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

net/publication/326995860

Tuning of Voltage Regulator Control in Distribution Systems with High

Renewable Penetration

Conference Paper · August 2018

CITATION READS

1 974

1 author:

Hossein Ghassempour Aghamolki

Eaton
9 PUBLICATIONS   128 CITATIONS   

SEE PROFILE

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

Volt-VAR management in Distribution Systems with High Renewable Penetration View project

All content following this page was uploaded by Hossein Ghassempour Aghamolki on 12 October 2018.

The user has requested enhancement of the downloaded file.

Tuning of Voltage Regulator Control in Distribution
Systems with High Renewable Penetration
Wei Ren, Senior Member, IEEE Hossein Ghassempouraghamolki, Member, IEEE
Eaton Corporate Research & Technology Eaton Corporate Research & Technology
Eden Prairie, MN, USA Eden Prairie, MN, USA
[email protected] [email protected]

Abstract— Rapid introduction of renewable generation based 3) Consumer’s environmental awareness: Many utilities
DER is challenging the traditional operation strategy of the are reviewing and reforming their interconnection process
radial distribution system. Switchable voltage regulation devices to deal with the long queue of DER connection request.
such as voltage regulator with pre-defined setting are
traditionally used to control the voltage within the ANSI High penetration of DER, especially of intermittent
standard limits. However, the ever-increasing penetration of renewable generation, brings many challenges to the operation
DER with inherent intermittency, turns the slow trend of the net of distribution systems. One that has been known by many is
load profile to a more fast dynamic trends which brings many the voltage regulation [3]. A few factors contribute to this
challenges to the operation of distribution systems including problem. 1) Due to the much smaller distance between
voltage regulation. In this paper we investigate some of the conductors, a feeder line usually has a rather low X/R
challenges that voltage regulator have to face in the feeders with impedance ratio as compared to that of a transmission line.
high-penetration DERs such as power quality concerns and high This characteristic results in high voltage sensitivity to change
number of tap operations and explore the effect of the voltage of real power. As the renewable DER varies its real power
regulator control parameters on those issues by leveraging output frequently, so are the voltages across the feeder. 2) As
CYME software’s long-term dynamic simulation toolbox (time the rated power of the DER increases, power can now flow in
series simulation) on the modified IEEE 33-node distribution the opposite direction from the feeder end towards the feeder
feeder. This study would help choosing the right parameters for head. When a traditional feeder voltage regulation scheme is
voltage regulator control to deal with the excessive switching
configured to have the feeder head voltage set close to the
problem without compromising on power quality.
higher limit with assumption that the voltage profile will
Index Terms— Volt/VAR Control, Voltage Regulator, follow a monotonic decreasing pattern towards the feeder end,
Distribution network, High DER penetration, PV intermittency, this reverse power flow now violates the assumption and
Power Quality, Flicker studies. could result in over-voltage at the feeder end. 3) The legacy
control of traditional voltage regulation devices such as
voltage regulators and capacitor banks is too slow, with tens of
I. INTRODUCTION
seconds delay that is suitable for the smooth feeder load
Distributed energy resources (DER), dominated by profile, to chase the voltage variation caused by renewable
renewable generation, are penetrating into electric systems intermittency without much improvement to feeder voltage
across the United State at an unprecedented speed. The NREL quality.
Renewable Electricity Futures Study [1] predicts that
renewable electricity generation will be more than adequate to Solutions to the above mentioned voltage regulation
supply 80% of the total U.S. electricity demand by 2050. problem have been discussed in many publications. These
Behind this fast trend stand three main driving forces. solutions can be categorized into the following groups.

