REVIEW OF THE CONCEPTS TO INCREASE DISTRIBUTED GENERATION INTO THE
DISTRIBUTION NETWORK
A Project
Presented to the faculty of the Department of Electrical and Electronic Engineering
California State University, Sacramento
Submitted in partial satisfaction of
the requirements for the degree of
MASTER OF SCIENCE
in
Electrical and Electronic Engineering
by
Daniel E Ilse
SPRING
2012
REVIEW OF THE CONCEPTS TO INCREASE DISTRIBUTED GENERATION INTO THE
DISTRIBUTION NETWORK
A Project
by
Daniel E Ilse
Approved by:
__________________________________, Committee Chair
Dr. Mohammad Vaziri
____________________________
Date
ii
Student: Daniel E Ilse
I certify that this student has met the requirements for format contained in the University format
manual, and that this project is suitable for shelving in the Library and credit is to be awarded for
the project.
__________________________, Department Chair
Dr. Suresh Vadhva
Department of Electrical and Electronic Engineering
iii
___________________
Date
Abstract
of
REVIEW OF THE CONCEPTS TO INCREASE DISTRIBUTED GENERATION INTO THE
DISTRIBUTION NETWORK
by
Daniel E Ilse
Distributed Generation (DG) level of penetration is expected to be increasing due to
the state and federal government’s mandates for utilization of renewable resources. State
of California’s renewable standards [1], and the U.S. solar installations rapidly growing
to 2.15GW in grid tied installed capacity in 2010 [2] are examples of such assertive
mandated programs. The current Distribution Network (DN) was not originally designed
for integration of DG at such high penetration levels. Improvements will need to be
made to DN to facilitate safe and reliable interconnection of DG at higher penetration
levels. This paper gathers and documents the major concerns such as; voltage,
protection, and power quality related to DG interconnections as well as the issues with
the existing design standards and criteria. Several different DN changes have been
proposed to help resolve these issues. The focus of this paper is about the exiting DN
iv
topology and DG integration issues, as well as documentation and discussion of the
proposed solutions. Additional areas of research are identified for further consideration.
_______________________, Committee Chair
Dr. Mohammad Vaziri
_______________________
Date
v
ACKNOWLEDGEMENTS
I would like to thank Professor M. Vaziri and the Power Engineering Department at
California State University, Sacramento. I applaud the dedication and patience of the
entire staff.
vi
TABLE OF CONTENTS
Page
Acknowledgements ..................................................................................................... vi
List of Figures ........................................................................................................... viii
Chapter
1. INTRODUCTION ................................................................................................. 1
2. TOPOLOGY OF NORTH AMERICAN SYSTEMS ............................................. 3
3. HOW DG IS CAPABLE OF DISRUPTING THE DN .......................................... 8
Capacitor Related Disruptions ........................................................................ 10
Tap Changer Settings Disruptions .................................................................. 10
Protection Disruptions .................................................................................... 13
Islanding Concerns.......................................................................................... 15
4. HOW THE IMPACT OF DG CAN BE MINIMIZED TO ALLOW FOR
MAXIMUM DG PENTRATION.. ....................................................................... 17
Upgrading Conductors .................................................................................... 17
Optimal Placement Planning........................................................................... 18
Voltage Control ............................................................................................... 18
Redesigning the Network to a Mesh System .................................................. 20
Islanding .......................................................................................................... 21
Fault Current Limiting Devices ...................................................................... 22
Adaptive Protection Relaying ......................................................................... 23
5. CONCLUDING REMARKS ................................................................................ 25
Biographies….. ........................................................................................................... 27
References. .................................................................................................................. 29
vii
LIST OF FIGURES
Figures
Page
1.
Vector representation of capacitor correction ......................................................... 5
2.
Example of a DN with switches, fuses and protection overlap shown ............... 7
3.
Example of LDC regulation issues ........................................................................ 12
4.
Inaccuracy of LDC Vs DG injection ..................................................................... 12
5.
Example of DN before DG is added ....................... …………………………. 14
6.
Example of DN with no upgrades and DG ........................................................... 14
7.
Example of DN with possibility of relay desensitization ................................... 15
viii
1
INTRODUCTION
In North America, the original design of the DN is based on uni-directional power
flow, known as “radial” feeders. This radial design is to keep the voltage in acceptable
parameters based on the minimum and the maximum load demands. Past investments in
the feeder to improve the characteristics is based on fixing a problematic area, or a
response to an increase in load [2]. Renewable targets created by state and federal
regulatory commissions have set targets for the percentage of renewable energy being
served to the consumer. California is leading the United States by requiring that 33% of
all retail electricity sales be from renewable resources by the year 2020 [1]. Often these
renewable resources are located in remote areas, and are limited in size. These small
remote generation sites are often connected with the DN without any detailed studies or
analysis. However, when DG is connected to the DN, it alters the performance of the
entire DN [39-40]. The DN is no longer a radial system with a unidirectional power flow
pattern. The problems associated with bi-directionality of the power flow are amplified
with increased levels of DG. Some of the known impacts of DG are: voltage quality,
protection equipment settings, desensitized relays, increased fault currents, increased
maintenance of “hunting” equipment, and even islanding portions of the DN[11][15].
Much research has been done in these various areas identifying the concerns and issues as
well as proposals for change of current standards and feeder design criteria [39]. The
objective of this paper is to review the current DN topology, describe the problems
2
associated with DG penetration, and to compare, analyze and document some of the
proposed solutions to increases in the amount of DG penetration levels on DN.
