See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/331023505

Flicker estimation associated with PV integrated distribution network

Conference Paper · April 2018

CITATIONS READS
5 1,367

5 authors, including:

Shahinur Rahman Masood Moghaddami

University of Alabama Florida International University
16 PUBLICATIONS 198 CITATIONS 41 PUBLICATIONS 837 CITATIONS

SEE PROFILE SEE PROFILE

Arif Sarwat Tèmítáyò O. Olówu

Florida International University Idaho National Laboratory
216 PUBLICATIONS 4,454 CITATIONS 69 PUBLICATIONS 668 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Inductive Power Transfer Systems View project

CRII: CPS: Internet-Inspired Autonomous Electric Vehicle (EV) Charging View project

All content following this page was uploaded by Mohamadsaleh Jafari on 22 March 2020.

The user has requested enhancement of the downloaded file.

 Flicker Estimation Associated with PV Integrated
Distribution Network
Shahinur Rahman, Masood Moghaddami, Arif I. Sarwat, Temitayo Olowu, Mohammadsaleh Jafaritalarposhti
Department of Electrical & Computer Engineering, Florida International University, USA
Email:{srahm026, mmogh003, asarwat, tolow003, mjafa006}@fiu.edu

Abstract— The interest for solar Photovoltaic (PV) systems PV power production relies on the weather patterns, seasonal
installation has been recently increasing due to their reducing cost variability and location [10]. The fast change of solar insolation
and high electricity price from the utilities. However, the PV level due to cloud movements creates significant variations in
power generation is sensitive to variable solar irradiation, which PV power production, which could result flicker at the
causes technical difficulties related to power quality and control.
Severity of those problems directly depends on PV penetration
connecting node.
level, grid design and geographic location. This paper explores the
flicker indices of grid-tied PV power plant, one of the major power Flicker is considered as synonym with light flicker, lamp
quality issues. The study has been carried out on a real grid- flicker or voltage flicker. According to International
integrated PV system, situated at Engineering Center, Florida Electrotechnical Commission (IEC) 61000-3-3:1994 [3],
International University (FIU). The field measurements of various flicker or light flickering is defined as the subjective impression
solar output power and flicker quantifying indices at the point of of light fluctuations on the eye due to voltage change.
common coupling (PCC) have been recorded using data Flickering in light output happens due to fluctuation in system
acquisition system and power quality meter respectively. The voltage, where this variability frequency is within 2 to 10Hz
collected data has been categorized and analyzed for voltage
fluctuations and flicker. Different approaches have been done to
range. Flicker causes a problem of visual perception from
investigate the influence of PV power plant on the flicker indices. human eye observation. Mainly, voltage variations are caused
Results show that solar output power has a marginal effect by load connection with electric power system, deviations in
towards flicker indices with existing penetration level. Load power consumption, specifically, reactive power imbalance and
connection and variation are the key indicators of high flicker second harmonics components [4]. Connecting or
value during observation time. disconnecting big loads or loads with inherent transient
response result power fluctuation in the distribution network.
Index Terms—Flicker, Point of Common Coupling (PCC), Short Loads such as, arc furnaces, welding machine, air conditioning
Term Flicker (Pst), Long Term Flicker (Plt).
equipment, lifts, pumps, fans, mine cranes, mills, piston
compressors, wind power generators, capacitor banks, and PV
I. INTRODUCTION generators cause continuous rapid voltage change. Significant
Solar Photovoltaic technology is one of the emerging variations in light output, a sense of visual irritation, image
sources in worldwide that helps to decrease carbon emission distortion and faults in electronic operation are the
and to fulfill electricity demand. Present growth of rooftop consequences of flicker in electricity distribution networks,
based PV systems and feed-in-tariff (FIT) have accentuated depending on the appearing frequency and fluctuation
more installations in commercial and residential applications magnitude.
[1]. According to Solar Energy Industries Association (SEIA)
report, the installed PV capacity has extended to 49.3 gigawatts Solar Photovoltaic plants, interconnected with the electric
(GW), enough to supply electricity 9.5 million American networks must maintain IEEE Standard 1453-2015 [5], for the
houses. Also, gradual decrease of installation cost has increased interconnection of distributed generation resources into the
average PV installation size over recent years. As a result, power grid. This paper intends to conduct a case study
penetration of installed PV systems has been increasing in the regarding to the influence of PV power generation on the flicker
low voltage (LV) distribution network. Therefore, the power level in the low voltage distribution network. This work
flow direction may change and as a result substantial changes investigates voltage fluctuation problems of a 1.41MW PV
to the design of the power system will be essential. The installed plant with existing 16% penetration level. Also, it describes
PV capacity becomes the determining factor for grid design flicker issues associated to the mentioned plant and possible
when PV system feeds more power into the distribution reasons behind it. Flicker indices measurement data for four
network than consumer demand. At high PV penetration level, months observation period was recorded and analyzed for
voltage variations and flicker could be one of the major flicker and power quality issues. Section II explores various
obstacles for any distribution system. A number of researchers flicker quantifying indices and standards to ensure quality
have considered the impact of PV penetration as a serious issue. power in the power system. The description of FIU PV power
It is found in recent publication that high penetration of PV plant and data measurement approach are presented in Section
initiates negative effects on voltage magnitude [2]. Additionally III. The different analysis approaches of flicker occurrence and
corresponding results are described in section IV. The outcomes
This work is supported by the National Science Foundation under Grant No. of the observation are summarized in Section V.
1553494. Any opinions, findings, and conclusions or recommendations
expressed in this material are those of the authors and do not necessarily reflect
the views of the National Science Foundation.
 TABLE I: Compatibility Levels of Flicker Indices Based on
TNB Technical Guidebook [2]

