Section 2.

The Potential Benefits of DG on Increased

Electric System Reliability

2.1 Summary and Overview

Electric system reliability is a measure of the system’s ability to meet the electricity needs of customers. It
is a term used by electric system planners and operators to measure aggregate system conditions, and as
an aggregate measure, it generally applies to entire service territories or control regions. As such, the
reliability of the electric system depends on the reliability of that system’s component parts, including, for
example, power plants, transmission lines, substations, and distribution feeder lines. To help ensure a
reliable system, planners and operators prefer having as much redundancy in these components as can be
justified economically.

System reliability is also dependent on events that affect daily operations, including the decisions made
by grid operators in real-time in response to changing system conditions. Operators like to have as much
real time, and location-specific information as they can get about system conditions, as well as the ability
to control power flows and dispatch power plants to enable effective response when problems occur.
Weather is the primary reason for reliability problems, and includes problems caused by lightening
strikes, high winds, snowfall, ice, and unexpectedly hot weather. The goal of both planners and operators
is to have as resilient a system as possible that can adjust to problems without causing major
consequences, and that when outages do occur, they are short-lived and affect the fewest number of
customers as possible.

DG has the potential to be used by electric system planners and operators to improve system reliability;
and there are a few examples of this being done currently. As discussed, DG is primarily used today as a
customer-side energy resource for services such as emergency power, uninterruptible power, combined
heat and power, and district energy. Utilities could do more to use the DG already in place, and they could
increase investment in DG resources themselves. However, successful business models for more
widespread utility use of DG are limited to certain locations and certain conditions.

There are currently two primary mechanisms being used today by utilities to access customer-side DG for
reliability purposes:
• Several utilities offer financial incentives to owners of emergency power units to make them
available to grid operators during times of system need.
• Several regions offer financial incentives or price signals to customers to reduce demand during
times of system need (e.g., demand response programs), and some participants in these programs
use DG to maintain near-normal on-site operations while they reduce their demand for grid-
connected power.

2-1
 Interest in these and other mechanisms to use DG
Madison Gas and Electric (MGE) owns and
to improve system reliability appears to be
operates backup generators at several business
growing, as concerns mount across the country
customers’ sites. These customers, who must
have a monthly demand of at least 75 kW, pay a about the adequacy of current resource plans (e.g.,
monthly fee based upon their maximum annual construction of new generation, transmission, and
demand to have the generation available if distribution facilities) to maintain the reliability of
power is interrupted. If the grid power fails, the the nation’s electric system. 16 There are several
backup units provide power within 30 seconds. reasons for these growing concerns. For example,
After the grid is restored, these units the electric system was generally designed to
automatically synchronize and then shut down provide reliable service by providing multiple
so that the customer does not incur another generators with a total capacity greater than the
service interruption. MGE, which takes
anticipated system peak demand, providing
responsibility for all environmental permits, can
overlapping transmission networks, and, in limited
also use these units to boost system reliability
during an electrical emergency. (Source: locations, including the ability to meet customer
Madison Gas and Electric 2006) electricity needs by managing power flows from
one distribution feeder to another. Planners
generally seek to build capacity in consideration of the single largest contingency, which is the sudden
loss of the largest generator, regional transmission line, or interconnection.

Problems in system adequacy, also called capacity deficiencies, can lead to outages if (1) system
operators activate emergency procedures such as rolling blackouts to avoid further system overload and
catastrophic failure, or (2) if the loss of a key system element results in serious overloads, cascading
equipment failure, and potentially widespread blackouts. While electric system planners and operators
work to avoid such events, the needs for generation, transmission and distribution (T&D) capacity
additions to meet increases in electricity demand have forced some utilities to take precautionary
emergency actions more routinely than in the past (Arthur D. Little, Inc. 2000).

The availability of redundant generating and transmission capacity has made those portions of the system
more robust than the distribution system. However, the recent restructuring of electric power markets and
regulations, and resulting increases in long-distance power transfers, have put pressure on traditional
strategies and procedures for maintaining system reliability. For example, the number of times that the
transmission grid was unable to transmit power for contracted transactions jumped from 50 in 1994 to
1,494 in 2002 (Apt et al. 2004).

In addition to redundant capacity, the electric system also uses operating procedures to provide reliable
service in the event of sudden disturbances. These procedures are needed because power flows reroute at
close to the speed of light whenever power system conditions change (e.g., due to changes in electricity
supply, demand, or weather-related events). For example, operators count on sufficient “spinning”
reserves to supply immediate replacement for any generation failure.

