of sensitive equipment and processes
a short circuit, estimating all involved parameters
affected by problems associated with the lack of a desired
automatically. This way, the techniques of Fourier
PQ (Power Quality). This concern arises mainly from the
Transform (FT), WT or TS are employed to extract the
growing presence of non-linear loads from the new
distinct characteristics of the signs under analysis and
equipment installed in the Electrical Power System (EPS),
subsequently use smart tools like Fuzzy System (FS),
which can generate and be sensitive to various problems,
Artificial Neural Network (ANN) or Neuro-Fuzzy for
causing damage to final consumers of electricity [1]. For
decision making of the classification of disturbances.
the industrial sector in particular, the increased efficiency
In the above context, from the practice, the analyzers and
in the use of electricity has resulted in the improvement
multimeters for PQ monitoring are embedded in a
and integration of automated processes, which in turn, may
structure in IES (Industrial Electrical System) only to
be sensitive to changes in voltage variations, and or cause
record, analyze and store data when an event is detected.
changes in voltage supply in a certain Point of Common
On the other hand, the Intelligent Electronic Devices
Coupling (PCC). This scenario denotes the potential

https://doi.org/10.24084/repqj14.250 145 RE&PQJ, No.14, May 2016

(IEDs) are included in IES to protect the loads or parts of probability of X having a value in the interval [a, b] as
the IES. As for the loads, in this specific case, the IEDs are shown in Fig. 1.
applied to protect the Three Phase Induction Motors
(TPIM), after the detection of an event that could damage
the system [4].
In this scenario, for the analyzers and multimeters, there is
no assignment of any function or immediate action task on
the load (in this case, on the TPIM). This is valid for either
a decision-making or an informative action for other
processes that might use the information provided to
Fig. 1. Probability determined from the area under f(x).
predict or prevent the incident situation, or future
situations, at the occurrence point. Fig. 2 illustrates a histogram in the form of bars,
Therefore, a method that may correlate characteristics of presenting the clusters of a set of data cells (bins). The
the signatures of the three phase voltage values, on the IES dynamic range of data is divided into a number of cells
under observation and hence on the TPIM is highly (bins) of the same "length". The "height" of each cell is
desirable. It is believed that the result of the correlation of calculated by counting the number of data contained in
these signatures allow obtaining PQ indicators focused on the range [8].
spotting possible operation problems in TPIM, which
enables it to predict and inform to IED one incident
situation and perhaps damaging to the system as a whole,
in a timely manner.
Thus, this research presents a study that uses the PDF
(Probability Density Function), represented by histograms,
as an analysis tool of PQ disturbances in an IES. The main
goal of the research is to obtain PQ indicators for an
Fig. 2. Histogram in the form of bars that approximates to the
informative action (warning sign) to IEDs through the PDF.
classification of disturbances conducted by the PDF.
In Fig. 2, the area of each bin is equal to the interval
2. Probability Density Function frequency. Thus, the area under f(x) over any interval is
equal to the interval probability.
In areas related to engineering, Stochastic Process (SP) or
Random Process (RP), refer to the random nature 3. Industrial Electrical System Analysed
phenomena that occur in temporal or spatial manner,
allowing its analysis through statistical tools. An RP is the To generate representative situations of Short Duration
result of a random experiment (random signal) associated Voltage Variations (SDVV), necessary for the analysis of
to a waveform as a function of time. the proposed algorithm, an IES was modelled considering
Consider a random voltage signal represented by the set of the real data of the IES and TPIM provided by a regional
its possible sample functions X(t, ζ1),..., X(t, ζn) within the industry, and using the DIgSILENT PowerFactory
sample space S, where its sample functions depend on the software [9]. Fig. 3 shows a representation of the IES
variation of parameters ζ1,..., ζn. Thus, a random variable X used for the simulations of the loads operation and fault
will be contained in the result of a random experiment by situations for the analysis of PQ disturbances. The
the X(ζ) number [5]. The probability distribution of a situations will be analyzed through the dynamic response
random variable X can be specified with the PDF, and f(x) of the TPIM against power disturbances in the power
can be used to describe the probability distribution of a network.
random variable X [6]. The probability of X being between The electrical system consists of: a 138.0 kV power
a and b is determined by the integral of f(x) of a and b, as substation; a 4.00 kV three phase induction motor of
shown in (1). 1,180 kW; two step-down power transformers (T1 and
T2) with relations 138.0/11.95 kV and apparent power of
௕
ܲሺܽ ൏ ܺ ൑ ܾሻ ൌ න ݂ሺ‫ݔ‬ሻ݀‫ݔ‬ (1) 12.5 MVA; a step-down power transformer (T3) with
௔ relations 11.95/0.22 kV and apparent power of
0.045 MVA; a step-down power transformer (TM) with
A histogram is an approximation of the PDF, where for relations 11.95/4.16 kV and apparent power 1.5 MVA;
each interval of the histogram, the bar area is equal to the and loads characterized by a unity power factor and
frequency of the measurements on the interval and the active power of 0.04 MW and 10 MW. It should be noted
frequency is an estimation of the probability of the that the T1 and T2 transformers have delta-star
measurement be in the interval. Thus, the area under f(x) connections with 17.25 Ω grounding resistance, and the
over any interval is equal to the real probability of the T3 and TM transformers present delta-star connections
measurement being in the interval [6]-[8]. with neutral solidly grounded. The TPIM was modelled
A PDF provides a simple description of the probabilities by using the asynchronous machine model of the
associated with a random variable. It is important to note DIgSILENT PowerFactory software [9], whereby,
that f(x) is used to calculate an area that represents the various electrical and mechanical parameters were