3
TOPOLOGY OF NORTH AMERICAN SYSTEMS
Historically, power is generated by large generation units and immediately
stepped up to high voltage power lines. These power lines traverse significant distances
from some remote areas to populated load centers where the voltage is stepped down to
MV or even Distribution Substations (DS) for connections to distribution feeders. The
main primary feeder branches into several primary laterals that then branch into several
sub-laterals to serve the distribution transformers. The most common type of primary
feeder is the radial-type feeder with uni-directional power flow. Generally, the main
feeder is three-phase, with the laterals either three-phase or single-phase. This topology
has low service reliability, due to any fault in the unidirectional power flow results in an
outage for the entire section of the feeder. The distribution voltages range from 4kV to
35kV with the most common primary distribution voltage in use in North America as
12.47kV [3].
In a normal DN, the highest voltage is within the DS, usually at the output of the
Load-Tap Changing transformers (LTC). The lowest voltage is usually at the furthest
customer’s site. Both the high and the low voltages must be kept within tightly regulated
standards of 0.95pu to 1.05pu (114V-126V on a 120V base) utilization voltage, as stated
by the American National Standards Institute (ANSI) Service Voltage Standard C84.1 [4]
[39].
To maintain the voltage level at a customer’s site, automatic voltage control is
provided by bus regulation at substation, individual feeder regulators at the substation,
and supplementary pole top line regulators along the feeder. A Voltage-Regulating Relay
4
(VRR) controls the tap changes based on the voltage settings, the bandwidth, and the time
delay. LTC’s are limited to raising or lowering the voltage levels at the point of the tap
changing transformer. Capacitors installed in a radial network, affect the voltages
between the transformer, and the capacitor. One of the most economical, and therefore
prevalent, ways to achieve voltage control is the use of LTC at the substation and shunt
capacitors on the feeders and at the stations [2].
The approximate formula for Voltage Drop (VD) calculations given by (1) shows
the relation between the parameters of the system and the load demand.
VD = IRR + IXXL

Where; IR and, IX are the real and the reactive components of the current required by the
load, R is the resistive component of the line and XL is reactive component of the line.
VD is related to the Resistive (R) and Reactive (XL) parameters of the line, with
IR and IX flowing through the conductors of the distribution feeder.
Corrective measures such as adding a shunt Capacitor with reactance (XC) as
suggest be (2), can reduce the effect of the XL, lowering the vector magnitude of VD.
This also will lower the magnitude of the current seen by the conductor and lower the line
losses caused by I2.

VD = IRR + IXXL - ICXC

5
Where; IR, IX, R, XL are the parameters a defined above, IC is the reactive current through
the and Xc is the reactive component of the capacitor.
IRXR
IRXR
IXXL-ICXC
VD
VD
IXXL

Figure 1.
Vector represenation of capacitor correction.
In certain feeders with small conductor diameter, high impedance, or long lines, R
and XL increase, making the VD become more sensitive to the amount of IR and IX of the
load demand. This high impedance line is often referred to as a weak network and
presents additional challenges [5].
The Line Drop Compensator (LDC) in the voltage regulators, actively monitors
the line current IL at the LTC and uses the R/X ratio of the line to predict the VD
suggested by (3) at a point with some distance (d) away from the regulator location. The
accuracy of this predictive measuring point is dependent on the loading before and after
the point at the distance (d).

VD = |IL| × Reff × cosθ + |IL| × Xeff × sinθ 
6
Where; |IL| is the magnitude of the line current, Reff is the calculated resistance to d, Xeff,
is the calculated reactance to the point at distance “d”, and θ is the angle difference
between the I and the V.
Distribution protection’s primary goals are; 1 - to detect and isolate the fault; 2 –
to minimize the number of customers affected by a fault, and 3 - to minimize the duration
of the fault, and if possible, 4 - to locate the fault. The most common fault experienced
on a DN system is a line to ground fault, which may be sustained or momentary. A
sustained fault refers to a conditions requiring repair of the line, such as the conductor
that has fallen to the ground due to an accident. Momentary or transient fault refers to a
condition that line may be re-energized without a need for any repairs. If a momentary
fault was to occur on a fused lateral, then it needs to be cleared before fuses are blown to
prevent interruption of service to the customers served by the lateral. To prevent fuses
from blowing for this transient fault, the protection scheme requires a high-speed tripping
breaker that can clear the fault before blowing of the fuse. Then an automatic reclosing
scheme can re-energize the system by closing the breaker again resume the service to the
customers. If the fault is a sustained or permanent type, then the fuse closest to the fault
should blow to limits the number of customers affected and to isolate the troubled line
location for repairs. To have the fuse closest to the permanent fault blow, the fuse and all
other interrupting devices in series with it up to the feeder breaker must be time
coordinated with one another. Usually the main feeder and feeder tie line are protected
by breakers or reclosers rather than fuses. The coordination of these individual
7
components requires knowledge of the loads at each of the various taps and loads, as well
as the characteristics of the protection devices should they fail [2]. Again, proper
coordination of protective devices are designed to; eliminate service interruptions,
minimize the extent of customers affected, and assist with the locating of the faults. As
one can see by Fig. 2, a minimum of expensive equipment is used to achieve the goals of
protection with a bidirectional power flow.
Distribution Substation
A
1
2
3
4
B
5
6
7
8
Legend
Fuse
Breaker
Load
Zone of
Protection
Figure 2.
9
Example of a DN with switches, fuses and protection overlap shown.