1) Government policies: Public Utilities Commission A. Through System Planning

(PUC) in various states are requesting utilities to more EPRI published a technical report [4] in 2012 that
proactively plan for high penetration of DER integration. discussed using a stochastic method to find out the maximum
The Distribution Resource Plan (DRP) in California and hosting capacity of distributed solar PV. The limit is evaluated
the Distributed System Implementation Plan (DSIP) in on line loading, reverse power, protection coordination, and
New York are two such examples. power quality (e.g., voltage variation, flicker, etc.). This type
2) Declining renewable cost: As the annual installation of of feature is available today in some commercial planning
photovoltaic (PV) generation in the U.S. increases from tools to help utilities with their system impact study.
<900MW in 2010 to >14,000MW in 2016, the PV
module price drops from $1.85/W to $0.37/W [2].
B. Using DER Local Control is first provided in Section II. Section III describes the
The new revision of IEEE Standard 1547TM allows DER to simulation study to be carried out in this paper, and Section 0
exercise various controls to mitigate feeder voltage issues. A summarizes the study result.
few commonly applied control strategies include: constant
leading power factor operation [5], varying reactive power II. REVIEW OF VOLTAGE REGULATOR & CONTROL
control [7], droop curve based volt/Var control [8], and direct A voltage regulator is a device (usually an auto-
voltage regulation. transformer) that provides constant voltage output under
varying input voltages and load currents. By standard (i.e.,
C. Using Additional Conditioning Devices IEC 60076-21:2011, IEEE Std C57.15-2009), it regulates the
When added cost is tolerable, additional power electronics voltage between -10% and +10% of the nominal value in 33
devices can help mitigate the problem too. Solid state steps of its tap changer – each tap gives 5/8% of voltage
transformer (SST) [9], battery energy storage system (BESS) adjustment. Based on the actual types of voltage regulator and
[10], and μSTATCOM [11] are a few that have been discussed its connection configuration, these numbers may differ in real
in literatures. life operation.
Each vendor may have their own way to design the VR
D. Using Centralized Control
control, but some basic functions are commonly available.
[12] discussed using centralized control to respond to Figure 2 illustrates the operation of a Type A VR responding
changing operation status of the system and provide maximum to varying source side voltage. A few important control
operating margin in order to accommodate higher DER parameters are listed in the following:
penetration level.
 Set voltage (V_Set) – the voltage reference of the VR
With the fast growing PV penetration, some utilities are control.
already experiencing problems in their actual feeder systems,
such as excessive operation of voltage regulator (VR) that  Bandwidth – when the feedback voltage is out of this
wears and tears the device much faster than its expected life band around the set point (i.e., V_Set-Bandwidth/2, to
time. Unfortunately, although the active local control of real V_Set+Bandwidth/2), a tap operation will be issued after
and reactive power of DERs has proven to effectively mitigate a pre-defined time delay.
the voltage variation at the point of interconnection, it does not
always guarantees that the VR will see reduced number of  Control time delay – the delay time before a tap operation
operations. For example, Figure 1 below illustrates a very is taken after the regulating voltage is out of bandwidth
simple feeder line with a PV generator connected to the end of  Line drop compensation R and X – these impedance
it. When PV’s output power varies between 40% to 100%, the values allow the VR to regulate voltage at a distance (e.g.,
voltage at the middle of the line still sees >3V variation even at load center) from where it’s measured. R and X
if the PV control regulates its terminal voltage perfectly at the represent the impedance of the feeder line in between.
nominal level (assuming there is no worry on control fighting
Line drop compensation
with other voltage regulation devices). If a VR is installed at disabled in this case. Load
this location with a standard control bandwidth of 2V, it will side voltage therefore Source side voltage
react to the voltage variation and move its tap positions all the overlaps the source side
voltage Load side voltage
time.
PV
Substation
R=0.2pu X=0.5pu Bandwidth
(2V)

Set voltage
(120V)

Tap changer delay

(2s)

Control time delay

(30s)

Figure 2: Real time operation of a voltage regulator

Figure 1: Voltage profiles of a simple feeder with varying PV
output power
This paper tends to focus on the traditional VR control and
seek potential remedies inside of it to relieve the excessive
switching problem. A quick review of traditional VR control
 III. STUDY OF VR CONTROL UNDER RENEWABLE B. Default VR Control Configuration
INTERMITTENCY As it has been discussed in section II, The main tunable
In this section, a feeder system is modeled and simulated control parameters in voltage regulators are voltage reference
to study the impact of varying different control parameters to set point, voltage control bandwidth, active time delay, and
the operation of VR under high renewable intermittency. Line drop compensation. For many reasons, utilities usually
Simulation results are post-processed to provide a quantitative have the tendency to keep all the default control parameters
way of evaluating the VR performance. It is hoped that this except voltage reference set point intact or keep them in the
result will provide useful insight to the readers of what to look predefined values found during the planning and
for when it’s necessary to change the VR settings. commissioning. Such approach is acceptable for the
traditional distribution system without any DER penetration.
A. Simulation Setup
However, as it has been discussed earlier, introduction of
IEEE 34 bus system (as shown in Figure 3) has been DER to the distribution system requires more attention to the
modified and modeled in CYMDISTTM with a 700kW PV reconfiguration of voltage regulator parameters. The usual
system connected to bus 840. To manifest renewable default setting of the voltage regulator has shown in Table 1.
intermittency, a cloudy winter day of solar irradiance profile The same parameters has been used as the baseline scenario
obtained from a real PV system in Los Angeles, CA is
of the simulation in this paper.
applied to the PV panel. Unity power factor is assumed at the
PV output as according to old IEEE Std 1547 TM for the
Table 1: VR control parameter setting for the default case
purpose of better demonstration of the problem. Also, a real-
world load profile with 5 min resolution obtained from Voltage Set Voltage Active Voltage
Galveston feeder in Texas has been used for long-term Point Bandwidth Time delay control node
dynamic simulation. A three phase VR is connected between 123 V 2V 30 s Regulator
bus 814 and bus 850 and its operation is to be studied. Terminal