https://doi.org/10.24084/repqj14.250 146 RE&PQJ, No.14, May 2016

employed. It is worth noting that the data entered in the Each characterized situation considered a sampling
model were taken from real motor tests in the field. frequency of 15,360 Hz (256 samples per cycle of the
signals under analysis), based on the IES fundamental
UTILITY
operating frequency of 60 Hz.
The following 10 different faults were simulated to
generate a representative database: three-phase short-
circuits; involving phase A and ground (phase A-ground);
phase B-ground; phase C-ground; phases A and B (phases
B1/BAT 138kV 138,00 kV AB); phases BC; phases CA; phases AB-ground; phases
BC-ground; and phases CA-ground for each of the IES
buses, considering the angle of fault incidence of 0°, fault
resistance of 0 Ω and duration of the fault (td) of 30
T1

T2
cycles (500 ms). Measurements were performed on the
B3/BMTM bus of Motor 1 for all cases of applied faults.
Each of the situations contains 2s duration that include
B2/BMT 11.95kV 11,95 kV 11,95 kV the time on a steady-state (pre-fault), the short circuit
application period (fault IES) and system restoration
(post-fault).
LOAD 1 LOAD 2
TM
T3

4. Methodology
The methodology is based on the recording and analysis
B4/BBT 0,22 kV
B3/BMTM 4,16 kV
of the short circuits applied on the IES resulting in SDVV
on bus B3/BMTM of Motor 1.
The initialization of the methodology consists of
simulations with high sampling rate (1 MHz) using
LOAD 3 M
~
Motor 1
DIgSILENT PowerFactory software in order to
accurately represent the analog signals of voltage and
Fig. 3. Single line diagram of the IES.
current found in the grid.
The acquisition and conditioning of the voltage signals
Table I shows the input parameters relating to the TPIM.
and sinusoidal three phase currents (instantaneous) are
Table I. – Parameters of Motor 1.
performed by transducers allocated close to the TPIM,
using a 60 Hz one cycle window, at a rate of 256 samples
Motor 1 per cycle. The displacement of the window on the signals
Nominal Voltage 4.00 (kV) considers ½ step of a cycle. For this research, a data
Nominal Power 1,180 (kW) window of a cycle is taken from the moment of the short
Power Factor 0.87 circuit application at 0.2 s in the IES bus.
Efficiency 87.4 (%) This post-fault window will contain the three phase
Nominal Frequency 60 (Hz) voltages recorded at the B3/BMTM bus of Motor 1 under
Speed 1,789 (rpm) analysis, as illustrated in Fig. 4.
Pole Pairs 2 The situation illustrated in Fig. 4 is due to the application
Connection Y of a phase-to-phase fault in the B1/BAT bus involving
Locked Rotor Current 6.70 (p.u.) phases A and B of the IES under analysis.
Locked Rotor Torque 1.25 (p.u.)
✑
Starting Torque 2.42 (p.u.) ✐✎
✘
✰❈❁▲❅ ✡
✰❈❁▲❅ ✢
Stator Resistance 0.105 (p.u.) ✐✎
✖
✰❈❁▲❅ ✣
◆❄❅ ✈❐◆✉