8
HOW DG IS CAPABLE OF DISRUPTING THE DN
Adding DG into a simple DN drastically changes the DN environment and
variables. One of the advantages is the reduction of demand on the main feeder and its
tie lines, resulting in the delay or postponement of equipment upgrades due to load
growth. Evenly dispersed generation also has the ability to support the local voltage,
making the voltage profile more constant for DN is that suffered poor voltage levels
before installation DG during peak loading scenarios. In reference [6], the authors were
able to show that under normal operating conditions, the possibility of 100% DG
integration resulted in voltages within operating limits. However, there are a large
number of effects that are viewed as negative for the; design, voltage profile and
protection of the DN. These negative effects limit the amount of generation that can be
adopted on a DN.
With small generation (usually below 25% of individual customers power
consumption), at only a few customers sites, few changes to the system are noticed. This
type of generation is non-exporting in nature, and maintains the systems bidirectional
power flow, reducing the load on the DN. The DG is generally incapable of islanding
and therefore does not need to be monitored for such problem. The DG has the effect of
improving voltages, and slightly destabilizes the protection equipment by changing the
demand values and short circuit ratings originally used in the protection study. The
losses in the DN are also reduced, benefiting the Distribution Network Operators (DNO)
9
costs [7]. If the generation is induction type, capacitor correction is recommended at the
Point of Common Coupling (PCC) to keep power factors near unity.
With the introduction of small DG capable of exporting power from a customer’s
site onto the lateral, some concerns although small are noticed. The voltage at the PCC
on a weak network can cause the voltage to rise on the lateral with DG’s; while nearby
laterals can be operating at significantly lower voltages. The voltage discrepancies can
make for challenging voltage control issues on the DN, possibly requiring additional
voltage regulating equipment, or the re-location of existing equipment. This size DG is
capable of islanding a portion of the feeder, and needs to be monitored to prevent
accidental islanding, and most importantly prevent the island from re-connecting with the
grid while out of phase. The bidirectional power flow disrupts the simple protection
equipment, requiring more protection equipment and complete review of the protection
system and coordination of devices. Losses in the DN begin to increase costs for the
DNO, and accelerate as more DG is added [7].
When larger number of small exporting DG’s, or large DG’s, that are capable of
reverse power flow on the laterals and main feeders are present, the DN requires a
complete re-design. The voltage regulating equipment will be incorrectly programmed,
the protection equipment is completely inadequate in components and coordination for
the power flows, and the system is a serious risk to the DG equipment and the DN
operator. Many islanding scenarios of the feeder sections are possible with high
10
probabilities. Such conditions require preventative measures to safeguard the DG and DN
equipment.
Capacitor Related Disruptions
Pole top switched capacitors correct power factor and improve voltage profile
between the capacitor and power source. Due to daily load curves, mixtures of switched
and fixed capacitors are often used on DN. Most capacitor banks are placed near the
location of reactive demand, or roughly 2/3 of the distance down the mains on evenly
loaded feeders. The addition of DG alters the effectiveness of the capacitors in several
ways. If the DG is near unity PF, or if it is a synchronous generator (which is normally
set at a fixed power factor schedule), the voltage at the PCC will usually rise, requiring
the reduction of fixed capacitive support, and even the possible need of switched reactors
to bring the voltage down. However, if the DG is an induction unit, the reactive support
of the capacitors may no longer be adequate [8]. In addition, the dynamic nature of DG’s
must be taken into account. Solar concentrating panels output greatly reduces with cloud
shadows and wind turbines are equally affected by wind variability. This very nonuniform output further stresses the voltage control schemes.
Tap Changer Settings Disruption
On-Load Tap Changer (OLTC) transformers controlled by Automatic Voltage
Control (AVC) relay are commonly used in the DN. As DGs are added to the DN,
voltage regulation problems can arise, altering the performance of the AVC relay. Both
11
the power factor of the DG and the power factor of the load flowing through the OLTC
can cause the AVC relay measured voltage to be shifted. This results in an inaccurate
voltage control. At the very minimum, any introduction of DG into a DN requires a reevaluation of the settings for the control equipment to verify that standards are upheld in
individual worse case scenarios and loading conditions [9].
A simplified example of this inaccurate voltage control is shown in Fig. 3. Using
equations (1) and (3), the actual VD and the value as determined by the LDC have been
calculated for the location at the distance “d”. Load1 and Load2 are each set as 2MW
with a Power Factor (PF) of 1. Load1, point “d”, and Load2 are located 4, 5, and 8 miles
from the substation respectively. The feeder conductor used is 266.8-kcmil all aluminum
conductor, with 37 strands and 53-inch geometric spacing. The impedance on the line is
z = 0.386 + j0.6611 Ω/mi. The line-to-line operating voltage is 13.2kV. The DG’s
output is varied from 0 to 4MW with PF=1 to determine its effect on the LDC’s accuracy.
Once the values are obtained, (4) is used to show the percent difference between the
actual VD and the one determined by LDC.

Diff = 100(VDActual- VDLDC)/(VDActual) 
Where; VDActual represents the true VD, and VDLDC represents the LDC’s estimated VD.
12
4 mi
8 mi
d
Substation
LDC
Load2
Load1
DG
Figure 3.
Example of LDC regulation issues.
Before the DG delivers power into the system, the LDC is accurate. As shown in
Fig. 4, once the DG begins to inject power at Load1, the accuracy of the LDC
deteriorates. This phenomena changes with the location of “d” with respect to the loads
and DG(s) on a given LDC monitored line.
Inaccuracy of LDC VD compared to actual VD as a funtion of
Generator output
120
100
%Diff
80
60
40
20
0
0
50
100
150
200
Generation percent of Load1
Figure 4.
Inaccuracy of LDC Vs DG injection.