Problems in system operational reliability can usually be classified as faults and failures. Faults are
caused by external events, such as tree contact, animal contact, lightning, automobile accidents, or
vandalism. Failures are caused by an equipment malfunction or human error not linked to any external
influence.

16
North American Electric Reliability Council 2006 Long Term Reliability Assessment – The Reliability of Bulk Power Systems in North
America October 2006

2-2
 “Both faults and failures can cause outages. These outages can be short, lasting less than
15 seconds and quickly resolved by automatic switching equipment. When a fault or a failure
results in a longer outage, it typically involves damage to equipment such as a transformer that
must be repaired or replaced before service can be restored. The time required for such remedies
can range from hours to days or weeks. Faults and failures, rather than capacity deficiencies, are
the causes of most outages. Outages created by faults and failures in generation are rare. While
transmission faults are somewhat more common, 94% of all power outages are caused by
faults and failures in the distribution system (Arthur D. Little, Inc. 2000).” (Emphasis added.)

DG offers the potential to increase system reliability, but it can also cause reliability problems, depending
on how it is used. Often the difference between improving the system and causing problems if a function
of how the DG is integrated with the grid, as noted in a review of critical power issues in Pennsylvania:

“In general, distributed generation can increase the system adequacy by increasing the variety of
generating technologies, increasing the number of generators, reducing the size of generators,
reducing the distance between the generators and the loads, and reducing the loading on
distribution and transmission lines. … Distributed generation can also have a negative impact on
reliability depending upon a number of factors that include the local electrical system composition
as well as the DG itself. These factors include DG system size, location, control characteristics
(including whether the DG is dispatchable), the reliability of the fuel supply, and the reliability of
the DG unit itself (Apt and Morgan 2005).”

2.2 Measures of Reliability (Reliability Indices)

Reliability indices are used by system planners and operators as a tool to improve the level of service to
customers. Planners use them to determine the requirements for generation, transmission, and
distribution capacity additions. Operators use them to ensure that the system is robust enough to
withstand possible failures without catastrophic consequences.

2.2.1 Generation

Reliability is measured using the available data, which varies across utilities and across system
components. One metric universal to all utilities is the loss-of-load probability (LOLP).

“Overall system reliability is often expressed as a loss-of-load probability, or LOLP. Although

based upon a probabilistic analysis of the generating resources and the peak loads, the LOLP is
not really a probability. Rather, it is an expected value calculated on either an hourly or daily
basis. A typical LOLP is “one day in ten years” or “0.1 days in a year.” This is often
misinterpreted as a probability of 0.1 that there will be an outage in a given year. Loss-of-load
probability characterizes the adequacy of generation to serve the load on the system. It does not
model the reliability of the transmission and distribution system where most outages occur
(Kueck et al. 2004).” (Emphasis added.)

Note that the LOLP is a function of the generation and peak loads – it does not include any failures in the
T&D systems.

2-3
2.2.2 Transmission

Transmission failures are relatively rare and indices are not typically used to keep track of transmission
line failure rates. However, at least one reliability council, East Central Area Reliability (now a part of
Reliability First along with other reliability coordinators), calculates an availability that is a function of
outage duration and number of circuits (East Central Area Reliability Coordination Agreement 2000).
Rather, the system is designed and operated so that there is always additional transmission capacity in
place to handle any unexpected line failures.

“The bulwark of reliability for bulk power transmission systems has long been the use of "worst
single contingency" design and operation– often referred to as the "n-1" principle or criterion. It's
kind of the "prime directive" of reliable power system operation. In short, it means that the
system is planned and operated in such a way that it can sustain the worst single disturbance
possible without adverse consequences– consequences like overloads on other facilities,
instability, or loss of firm customer load. The contingency is usually the sudden outage of a key
high voltage transmission line or major generating unit (Loehr 2001).”

2.2.3 Distribution

Other reliability metrics are based upon customer outage data, and the vast majority of these outages
reflect faults and failures in the distribution system. These data describe how often electrical service was
interrupted, how many customers were involved with each outage, how long the outages lasted, and how
much load went unserved. Industry indices are defined in Institute of Electrical and Electronics Engineers
(IEEE) Standard 1366. 17 The most commonly used are listed here.

SAIFI, or system average interruption frequency index, is the average frequency of sustained
interruptions per customer over a predefined area. It is the total number of customer interruptions divided
by the total number of customers served.