C. Evaluation Criteria
In evaluating the performance of VR operation, the
following four criteria are considered.
 Number of VR operations:
The tap changer of a VR has a typical life of 20 to 30
years for around 100,000 operations with proper
maintenance (although field experience verifies that this
number can sometimes go up to as high as 1,000,000
operations before actual mechanical failure). Excessive
Figure 3: Circuit diagram of IEEE 34 bus system operation of VR’s tap changer wears and tears the
apparatus faster than expected and shall be avoided.
 Voltage violation:
It is the responsibility of utilities to keep the customer’s
service voltage within an acceptable range. ANSI C84.1
specifies a guideline for this range as summarized in
Table 2.
Table 2: ANSI C84.1 voltage range for 120V base
Service Utilization
Figure 4: PV irradiance profile to represent a cloudy day Min Max Min Max
Range A
-5.0% +5.0% -8.3% +4.2%
(Normal)
Range B
-8.3% +5.8% -11.7% +5.8%
(Emergency)

However, ANSI C84.1 does not clearly define what is

“normal” voltage and what is “emergency” voltage. In
this paper, we take PG&E’s practice [13] and consider
two minutes as the minimum period for normal range
evaluation.
Figure 5: Galveston’s feeder Load profile
 Flicker:
 Voltage flicker is one of the most important metric to
power quality. It describes the level of voltage variation at
certain frequency range that may result in light fluctuation
noticeable by human eyes. IEC 61000-4-15:2010 defines
the short-term and long-term flicker severity to assess the
negative effect of this phenomenon. IEEE STd 1453-2011
adopts the method for testing and measurement of flicker.
A flicker meter is implemented according to the IEC
standard and applied to VR performance evaluation.
According to the standard, the compatibility levels in low
voltage power system are provided in Table 3.
Table 3: Compatibility levels in LV power systems
Compatibility Levels
Pst 1
Plt 0.8 Figure 6: Time series plot (Test 2)
D. Tuning of Control Parameters
Table 4 below summarizes the simulations performed on
the test feeder. Three control parameters are considered and
varied in five test cases.
Table 4: Simulation test setup
Test ID Bandwidth Time delay LDC
1 2V 30s Disabled
2 4V 30s Disabled
3 2V 45s Disabled
4 2V 60s Disabled
5 2V 30s Enabled
IV. SIMULATION RESULT
A. Time series plots of all node voltages and VR taps
Figure 5 to Figure 9 illustrate the simulation results in Figure 7: Time series plot (Test 3)
time series from the five test cases. The red dotted lines show
the normal service voltage range (Range A) of +/-5% on the
120V base.

Figure 8: Time series plot (Test 4)

Figure 5: Time series plot (Test 1)

