1

Combined Effect of CVR and DG Penetration

in the Voltage Profile of Low Voltage Secondary
Distribution Networks
Abdullah Bokhari, Member IEEE, Ashhar Raza, Marc Diaz-Aguiló, Francisco de León, Fellow, IEEE,
Dariusz Czarkowski, Member, IEEE, Resk Ebrahem Uosef, Member, IEEE, and David Wang, Senior
Member, IEEE

 Penetration is a percentage/dynamic measure of the amount

Abstract—In this paper the voltage profile of secondary net- of power delivered/generated by interconnected DG compared
works under conservation voltage reduction (CVR) and distrib- with the total generation resources on a power system for a
uted generation (DG) penetration is studied for the first time. specific time of loading [1]. Penetration is not a static measure
Three networks in New York City, modeled in detail, are used as
study cases. Interconnection of DG is proposed to eliminate local-
as small percentage of DG penetration during peak load could
ized low voltage violations due to voltage reduction of 4%, 6%, be a high level of penetration under light load conditions.
and 8% from the normal schedule. The selection of the type of Different types of DG have the potential to substantially affect
DG is based on the requirements imposed by the various inter- system performance. For instance, conventional type synchro-
connection standards, most notably IEEE 1547, public service nous generators can have a greater effect on customer voltage
commission, and local utility regulations. It is found that a small than inverter-based DG or induction generators. However,
percentage of DG penetration would alleviate voltage violations.
The study shows that DG installed in distributed networks im-
regulation, cost, and reliability impose limitations on synchro-
prove voltage regulation, allowing utilities to use deeper voltage nous DG deployments in distribution systems as the short-
reductions during critical conditions. It is also shown that the circuit capacity of the installed breakers may be exceeded.
network power factor reduces when penetration of DG is high The compromise between DG interconnection requirements
and thus the line drop compensation needs to be adjusted for the for the avoidance of islanding and the security of the EPS have
new power demand. been studied in [6]–[8]. Numerous studies have investigated
Index Terms— Conservation voltage reduction (CVR), distrib-
the optimal placement of distributed generation in power sys-
uted power generation (DG), DG allocation, DG penetration, tems [9]–[11].
energy conservation, load model, secondary network, voltage Benefits of DG interconnection can be summarized as [1],
profile, ZIP coefficients. [9], [10]:
• Standby/backup power availability and reliability,
I. INTRODUCTION • Peak load shaving,

A S the penetration of Distributed Generation (DG) in Elec-

tric Power System (EPS) increases, so the reliability and
economic benefits. Utility regulators have been a driving force
• Combined heat and power,
• Sales of power back to utilities or other users,
• Renewable energy,
toward accelerating the implementation of DG [1]. The DG • Power quality, such as reactive power compensation and
interconnection requirements began with the IEEE Standard voltage support,
929 in 1988 [2]. Uniform mandatory interconnection require- • Dynamic stability support.
ments at the point of common coupling (PCC) were developed Voltage variation studies when a significant portion of the
in 2003 for all types of DG in the IEEE Standard 1547 [3]. total generation is DG have been performed in [11]–[15].
Due to the large variations in distribution system configura- Previous efforts introduced a comprehensive analysis of the
tions and situations where DG may be connected, a series of possible impacts of different penetration levels of DG on volt-
standards were developed as the guide on impact studies for age profiles in low-voltage secondary distribution networks
DG interconnection [4], [5]. The recommendations for DG [16]. The work was aimed to explore the maximum amount of
interconnection with secondary networks are given in the DG that secondary distribution networks can withstand in a
IEEE Std. 1547.7 [4]. probabilistic fashion. A field-validated load model for the
calculation of Conservation Voltage Reduction (CVR) in
A. Bokhari, A. Raza, M. Diaz-Aguiló, F. de León and D. Czarkowski are several secondary networks was presented in [17]. Both stud-
with the Department of Electrical and Computer Engineering, NYU Polytech- ies ([16]-[17]) concluded that the implementation of DG or
nic School of Engineering, Brooklyn, NY 11201 USA (e-mails: abdul-
[email protected]; [email protected]; CVR will provide energy and economic savings for the utility
[email protected]; [email protected]; [email protected]) and the customers. Many power utilities are moving towards
R. Uosef and D. Wang are with Consolidated Edison Inc., New York, NY implementing CVR [18]–[22]. Benefits of CVR in terms of
10003 USA (e-mail: [email protected]; [email protected]).
 2