Stator Reactance 0.02 (p.u.) ✐✎

✔

Magnetization Reactance 3.15 (p.u.) ✐✎

✒
▼
❁❇❅ ✭❁❇■❉

Rotor Resistance 0.007 (p.u.) ✐

✍
✐✎✒
Rotor Reactance 0.421 (p.u.)
▼
✶❏●

✍
✐✎✔
Rotor Type Double Cage
✍
✐✎✖

✍
✐✎✘
The simulation of loads 1, 2 and 3 connected to bus B2/ ✍
✑
✐✎✑✖ ✐✎
✑✘ ✐✎
✒✐ ✐✎
✒✒ ✐✎
✒✔ ✐✎
✒✖ ✐✎
✒✘ ✐✎
✓✐
BMT and B4/BBT comprises the modeling of the general ✴❉❍❅ ✈▲✉

load module available in the library of DIgSILENT Fig. 4. Window of one cycle as an example taken from the
PowerFactory software, where the provided active power applied fault.
is considered an input parameter [9].
Representative operating conditions that characterized Fig. 5 shows the graph of the three phase voltages
interruptions, sags and swells were generated on this containing a data window of a cycle with 256 pre-fault
system, which allowed the characterization of the desired samples recorded on the B3/BMTM bus of Motor 1
phenomena as established by the IEEE Standard 1159- regarding the phase-to-phase fault applied to the B1/BAT
1995 [10]. bus.

https://doi.org/10.24084/repqj14.250 147 RE&PQJ, No.14, May 2016

 ✑
✰❈❁▲❅ ✡
✐✎
✘ ✰❈❁▲❅ ✢
✰❈❁▲❅ ✣ Start
✐✎
✖
◆❄❅ ✈❐◆✉

No If Phases VA, VB and VC Yes Voltage Swell Warning Sign

✐✎
✔ >= 1.1 pu
Obtain the histogram
✐✎
✒
❁❇❅ ✭❁❇■❉▼

considering the voltages

✐ VA, VB and VC No

✍
✐✎✒ Warning Protection
Search for the minimum If Phases VA, VB and VC Yes Voltage Sag Sign
▼

IED
✶❏●

✍
✐✎✔
and maximum magnitude <0.9 and > 0.1 pu Motor 1
✍
✐✎✖ bins in the histogram with
the highest frequencies for
✍
✐✎✘ No
phases A, B and C
✍
✑
✐✎
✑✘✔ ✐✎
✑✘✖ ✐✎
✑✘✘ ✐✎
✑✙✐ ✐✎
✑✙✒ ✐✎
✑✙✔ ✐✎
✑✙✖ ✐✎
✑✙✘ ✐✎
✒✐✐
✴❉❍❅ ✈▲✉ If Phases VA, VB and VC Yes Voltage Interruption Warning Sign
Calculate the absolute
< 0.1 pu
Fig. 5. Windowed three phase signal of a pre-fault cycle. values for the three phases

No
Obtain the lowest value
Fig. 6 shows the pre-fault histogram of phase A. After for the three phases
collecting the post-fault window of 1 cycle with ½ cycle If Phases VA, VB and VC
<1.1 and > 0.9 pu
Yes Steady State

step of three phase signals under consideration, the events Classify the
Fault Type
are classified through the analysis of the histograms
approaching the PDF, and the classification of each event Fig. 8. Flowchart of the algorithm to classify the type of
will be through characteristic values for each type of fault occurring event.
applied to the IES.
As a first step, the algorithm performs simultaneously the
histogram calculation by the values of the ratio between
✒✐
the frequency (number of occurrences) and the
magnitude of the signal under analysis at bus B3/BMTM
✑✕
on Motor 1, distributed into 50 bins (bars) for the three-
✦❒❅❑◆❅■❃❙