Short term flicker Long term flicker

severity (Pst) severity (Plt)
LV System 1.0 0.8
11kV-33kV 0.9 0.7
Above 33kV 0.8 0.6

III. SOLAR PHOTOVOLTAIC PLANT: CASE STUDY

In order to analyze voltage fluctuations and flicker due to
grid-tied PV connection, a comprehensive study was conducted
on a 1.41MW AC PV power plant, tied to a distribution feeder
network at Miami, Florida (Fig. 1). Three types of solar panels
Fig. 1: View of 1.41MW PV Power Plant, Located at EC, Florida International are installed in this PV power plant: Jinko Solar, CSUN, and
University (FIU) Canadian Solar. The plant comprises of 4,460 PV modules of
three different types, each with rated power around 315W. 46
SUNNY TRIPOWER 24000TL-US smart string inverters (each
II. FLICKER PHENOMENON QUANTIFICATION of 24kW size), convert DC Power generated by PV to AC. A
Flicker quantification is related to a number of technical, local weather station records irradiance and both ambient as well
psychological and physiological aspects that arise as module temperature. An Elkor production meter records the
simultaneously. Flickermeter, regulated by Standard IEC plant’s net energy production. Each inverter module has
61000-4-15:2010 is used for flicker measurement. It measures different breakers and panel boxes for testing and disconnection.
the level of voltage change as well as replicates the reaction of A RevolutionR Wireless Power Quality Recorder, connected to
human eye and response of light source. According to standards the lower side of transformer at point of common coupling
IEC 61000-4- 30:2008 and EN 50160:2010, the main indices (PCC), records flicker indices (IFL, Pst and Plt) and other power
which quantify flicker levels and therefore, quantify electricity quality data with a resolution of 1 minute interval. An AC
quality are: instantaneous flicker (IFL), percentile short-term disconnect box is connected to the AC output of the inverter and
flicker (Pst), and percentile long-term flicker (Plt) [6]. Higher has necessary protection devices such as switch and fuse. A Data
magnitudes leads to disturbing cognizance caused by voltage Acquisition System (DAS) measures the multivariate time-
and luminance variation of light sources, approved by more than series data from inverters, meter and weather station and
50% of the test persons. IFL quantifies sudden voltage securely stores these values in a cloud server. The plant is
fluctuation. Phenomenon that cause a voltage sag, immediately currently operating at 16% penetration level, with a peak
cause a change in the IFL data. Pst is determined by the values generation of 1MW on a clear, bright sunny day and a peak
of cumulative weighted probability of flicker perception levels circuit demand of 6.7MW on the feeder under study. The plant
at 0.1%, 1%, 3%, 10% and 50% of every 10 minutes interval is located at a distance of 1.7 miles from the feeder’s substation.
using (1). Although the substation services eight feeders, this paper deals
with the impact of the PV plant on one feeder alone. At the
ܲ௦௧ ൌ  ඥͲǤͲ͵ͳͶܲ଴Ǥଵ ൅ ͲǤͲͷʹͷܲଵ ൅ ͲǤͲ͸ͷ͹ܲଷ ൅ ͲǤʹͺܲଵ଴ ൅ ͲǤͲͺܲହ଴ substation, there is a 138kV/23kV step-down transformer and
………. (1) circuit breakers, tapping into different feeders for power
where, P0.1, P1, P3, ....etc remark as the ratio of relative distribution. At the plant, a 480V/277V, 2000 kVA oil-cooled
voltage difference (∆V = Vmax - Vmin) to the base voltage (Vnom). step-down transformer is present. It is noteworthy that the
Plt is measured as an average of Pst levels over 2 hours duration concerned feeder, which serves 700 residential and 200
using equation (2). commercial customers, has three capacitor banks, one voltage
ଵଶ regulator, and two transformers [7].
య
ͳ
ܲ௟௧ ൌ ඩ ෍ ܲ௦௧ଷ IV. FLICKER ANALYSIS AND RESULTS
ͳʹ
௜ୀଵ
The PQ meter recorded flicker data (IFL, Pst and Plt) with
………. (2) 1min resolution at the point of PV interconnection from August
The utility settings of flicker indices depends on the 18, 2017 to December 27, 2017. To characterize flicker for
agreement with the consumer. The power quality guidelines of different day operations, four months recorded data has been
interconnection are emphasized in TNB Technical Guidebook used. In addition, the solar irradiation and PV output power data
on grid connected PV system to low and medium voltage was also captured by data acquisition system. The flicker (IFL,
Networks [2]. It incorporates the standards such as IEC and Pst and Plt) recorded at output edge of PV inverter for
IEEE and provides a recommended practice to compromise corresponding operating days are presented in Fig. 2. The
safety and operation issues. Based on the TNB Technical severity of flicker is illustrated by different Pst (10 min) and Plt
Guidebook, the permissible flicker levels are shown in Table I. values (2 hour).
The limiting value of Pst and Plt are 1.0 and 0.8 for low voltage
distribution network.
 values along with PV output power of these three sunny and
three cloudy days are plotted in Fig. 3. Green represents the
power output (kW) and red, yellow and blue mark as the flicker
profile of three phases for these two exemplary days. Result
corresponds to the recorded Pst values only because of higher
temporal resolution compared to the Plt values. High spike
flicker values were observed at early morning on 12/10/2017
(sunny day) when PV did not generate power. These higher
flicker values may be because of fault, because of loads variation
or internal grid scenario. Some occasional spikes were observed
during day time (both sunny and cloudy days). The exact reason
of flicker in different weather conditions is not clear in Fig. 3.
(a)