250
13
Protection Disruptions
The total fault current is the vector summation of the individual fault current
contributions from each of the sources. This is primarily the short circuit ratings of the
feeder, and all rotating loads. With the introduction of DG, the total fault current rises
while, the contributions from each source decreases [15], [40]. Depending on the type
and size of the cumulative DG, the larger the contribution, it could have to the fault
current [10]. As the total fault current increases, the interrupting rating of the protective
equipment must be re-evaluated for their adequacy [40]. Protection equipment is installed
and set according to the current load profile and short circuit ratings. These settings are
altered with the introduction of DG, and require the addition of more protection
equipment, with updated protection schemes [11]. Currently to limit the need for large
changes to the DN, National Standard IEEE 1547 [12] requires all generation to
disconnect in the presence of a fault or disturbance on the feeder. As DG increases, the
need to keep more generation online during nearby faults becomes more advantageous
for system stability and reliability [13].
For example, in Fig. 5, the power flows from the substation to a main feeder
equipped with other lateral feeders protected by breakers at each of the relays R1 through
R3. If a fault occurred at location “F”, the protection scheme would be designed to trip
R2 before R1, with R1 providing backup protection in case of miss-operation by R2.
14
1
Substation
2
R1
Load 1
Figure 5.
F
3
R2
Load 2
R3
Load 3
Load 4
Example of DN before DG is added.
However, if exporting capable DG is added in Fig. 6, then a fault at location “F1”
must also be detected by each of new DG sources, namely G2 and G3. As the DG
penetration levels increase, the contributions from each source will decrease, thus making
fault detection less sensitive with longer clearing times from each of the sources [15].
This is highly undesirable.
1
Substation
F1
R1
Load 1
2
3
R2
Load 2
Load 3
G2
Figure 6.
R3
Load 4
G3
Example of DN with no upgrades and DG.
The changing system conditions such as variability, present further issues. As G2
and G3 change dynamically, due to lack of wind or clouds on concentrating solar panels,
the maximum and minimum fault values also change. As the fault values drift, the
15
protection scheme may not operate properly [11]. In Fig. 7, sympathetic tripping of G2
or G3 can also occur when fault “F2” occurs, and the DG’s contribution to the total
instantaneous fault current exceeds the DG’s relay setting before the relay R2 or R3
senses and opens [14].
The DG is also capable of rendering a relay completely insensitive. In Fig. 7, if
fault F2 occurs, the G4 generator adds to the total fault current, while the R3 relay sees
lower fault values of fault current due to the reduced contributions from other sources
depending on the type of DG and the distances between relay, generator and load [15].
This desensitization of the relay R3 causes a slower response, or even a lack of tripping,
until the unit G4 trips [14].
1
Substation
2
R2
R1
Load 1
3
Load 2
F2
G4
Load 4
Load 3
G2
Figure 7.
R3
G3
Example of DN with possiblity of relay desensitization.
Islanding Concerns
Islanding is possible when the DG is independently capable of supporting any
portion of the DN. Human safety issues arise, as equipment may be energized when the
servicemen believe it is de-energized. The island may not be operating correctly per
16
operational standards [4], with the possibility for improper grounding, voltage, frequency
or adequate protective equipment. Reconnection to the rest of the grid when out of phase
can cause fault like conditions and stresses to the equipment [16 17]. Passive islanding
detection and disconnection can be accomplished by monitoring voltage, frequency and
other system parameter variations and rate of changes at the DG’s PCC. However, if the
DG and the islanded DN load are balanced, then the island can become a Non-Detection
Zone (NDZ) [18]. Active anti-islanding monitoring is accomplished by the introduction
in fluctuations in frequency, phase shifting, reactive power injection, current injection,
positive feedback and other methods. These small fluctuations result in negligible
changes when the system is operating normal, yet create significant detectable changes
when the DG is part of an island. Hybrid approaches that combine both active and
passive monitoring techniques will provide more robust systems of detection for the
various islanding conditions [16-17].
17
HOW THE IMPACT OF DG CAN BE MINIMIZED TO ALLOW FOR MAXIMUM
DG PENETRATION
Strategic investments in enabling technologies can greatly increase the DN’s
ability to utilize an increase in DG. There have been a host of ideas to increase the
capacity, and still maintain quality of service. Some concepts are simplistic, and some are
revolutionary. The main proposals published for this purpose including their benefits
and/or shortcomings are presented in the following sub-sections.
Upgrading Conductors
One of the original solutions to improve the DN power transfer capability has
been the upgrade of feeder conductors. As an example, each DG can be put on its own,
properly sized radial feeder, serving the local High Voltage / Medium Voltage (HV/MV)
substation. While this solution works, it is inherently expensive and limits the realization
of DG in remote locations where renewable resources such as wind and solar are more
likely to be found. If the DG is placed on a weak network, and is incapable of realizing
its full potential during low local consumption, the conductors between the DG and the
HV/MV substation necessitate upgrading to increase the capacity of the DN to accept
more DG. This is also an unattractive solution, as the cost to the DG may outweigh the
gain in capacity increase.
18
Optimal Placement Planning
Individual DG related problems could be mitigated by technically proper
placement of DG on feeders. However, there is concern that poorly placed, or incorrectly
sized DG in a DN could limit the placement of additional DG [19]. Optimal Power Flow
(OPF) is a widely available tool that most DNO’s use in case studies. The OPF software
can be utilized to determine the maximum capacity of the DG on DN systems. The
program can also determine the optimal placement of DG in the DN. To utilize this
readily available tool, DG’s are modeled as negative loads. By using this method, the
capacity evaluation can be modeled as a load addition problem, with the voltage and
thermal limits of the lines taken into account. The system is at its capacity when the cost
associated with negative load is minimized. This method can help point out where on the
DN a particular DG unit would be most beneficial. Or, the locations where additional DG
in would degrade the remaining network capacity. For example, a 1MW increase in a
poor location can lower the DN capacity for DG at another bus by more than 2MW.