phases.
✑✐
Further, the algorithm searches for the minimum and
✕
maximum magnitude values contained in the bins with
the largest number of occurrences.
✐
✍✑ ✍
✐✎✙✍
✐✎✘✍
✐✎✗ ✍
✐✎✖✍
✐✎✕✍
✐✎✔✍
✐✎✓✍
✐✎✒✍
✐✎✑ ✐ ✐✎
✑ ✐✎
✒ ✐✎
✓ ✐✎
✔ ✐✎
✕ ✐✎
✖ ✐✎
✗ ✐✎
✘ ✐✎
✙ ✑
Thus, after the minimum and maximum values are found,
✭❁❇■❉
▼◆❄❅ ✈❐◆✉ the absolute value of the amplitude is then calculated.
Fig. 6. Histogram of phase A of the pre-fault signal – With the absolute values of minimum and maximum
B3/BMTM – M1.
magnitude, the algorithm seeks their lowest values and
then moves to the event classification step.
The histogram that represents the ratio of the frequency
Classification is performed by comparing the lowest
(occurrence number) versus the magnitude of the signal
under analysis in the B3/BMTM bus in Motor 1, enables amplitude value found to the threshold set in Fig. 9 [10].
the classification of the disturbance by observing the larger After performing the classification, the algorithm sends a
number of samples (occurrences) found in maximum warning sign to the alarm relays of the IED.
magnitudes of the signal in question, as shown in Fig. 7. The classification of events is based on a range of values
After execution and analysis of various simulations, an as shown in Fig. 9. It is possible to classify the type of
algorithm was developed through MATLAB software, disturbance faced by Motor 1 from the analysis of each
which allows the analysis of the maximum magnitude phase of the system simultaneously considering only one
values contained in the histogram bins for each voltage cycle of the faulty signal.
phase under analysis.
Voltage (pu)

1.80
✒✐

Swell
✑✕
✦❒❅❑◆❅■❃❙

1.10
Steady
✑✐ 0.95
State
0.90
✕
Sag
✐
✍✑ ✍
✐✎✙✍
✐✎✘✍
✐✎✗ ✍
✐✎✖✍
✐✎✕✍
✐✎✔✍
✐✎✓✍
✐✎✒✍
✐✎✑ ✐ ✐✎
✑ ✐✎
✒ ✐✎
✓ ✐✎
✔ ✐✎
✕ ✐✎
✖ ✐✎
✗ ✐✎
✘ ✐✎
✙ ✑
✭❁❇■❉
▼◆❄❅ ✈❐◆✉ 0.10
Interruption
Fig. 7. Histogram reflecting the frequency (occurrence 0.00
number) versus the magnitude of the signal under analysis. Phase Fig. 9. Range of values considered to associate the events to
A - Post-fault – B3/BMTM – M1. SDVV (p.u.) [13].

Thus, the classification of each event associated with For future work, it is intended to apply the PDF with
SDVV in the IES under analysis, was developed according other statistical tools that can correlate the three-phase
to the flowchart of Fig. 8.

https://doi.org/10.24084/repqj14.250 148 RE&PQJ, No.14, May 2016

signals of voltages registered with indicators that allow the horizontal axis shows the maximum magnitude of the
prediction or forecast of the disturbance and the TPIM signal at 0.10 p.u. for phase B and 0.08 p.u. for phase C.
performance and then alert the protection IED indicating
✓✕
the bus, the severity and the development (evolution) of
✓✐
the faced disturbance.
✒✕

5. Results

✦❒❅❑◆❅■❃❙
✒✐

✑✕
This section presents the results obtained from the analysis
✑✐
tool for classifying PQ disturbances in the IES. The
dynamic response of Motor 1, simulated in adverse ✕

operational conditions of the IES, enabled the ✐

✍
✐✎✐✖ ✍
✐✎✐✔ ✍
✐✎✐✒ ✐ ✐✎
✐✒ ✐✎
✐✔ ✐✎
✐✖ ✐✎
✐✘ ✐✎
✑✐
synthetization of the results in several characteristic ✭❁❇■❉
▼◆❄❅ ✈❐◆✉
patterns for each applied short circuit. Fig. 12. Histogram Graph and PDF for phase B monitored
on bus B3/BMTM - M1.
A. Analysis of voltages in phases A, B and C resulting
from a three phase fault involving phases "ABC" on Thus, the smallest value for each phase of the voltage
the bus B1/BAT 138 kV signal was obtained from the algorithm, wherein phase A
presented 0.061 p.u., phase B 0.068 p.u. and phase C
Fig. 10 shows a cycle of the post-fault voltage signal with 0.080 p.u.. Respecting the range of values considered to
window displacement of ½ step cycle regarding phases A, associate the events to SDVV (Fig. 9), the disturbance
B and C monitored on bus B3/BMTM in Motor 1 for a was rated as a voltage interruption.
three phase short circuit applied to bus B1/BAT 138 kV. ✔✕