(b)

Fig. 3: Time vs PV Output Power and Short Term Flicker Index (Pst) for Three
Typical Sunny and Cloudy Days.

B. Flicker Study Based on PV Generation

To determine the PV plant dependency on flicker in the low
voltage distribution network, scatter plot of PV output vs flicker
indices (Pst, Plt) has been displayed in Fig. 4. Maximum Pst
values were below 1.0 (black line in Fig. 4(a)). Therefore, Pst
crossed IEEE Standard 1453-2015 voltage fluctuation limit at
(c)
Fig. 2: Measured Flicker Indices at PV Inverter Output: (a) IFL, (b) P st (c) Plt
times of observation period. Same result has been found for long
term flicker severity (Fig. 4(b)). Most of the measurement time,
It is observed in Fig. 2(a) that the IFL was frequent and heavy Plt was under 0.8 (black line in Fig. 4(b)) but sometimes it
in September and first few days of October and December due crossed the Standard limit. Flicker profile did not follow definite
to rapid voltage fluctuations. The measured Pst and Plt were trend with increasing PV output. It is concluded from this
exceeded the utility prescribed flicker emission limits as per subsection that the flicker level is nearly independent of PV
IEEE 1453-2015 standard of voltage fluctuation (showed in Fig. output power.
2(b) and 2(c)). Primarily, it is assumed that the observed flickers C. Statistical Analysis of Flicker Dependency on PV Plants
were generated by either PV systems or loads, or both. The To identify the effect of PV plants on flicker profile, the
harmonics generated by non-linear loads, have frequency correlation analysis between flicker and PV output power is a
components below the line frequency which cause flicker in the good approach. Correlation coefficient, +1 and -1 indicate
distribution system. Another possible reason is cloud strongest positive and negative correlation respectively. 0
movements over the solar panel which reduces the solar input. indicates no correlation on that scale. Hence, values approach to
Due to low solar insolation, the output of the PV panels would zero, reveals weak correlation. Correlation Coefficient of three
be low which results in temporary increase in the harmonics phase flicker indices (both Pst and Plt) and PV System are
generated by the inverter. calculated and showed in Table II. It can be concluded from the
low values of correlation coefficient that there is insignificant
A. Comparison of Flicker Index on Cloudy Day with Clear Sky relation between PV output and flicker indices.
Day
Solar output power of three most sunny and cloudy days are
analyzed and compared in this subsection. The inherent flicker
 (a) (a)