SAIDI, or system average interruption duration index, is commonly referred to as customer minutes of
interruption or customer hours, and is designed to provide information as to the average time the
customers are interrupted. It is the sum of the restoration time for each interruption event multiplied by
the number of interrupted customers for each interruption event divided by the total number of customers.

CAIDI, or customer average interruption duration index, is the average time needed to restore service to
the average customer per sustained interruption. It is the sum of customer interruption durations divided
by the total number of customer interruptions.

A reliability index that considers momentary interruptions is MAIFI, or the momentary average
interruption frequency index.

MAIFI is the total number of customer momentary interruptions divided by the total number of customers
served. Momentary interruptions are defined in IEEE Standard 1366 as those that result from each single
operation of an interrupting device such as a recloser.

17
The equations used to calculate these indices are included in Definitions and Terms.

2-4
Unfortunately, it is very difficult to compare these indices from one location to another or from one
utility to another because of differences in how they are calculated. Some utilities exclude outages
due to major events, or normalize their results for adverse weather. For the SAIDI calculation, some
utilities consider an outage over when the substation is returned to service and others consider it over
when the customer is returned to service, a difference in approach that can change the SAIDI by a factor
of two. Some utilities use automatic data collection and analysis while others rely on manual data entry
and spreadsheet analysis.

Depending upon the utility, momentary outages may be classified as a power quality event rather than a
reliability event. Less often used indices include ASIFI, the Average System Interruption Frequency, and
ASIDI, the Average System Interruption Duration. Both of these factors incorporate the magnitude of the
load unserved during an outage. However, less than 10% of utilities track these indices (McDermott and
Dugan 2003). Considering that the data collection and reporting of reliability indices vary over a broad
range, their usefulness in assessing DG effects may be limited.

Another common reliability index is referred to as “nines.” This index is based upon the expected minutes
of power availability during the year. For example, if the expected outage is 50 minutes per year, the
power is 99.99% available or four nines. However, if this index is calculated using the LOLP it won’t
reflect outages in the T&D systems. If the nines are calculated based on the SAIDI, the nines index will
give some indication of the average system availability, but not the availability for any particular
customer.

“Conventional bulk supply systems, from a service interruption perspective, deliver power with
reliability in the range of 99.0% up to 99.9999% (also referred to as “two nines” up to “six nines,”
respectively) and average reliability being about three to four nines, or 99.9% to 99.99%. Rural
electric customers typically experience the least reliable power in the range of two or three nines.
Urban customers served by networks typically have the highest reliability with five or six nines
(Gellings et al. 2004).”

Considering that the data collection and reporting of reliability indices vary over a broad range, their
usefulness in assessing DG effects may be limited.

2.3 DG and Electric System Reliability

DG can be used by electric system planners and operators to improve reliability in both direct and indirect
ways. For example, DG could be used directly to support local voltage levels and avoid an outage that
would have otherwise have occurred due to excessive voltage sag. DG can improve reliability by
increasing the diversity of the power supply options. DG can improve reliability in indirect ways by
reducing stress on grid components to the extent that the individual component reliability is enhanced.
For example, DG could reduce the number of hours that a substation transformer operates at elevated
temperature levels, which would in turn extend the life of that transformer, thus improving the reliability
of that component.

2.3.1 Direct Effects

DG can add to supply diversity and thus lead to improvements in overall system adequacy. DG’s
contribution is often assessed by comparing the DG solution to the traditional solution. In this traditional

2-5
comparison, emphasis is often placed upon the reliability of the DG system itself, and the argument is
sometimes made that the DG capacity cannot be counted because it is not 100% reliable. However, there
are two other factors that must be taken into consideration for this comparison to be useful. First, multiple
DG units provide an element of diversity that has an improved reliability compared to a single unit, and
second, the traditional alternatives are also not 100% reliable.

“Multiple analyses have shown that a distributed network of smaller sources provides a greater
level of adequacy than a centralized system with fewer large sources, reducing both the
magnitude and duration of failures. However, it should also be noted that a single stand-alone
distributed unit without grid backup will provide a significantly lower level of adequacy (Apt and
Morgan 2005).”

Traditionally, as load on a feeder grows, additional supply must be provided to maintain system
reliability. The additional supply is usually provided to the load by adding another feeder or increasing
the capacity of the local substation.