✐✎
✑ ✔✐
✰❈❁▲❅ ✡
✐✎
✐✘ ✰❈❁▲❅ ✢ ✓✕
✰❈❁▲❅ ✣
✐✎
✐✖
◆❄❅ ✈❐◆✉

✓✐
✐✎
✐✔
✦❒❅❑◆❅■❃❙

✒✕
✐✎
✐✒
▼
❁❇❅ ✭❁❇■❉

✒✐
✐

✍✐✎
✐✒ ✑✕

✍✐✎
✐✔ ✑✐
▼
✶❏●

✍✐✎
✐✖
✕
✍✐✎
✐✘
✐
✍
✐✎✑ ✍
✐✎✑✐ ✍
✐✎✐✘ ✍
✐✎✐✖ ✍
✐✎✐✔ ✍
✐✎✐✒ ✐ ✐✎
✐✒ ✐✎
✐✔ ✐✎
✐✖ ✐✎
✐✘
✭❁❇■❉
▼◆❄❅ ✈❐◆✉
✐✎
✒✓✑ ✐✎
✒✓✓ ✐✎
✒✓✕ ✐✎
✒✓✗ ✐✎
✒✓✙ ✐✎
✒✔✑ ✐✎
✒✔✓ ✐✎
✒✔✕
✴❉❍❅ ✈▲✉ Fig. 13. Histogram graph and PDF for phase C monitored on
Fig. 10. A cycle of the post-fault voltage signal regarding bus B3/BMTM - M1.
phases A, B and C monitored on bus B3/BMTM - M1 to a three-
phase short circuit applied to B1/BAT 138 kV. B. Analysis of voltages in phases A, B and C due to a
fault involving phases "C" and "A" on bus B2/BMT1
Upon such situation applied, it is possible to sort the type 11.95 kV
of disturbance through the analysis of the phase histogram.
Fig. 11 shows the histogram for the frequencies of Fig. 14 shows the voltage post-fault signal windowed (1
occurrence of magnitude of the voltage signal of phase A, cycle) at ½ cycle step related to phases A, B and C
where the bin on the horizontal axis shows the maximum monitored on bus B3/BMTM in Motor 1, for a fault
signal magnitude of 0.08 p.u.. situation involving phases C and A, applied on bus
✓✕
B2/BMT1 11.95 kV.
✓✐ ✐✎
✘
✰❈❁▲❅ ✡
✐✎
✖ ✰❈❁▲❅ ✢
✒✕ ✰❈❁▲❅ ✣
◆❄❅ ✈❐◆✉

✐✎
✔
✦❒❅❑◆❅■❃❙

✒✐
✐✎
✒
▼
❁❇❅ ✭❁❇■❉

✑✕
✐
✑✐
✍
✐✎✒
▼
✶❏●

✕ ✍
✐✎✔

✐ ✍
✐✎✖
✍
✐✎✐✘ ✍
✐✎✐✖ ✍
✐✎✐✔ ✍
✐✎✐✒ ✐ ✐✎
✐✒ ✐✎
✐✔ ✐✎
✐✖
✭❁❇■❉
▼◆❄❅ ✈❐◆✉
✍
✐✎✘
✐✎
✒✔✔ ✐✎
✒✔✖ ✐✎
✒✔✘ ✐✎
✒✕✐ ✐✎
✒✕✒ ✐✎
✒✕✔ ✐✎
✒✕✖ ✐✎
✒✕✘
Fig. 11. Histogram graph and PDF for phase A ✴❉❍❅ ✈▲✉

monitored on bus B3/BMTM - M1. Fig. 14. Windowed signal to a phase-to-phase fault situation
involving phases C and A monitored in B3/BMTM - M1 and
Figs. 12 and 13 show the histogram for the frequencies of applied on the bus B2/BMT1 11.95 kV.
magnitude occurrence of the voltage signal for phases B
and C, respectively. It is noted that the bin in the

https://doi.org/10.24084/repqj14.250 149 RE&PQJ, No.14, May 2016

Upon such situation applied, it is possible to sort the type 6. Conclusion
of disturbance through the analysis of the phase histogram.
Fig. 15 shows the histogram for the frequencies of This paper presented a new stochastic-based
occurrence of the voltage signal magnitude of phase A. It methodology for classifying short duration voltage
is noted that the bin has a maximum magnitude of the variations sensed by three phase induction motors in
signal close to 0.04 p.u.. industrial plants. The main idea is to use it in a
✓✕
monitoring system integrated with the motors’ IED so
that an incident and harmful abnormal condition can be
✓✐
managed in a timely way to avoid damage to the
✒✕ machine.
The stochastic tool (PDF) used in the research is easy to
✦❒❅❑◆❅■❃❙