energy savings and loss reduction have been studied in [23]– typical structure of an isolated spot network. For reliability
[25] while different implementation methods of CVR are purposes, the distribution system of New York City and the
described in [26]–[27]. downtown core of many cities in North America, use large
A recent study on peak demand reduction and energy con- interconnected low voltage (208/120 V) networks to supply
servation favored volt/var optimization via power factor cor- loads of hundreds of MW. This is different from most other
rection over CVR via active voltage regulation [28]. The study locations where the systems are mostly radial and supply loads
used load model-based approaches for the application of CVR of only a few hundreds of kW.
using two load categories: with and without thermal cycles. A
comparison of the polynomial static load model against physi-
TABLE I
cal load model gave credit to the later model when the dynam- NETWORK TOPOLOGY AND DEMAND
ic load behavior is considered. Another study highlighted the
Network Sutton Madison Sq. Yorkville
role of feeder characteristics for CVR application [29]. It was
High voltage 69 kV 138 kV 138 kV
concluded that short feeders on a densely populated networks No. of substation 7 5 (one spare) 4 (3-winding
would be most convenient to achieve the economical goal of transformers (69/13.8 kV) (138/13.8 kV) transformers)
CVR. A counter opinion was presented in [30]. Reference [31] (138/13.8 kV)
shows that CVR provides energy and economic savings for the No. of breakers 27 45 45
No. of network
utility and the customer. The results of [16] and [17] led to a transformers
224 462 542
challenge to study the behavior of low-voltage distribution No. of primary feeders 12 24 29
networks with a combined effect of CVR and DG penetration. Light load demand 47 MW 90.7 MW 118.4 MW
The interaction of DG implemented in a secondary grid can Peak load demand 141.7 MW 307 MW 250 MW
become more challenging when the EPS is under different
operating voltage conditions; for example, CVR or in periods TABLE II
of stress in the network due to contingencies. This becomes NETWORKS TOTAL LOAD AND LOAD COMPOSITION
more pronounced with a higher DG penetration as the network Network Sutton Madison Yorkville
power factor reduces. This causes further reduction in the line LT loads
Load points

284 1102 2272

Structure/

drop compensation (LDC) setting compromising the voltage (120/ 208V)

No. of

1122

2282
311
limits. Thus, research on the integration of customer genera-
tion in a distributed network with different types of intercon- HT loads (460V) 27 20 10
nected DG is needed to determine the impact on the steady Small residential 6.3 % 11.1 % 16.3 %
state behavior of the system.
composition

Large residential 3.9 % 9.7 % 0.0 %

The main contribution of this paper is to show how a small
Load

Small commercial 1.9 % 3.8 % 16.4 %

percent of DG penetration can alleviate voltage violations Large commercial 87.9 % 75.3 % 61.2 %
when CVR is applied. This allows further reducing the voltage Industrial 0.0 % 0.0 % 6.1 %
and therefore increasing the energy savings. The study is per-
formed on several secondary networks in New York City
taking into account the behavior of different types of DG dis-
tributed in realistic scenarios.
All simulations are performed with the open source simula-
tion package developed by EPRI: OpenDSS [32]. The net-
works and DG models were validated against New York City
utility records and the models developed in previous studies
[16]-[17].

II. NETWORK MODELING

A. Topology of the Networks under Study
The networks under study are: Madison Square, Sutton, and
Yorkville, all located in Manhattan. The selection of networks
was made to test different load compositions and varied num-
ber of customers. Some details of the networks are described Fig. 1. Illustration of a LV secondary network including: high voltage, substa-
tion, loads, transformers, DG, and a typical structure of an isolated spot net-
in Table I. work. In NYC the low voltage networks operate at 208/120 V and isolated
Power is fed into the low voltage grid network serving low spot networks are fed at 460 V.
tension (LT) customers at 120/208 V and a small percentage
of high tension (HT) local building buses (spot networks at The three networks selected are of varied sizes and de-
460 V). Detailed description of the load composition of the mands: a small network (Sutton), a medium network (Madison
three networks is given in Table II. Fig. 1 shows a simplified Square), and a large network (Yorkville) with different load
topology of the network with loads, transformers, and the compositions.
 3

TABLE III
SAMPLE OF THE IMPLEMENTED VOLTAGE SCHEDULE
No CVR 2.25% reduction 4 % reduction 6% reduction 8 % reduction
Total

Reduction

Reduction

Reduction

Reduction

Reduction
13.8 kV
Feeder Bus Feeder Bus Feeder Bus Feeder Bus Feeder Bus
Bus Load

%
Voltage (kV) Voltage (kV) Voltage (kV) Voltage (kV) Voltage (kV)
(MW)

0-50 13.3 0% 13.0 2.26 % 12.76 4.06 % 12.50 6.02 % 12.23 8.05 %
51-90 13.4 0% 13.1 2.24 % 12.86 4.03 % 12.59 6.01 % 12.32 8.02 %
91-130 13.5 0% 13.2 2.22 % 12.96 4.00 % 12.69 6.00 % 12.42 8.00 %
131-170 13.6 0% 13.3 2.21 % 13.06 3.97 % 12.78 5.99 % 12.51 7.98 %
171-210 13.7 0% 13.4 2.19 % 13.16 3.94 % 12.88 5.99 % 12.61 7.96 %