The capacity contribution that can be made by multiple DG units is shown in Figure 2.1 for a simplified
case where all the DG units are the same size and have the same forced outage rate (Hadley et al. 2003).
Figure 2.1 indicates that as the reliability criteria is relaxed from 0.9999 to 0.999, for an unchanged DG
unit forced outage rate of 2%, the number of DG units that can be counted as “available” increases.
Figure 2.1 also shows that as the DG unit forced outage rate increases from 3% to 6% for a fixed
reliability criteria (.99999 in this example), the number of DG units that can be counted as “available”
decreases.

As shown, the diversified system reliability is a function of the reliability of individual units, among other
factors. A study of actual operating experience determines how DG units perform in the field (Energy and
Environmental Analysis, Inc. 2004a). Study results include forced outage rates, scheduled outage factors,
service factors, mean time between forced outages, and mean down times for a variety of DG
technologies and duty cycles. The availability factors collected during this study are summarized in
Figure 2.2. Although the sample size for the DG equipment was smaller than that for the central station
equipment, the availability of the DG is generally comparable to that of central station equipment.

Other statistical techniques, such as Monte Carlo simulations, can be used to assess DG in more
complicated cases. One such study evaluated a case with several DG systems running in parallel within a
central system and calculated the system margin and the average amount of unsupplied loads. The results
showed that DG can enhance the overall capacity of the distribution system and be used as an alternative
to the substation expansion to meet expected demand growth (Hegazy et al. 2003). Several other analysts
have also created models that acknowledge this more complete and complex situation of diversified
sources, each with their own reliability characteristics (Chowdhury et al. 2003). From Apt and Morgan
(2005):

2-6
 Figure 2-1. The Availability of DG Units is A Function of the Number of Units,
Specified Reliability Criteria, and the Equipment Forced Outage Rate 18

18
Created by ORNL based on an equation shown in S.W. Hadley et al, “Quantitative Assessment of Distributed Energy Resource Benefits,”
ORNL/TM-2003/20, Oak Ridge National Laboratory, May 2003

2-7
 Figure 2-2. A Comparison of Availability Factors for DG Equipment
and Central Station Equipment

Source NERC GAR 1997-2001

“In addition to changing the adequacy of the system at the individual facility or distribution
system level, it is possible that widespread use of grid-connected DG could affect the adequacy of
the overall power system. Models comparing centralized with completely distributed system
architectures show a dramatic improvement in adequacy for the distributed systems, particularly
under stress conditions. Zerriffi et al. (2005) compared the results of transmission system failures
on two 2,850 MW peak load systems. The first was a central generation system with
32 generators with capacities from 12 to 400 MW. The second met the load with 500 kW natural
gas fired distributed generators. In reliability models run with failure rates appropriate to current
generation and transmission components, the distributed generation system had roughly 25 times
the reliability of the central generation system. 19 (These results compare a central generation
system with 20% more capacity than load to a DG system with 1.6% more capacity than load
[Zerriffi et al. 2005].)”

“An examination of systems with mixed centralized and distributed generation shows that the
potential reliability benefits depend on a mix of factors, particularly the reliability characteristics
of the centralized generating technologies being replaced versus those being kept, the reliability
characteristics of the distributed technology, and the degree of DG penetration (Zerriffi 2004).”

Brown and Freeman (2001) made a detailed model of four utility feeders, connected with normally open
tie points. In this test system, based upon an actual utility system, SAIDI improvements ranged from 5%

19
The reliability was measured in this study using a Loss of Energy Expectation (MWh/year)

2-8
to 22% with the addition of DG on just one of the four feeders. The reliability of the other feeders was
improved because feeder tie operations that were previously blocked by high load levels became possible
after the DG was added to serve a portion of the load (Brown and Freeman 2001).

Hegazy et al. (2003) modeled a feeder with five DG systems of varying failure and repair rates using a
Monte Carlo technique. Using the unserved load as a reliability measure, the results showed that DG can
enhance the overall capacity of the distribution system and can be used as an alternative to the substation
expansion in case of expected demand growth (Hegazy et al. 2003).

2.3.2 Indirect Effects

DG has the potential to reduce the number of outages caused by overloaded utility equipment. For
example, during peak load situations, higher currents may lead to thermal loss-of-life in transformers and
other equipment, which in turn may lead to service interruptions. These outages are usually caused by
sudden equipment failures that lead to increased loads on the remaining equipment. Such overload
failures account for about 10% to 30% of all outages, depending on the utility and the region. DG can be
used to reduce the number of times per year when distribution equipment is used near nameplate ratings,
and thus could reduce the frequency of equipment failures and subsequent outages (EPRI 2004;
McDermott and Dugan 2003).

