10.

1 MANAGING TWO-Q DEMAND ON THE CONSUMER SIDE

A host of measures, often lumped together as "distributed resources," can be applied to re-
shape or reduce the demand of electric consumers, as well as to mitigate the impact
interruptions and poor voltage have on their usage. These include conservation and load
management methods that were traditionally known as Demand Side Management (DSM)
as well as Distributed Generation (DG) and a number of power quality methods (PQM).
Although diverse in nature and purpose, they all share two characteristics. They try to
improve the value the consumer receives from energy in some manner, and they are
implemented at or near the actual site of energy consumption.
This chapter summarizes these methods and the interactions they have with distribution
system planning. Section 10.2 begins with a brief review of demand, energy, and reliability
control and management methods and technologies, including both what were traditionally
called demand side management and newer power quality improvement systems. Section
10.3 looks at conservation voltage reduction, a utility-implemented mechanism for energy
and peak reduction that is complicated to analyze but potentially very useful. Distributed
generation (DG) is then covered in section 10.4. Section 10.5 discusses energy storage
systems. Planning methods to evaluate the appropriateness of these methods are then
covered in section 10.6. This chapter is only a summary of key points. For more detail,
readers should see the References.
Counting "Nega-Watts"
The energy not used and the peak load reduced due to a distributed resource program
involving DSM are often referred to as "nega-watts." One problem with DSM in the 20
th
century, and continuing today with energy management and demand interruption programs,
was accurately counting these nega-watts. Reductions and benefits from many programs in
the 1980s and 1990s were sometimes overestimated by mistake (see Chapter 29).
Monitoring technologies have improved somewhat since the 1980s and 1990s, but
determining how many kilowatts of power usage are avoided, or how much the peak-
demand-that-would-have-occurred is really cut, is still often a challenge with some
distributed resources. Generally, only verification with DG is straightforward (the power
produced by the generator can be metered).
331
10
Distributed Resources
Copyright 2004 by Marcel Dekker, Inc.
332 Chapter 10
Planners and utility managers considering energy and peak load reduction measures for
their systems today need to carefully consider verification and tracking issues from the
beginning of any program. Measures to make certain both that the reductions are real and
that they last for the period the utility has planned (often many years) need to be planned
carefully. Monitoring and tracking systems are a critical part of any distributed resource
program and need to be an integral part of the design of any program and one of the first
aspects that should be designed and approved.
Who Benefits from Peak Load and Energy Reductions?
Who benefits has a good deal to do with who favors energy management and who will
take an initiative to employ it. Measures that reduce total energy usage generally have a
perceived value only to the consumer, not the utility. They reduce the consumer's energy
bill, certainly an advantage in the eyes of most consumers. But conversely they reduce
utility revenues, which is certainly not an advantage from its perspective.
Thus, usually it is energy reduction measures, as opposed to peak load reduction
measures, that are most popular in a purely market-driven energy management
environment. Items like insulation, compact fluorescent lights, and other measures that
more than pay back their investment with energy savings have been steadily but slowly
growing in market acceptance purely on the basis of their benefit to the consumer.
By contrast, reductions in peak demand level are valuable mostly to just the utility. It
can benefit because it does not need to invest in capacity to meet high, perhaps growing,
peak loads nor pay for losses and operating costs driven high by peak demand levels.
Residential and small business energy consumers, who pay only for energy, not demand,
see no savings from peak load reduction measures and have no incentive to implement it.
Large commercial and industrial energy consumers do pay for demand, and therefore are
usually directly interested in demand reduction programs. Regardless, since the majority of
energy usage in many systems is residential, a utility may find that it has to promote peak
reduction programs by using financial incentives and marketing them in a way that mirrors
the "old fashioned" utility-sponsored DSM programs of the 1980s and 1990s. An important
difference though is that modern utilities pretty much have a free hand to price these
incentives and set up their peak load management programs in a market-respond ve way
10.2 ENERGY AND DEMAND MANAGEMENT METHODS
Many distributed resource approaches aim to reduce energy and/or the peak demand. Figure
10.1 illustrates the general concept: a decrease in the area under the load curve indicates
reduced peak load and load at time of peak (same as peak load in this case). Many of the
modern methods employed were referred to as demand side management (DSM) during the
1980s and 1990s. Whether aimed at reducing energy usage or cutting peak demand levels,
these methods are often referred to as energy management methods.