  In all test cases, the flicker severity factors are rather
low and very little differences exist from the different
VR control settings
V. CONCLUSION
Utilities across the U.S. have already experienced more
than expected operation of their VR assets in the distribution
feeder and are seeking potential solutions. Although remedies
may exist in DER control as suggested by the new revision of
IEEE Std 1547, it’s not always possible for utilities to fully
rely on the DER owners to address the issue.
This paper evaluates the impact of a few commonly
available VR control parameters to the number of tap changer
operations with consideration of quantitative power quality
Figure 9: Time series plot (Test 5) metrics. As the simulation results reveal, there are margins
B. Tabulated Result of VR Evaluation for the adjustment of VR control to significantly reduce the
wear and tear of tap changers while still maintaining good
Table 5 compares the VR evaluation result based on the
power quality. The analysis method introduced in this paper
three criteria discussed in Section III. In the table, “N_Tap” is
may provide useful insights to utilities in deciding their field
the total number of tap changer operations of the VR in all
operation strategies.
three phases, “Voltage Violation” evaluates the time series
voltages at all nodes (in all three phases) and accumulate the REFERENCES
time of any voltage outside of the ANSI C84.1 normal range [1] National Renewable Energy Laboratory (NREL), “Renewable
for more than 2 minutes. The short term and long term flicker Electricity Future Study,” https://www.nrel.gov/analysis/re-
severity factor (Pst and Plt) are calculated on each node for futures.html
every 10 minute and 2 hour time segment. At the end, the [2] GTM Research and Solar Energy Industries Association (SEIA),
“U.S. Solar Market Insight Reports,” www.greentechmedia.com
maximum values of all calculated Pst and Plt factors are [3] Y. Liu; J. Bebic; B. Kroposki; J. de Bedout; W. Ren, “Distribution
taken and documented as “Flicker Pst_Max” and “Flicker System Voltage Performance Analysis for High-Penetration PV,”
Plt_Max” in the table. 2008 IEEE Energy 2030 Conference, Year: 2008, Pages: 1- 8
Table 5: VR performance evaluation of test cases [4] J Smith, EPRI, “Stochastic Analysis to Determine Feeder Hosting
Capacity for Distributed Solar PV”, Technical Report, December
Voltage 2012
N_Tap (all Flicker Flicker
Test ID Violation [5] Rich Seguin, Jeremy Woyak, David Costyk, Josh Hambrick, “High-
phases) Pst_Max Plt_Max
(min) Penetration PV Integration Handbook for Distribution Engineers,”
1 53 0 0.032 0.019 Technical Report, NREL/TP-5D00-63114, January 2016
[6] Pedro M. S. Carvalho; Pedro F. Correia; LuÍs A. F. M. Ferreira, “
2 4 0 0.031 0.018 [7] Distributed Reactive Power Generation Control for Voltage Rise
3 53 0 0.032 0.019 Mitigation in Distribution Networks,” IEEE Transactions on Power
4 36 0 0.032 0.019 Systems, Year: 2008, Volume: 23, Issue: 2, Pages: 766 – 772
5 227 0 0.035 0.021 [8] J. W. Smith; W. Sunderman; R. Dugan; Brian Seal, “Smart inverter
volt/var control functions for high penetration of PV on distribution
systems,” 2011 IEEE/PES Power Systems Conference and
Some observations are made as the following: Exposition, Year: 2011, Pages: 1 – 6
 The default setting of the VR control results in many VR [9] Georgios C. Kryonidis; Andreas I. Chrysochos; Charis S. Demoulias;
operations in the simulated time frame of 8 hours. Due to Grigoris K. Papagiannis, “Performance of solid-state transformers on
voltage regulation of active distribution networks,” 2017 IEEE
the intermittency of PV generation, the VR needs to Manchester PowerTech, Year: 2017, Pages: 1 – 6
constantly adjust its tap position [10] Arthur K. Barnes; Juan C. Balda; Andrés Escobar-Mejía; Scott. O.
 Increasing the control bandwidth significantly reduces Geurin, “Placement of energy storage coordinated with smart PV
inverters,” 2012 IEEE PES Innovative Smart Grid Technologies
the total number of VR operations while adding more (ISGT), Year: 2012, Pages: 1-7
time delays has only limited impact [11] Ehab Shoubaki; Somasundaram Essakiappan; Pankaj Bhowmik;
 Enabling the line drop compensation (LDC) logic in this Madhav Manjrekar; Johan Enslin; Stuart Laval; Aleksandar
VR results in many more tap changer operations as the Vukojevic; Jason Handley, “Distributed μ-STATCOM for voltage
support and harmonic mitigation on low voltage networks,” 2017
VR is trying to regulate the bus where the voltage is the IEEE Applied Power Electronics Conference and Exposition (APEC),
most volatile. This situation, however, may change when Year: 2017, Pages: 925 – 930
the PV system employs Volt/Var control and holds its [12] Panagis N. Vovos; Aristides E. Kiprakis; A. Robin Wallace; Gareth P.
terminal voltage more constant Harrison, “Centralized and Distributed Voltage Control: Impact on
Distributed Generation Penetration,” IEEE Transactions on Power
 In all tests cases, no voltage violation was observed in Systems, Year: 2007, Volume: 22, Issue: 1, Pages: 476 – 483
this studied feeder system even though certain node [13] PG&E, “Voltage Tolerance Boundary,” January 1999,
voltages may exceed the limits transiently (as can be https://www.pge.com/includes/docs/pdfs/mybusiness/customerservice
seen in Figure 9) /energystatus/powerquality/voltage_tolerance.pdf

View publication stats