Furthermore, using the OPF allows the identification of limiting factors, allowing the DN
owner to accurately determine where strategic investments would allow additional DG to
be added to the network. This is, however, a single point analysis, and would have to be
repeated each time a DG is connected to the system [20].
Voltage Control
Voltage control is one of the most widely studied topics, and one of the most
mature systems in the DN. The equipment includes capacitors, reactors, tap changing
19
transformers, Static Var Compensators (SVC), Static Synchronous Compensators
(STATCOMS), and other devices. Many algorithms and fuzzy logic systems have been
proposed and analytically shown that the existing voltage control equipment, with the
proper modeling, may be capable of handling the increased dynamics of intermittent or
increased DG [5,9,21-31]. However, all of these proposals will have to be tested and
implemented in the field to prove practical viability. With the addition of FACTS
devices further control may be possible [32-33]. As research in this area is so broad and
widely studied, only a brief summary of two cost effective DG enabling mitigation
techniques is required.
Specific case studies have shown that local voltage rise can be mitigated with the
addition of a reactor at the DG’s PCC, and additional capacitors at the HV/MV
substation. Although this is an inexpensive capacity increasing solution, this topology
greatly increased the active power losses for the DN. The study also determined that the
lowering of the upstream HV/MV stations voltage facilitated the most efficient network
transfer capacity. Though this is one of the most cost-effective methods, it was noted that
many DN operators would be unwilling to accommodate such a request [8].
DG capable of variable reactive injection support, coupled with utilization of
SCADA communication on the distribution network, would facilitate the coordination
between the voltage controlling equipment and ultimately a better system voltage profile.
With this more complete information, all available voltage controlling equipment in the
DN can be managed by a single algorithm. This centrally coordinated voltage algorithm
20
allows for more accurate usage of each individual component and keeps individual
voltage controlling equipment from “hunting” of the regulating equipment. “Hunting”
refers to a problem encountered when each regulating component utilizes only local
voltage and current readings to control its output independent of other voltage controlling
equipment. The independent local controls often causes initiate unnecessary steps
depending on the corrective actions taken by other voltage regulators. With variable
reactive support from the DG, even fewer actions are required by the existing voltage
controlling equipment.[34] The cost for establishment of a centralized control system
and associated communication scheme for this purpose will be significant.
Redesigning the Network to a Mesh System
As more DG is introduced, the power becomes bidirectional, much like the mesh
networks of the HV power grid. This complete re-design also creates its own problems.
The paralleling of the transmission network and the DN increases the short circuit duties
on the system. The increased levels of fault currents alone will require equipment with
higher interrupting rating which can easily go beyond the limits for distribution
equipment, thus making the option impractical. Another meshed option may be balanced
looped secondary sub-feeders, connecting consumers and producers. Despite the
increased fault values, one of the advantages of such a system is for isolation of the
faults. If a fault is on the main feeder, the normally open breakers, and closed breakers,
can be rearranged to keep as much of the system energized as possible. If the fault is
inside one of the loops, the fault is localized, and other loops are not affected. If the loop
21
is equipped with fault locating relays, the fault can be further isolated, splitting the loop
into two radial sub-feeders. The study showed an increase from 47% maximum
reachable DG insertion for a secured feeder, to 67% depending on the number of
secondary loops. The authors verified that this topology is capable of service continuity
and then argue that the cost savings due to the reduced power losses can be comparable to
the secured feeder [35]. This point will have to be verified by future research and field
implementation.
Islanding
Islanding falls into two categories; intentional and unintentional. Unintentional
islanding can have disastrous effects as the island may be out of phase during the
reconnection, system operators may be unaware that the equipment is energized, and
systems of protection and voltage regulation will be inherently compromised. Standards
IEEE 1547 [12] has a maximum delay of 2 seconds for a DG to detect and disconnect
form an unintentional island. However, passive and active monitoring may not be
enough for a timely detection and isolation of an unintentional island. Hybrid approaches
that utilize passive and active monitoring are able to improve the speed of detection,
decrease the threat of NDZ’s, and maintain little change in the power quality of the DG
[16]. Some promising research has been completed in comparing the behavior of
islanded and non-islanded Automatic Load Frequency Controllers (ALFC). The actions
of the ALFC coupled with learned behavior of self-organizing map neural networks have
shown to detect the difference between islanded and non-islanded operation within
22
200ms and 97.98% accuracy. The detection would signal the relay to disconnect at the
PCC far faster than the required 2 seconds [36].
With intentional islanding, regulating and protecting the equipment is the primary
concern. The benefits gained in terms of service reliability and deferment of revenue
losses from larger area blackouts is promising. However, the infrastructure changes are
also significant. Most synchronous generators are capable of islanded operation due to
the nature of their excitation systems and the ALFC. However, induction generators
absorb reactive power from the system thus requiring voltage support from equipment
such as capacitors. Inverters are currently designed and set as constant current devices,
not as voltage control devices. If the inverter based DG’s within the detected island are
capable of sensing the island conditions and switching to a voltage control scheme, it is
possible to intentionally control the island with load shedding schemes in place to keep
high priority loads in service inside the island. This controlled island must also be able to
re-synchronize with the grid as part of the scheme with the inverter based DG’s reverting
back to the constant current mode of operation [37].