(b)
Fig. 4: Flicker Index as a Function of Solar Output Power (a) Pst (b) Plt (b)
Fig. 5: Linear Regression Line Analysis of Flicker Index and PV System (a) Pst
(b) Plt

TABLE II: Correlation Coefficient of Three Phase Flicker

Indices and PV System.

Correlation Phase A Phase B Phase C

Coefficient
R (Pst) 0.0067 0.0073 0.0059
R (Plt) 0.0154 0.0180 0.0092

Fig. 5 presents linear regression line analysis of Pst and Plt

under mentioned grid connected PV system. All Pst values are
analyzed under different solar PV outputs. Small positive slope
of Pst regression line equation in Fig. 5(a) indicates a slight
positive relation between flicker and PV output. Similarly,
regression line of Plt has been constructed based on real flicker
Fig. 6: Classified Flicker Indices into Three Cluster Groups [Pst (left) and Plt
and PV power data, measured at PCC. The equation in Fig. 5(b) (right)]
proves a marginal positive relation between Plt and PV output. It
can be concluded from the correlation and linear regression line TABLE III: Cluster based Events Count of Short Term Flicker
analysis that the flicker produced from PV systems is lower than (Pst)
the one created by loads of the investigated GC-PV system with Cluster Range No of Events No of Events No of Events
existing penetration level. (Phase A) (Phase B) (Phase C)
D. Flicker Data Classification using Data Clustering 0< Pst ≤ 0.5 117927 117987 117997
Technique 0.5< Pst ≤1.0 30 60 60
Pst>1.0 260 170 160
A clustering method has been applied to estimate the flicker
severity at PCC during observation period. All Pst values have
TABLE III: Cluster based Events Count of Short Term Flicker
been divided into three groups as following using clustering
(Plt)
algorithm. Cluster 1 (safe region): 0< Pst ≤ 0.5; Cluster 2 (close Cluster Range No of Events No of Events No of Events
to threshold value): 0.5< Pst ≤1.0; Cluster 3 (Above threshold (Phase A) (Phase B) (Phase C)
limit): Pst>1.0. Similarly, all Plt values have been classified into 0< Plt ≤ 0.4 115897 116847 117017
three clusters as follows: Cluster 1 (safe region): 0< Plt ≤ 0.4; 0.4< Plt ≤0.8 1330 490 600
Cluster 2 (close to threshold value): 0.4< Plt ≤0.8; Cluster 3 Plt>0.8 1080 880 600
(Above threshold limit): Plt >0.8.
 All flicker values for each phase have been counted and higher load variations have adverse effects on voltage profile
categorized into different corresponding groups. The classified and thus, on flicker. Each phase has different no of events per
number of events both for Pst and Plt are shown as a tabular form hour above threshold (both Pst and Plt), because of different
in Table III and IV respectively and as a chart in Fig. 6. Flicker loading scenario for each phase. It concludes that the primary
indices (both Pst and Plt) were within safe level for almost 99% reasons of flicker during observation time are load pattern and
of observation time. The events mentioned in cluster 3, violated power system internal issues. PV power plant has marginal
the permissible voltage fluctuation limits in accordance to the contribution on the upstream network flicker level with existing
IEEE 1453-2015 standard for less than 1% time, which are total penetration level. This conclusion is only valid under present
of around 260 times for Pst and 1080 times for Plt in phase A. planning and configuration of grid. These statements will not be
From Fig. 6, it is noticed that cluster 1 for both Pst and Plt (green applicable, if this scenario alters. PV Systems may generate
portion of pie chart) has the highest count recorded. In the worst more voltage fluctuation during high PV penetration.
conditions, cluster 3 of Plt for Phase A has been counted a total
of 1080 events, which is insignificant considering nearly 0.2
million observation data during four month study. V. CONCLUSION
Grid connected PV power plants can adversely affect the
power quality of distribution network, which must be within
regulation limits according to IEEE Standard. This article has
investigated the influence of 1.41 MW PV power plant, located
at FIU, Engineering center campus on the distribution network
flicker level with 16% PV power penetration. It is observed that
permissible value of the Pst (1.0) and Plt (0.8) for better power