B. Network Model   V 2 V  
P  P0  Z p  i   I p  i   Pp  (1)
In a previous study carried out by the authors, a polynomial   V0   V0  
static load model with ZIP coefficients was used to represent
the power consumed by a load as a function of voltage [17], Subject to Z p  I p  Pp  1 (2)
[33]. ZIP parameters are the coefficients of a load model com-  V  2

prised of constant impedance Z, constant current I, and con- V 
Q  Q0  Z q  i   I q  i   Pq  (3)
stant power P loads. ZIP based load models were developed   V0   V0  
for residential, commercial and industrial loads [33]. The
models were validated in the field for the networks under Subject to Z q  I q  Pq  1 (4)
study. Experimentally validated network models are used to
where P and Q are the active and reactive powers at operating
analyze the behavior of the distribution networks under the
voltage (Vi); P0 and Q0 are the active and reactive powers at
combination of CVR and DG penetration. The DG models
rated voltage (V0); Zp, Ip and Pp are the ZIP coefficients for
used in the study are selected from the OpenDSS library and,
active power; Zq, Iq and Pq are the ZIP coefficients for reactive
also, they have been validated against EMTP results in [16].
power.
Using actual data, the network model was built in OpenDSS.
The networks under study are highly integrated with diverse
The data includes primary feeders, transformers, network
residential, commercial, and industrial loads. Each load is
protectors, and secondary mains with each customer repre-
classified into one of the four following categories: small or
sented as a ZIP coefficients load. The behavior of the DG
large residential, commercial, or industrial. Each load is then
(synchronous generators and inverter-based DG) is considered
represented with the appropriate ZIP coefficients model.
using the existing models from the OpenDSS library. Capaci-
tors are modeled based on the network load demand. As an
III. DG STUDY UNDER CONSERVATION VOLTAGE REDUCTION
example, Sutton network has two switching capacitors, one of
them is connected at medium load (50% to 75% of demand), A. Network Model Validation
two are connected at peak load, and no capacitors are connect- The analysis presented here is based on the detailed three-
ed at light load. phase model developed in [17] using network characteristic
The network voltage is controlled exclusively from the area and real data records (for 2010). The results of steady state
substation on-load tap changer transformers. CVR is imple- (power flow) simulations under DG penetration were verified
mented by reducing voltage at the substation by controlling against EMTP time-domain simulations reported in [16]. Re-
line drop compensation (LDC) mechanism. A lower LDC production of several events and DG penetrations of the same
setting at the substation allows voltage reduction to be imple- network were compared and validated. Fig. 2 shows the volt-
mented. Table III shows a sample voltage schedule with the age profile comparison between OpenDSS and EMTP for the
voltage reduction level for various network demands. base case (with no DG) and the worst case scenario reported in
C. Load Models [16].
To obtain reliable results, a voltage-sensitive load model was B. Voltage Violation Study
used for all networks. Both watts and vars vary with voltage
The application of CVR in highly meshed secondary networks
based on typical residential, commercial, and industrial cus-
is known to have a satisfactory impact on energy savings and
tomers in New York City. The loads connected on the second- losses [17]. However, voltage reduction can produce under-
ary network are represented as static load model with their voltage violations at some loads. Utilities are mandated to
polynomial ZIP coefficients. The models have been obtained keep voltage values within acceptable ranges across all the
from numerous voltage reduction tests performed in the labora- nodes in the network, both on the primary and secondary
tory on many domestic appliances performed on typical resi- sides. For the purpose of this study, voltage reduction simula-
dential, commercial, and industrial customers in New York tions of each network were performed to identify all
City. These experiments are described and documented in [33]. loads/structure points with violations on the peak hour of the
The polynomial expressions for active and reactive powers year. Voltage reduction operations are performed for voltage
of the ZIP coefficients model are: levels of 2.25%, 4%, 6% and 8% and voltage violation of 5%
 4

(under 114 V) and 10% (under 108 V) are monitored for all Fig. 4 shows voltage violations exceeding 10% (under 108 V)
loads. when 8% voltage reduction is used. These figures are shown
for the peak-load hour of the year. The plots show that the
voltage violations are localized in small geographical areas.
Then an investigation was launched to find if the problems can
be solved with a small percentage of DG penetration. This
stems from the fact that the interconnection of DG is known to
produce localized overvoltages. A win-win situation is ex-
pected since both techniques (CVR and DG) save energy, but
their potential bad side effects may cancel each other.

Fig. 2 Comparison of results for customer voltage profile at 120 V from

OpenDSS and EMTP for the worst case scenario (left) and base case with no
DG (right) reported in [16].

Fig. 4. Geographical voltage distribution in Yorkville network for 8% voltage

reduction of the peak hour of the year. Eight voltage violations are detected
exceeding 10% (under 108 V) out of 2282 structure points. Underlying map
©2014 by Google.