Fault Current Limiting Devices
Higher levels of DG penetration increase the short circuit current (ISC) in the DN
[10],[15],[40]. There are ways to mitigate this increase through the use of
Superconducting Fault Current Limiting Devices (SFCL). The two primary types are
resistive and inductive superconductors. In a resistive type, the resistance is nearly
negligible, until a large fault creates enough heat to quench the superconductor, then the
23
SFCL becomes resistive and limits the fault. An inductive type is like a transformer with
a closed superconductor ring for a secondary winding. During normal operation, the
resistance of the secondary is negligible, however, in the presence of a fault the
superconductor is quenched and the impedance of the primary rises sharply. With all
protection equipment, there is very short delay (fractions of a cycle) between a fault and
when the circuit breaker is initiated to trip. With the SFCL in use, the current is limited
(Ilim) for a duration equal or slightly greater than the automatic switching device by
design. In [38] the installed SFCL results showed that; without the SFCL the ISC were
1650Amps peak, and did not diminish with time. With the SFCL in place, the tests were
repeated, and the Ilim was 960Amps at the beginning of the fault and 310Amps when the
fault was cleared. This diminishing of fault current was due to the temperature dependant
resistive properties of the SFCL. Such devices can theoretically allow more DG to be
added into the DN while limiting the negative impacts. Further research and field tests
are required for verification of results in practical systems.
Adaptive Protective Relaying
By the existing rules and criteria, each of the contributing energy sources such as
the substation or any interconnected DG is responsible to detect the faults on the DN
system and interrupt its own contribution. This removes the benefits that DG can offer a
system support during a nearby fault that may be isolated by other devices [13]. The DN
protection system could be designed to adapt to the changing environment, using a host
of new equipment and communication. The configuration consists of relays with
24
communication links to downstream and adjacent current protective equipment. Each
relay would be equipped with both adaptive primary and secondary functions. If a fault
occurs in a primary zone of protection, the loads and DG currents upstream of the relay
are measured and sent to the relay. The relay is capable of identifying the fault using
various analytical techniques, and isolates the faulted zone with the appropriate breakers.
If a fault occurs in the backup zone, the loads and DG currents downstream of the relay
are transformed to branch currents, whereby the changes associated with the loads and
DG’s can be eliminated. This system has shown to be highly effective, and able to
improve the protection performance under changing load and generation [14]. Additional
research and studies are needed to verify practicality of this concept.
25
CONCLUDING REMARKS
Renewable targets created by state and federal regulatory commissions, and
residential installation of co-generation have begun to alter the dynamics of DN. A
number of valid concerns and issues related to Voltage Regulation [11], Protection [3940], and existing standards [15] have been reported for the DN by the research as the
penetration level of DG increases. Voltage concerns are related to higher than normal
voltages seen by the customers located near DG. Relay desensitization is experienced due
to reduction in levels of fault current contributions associated with interconnection of
additional DG. As shown in Fig.4 LDC’s lose the ability to accurately predict the VD
with the introduction of DG, and may cause voltages to rise or fall outside of the
operating standards [9]. Despite the issue and concerns, some research has also shown
that strategic investments in enabling technologies can greatly increase the DN’s ability
to utilize an increase in DG. National Standard IEEE 1547 [12] is a stepping-stone to
furthering the possibilities of DG, as separate standards focused on high DG penetration
are needed. Optimization of generation location is currently a tedious process and further
research on rapid analyzing for multiple generators and locations would be beneficial.
Voltage control, despite how widely it has been studied, will need further research and
investigation as multiple solutions to other problems alter the existing topology. Mesh
topology offers solutions to many problems, and introduces a host of new scenarios to
study. Investigations will include; transitioning from radial to mesh, the cost,
maintenance, short circuit studies, voltage control, and system protection. The use of
hybrid detection techniques has shown effective reduction of the NDZ. Using the ALFC
26
for island detection although considered a novel approach, it lacks verification through
additional research and case studies. FCL’s have shown to be effective in test
environments. While the installation at practical sites, cost minimization, and its effect
on the system need further investigation. Advancements in relay techniques focused on
the interaction of varying levels of generation also needs further research to verify their
adequacies for protection of the DN with higher levels of interconnected DG.
27
Biographies
Daniel E Ilse received BS in Electrical Engineering – 2006 from Califonia State
University, San Jose and currently pursuing a master’s degree in Electrical Engineering
from California State University, Sacramento with a specialization in power systems.
Mohammad Vaziri received BS EE- 1980, MS EE -1990, and Ph.D. EE - 2000 degrees
form UC, Berkeley, CSU, Sacramento, and WSU, Pullman WA, respectively. He has 24
years of professional experience at PG&E, and CAISO, and over 16 years of academic
experience at CSU Sacramento, San Francisco, and WSU Pullman. Mohammad has
authored and presented technical papers and courses in US, Mexico, and Europe. He is a
senior member of the IEEE, actively involved in funded research for integration of
renewable resources, and serves as member/advisor on various technical committees.
Currently, he is a faculty in the Electrical and Electronic Engineering department at the
CSU Sacramento and a part time technical consultant. Dr. Vaziri is a registered
professional engineer in the state of California, and his research interests in Power
Systems are; Operation, Planning, Protection, and Integration of Distributed Generation.
Suresh Vadhva received BS - 1974, MS -1976, and Ph.D. - 1982 degrees form
University of Tulsa, University of New Mexico, respectively. Currently he is Chair and
Professor of Electrical & Electronics Engineering department and Computer Engineering
Program Coordinator at California State University, Sacramento. He is also Associate
28
Director of Research at the California Smart Grid Center, Sacramento. He is member of
organizing committee and financial chair for IEEE IRI 2008 -2010 conferences.
Previously he was chair of Computer Science Department at the University of Montana.
29
References
[1]
Arnold Schwarzenegger. (2008, November 17) Executive Order S-14-08, Web.