quality is not always followed during observation time. 0.72%
and 0.16% of Plt and Pst crossed the standard limit, though these
numbers are insignificant. Different approaches are taken into
account to determine the impacts of the PV power feed on the
flicker level. Calculated correlation coefficient between flicker
value and PV output shows marginal relationship between them.
Fig. 7: Time vs Probability Count of Pst Above Threshold Value According to the conducted study, it can be concluded that a
single reason is not responsible to produce noticeable flicker
during observation period. It is an aggregated effect resulting
from several flicker sources and contributing factors, such as
network configuration, atmospheric conditions and PV plant
location. Although flicker effects during observation time seems
to be lower, these values may cross the standard limits and affect
consumer at high PV penetration and high rate changes in solar
irradiation. Proper protective scheme should be considered
during planning and implementation stage.
REFERENCES
[1] US Solar Photovoltaic System Cost Benchmark: Q1 2017 - NREL. Online
Available: https://www.nrel.gov/docs/fy17osti/68925.pdf
Fig. 8: Time vs Probability Count of Plt Above Threshold Value [2] A. H. Faranadia, A. M. Omar, S. Z. Noor, "Voltage flicker assessment of
15.3kWp grid connected photovoltaic systems", 2017, IEEE 8th Control
and System Graduate Research Colloquium (ICSGRC), p. 110 – 115
E. Time Based Analysis of Flicker Indices above Threshold [3] Gabriel Tanasescu, Stefan Busoi, Bogdan Gorgan, Petru V. Notingher,
Value "Photovoltaic power plants contribution to the flicker level in the medium
Figs. 7 and 8 present time base analysis of flicker indices voltage energy network. Case study", 2016 Conference on Diagnostics in
which shows the time when flicker values were above threshold. Electrical Engineering (Diagnostika), p. 1 – 4
[4] Zai Peng Goh, Mohd Amran Mohd Radzi, Hashim Hizam, Noor Izzri
The horizontal and vertical axis represents time intervals and the Abdul Wahab, "Investigation of severity of voltage flicker caused by
events count when indices were above threshold value according second harmonic", IET Science, Measurement \& Technology, 2017,
to IEEE Standard. The recorded number of daily flicker data Volume: 11, Issue: 3, Pages: 363 - 370
above threshold values of four months period has been [5] IEEE Recommended Practice for the Analysis of Fluctuating Installations
on Power Systems IEEE Std 1453-2015 (Revision of IEEE Std 1453-
categorized and superimposed on 24 hours time framework. 2011), 2015, Pages: 1 - 74
High value of flicker, both Pst and Plt, were observed during both [6] Annapoorna Chidurala, Tapan Saha, N. Mithulananthan, "Field
PV generation time (from 7:00 to 19:00) and when PV did not investigation of voltage quality issues in distribution network with PV
generate power. Thus, flickers that appear on low or even no PV penetration", 2015 IEEE PES Asia-Pacific Power and Energy Engineering
Conference (APPEEC), p. 1 - 5
power injection instants, can be attributed to loads. It is obvious [7] Arash Anzalchi, Aditya Sundararajan, Amir Moghadasi, Arif Sarwat:
in Fig. 7 and 8 that the events in which flicker levels are above "Power quality and voltage profile analyses of high penetration grid-tied
threshold value during daytime, were observed due to load photovoltaics: A case study", IEEE Industry Applications Society Annual
variations. The connected nonlinear loads, such as motors, Meeting, 2017, pp. 1-8
[8] Andreas Spring, Georg Wirth, Gerd Becker, Robert Pardatscher, Rolf
industrial and agricultural machineries, electronics device, Witzmann, "Effects of Flicker in a Distribution Grid with high PV
 Penetration", 28th European Photovoltaic Solar Energy Conference and system”, 2015 IEEE International Conference on Industrial Technology
Exhibition. (ICIT).
[9] Sundararajan, A.I. Sarwat, "Roadmap to Prepare Distribution Grid-Tied
Photovoltaic Site Data for Performance Monitoring", International
Conference on Big Data, IoT \& Data Analytics, December 2017
[10] Joan D’Rozario, Sabrina Shams, Shahinur Rahman, Ahmed Sharif,
Enamul Basher, “Cost effective solar-biogas hybrid power generation

View publication stats