C. Overview of DG Interconnection under CVR

Fig. 3. Geographical voltage distribution in the Yorkville network for 4% The operation of DG has an influence on the distribution
voltage reduction during the peak hour of the year. Twenty Six voltage viola- system voltage levels by changing the current levels on the
tions are detected exceeding 5% (under 114 V) out of 2282 structure points. system [8]. This influence is defined by the size, type, and
Underlying map ©2014 by Google.
location of the DG, the network topology, DG operation strat-
The utility of New York City regulates the minimum volt- egy, and the characteristics of the distribution system. The
age on distribution feeders so that the delivery voltage at the operation of the generator should not cause the distribution
customer’s meter will stay within ±5% of nominal (i.e. 120 V system voltage (utilization voltage) to go outside of the steady
± 5% or 126 V to 114 V) during normal operating conditions state voltage limits specified by ANSI Std. C84.1. The Public
and 10% below nominal voltage (108 V) for emergency condi- Utilities Commission establishes service voltage (customer
tions [34]. The national standard related to these voltage levels voltage) limits for the utility. However, during severe voltage
is ANSI C84.1 where 114 V (95%) is defined as the minimum reduction (or contingency), service voltage supplied by the
service voltage and 108 V (90%) is defined as the minimum utility could go below specified limits for customers connect-
utilization voltage [35]. In this study, we have computed volt- ed at the end of feeder due to voltage drop.
age violations for both of these levels for loads with a voltage The interconnection of DG must meet the basic requirements
base of VLN = 120 V. imposed by the various standards, most notably IEEE 1547
The investigation aimed at identifying voltage violations of [3]-[5], public service commission [36], and local utility regu-
5% and 10% under different voltage reduction levels for the lation [37], while providing a foundation on which higher
three networks. Fig. 3 shows the voltage violations exceeding levels of penetration can be built. As dictated by Consolidated
5% (under 114 V) when a 4% voltage reduction is applied. Edison Inc. of New York, the default voltage operating range
 5

for DG shall be from 88% to 110% of nominal voltage magni- DG) described in Subsection III. A, and compared with the
tude and be operated in a manner that does not cause the volt- results with DG penetration presented in Subsection III. C. In
age regulation to go outside the applicable limits. addition, these simulations are done for all voltage reduction
DG allocation with constraints of maximum 2 MW output levels. With proper DG allocation, the utility can implement
power or less on each DG is considered in this study. Note that reduction in voltage that was not acceptable (due to voltage
no power can be exported from the secondary network to violations) for the case without DG implemented. For exam-
primary because network protectors will trip. ple, some medical equipment such X-ray and MRI machines
have a small range of operating voltage which makes them
D. DG Allocation Approach
sensitive to voltage variations. DG could be an inexpensive
Following are the key operations performed to obtain the solution to health care facilities and hospitals since no medical
minimum DG penetration required to solve localized voltage equipment will dropout due to CVR implementation during
violations: emergency situations.
1) Look up for the geographical and electrical location of TABLE IV
structures under low voltage violation. SUMMARY OF IDENTIFIED VOLTAGE VIOLATION AND ALLOCATED DGS
2) Low voltage structures that are electrically close to each Network Sutton Madison Sq. Yorkville
other are treated together. Voltage violation
5% (114V) 5% (114V)
5% (114V),
3) One DG is installed for a group of structures to reduce level 10% (108V)
No. of violation
the overall number of DG. 62 (6%) 3 (4%) 26 (4%), 8 (8%)
(CVR %)
Only two types of DG systems are used; inverter type and No. of allocated DGs 29 1 9
synchronous machine type. The inverter type DG is operated Total allocated DGs
at a unity power factor, and the synchronous machine type 3.4 MW 250 kW 1.25 MW
output power
DG is operated at power factor 0.9 leading. Structure points % of DGs to peak
2.3% 0.08% 0.5%
that have lower demands of less than 100 kW are allocated load demand
inverter type DG, with lowest DG size not less than 50 kW.
Structure points with heavier loads are allocated synchronous
machines, with a limit of 2 MW. Low voltage structures that
are electrically connected are not allocated separate DGs,
rather a single DG is installed for all the structure points that
are electrical neighbors. This helps reducing the overall num-
ber of DGs, and also reducing the cost of installation and
maintenance. However, if a particular group of electrically
close structure points have a combined load value more than 2
MW, more than one DG of similar type are connected in order
to improve the voltage profile.
For Yorkville network, nine DGs were allocated in the low
voltage distribution network with a total power of 1.25 MW
representing 0.5% of the total peak demand. Voltage reduction
of 4% and 8% were simulated with DG penetration to solve
voltage violations exceeding 5% (under 114 V) and over 10%
(under 108 V). A similar DG allocation approach was applied Fig. 5. Geographical voltage distribution in Madison Square network for 4%
on the Madison Square network to solve the over 5% and 10% voltage reduction of the peak hour of the year with three voltage violations
voltage violations resulted from the 4% and 8% voltage reduc- detected exceeding 5% (under 114 V). Only one DG allocated on the structure
point under voltage violation. Underlying map ©2014 by Google.
tion. This network is robust to voltage violation with only 3
voltage violation clustered in one location. Fig. 5 shows the
voltage map with voltage violation over 5% (under 114 V) in
the Madison Square network when 4% voltage reduction op-
eration is conducted. Only one DG of 250 kW (0.08% of peak
demand) was needed to remove the 5% and 10% voltage vio-
lations. Finally, the smallest network (Sutton) has a weak
characteristic with 62 voltage violations of over 5% (under
114 V) when 6% voltage reduction was applied. 29 DGs with
a total power of 3.4 MW (2.3% of the total peak demand)
were used to solve voltage problems. Results for Sutton net-
work are shown in Fig. 6. The results for the three networks
under study and allocated DGs are summarized in Table IV.
E. Simulation Results of the Proposed DG Allocation Fig. 6. Geographical voltage distribution in Sutton network for 6% voltage
In this section, load-flow simulation results showing the reduction of the peak hour of the year with 62 voltage violations detected of
voltage profile of all loads for each network are presented. The over 5% (under 114 V). Twenty nine localized DG allocated on structure
points under voltage violation. Underlying map ©2014 by Google.
results are obtained for the voltage violation study (with no
 6