16 Aptil 2012. [Online]. Available: http://www.energy.ca.gov/33by2020/
[2]
IREC. “U.S. Solar Market Trends 2010,” Interstate Renewable Energy Council,
June 2011. Web. 16 April 2012. [Online]. Availible: http://irecusa.org/wpcontent/uploads/2011/07/IREC-Solar-Market-Trends-Report-revised070811.pdf
[3]
T. Gonen, Electric Power Distribution System Engineering, 2nd ed., CRC Press,
Boca Raton: Florida. 2008, pp.371-556.
[4]
American National Standard for electric power systems and equipment-Voltage
Ratings (60 Hz), ANSI C84.1-1995, Range A. pp. 3, Rosslyn, VA. : National
Electrical Manufacturers Association, 2006
[5]
Kiprakis, A.E.; Wallace, A.R.; , "Maximising energy capture from distributed
generators in weak networks," Generation, Transmission and Distribution, IEE
Proceedings- , vol.151, no.5, pp. 611- 618, 13 Sept. 2004
[6]
Richardson, Peter; Keane, Andrew; , "Impact of high penetrations of microgeneration on low voltage distribution networks," Electricity Distribution - Part
1, 2009. CIRED 2009. 20th International Conference and Exhibition on , vol.,
no., pp.1-5, 8-11 June 2009
[7]
Quezada, V.H.M.; Abbad, J.R.; Roman, T.G.S.; , "Assessment of energy
distribution losses for increasing penetration of distributed generation," Power
Systems, IEEE Transactions on , vol.21, no.2, pp. 533- 540, May 2006
30
[8]
Repo, S.; Laaksonen, H.; Jarventausta, P.; Huhtala, O.; Mickelsson, M.; , "A case
study of a voltage rise problem due to a large amount of distributed generation on
a weak distribution network," Power Tech Conference Proceedings, 2003 IEEE
Bologna , vol.4, no., pp. 6 pp. Vol.4, 23-26 June 2003
[9]
Salman, S.K.; Wan, Z.G.; , "Voltage control of distribution network with
distributed/embedded generation using fuzzy logic-based AVC relay,"
Universities Power Engineering Conference, 2007. UPEC 2007. 42nd
International , vol., no., pp.576-579, 4-6 Sept. 2007
[10]
Baghaee, H.R.; Mirsalim, M.; Sanjari, M.J.; Gharehpetian, G.B.; , "Effect of type
and interconnection of DG units in the fault current level of distribution
networks," Power Electronics and Motion Control Conference, 2008. EPE-PEMC
2008. 13th , vol., no., pp.313-319, 1-3 Sept. 2008
[11]
Khoan Tran; Vaziri, M.; , "Effects of dispersed generation (DG) on distribution
systems," Power Engineering Society General Meeting, 2005. IEEE , vol., no.,
pp. 2173- 2178 Vol. 3, 12-16 June 2005
[12]
IEEE standard for Interconnecting Distributed Resources with Electrical Power
Systems, IEEE STD 1547-2003, New York, NY: IEEE Press, 2003
[13]
Martinez-Velasco, J.A.; Martin-Arneclo, J.; , "Distributed generation impact on
voltage sags in distribution networks," Electrical Power Quality and Utilisation,
2007. EPQU 2007. 9th International Conference on , vol., no., pp.1-6, 9-11 Oct.
2007
31
[14]
Hussain, B.; Sharkh, S.M.; Hussain, S.; , "Impact studies of distributed generation
on power quality and protection setup of an existing distribution network," Power
Electronics Electrical Drives Automation and Motion (SPEEDAM), 2010
International Symposium on , vol., no., pp.1243-1246, 14-16 June 2010
[15]
Vaziri, M.; Vadhva, S.; Oneal, T.; Johnson, M.; , "Smart grid, Distributed
Generation, and standards," Power and Energy Society General Meeting, 2011
IEEE , vol., no., pp.1-8, 24-29 July 2011
[16]
Abdolrasol, M.G.M.; Mekhilef, S.; , "Robust hybrid anti-islanding method for
inverter-based distributed generation," TENCON 2010 - 2010 IEEE Region 10
Conference , vol., no., pp.13-18, 21-24 Nov. 2010
[17]
Abdolrasol, M.G.M.; Mekhilef, S.; , "Three phase grid connected anti-islanding
controller based on distributed generation interconnection," Power and Energy
(PECon), 2010 IEEE International Conference on , vol., no., pp.717-722, Nov. 29
2010-Dec. 1 2010
[18]
Vieira, J.C.M.; Freitas, W.; Wilsun Xu; Morelato, A.; , "An Investigation on the
Nondetection Zones of Synchronous Distributed Generation Anti-Islanding
Protection," Power Delivery, IEEE Transactions on , vol.23, no.2, pp.593-600,
April 2008
[19]
Keane, A.; O'Malley, M.; , "Optimal Allocation of Embedded Generation on
Distribution Networks," Power Systems, IEEE Transactions on , vol.20, no.3, pp.