2.4 Simulated DG Impacts on Electric System Reliability

Simulation modeling is a valuable tool that can be used to explore the potential impacts of DG on electric
systems. For example, a Virtual Test Bed simulation platform suite was constructed in one detailed study
to examine both power quality and reliability issues associated with DG installations (GE Corporate
Research and Development, 2003). The Virtual Test Bed models the utility’s power delivery system, the
loads, and the DG. In this study, parametric analysis is used to examine the influence of the amount of
DG on a feeder, the location of the DG relative to the loads, (lumped at the beginning, middle, or end of
the feeder, or uniformly distributed along the feeder), inverter-based and rotating DG technologies, DG
local voltage regulation strategies (either operation at a power factor of 1.0 or the DG provides voltage
regulation based on local conditions), two radial feeder lengths, and the presence or absence of capacitor
banks on the feeder.

The analysis of protection and reliability in this study included: transient response and fault behaviors
(capacitor switching and fault behaviors); reclosing; anti-islanding scenarios; and power systems
dynamics and stability. Some of the conclusions from this analysis, which focused on the behavior of DG
units with power electronics, were that:

“A fault analysis found that the fault current contribution of a standard induction motor is usually
much larger than that of current controlled inverter-DG. … the DG, in this example, provides some
damping to high-frequency oscillations. Other findings include:
• Local distribution system dynamics are most affected by DG trips.
• Distributed generation controls do not have a major impact on local dynamics when the
connection to the host utility is maintained.

2-9
 • Anti-islanding schemes (of the type tested here) appear to be effective at destabilizing islands
containing multiple DG units and loads with relatively complex dynamics.
• Voltage and power regulation tend to act contrary to the anti-islanding schemes.
• Widespread penetration of DG units at the load appears to be benign with respect to system
response to bulk system disturbances.
• Anti-islanding schemes (of the type tested here) appear to have little impact on system response
to bulk system disturbances.
• Aggressive tripping of DG units in response to under voltages appears to present a substantial
hazard to the bulk system, and was shown to bring down the entire U.S. western system in one
extreme case (GE Corporate Research and Development, 2003).”

Another analyst used a probabilistic reliability model to compare the options of adding DG or adding
another feeder to a local distribution network. Using the Expected Energy Not Served as the reliability
index, this model is able to optimize both the size and location of alternative DG units. The input for this
model includes values for the annual failure rate of each system component, the repair time, and
switching times. For example, for the network studied, substations were given failure rates of
0.02 occurrences per year, line sections of 0.04 to 0.12 occurrences per year, and DG of 5 occurrences per
year, with repair times of 4 hours for the network resources and 50 hours for the DG resources. For this
network, an additional feeder was able to reduce the Energy Not Served from over 17 MWh per year to
less than 5 MWh per year. Three possible DG configurations were identified that provided that same
level of reliability (Chowdhury et al. 2003). This study is enlightening because it recognizes that DG can
improve system reliability even if it is not 100% reliable itself, that is, that physical assurance
requirements are no more appropriate for DG resources than for any other network resource used to
provide reliable service.

In 2003, Oak Ridge National Laboratory (ORNL) performed a study entitled “Quantitative Assessment of
Distributed Generation Resource Benefits.” In this study, ORNL quantified the benefits of system
reliability in terms of a reduction in the LOLP of DG (Hadley et al. 2003). Reliability of the
Pennsylvania/New Jersey/Maryland Interconnection (PJM) system was simulated across multiple
scenarios of differing generation unit sizes. The study shows that improvement in the LOLP is achieved
when generation expansion needs are met with ten small plants compared to a single large plant of the
same size. For example, in one scenario, generation expansion was designed to be met by a new 100 MW
single unit and in the alternative scenario as ten 10 MW units. Many other paired scenarios of single or
multiple units of generation capacity were also analyzed.

The study results indicate that the LOLP for each pair of scenarios was always lower in the scenario with
the higher number of units. This suggests that a system in which capacity expansion is comprised of
many DG units, rather than one central station power plant, can provide more reliable service to
customers. The study draws the following conclusions:

“Based on the … analysis there is a small but positive value to having capacity added at the unit
size of DG as opposed to typical central station size. The main beneficiary may be society. If
reserve margins are fixed by PJM at a certain percentage of demand, or by the largest single
contingency, then society will benefit by increased reliability at the same amount of capacity.
This can also lead to lower electricity prices since high cost plants will not be called upon as

2-10
 often. If, however, the ISO chooses to lower the required reserve margins, then utilities may
benefit by not having to have as much reserve capacity on hand, through either ownership or the
capacity market (Hadley et al. 2003).”