A Rose by Any Other Name
During the 1980s and 1990s many utility regulatory agencies in the United States and other
nations required electric utilities to promote energy management methods, referred to as
demand side management, wherever it was cost effective. Definitions of what "cost
effective" meant and how it was determined varied from state to state, but qualitatively the
concept was the same in nearly every regulated venue. The electric utility was considered to
be a steward of energy usage in its service territory and was expected to accept as part of its
role the responsibility for seeing that its customer used, or at least had the option to use,
energy in an efficient and cost effective manner. It was required to offer programs that
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 333
encouraged energy conservation, peak demand shaving, and waste reduction, when and
where these programs passed certain evaluation tests of efficacy and economy.
For example, an electric utility might be ordered to effect a program to help its
customers insulate their homes to reduce energy usage. The utility company would be
required to provide free advice as well as building and insulation evaluation services, to
make available installation of insulation in homes and small businesses, and to offer
financing of these to their consumers. Many states set targets for utilities based on studies
of the potential reductions possible from such programs and required that the utilities
offer incentives to customers to get them to participate.
Thus, many residential energy consumers found themselves offered a reduction in
electric rates as an added incentive to insulate their homes or participate in other
programs with a similar spirit, programs which would ostensibly save them money even
without that additional incentive. An entire industry grew up around these and similar
utility DSM programs. Electric utilities found themselves in the business of actively
promoting and even implementing measures to reduce the sales of the product they
provided.
Overall, these DSM programs were not a great success. Often the targets set by state
commissions, and claimed to have been attained by utilities, were of doubtful legitimacy, or
the program plans were simplistic to the point of guaranteed failure.
1
Over the last two
decades of the twentieth century, the term "DSM" gained a particularly bad reputation with
utilities, not so much because of a lack of merit in the concepts or technologies employed,
but because of the design, evaluation, and implementation that grew out of the regulatory-
utility framework that usually gave birth to these programs.
But in the author's opinion, the major flaw with the entire DSM industry of the late 20
th
century was that it was not market driven: State commissions drove utilities to implement
programs. Formulae used by commissions might say that the DSM programs were cost
effective, but the homeowners and businessmen clearly did not really think so. If they had,
they would have embraced such programs wholeheartedly.
This situation was often exacerbated by a particularly obnoxious and arrogant attitude
that dominated some DSM programs. Utility DSM departments and regulatory
commissions alike were sometimes staffed by fervent "energy conservation fanatics" who
believed strongly in their cause and also were certain that consumers did not know what
they were doing with respect to energy usage. Basically, consumers then, and today, knew
one thing that dominated a lot of their energy usage decisions: electric energy (and gasoline,
for that matter) is cheap enough to waste. To a majority of consumers, the matter was not
that important, and the hassle associated with managing energy to the standards that the
commissions mandated to utilities was just not justified in their eyes by the savings. But
none of that should blind modern utility planners to the fact that many of those DSM
programs were basically good ideas, measures that improved the quality of the end-use,
reducing its cost, and offered improved value. In a de-regulated power industry where
consumers have a choice of energy options, some of these programs - those that justify
1
See Chapter 29 on the ways that apparently objective planning studies can be either deliberately or
inadvertently corrupted to produce optimistic results. In no venue has the author ever seen so much
misrepresentation and mistake as in the evaluation of utility DSM potential and its accomplishment
during the 1980s. Errors in "nega-watt" estimates, some deliberate but most due to a combination of
optimism and naivete, often were close to 100%. Many of the examples discussed in Chapter 29 are
taken from detailed "audits" of such programs done by the author once a utility or commission began
to see that results did not match claims.
Copyright 2004 by Marcel Dekker, Inc.
334 Chapter 10
Normal load curve.
Reduction due to
energy efficiency
improvement
Figure 10.1 Energy reduction methods reduce the total energy usage, usually by lowering energy
requirements every hour, as shown here for a daily energy curve.
Table 10.1 Basic Types of Energy and Peak Load Reduction Methods
Type of Consumer
Method R C I
Appliance upgrade
Distributed generation
Building insulation
End-use storage
Fuel switching
Interlocking
Lighting efficiency
Load control
Motor/equip upgrade
Renewable energy
Site automation
UPS and PQ devices
CVR
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Reduces
Energy Peak
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Driven by Improves Consumer Improves
Market Industry End Use? Hassle? Reliability?
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
No
Yes
Yes
Marginally
Maybe
No
Maybe
No
A bit
No
Perhaps
Yes
No
No
Can be
Maybe
Maybe
Can be
Can be
Minor
Can be
Minor
Maybe
Yes
No
No
No
Yes
No
Yes
Maybe
No
No
No
No
Yes
Maybe
Yes
No
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 335
themselves to consumers - will see a resurgence as the industry begins to adjust to the
opportunities created by de-regulation. The modern equivalent of DSM programs will
succeed as market-driven services, when packaged and promoted by savvy energy service
companies.