✒✐
be implemented and provides a fast and reliable
✑✕
classification of the disturbances, regardless the system
✑✐ bus to which the fault is applied. In addition, the results
✕ showed that the voltage variations were classified
correctly in all the cases tested in the paper.
✐
✍
✐✎✐✘ ✍
✐✎✐✖ ✍
✐✎✐✔
✭❁❇■❉
▼
✍
✐✎✐✒
◆❄❅ ✈❐◆✉
✐ ✐✎
✐✒ ✐✎
✐✔
For future work, it is intended to apply the PDF with
Fig. 15. Histogram graph and PDF of phase A monitored on other statistical tools that can correlate the registered
bus B3/BMTM - M1. three phase voltage signals with indicators that allow to
predict or forecast the disturbance and the TPIM
The histogram of Figs. 16 and 17 shows that the maximum performance and alert the protection IED, by indicating
magnitude of the signals with a higher frequency are close the bus, the severity and development (evolution) of the
to 0.60 p.u. for voltage signals of phases B and C. disturbance.
✒✕
Acknowledgement
✒✐
The authors would like to thank the University of São
✑✕ Paulo and the São Carlos School of Engineering for the
✦❒❅❑◆❅■❃❙

infrastructure provided for the development of this work,

✑✐ as well as the São Paulo Research Foundation (FAPESP)
and the National Council for Scientific and
✕
Technological Development (CNPq) for the financial
support provided in different parts of this research.
✐
✍
✐✎✖ ✍
✐✎✔ ✍
✐✎✒ ✐ ✐✎
✒ ✐✎
✔ ✐✎
✖
✭❁❇■❉
▼◆❄❅ ✈❐◆✉
Fig. 16. Histogram graph and PDF of phase B monitored on References
bus B3/BMTM - M1.
[1] R. C. Dugan, M. F. McGranaghan, S. Santoso and H. W.
The type of disturbance was classified by analyzing the Beaty, Electrical Power Systems Quality, McGraw-Hill,
histogram graphs of phases A, B and C; obtaining voltage 2012.
values of 0.04 p.u. for phase A (rated as a voltage [2] D. G. Lieberman, R. J. T. Troncoso, R. A. O. Rios, A. G.
interruption) and 0.59 p.u. for phase B and 0.65 p.u. for Perez, and E. C. Yepez, “Techniques and methodologies for
phase C (rated as a voltage sag), respectively. power quality analysis and disturbances classification in
For the moment, the proposed methodology was able to power systems: a review,” IET Generation, Transmission
and Distribution 5(4), 519-529, 2011.
sort 100% of the events related to 30 standards presented [3] W. Fan, Y. Liao, T. Laughner, B. Rogers, G. Pitts, et al.,
and analyzed. “Intelligent data analysis for power systems,” In IEEE PES
✒✕
General Meeting, 2012.
[4] R. Bernardes and F. Ayello, “PQMS – Power quality
monitoring system: improve power systems through ieds ,”
✒✐
in 20th Internacional Conference on Electricity Distribution,
IEEE, 2009.
✑✕
✦❒❅❑◆❅■❃❙

[5] S. Haykin and M. Moher, Comunication Systems, Wiley,

2009.
✑✐
[6] D. C. Montgomery and G. C. Runger, Applied Statistics and
Probability for Motorers, LTC, 2012.
✕
[7] J. L. Devore, Probability and Statistics for Motorering and
the Sciences, Cengage Learning, 2015.
✐
✍
✐✎✖ ✍
✐✎✔ ✍
✐✎✒
✭❁❇■❉
▼
✐
◆❄❅ ✈❐◆✉
✐✎
✒ ✐✎
✔ ✐✎
✖ [8] Z. Gang, A. Duffy, H. Sasse and W. Lixin, “The use of
Fig. 17. Histogram graph and FDP of phase C monitored on Probability Density Funtions to improve the interpretation
bus B3/BMTM - M1. of FSV results,” in IEEE International Symposium on
Electromagnetic Compability (EMC), pp. 685-689, 2012.
[9] DIgSILENT PowerFactory, “Power System Analysis
Software,” Available: http://www.digsilent.de, 2010.
[10] IEEE 1159, “Recommended Practice for Monitoring
Electric Power Quality,” 1995.

https://doi.org/10.24084/repqj14.250 150 RE&PQJ, No.14, May 2016