The study also indicates that DG units can be used to improve system reliability even though each
individual unit is less than 100% reliable. That is because the same rules of redundancy and diversity that
applies to central station plants, or any other component of the power system, also applies to DG.

2.5 Possible Negative Impacts of Distributed Generation on Reliability

In light of the many potential benefits associated with DG, there has been a large body of work devoted to
addressing a number of concerns with regard to the impact of DG on system stability and safety.
Standards agencies, such as the IEEE, have promulgated interconnection standards to protect both the grid
and the DG equipment. Some states have instituted interconnection rules that serve the same purpose.
However, some of the equipment required to meet these standards or other utility-imposed rules can be
costly, especially if used for smaller scale DG projects. Research is on-going to find better solutions and
to optimize the use of DG in the grid.

Some researchers are also examining possible common cause failure modes that could become important
if the use of DG grows. One DG failure mode, the loss of local natural gas supply, is also important for
central generation as more central station power plants use that relatively clean fuel.

2.5.1 Traditional Power System Design, Interconnection and Control Issues

The electric system has been designed to accept power input from large generating stations that are
synchronized with each other and the rest of the grid. That is, the wave form of the electricity produced
by each central generator matches the wave form of the electricity traveling on the grid. Large
transmission lines carry this electricity to substations, where smaller distribution lines carry the electricity
to customers. The vast majority of these distributions systems were designed for one-way flow of
electricity (called radial), from the substation to the customer. This design is reflected in the protection
devices that open and close switches when a tree limb falls on a power line or when lightning strikes a
part of the system. A few urban distribution systems have been designed for two-way flow through the
lines (called network), so that if one line fails another line can be used to deliver electricity to the
customers. Network systems are more complex to operate, but many of their design features may be
useful as DG systems are added in greater numbers to radial systems.

2.5.2 Fault Currents

A fault occurs when electricity travels along unintended pathways, for example along a tree branch that
falls across two wires. Most faults on overhead distribution lines are temporary, such as an arcing current
to the ground that might be initiated by a lightning strike. These temporary faults can be corrected by
simply turning off the current to the affected wire(s) and letting the arc extinguish. Because the system
itself has not been damaged, the current can then be turned on again. Automatic protection systems are
designed to do just that, turn off the current when a fault occurs and then turn it back on after the arc is
gone so that customer service interruptions are as short as possible. If a DG unit is providing power to the
system at a location between the protective switch and the fault, and no appropriate communication or
protection equipment has been installed, it can continue to provide current to the fault so that the fault

2-11
continues. The longer a fault lasts, the more likely it is to cause damage to both the distribution system
and to customer equipment (Dugan and McDermott 2002).

“Distributed units can provide voltage support on distribution feeders. However, this can
complicate service restoration after a fault. If the load becomes dependent upon the distributed
unit for voltage but the DG unit must disconnect due to a fault, the utility may not be able to
maintain voltage at acceptable levels as the fault is cleared, necessitating changes in procedures
and possible delays in restoring power (Kashem and Ledwich 2005).”

Distribution-level instabilities can also be related to DG, as explored by Cardell and Tabors (1998).

“Cardell and Tabors (1998) found that installing generation at the distribution level can decrease
the stability of the system. This is the result of changes in designed power flow direction as well
as in the electrical characteristics of the lines themselves …, which can affect the degree to which
connected generators and loads can interact with one another. Under certain combinations of
distributed generation technologies, the system can become unstable when a disturbance (such as
a line or generator outage) is introduced. …. The authors argue that these results show the need
for new methods to control and stabilize systems that have numerous distributed generators.”

A general description of the issues here is adapted from Apt and Morgan (2005).

Location. DG units located upstream of a system failure point cannot mitigate the impact on
customers located downstream of the failure location. The DG placement on a distribution feeder
can also determine whether there will be stability and power flow problems.

Dispatchability. Intermittent resources, such as photovoltaics or wind, can aid in reducing power
needs, but can have a negligible impact on reliability needs due to their lack of dispatchability.
Similarly, a DG unit that is tied to a thermal load may not be independently dispatchable.

Controllability. Technologies with fast switching times can potentially provide a wider variety
of reliability support. On the other hand, if a technology is installed that has a slower response
time, it may be necessary to modify the operation of other components in the system, potentially
degrading one measure of reliability even as another is increased.