Fig. 7(a) shows the voltage profile of all loads in Yorkville tions. It was also shown that 100% of the load could be fed
network. With 0.5% (1.25 MW) DG penetration of the total from DGs when allocated in a way that the load is negated. In
network peak demand (250 MW), 26 violations of 5% (under this section, it is shown how voltage reduction can be applied
114 V) and 8 violations of 10% (under 108 V) for 4% and 8% under high DG penetration.
voltage reduction levels, respectively, are now removed. Simi- With no DG, the total peak load demand of the Sutton net-
lar analysis is shown for Madison Square network during peak work is 141.7 MW and the reactive power demand is 72.74
demand (307 MW) with one DG to solve violations of 5% and Mvar, giving a power factor of 0.89 lagging. The substation
10% occurred in 4% and 8% voltage reductions. Finally, the transformers setting is at 13.6 kV (see Table III). Let us as-
proposed DG allocation is also applied to Sutton network sume a total power supplied by DGs at 24.73 MW and 3.17
(141.7 MW peak demand) to solve 62 violations of under 114 Mvar at a power factor of 0.99 leading (which corresponds to
V for the 6% voltage reduction using only 2.3% DG penetra- 50% of light load). The new power demand seen by the sub-
tion. station is 116.97 MW and 69.57 Mvar at a power factor of
0.86. The power factor of the network has lowered from 0.89
to 0.86 due to the high penetration of DGs. The substation
transformer setting for this demand is 13.5 kV (see Table III).
However, the original tap settings were designed assuming a
power factor of 0.89. At 0.86 power factor, more reactive
power is supplied (in proportion) than originally foreseen,
which causes a larger voltage drop in the feeders and offsets
the effect of DG.
From Fig. 8(a) it can be seen that the voltage profile at this
DG penetration level is becoming flatter, i.e. structure points
that had lower voltages previously have a higher voltage now,
while the structure points which had higher voltages previous-
(a) ly now have a lower voltage. The decrease in voltage of struc-
ture points that were previously higher is caused by the lower-
ing of transformer taps. This scenario is more favorable for a
utility since the difference between highest voltage and lowest
voltage is reduced, which allows the utility to control the volt-
age of the loads more effectively. The phenomenon of flatten-
ing of voltage profile is favorable at normal operation with no
voltage reduction. However, when 8% voltage reduction on
peak load demand hour is applied, more structure points vio-
late the low voltage limit as can be seen in Fig. 8(a). There-
fore, adding more DG will not improve voltage profile if no
modification is made to the LDC settings.
(b) Fig. 8(b) shows improvements in the voltage profile for
the same DG allocation by modifying the tap setting such that
it considers the new power factor of the load in addition to the
active power demand. The aforementioned cases show that
distribution networks have not been designed for connecting
large percentages of DG. This issue reveals that modification
of the substation transformers setting is needed to achieve the
desired results for large DG penetrations.
Fig. 9 shows the power triangle of the network and the in-
crease of the power angle due to high DG penetration. The
original tap setting was designed assuming a power factor of
0.89. With DG penetration, the power factor of the system
(c) becomes smaller; hence a higher voltage at the substation is
Fig. 7 Comparison of voltage profile for loads at 120 V without DG penetra- needed to compensate for the increased proportion of reactive
tion (dash dotted lines) and with DG penetration. Results are shown for the power. This effect is further pronounced when the DG pene-
base case with no voltage reduction, 4% and 8% CVR; (a) for Yorkville; (b) tration is increased.
Madison Square; (c) Sutton with no voltage reduction and 6% CVR.

IV. EFFECT OF HIGH DG PENETRATION ON POWER FACTOR

In [16], it was shown that high penetration of randomly al-
located DG results in over voltage and under voltage viola-
 7

The study has also revealed new issues related to line drop
compensation (LDC) settings when DG penetration increases.
When the network’s power factor is reduced under high DG
penetration, mitigation of the effects of the previous tap
scheduling is needed to control the voltage of the loads effi-
ciently.