1640- 1646, Aug. 2005
32
[20]
Harrison, G.P.; Wallace, A.R.; , "Optimal power flow evaluation of distribution
network capacity for the connection of distributed generation," Generation,
Transmission and Distribution, IEE Proceedings- , vol.152, no.1, pp.115-122, 10
Jan. 2005
[21]
Conti, S.; Greco, A.M.; , "Active MV Distribution Network Planning Coordinated
with Advanced Centralized Voltage Regulation System," Power Tech, 2007 IEEE
Lausanne , vol., no., pp.2105-2109, 1-5 July 2007
[22]
O'gorman, Raymond; Redfern, Miles; , "The impact of distributed generation on
voltage control in distribution systems," Electricity Distribution, 2005. CIRED
2005. 18th International Conference and Exhibition on , vol., no., pp.1-6, 6-9
June 2005
[23]
Richardot, O.; Viciu, A.; Besanger, Y.; Hadjsaid, N.; Kieny, C.; , "Coordinated
Voltage Control in Distribution Networks Using Distributed Generation,"
Transmission and Distribution Conference and Exhibition, 2005/2006 IEEE PES ,
vol., no., pp.1196-1201, 21-24 May 2006
[24]
Kim, T.E.; Kim, J.E.; , "Voltage regulation coordination of distributed generation
system in distribution system," Power Engineering Society Summer Meeting,
2001 , vol.1, no., pp.480-484 vol.1, 2001
[25]
Jauch, E.T.; , "“Conservation biased” distribution Volt/Var/(kW) management,"
Power & Energy Society General Meeting, 2009. PES '09. IEEE , vol., no., pp.17, 26-30 July 2009
33
[26]
Su, C.L.; , "Comparative analysis of voltage control strategies in distribution
networks with distributed generation," Power & Energy Society General Meeting,
2009. PES '09. IEEE , vol., no., pp.1-7, 26-30 July 2009
[27]
Fila, M.; Taylor, G.A.; Hiscock, J.; Irving, M.R.; Lang, P.; , "Flexible voltage
control to support Distributed Generation in distribution networks," Universities
Power Engineering Conference, 2008. UPEC 2008. 43rd International , vol., no.,
pp.1-5, 1-4 Sept. 2008
[28]
Senjyu, T.; Miyazato, Y.; Yona, A.; Urasaki, N.; Funabashi, T.; , "Optimal
Distribution Voltage Control and Coordination With Distributed Generation,"
Power Delivery, IEEE Transactions on , vol.23, no.2, pp.1236-1242, April 2008
[29]
Triggianese, M.; Morren, J.; de Haan, S.W.H.; Marino, P.; , "Improved and
extended DG capability in voltage regulation by reactive and active power,"
Power Engineering, Energy and Electrical Drives, 2007. POWERENG 2007.
International Conference on , vol., no., pp.583-588, 12-14 April 2007
[30]
Carvalho, P.M.S.; Correia, P.F.; Ferreira, L.A.F.; , "Distributed Reactive Power
Generation Control for Voltage Rise Mitigation in Distribution Networks," Power
Systems, IEEE Transactions on , vol.23, no.2, pp.766-772, May 2008
[31]
Smith, J.W.; Brooks, D.L.; , "Voltage impacts of distributed wind generation on
rural distribution feeders," Transmission and Distribution Conference and
Exposition, 2001 IEEE/PES , vol.1, no., pp.492-497 vol.1, 2001
34
[32]
Yuvaraj, V.; Deepa, S.N.; Rozario, A.P.R.; Kumar, M.; , "Improving Grid Power
Quality with FACTS Device on Integration of Wind Energy System," Modelling
Symposium (AMS), 2011 Fifth Asia , vol., no., pp.157-162, 24-26 May 2011
[33]
Roy, N.K.; Hossain, M.J.; Pota, H.R.; , "Voltage profile improvement for
distributed wind generation using D-STATCOM," Power and Energy Society
General Meeting, 2011 IEEE , vol., no., pp.1-6, 24-29 July 2011
[34]
Kulmala, A.; Mutanen, A.; Koto, A.; Repo, S.; Järventausta, P.; , "RTDS
verification of a coordinated voltage control implementation for distribution
networks with distributed generation," Innovative Smart Grid Technologies
Conference Europe (ISGT Europe), 2010 IEEE PES, vol., no., pp.1-8, 11-13 Oct.
2010
[35]
Alvarez-Herault, M.-C.; Picault, D.; Caire, R.; Raison, B.; HadjSaid, N.; Bienia,
W.; , "A Novel Hybrid Network Architecture to Increase DG Insertion in
Electrical Distribution Systems," Power Systems, IEEE Transactions on , vol.26,
no.2, pp.905-914, May 2011
[36]
Darabi, A.; Moeini, A.; Karimi, M.; , "Distributed generation intelligent islanding
detection using governor signal clustering," Power Engineering and Optimization
Conference (PEOCO), 2010 4th International , vol., no., pp.345-351, 23-24 June
2010
35
[37]
Balaguer, I.J.; Qin Lei; Shuitao Yang; Supatti, U.; Fang Zheng Peng; , "Control
for Grid-Connected and Intentional Islanding Operations of Distributed Power
Generation," Industrial Electronics, IEEE Transactions on , vol.58, no.1, pp.147157, Jan. 2011
[38]
Dalessandro, Rossella; Ascade, Massimo; Bartalesi, Daniele; Bocchi, Marco;
Rossi, Valerio; Martini, Luciano; , "Development, installation, field test and
operational experience of a superconducting Fault Current Limiter in a distributed
generation Test Facility," Electricity Distribution - Part 1, 2009. CIRED 2009.
20th International Conference and Exhibition on , vol., no., pp.1-4, 8-11 June
2009
[39]
M. Vaziri, S. Vadhva, C. Hoffman, K. Yagnic, S. Ghadiri “Standard, Rules, and
Issues for Integration of Renewable Resources” Proceedings of the IEEE PES
General Meeting – 978-1-4244-8357-0/10 Summer 2010, July 25-29 2010.
Minneapolis, Minnesota
[40]
G. Kaur, M. Vaziri, “Effects of Distributed Generation (DG) Interconnections on
Protection of Distribution Feeders ” Proceedings of the IEEE PES General
Meeting – Summer 2006, May, 2006. Montreal Canada 1-4244-0493-2/06/$20.00
©2006 IEEE