Fuel and Unit Reliability. The reliability characteristics of the distributed resource itself,
including the reliability of the fuel supply, will also determine its contribution to system
reliability (Apt and Morgan 2005).

2.6 Approaches to Valuing DG for Electric System Reliability

The economic benefits of using DG to improve electric system reliability can be estimated by determining
the avoided costs of traditional forms of investment in electric reliability. Under this approach, the net
benefits of installed DG to the utility is the benefit from deferred generation and T&D investments, net
the costs associated with installing, operating, maintaining, administering, coordinating, scheduling, and
dispatching DG units. Not many utilities assess DG in this way when considering expansions and/or
upgrades in T&D equipment. If many did it is likely there would be more instances where the benefits of

2-12
DG would outweigh the costs, although it is important to remember that the financial attractiveness of DG
is highly dependent on local conditions, costs, and resources.

Ownership and type of business model is an important consideration in the valuation of the potential
benefits of DG. For example, when used for reliability purposes, utilities generally require customer-
owned DG to provide performance guarantees and/or physical assurances that the units will be reliable
and available when needed, especially at the time of the peak demand. Such guarantees are normally not
required for investments in utility-owned generation, transmission, and distribution equipment. These
requirements add to the costs and risks of DG ownership.

In certain situations it is possible that there could be a cost justifiable basis for utilities to offer DG
owners capacity payments for units that are able to be dispatched by grid operators during times of system
need. Such payments could support the acquisition of redundant DG units to ensure availability and
address utility interests in performance guarantees.

Energy and Environmental Economics (E3) developed an approach for evaluating the economic potential
for renewable DG applications for municipal utilities (Energy and Environmental Economics, Inc., 2004).
The study used estimates of value-of-service (VOS) and unserved energy to assess the economic benefits
of DG for specific grid locations. The E3 approach is similar to the LOLP methodology used in Hadley et
al. (2003), but the E3 approach included an explicit VOS component, which is intended to quantify the
value of improved reliability.

The E3 methodology comprises two steps. The first step is to compute a weighted VOS based on the
proportion of each customer class served on the feeder or system affected by the DG, and the VOS for
each customer class, on a kWh basis. The VOS estimates are derived from studies that query customers
about how much they would be willing to pay to avoid an outage. The VOS estimates are usually much
higher than standard electricity rates, which can be interpreted to mean that most customers are willing to
pay more for electricity than they currently do. The report cites VOS values in the range of $5 to $30
dollars per kWh in historical survey studies (Energy and Environmental Economics, Inc. and Electrotek
Concepts, Inc., 2005). Figure 2.3 provides a range of the VOS values used in this study; note the
logarithmic scale used to portray the wide range of values from less than $1 to almost $100/kWh
unserved.

The second step calculates the change in unserved energy. In this example, unserved energy is calculated
using an in-depth engineering analysis designed to calculate the number of hours in which a defined
system will exceed the emergency ratings on a particular distribution feeder. This value is calculated for
two contrasting cases. The first is a status quo case and the second reflects the introduction of a number of
small renewable DG facilities.

2-13
 Figure 2-3. Range of Vos Values Used in Municipal Planning Study

Source Energy and Environmental Economics, Inc. and

Electrotek Concepts, Inc., 2005

The E3 study presents results for a number of detailed DG scenarios, including various levels of
installation of photovoltaic systems, combined heat and power additions at critical facilities or substation
sites, and various configurations of peaking DG units. Each case presented positive results associated with
installation of DG as summarized in Table 2.1.

Table 2.1. Value of Reliability Improvement (Year 2004)

Note that the study authors do not explicitly address the comparative costs of competing DG options or
alternative investment options designed to provide identical reliability. This addition to the methodology
is discussed below.

2.7 The Value of Electric Reliability to Customers

One of the reasons why customers value electricity so highly is that the cost of electric system failures can
be significant. One way to value DG-related improvements in the reliability of electric systems is to
determine the value of higher reliability to customers. Value-of-service is one methodology to determine
the value of reliability to customers. Another approach is to assess the outage costs to customers. There
are a number of recent studies of outage costs; however there are no recent studies that use outage costs to
determine the value of DG to improving electric system reliability.

2-14
Recent studies generally indicate that outage costs can be as high as 100 times the average price of
electricity, depending on the type of customer. Some surveys indicate the cost to be between $0.25/kWh
to approximately $7/kWh. For example, Navigant Consulting estimates the reliability benefit from
avoided downtime at $1/kWh (Navigant Consulting 2006). A recent study by Sentech involved the review
of a set of commonly cited power outage cost data ranging from $41,000/h for cellular communications to
$6,500,000/h for brokerage operations. The Sentech study sought “to assess the cost of power outages to
businesses in the commercial and industrial sectors using the best and most current data available, short of
surveying a statistically significant pool of building owners.”