VI. REFERENCES
[1] N. Miller and Z. Ye, “Report on Distributed Generation Penetration
Study”, National Renewable Energy Laboratory, Golden, CO,
NREL/SR-560 – 34715, Aug. 2003.
[2] IEEE Recommended Practice for Utility Interface of Residential and
Intermediate Photovoltaic (PV) Systems, ANSI/IEEE Std. 929-1988,
1987.
[3] IEEE Standard for Interconnecting Distributed Resources with Electric
(a) Power Systems, IEEE Std 1547-2003, July 2003.
[4] IEEE Recommended Practice for Interconnecting Distributed Resources
with Electric Power Systems Distribution Secondary Networks, IEEE
Std. 1547.6 - 2011, Sept., 2011.
[5] IEEE Guide for Conducting Distribution Impact Studies for Distributed
Resource Interconnection, IEEE Std. 1547.7 – 2013, Feb. 28, 2014.
[6] R. A. Walling, and N. W. Miller, “Distributed Generation Islanding –
Implications on Power System Dynamic Performance.” Proceedings of
the IEEE/PES Summer Power Meeting, Chicago, July 2002.
[7] Z. Ye, R. Walling, L. Garces, R. Zhou, L. Li, and T. Wang “Study and
Development of Anti-Islanding Control for Grid-Connected Inverters”,
National Renewable Energy Laboratory, Golden, CO, NREL/SR-560-
36243, May 2004.
[8] L. Yu, D. Czarkowski, and F de León, "Optimal Distributed Voltage
Regulation for Secondary Networks with DGs," IEEE Trans. Smart
Grid, vol. 3, no. 2, pp.959 – 967, June 2012.
[9] H. A. Gil and G. Joos, "Models for Quantifying the Economic Benefits
of Distributed Generation," IEEE Trans. Power Sys., vol. 23, no. 2, pp.
327 – 335, May 2008.
(b) [10] C. Pathomthat and R. Ramakumar, "An approach to quantify the tech-
Fig. 8. Sutton network voltage profile for customers at 120 V during peak load nical benefits of distributed generation," IEEE Trans. Energy Conv., vol.
hour; (a) regular tap setting; (b) modified tap setting. Results for base case and 19, no. 4, pp. 764 – 773, Dec. 2004.
CVR with no DG penetration are in solid lines. [11] M. A. Mahmud, M. J. Hossain, and H. R. Pota, "Voltage Variation on
Distribution Networks With Distributed Generation: Worst Case Scenar-
io," IEEE Sys. Journal, doi: 10.1109/JSYST.2013.2265176, June 2013.
[12] M. L Doumbia and K. Agbossou, "Voltage Variation Analysis in Inter-
connected Electrical Network - Distributed Generation," 2007 IEEE
Canada Elect. Power Conf., EPC 2007, pp. 525 – 530, 25 – 26, Oct.
2007.
[13] T. Lee, S. Hu, and Y. Chan, "D-STATCOM With Positive-Sequence
Admittance and Negative-Sequence Conductance to Mitigate Voltage
Fluctuations in High-Level Penetration of Distributed-Generation Sys-
tems," IEEE Trans. Ind. Electronics, vol. 60, no. 4, pp.1417 – 1428,
Fig. 9. Illustration of the network power triangle with no DG penetration April 2013.
(solid lines), total power supplied by all DGs (dash-dotted line), and the new [14] T. Senjyu, Y. Miyazato, A. Yona, N. Urasaki, and T. Funabashi, "Opti-
network power triangle DG penetration (dashed lines). mal Distribution Voltage Control and Coordination ith Distributed Gen-
eration", IEEE Trans. Power Del., vol. 23, no. 2, pp. 1236 – 1242, April
2008.
V. CONCLUSIONS [15] J. M. Sexauer and S. Mohagheghi, "Voltage Quality Assessment in a
Distribution System With Distributed Generation—A Probabilistic Load
A new technique to solve voltage violations in a highly Flow Approach," IEEE Trans. Power Del., vol. 28, no. 3, pp. 1652 –
meshed network when CVR is implemented using a small 1662, July 2013.
percentage of DG penetration is investigated. It is shown that a [16] P. Chen, R. Salcedo, Q. Zhu, F. de León, D. Czarkowski, Z. Jiang, V.
Spitsa, Z. Zabar, and R. E. Uosef, "Analysis of Voltage Profile Problems
win-win situation exists when combining DG and CVR. On
Due to the Penetration of Distributed Generation in Low-Voltage Sec-
one hand, in an unregulated secondary network, the limit of ondary Distribution Networks," IEEE Trans. Power Del., vol. 27, no. 4,
the voltage reduction is given by the low-voltage violations. It pp. 2020 – 2028, Oct. 2012.
has been shown that in secondary networks, loads experienc- [17] M. Diaz-Aguilo, J. Sandraz, R. Macwan, F. de León, D. Czarkowski, C.
Comack, and C. Wang, "Field-Validated Load Model for the Analysis of
ing voltage violations are strongly correlated and usually oc- CVR in Distribution Secondary Networks: Energy Conservation," IEEE
cur in localized pockets in the network. On the other hand, it is Trans. Power Del., vol. 28, no. 4, pp. 2428 – 2436, Oct. 2013.
known that the interconnection of DG produces localized [18] B. Scalley and D. Kasten, “The effects of distribution voltage reduction
on power and energy consumption,” IEEE Trans. Educ., vol. 24, no. 3,
over-voltages. Small amounts of DG can the alleviate voltage pp. 210 – 216, Aug. 1981.
violations, therefore allowing deeper voltage reductions and as [19] D. Kirshner, “Implementation of conservation voltage reduction at
a consequence larger energy and economic savings. commonwealth Edison,” IEEE Trans. Power Sys., vol. 5, no. 4, pp. 1178
– 1182, May 1990.
 8

[20] D. Lauria, “Conservation Voltage Reduction (CVR) at northeast utili- system modeling and analysis, power theory, renewable integration, smart
ties,” IEEE Trans. Power Del., vol. PWRD-2, no. 4, pp. 1186 – 1191, grid, and electrical machines.
Aug. 1987.
[21] K. Matar, “Impact of voltage reduction on energy and demand,” M.S. Ashhar Raza received the B.Tech degree in electrical engineering from
thesis, College Eng. Tech., Ohio Univ., Athens, OH, 1990. Aligarh Muslim University, Aligarh, India in 2011 and M.S degree in electri-
[22] V. Dabic, S. Cheong, J. Peralta, and D. Acebedo, “BC Hydro’s experi- cal engineering from NYU Polytechnic School of Engineering, Brooklyn, NY,
ence on voltage VAR optimization in distribution system,” presented at USA in 2014 where he is currently pursuing a Ph.D. degree.
the IEEE Power Energy Soc. Trans. Dist. Conf. Expo, New Orleans, LA, His research interests are in distributed generation system, steady state
2010. analysis, fault analysis and modeling of distribution system.
[23] K. P. Schneider, F. K. Tuffner, J. C. Fuller, and R. Singh, “Evaluation of
conservation voltage reduction (CVR) on a national level,” Pacific Marc Diaz-Aguiló was born in Barcelona, Spain. He received the M.Sc.
Northwest National Laboratory, Richland, WA, Rep. no. PNNL-19596, degree in telecommunications engineering from the Technical University of
Jul. 2010. Catalonia (UPC), Barcelona, Spain, in 2006 and the M.Sc. degree in aerospace
[24] V. J. Warnock and T. L. Kirkpatrick, “Impact of voltage reduction controls engineering from a joint program between Supaero, Toulouse France,
on energy and demand: Phase II,” IEEE Trans. Power Sys., vol. PWRS- and the Massachusetts Institute of Technology, Cambridge, MA USA, in
1, no. 2, pp. 92 – 95, May 1986. 2008, and the Ph.D. degree in aerospace simulation and controls from the
[25] S. Lefebvre, G. Gaba, A.-O. Ba, D. Asber, A. Ricard, C. Perreault, and Technical University of Catalonia, Barcelona, Spain, in 2011.
D. Chartrand, “Measuring the efficiency of voltage reduction at Hydro Currently, he is a Postdoctoral Researcher at the Polytechnic Institute of
Qubec distribution,” 2008 IEEE Power and Energy Society General New York University, Brooklyn, NY, USA. His research interests are in
Meeting – Conv. and Del. of Elect. Energy in the 21st Century, Pitts- power systems, controls, smart-grid implementations, and large systems
burgh, PA, July 2008. modeling and simulation.
[26] J. G. De Steese, S. B. Merrick, and B. W. Kennedy, “Estimating meth-
odology for a large regional application of conservation voltage reduc- Francisco de León (S’86–M’92–SM’02–F’15) received the B.Sc. and the
tion,” IEEE Trans. Power Sys., vol. 5, no. 3, pp. 862 – 870, Aug. 1990. M.Sc. (Hons.) degrees in electrical engineering from the National Polytechnic
[27] T. L. Wilson, “Measurement and verifications of distribution voltage Institute, Mexico City, Mexico, in 1983 and 1986, respectively, and the Ph.D.
optimization results for the IEEE Power and Energy Society,” presented degree in electrical engineering from the University of Toronto, Toronto, ON,
at the IEEE Power Energy Soc. Gen. Meeting, Minneapolis, MN, Jul. Canada, in 1992.
2010. He has held several academic positions in Mexico and has worked for the
[28] D. Pinney, “Costs and Benefits of Conservation Voltage Reduction CVR Canadian electric industry. Currently, he is an Associate Professor with the
Warrants Careful Examination,” National Rural Electric Cooperative Department of Electrical and Computer Engineering at New York University.
Association, Arlington, VA, DE – OE0000222, Nov. 2013. His research interests include the analysis of power phenomena under non-
[29] B. W. Kennedy, and R. H. Fletcher, "Conservation voltage reduction sinusoidal conditions, the transient and steady-state analyses of power sys-
(CVR) at Snohomish County PUD," IEEE Trans. Power Sys., vol. 6, no. tems, the thermal rating of cables and transformers, and the calculation of
3, pp. 986 – 998, Aug 1991. electromagnetic fields applied to machine design and modeling.
[30] M. A. Peskin, P. W. Powell, and E. J. Hall, "Conservation Voltage Prof. de León is an Editor of the IEEE Transactions on Power Delivery and
Reduction with feedback from Advanced Metering Infrastructure," 2012 the IEEE Power Engineering Letters.