Downtime cost components were categorized as either tangible or intangible as shown in Figure 2.4.The
study used existing literature based on surveys of actual end users that covered outages of 20 minutes, 1
hour and 4 hours in duration. The data from the surveys show that the duration of an outage has a large
effect on estimated downtime costs. Although all sub-sectors estimate similar downtime costs during
short outages, as the duration increases, the costs identified by different commercial sub-sectors begins to
vary more widely.

At the 20 minute duration, almost all commercial sub-sectors have comparable downtime costs.
However, as an outage persists and food spoilage sets in, costs for restaurants (food service) and grocery
stores (food sales) increase faster than for other sectors.

The next two figures from the Sentech study provide another way to illustrate these changes in the
distribution of costs for commercial sub-sectors over the duration of a blackout. One can see that the share
of costs experienced by food service and sales grows until it accounts for the majority of costs after four
hours of outage duration. These figures also illustrate that offices incur large costs during the initial
minutes of a blackout, but subsequent losses are much smaller. Presumably, this is because of the high
cost of data loss and damage to computer equipment that occurs during the initial moments of a blackout;
more data collection and analysis would be needed to confirm this assumption.

Figure 2-4. Costs Considered in Sentech Outage Cost Study

2-15
 Figure.2-5. Commercial Sub sector Power Outage Costs

Figure 2-6. Sentech Study Outage Costs after 20 Minutes and After 4 Hours

Lawrence Berkeley National Laboratory (LBNL) recently conducted a study of the costs of power
outages to the U.S. economy (LaCommare and Eto 2004). The study estimates annual losses to the U.S.
economy from momentary and sustained power outages to be about $79 billion annually, with 72% of
those costs affecting the commercial sector, 26% industrial, and 2% residential. The study reports that
during a reliability monitoring program, several participants contributed business information to help
explain the sources of outage costs:

“…valuable insight on the often-cited statistic that an outage costs silicon-chip fabricators $1
million per event…The determining factor is whether the downtime results in the firm missing a
deadline for delivery of chips that have already been sold. He pointed out that, in 2003, many
firms were running at less than full capacity. Under these conditions…costs of materials lost as a

2-16
 result of the outage were minimal in comparison to the financial penalties that would be
associated with missing shipping delivery dates. The chip fabricator participating in our study
reported that outages of even a few minutes could sometimes lead to 1 to 1.5 days of downtime,
causing the firm to forego $500,000 per day in revenues. …. A related example was provided by
the manufacturer of silicon-chip fabrication equipment…the manufacturer must conduct a
continuous, 1,000-hour factory test, which takes about six weeks. Any interruption during this
period requires restarting the entire test from the beginning.…This firm reported that it had
recently made a $2.5-million investment in equipment to improve electricity reliability that paid
for itself in nine months, which translates into an implied cost per outage of $350,000 per
event…The monetary penalties for missing deliveries are especially high in the financial services
industry. For these firms, “missed” deliveries refer to financial transactions that cannot be
executed…Stringent financial penalties, based in part on the value of foregone or inaccurate
transactions, result from exceeding pre-specified limits…We were told of a financial
clearinghouse in Texas that had experienced a $12- million loss as the result of a 30-minute
outage caused by a lightning strike.” (LaCommare and Eto 2004).

2.8 Major Findings and Conclusions

Electric system reliability is an aggregate measure used by electric system planners and operators to
evaluate the level and quality of service to customers. One of the traditional approaches to achieving a
reliable system involves building sufficient redundancy to ensure continued operations even with the loss
of the largest generator or transmission line. Another involves monitoring grid operations and making
adjustments to changing conditions to prevent momentary problems from cascading into local or regional
outages. DG units can be used by electric system planners and operators to augment these traditional
approaches to electric system reliability. While mostly customer-owned, some existing DG units are made
available to utilities for operations during times of system need through various incentives and pricing
approaches, including demand response. Studies show that in many instances utilities could make greater
use of DG directly, and deploy units to provide peak power, voltage and VAR support, or other ancillary
services to meet electric system reliability needs. However, most utilities do not own or operate DG units
in this way. And, there are no standard models, tools, or techniques for utilities to evaluate DG and
incorporate DG resources into electric system planning and operations.

2-17