IEEE PES Trans. and Dist. Conf. and Expo., pp. 1 – 8 , 7 – 10, May
2012. Dariusz Czarkowski (M’97) received the M.Sc. degree in electronics from
[31] J. Sandraz, R. Macwan, M. Diaz-Aguiló, J. McClelland, F. de León, D. the AGH University of Science and Technology, Cracow, Poland, in 1989, the
Czarkowski, and C. Comack, “Energy and Economic Impacts of the M.Sc. degree in electrical engineering from Wright State University, Dayton,
Application of CVR in Heavily-Meshed Secondary Distribution Net- OH, USA, in 1993, and the Ph.D. degree in electrical engineering from the
works”, IEEE Trans. Power Del., vol. 29, no. 4, pp. 1692 – 1700 August University of Florida, Gainesville, FL, USA, in 1996.
2014. In 1996, he joined the Polytechnic Institute of New York University,
[32] R. C. Dugan, and T. E. McDermott, “An open source platform for Brooklyn, NY, where he is currently an Associate Professor of Electrical and
collaborating on smart grid research,” 2011 IEEE Power and Energy So- Computer Engineering. He is a coauthor of Resonant Power Converters
ciety General Meeting, pp.1–7, July 2011. (Hoboken, NJ: Wiley, 2011). His research interests are in the areas of power
[33] A. Bokhari, A. Alkan, R. Dogan, M. Diaz-Aguilo, F. de León, D. electronics, electric drives, and power quality.
Czarkowski, Z. Zabar, L. Birenbaum, A. Noel, and R. E. Uosef, "Exper-
imental Determination of the ZIP Coefficients for Modern Residential, Resk Ebrahem Uosef (M’01) received the B.Sc. and M.Sc. degrees in elec-
Commercial, and Industrial Loads," IEEE Trans. Power Del., vol. 29, trical engineering from Alexandria University Faculty of Engineering, Alex-
no. 3, pp.1372 – 1381, June 2014. andria, Egypt, in 1979 and 1981, respectively, and a second M.Sc. degree in
[34] Low Tension A.C. Service Voltage Limits, Consolidated Edison Compa- electrical engineering, and the Ph.D. degree in electrical engineering from
ny of New York, Inc., Specification EO-2065, New York, NY, Rev. 6, Polytechnic University, Brooklyn, NY, USA, in 2007 and 2011, respectively.
Jan. 2011. He was an Engineer in a Hydropower Generating Station in Egypt, and
[35] American National Standard for Electric Power Systems and Equipment. then he was the Owner of a consulting firm for an electric construction com-
Voltage Ratings (60 Hertz), ANSI Standard C-84.1-2011, 2011. pany in Egypt. He joined Con Edison’s Distribution Engineering Department,
[36] New York State Standardized Interconnection Requirements and Appli- New York, USA, in 2003 and is currently responsible for Con Edison’s distri-
cation Process for New Distributed Generators 2 MW or Less Connect- bution system design and analysis.
ed in Parallel with Utility Distribution Systems,” New York State Public Dr. Uosef is a Registered Professional Engineer in the State of New York.
Service Commission, New York, April 2013.
[37] Handbook of General Requirements for Electrical Service to Dispersed David Wang (S’90–M’90–SM’07) received the B.S. degree in electrical
Generation Customers, Consolidated Edison Company of New York, engineering from Shanghai University of Engineering Science, Shanghai,
Inc., Specification EO-2115, New York, NY, Rev. 8, March 2006. China, in 1988, the M.S. degree in electrical engineering from New Jersey
Institute of Technology, Newark, NJ, USA, in 1990, the M.S. degree in com-
Abdullah Bokhari (S’12–M’15) received the B.Sc. degree in electrical puter science from New York University, New York, USA, in 1998, and the
engineering from King Saud University, Riyadh, Saudi Arabia, in 2004, the Ph.D. degree in electrical engineering from Polytechnic University, Brooklyn,
M.Sc. degree and Ph.D. degree in electrical engineering from New York NY, USA, in 2006.
University, Polytechnic School of Engineering, Brooklyn, NY, USA, in 2009 He joined Con Edison’s R&D Department in 1991 and currently he is a
and 2014 respectively. Technical Expert in the Distribution Engineering Department responsible for
He was a Research Assistant since 2011 and a Postdoctoral Researcher in the development of Con Edison’s distribution system design and analysis
2014 with New York University, Polytechnic School of Engineering. He has software.
held several industry positions and worked for ABB Automation, Riyadh,
Saudi Arabia, and Public Service Electric and Gas (PSE&G), New Jersey,
USA. His research interests include distributed generation system, power