Types of Energy Management Methods
It is possible to list dozens, even hundreds of different energy management methods, but for
the most part, consumer site energy management methods fall into broad categories, as
shown in Table 10.1. The table represents a modern, rather than a traditional (1980-1990)
DSM perspective. It shows the type of consumer for which each category mostly applies,
along with the author's assessment, based on experience, of other salient characteristics.
Size of the "X" in a column indicates the relative strength or scale of that column's
characteristic with respect to that energy management type. Each of these categories is
briefly discussed in the remainder of this section.
Appliance upgrades
As discussed in Chapters 2 and 3, "appliance" is a category of electrical device that covers
all manner of equipment that performs some useful end-use function or purpose for home or
business. "Household appliances" such as refrigerators, microwave ovens, washer-dryers,
and dishwashers fall into this category, as do "appliances" defined by a broader
interpretation such as heat pumps, water heaters, air conditioners, garage door openers,
computers and televisions, elevators, and so forth.
2
For purposes of this discussion, lighting
is not considered an appliance (it is handled separately, below) but in many discussions of
appliances in the broadest sense, it is considered a member of this family.
Devices in many appliance categories vary considerably in their energy efficiency. For
example, available window air conditioners on sale near the author's home, at the time of
this writing, vary from a SEER (seasonable energy efficiency ratio, a measure of efficiency
in annual energy usage) of 8 to more than 11, a difference of more than 35%. Appliance
upgrade means the replacement of appliances with equipment of similar end-use (i.e., AC
for AC unit) but a higher energy efficiency.
Energy and peak reductions. An "energy efficient" appliance is one that uses less
energy in the course of the year, not necessarily one that uses less energy when running
(i.e., has a lower contribution to utility peak load). Energy efficiency standards are based on
annual energy use, not device load. For example, SEER replaced EER (energy efficiency
ratio, a measure of the electricity an AC unit actually used when operating) as the measure
of AC energy efficiency in late 1980s and early 1990s.
Regardless, many appliance upgrade programs will reduce both energy and peak
demand by roughly equal amounts. For example, some energy efficient window AC units
both use less energy during an entire summer season and have a lower load when running.
But very often an "energy efficient" appliance will not provide a reduction in peak
demand just a reduction in energy usage over a period of a day of more. Planners should
never automatically assume they render a reduction in T&D peak loads proportional to their
energy reduction. In fact, some "energy efficient" appliances increase peak demand. On-
demand water heaters have no storage tank to gradually lose heat as most residential waters
2
By contrast, equipment used in manufacturing, process industries, and similar industrial venues are
not appliances because most of it does not cleanly fit the "end-use" definition. Machinery such as
rolling mills, welders, and so forth performs a useful function but generally not one that is directly
associated with a completed end-use.
Copyright 2004 by Marcel Dekker, Inc.
336 Chapter 10
heaters do, and thus are more energy efficient over the course of a year. But they work by
using a large heating element buried directly in the pipes leading to hot-water outlets, which
uses from 15 to 25 kW, five times the load of a storage-tank residential water heater.
Similarly, some types of variable-speed air conditioners and heat pumps actually use more
electricity when running during peak periods. They gain their energy efficiency over a year
because they can run at variable speeds to efficiently vary output (rather than cycling on
and off like normal units, which wastes a small amount of energy) But motor-compressor
units that can be operated at variable speed are not as inherently efficient as constant speed
units, and thus these units are less efficient (i.e., require more power) when running during
hot (peak) periods.
Energy efficiency technologies. There are two ways than an appliance's energy
efficiency can be increased so that an appliance can obtain added efficiency from one or
both manners. First, the electro-mechanical efficiency of the device - the efficiency of its
machinery in turning electricity into an end-use product (e.g., cool air, hot water, clean
dishes) can be increased. This might be done by employing more efficient motors or pumps,
or heat exchangers, or any of dozens of other "tricks" that appliance design engineers can
employ to make the device more efficient in turning electricity into the gross end-use
product produced by the device. Usually, these types of changes are responsible for any
really significant possible improvements that can be made, those representing 20, 30, or
40% improvements in energy, and peak load reductions of noticeable degree.
A second form of improvement can be effected by increasing the device's efficiency
with respect to its environment or the actual duty cycle it will perform. For example, a
dishwasher can be fitted with insulation and a vent to the outside so that heat does not
escape into the kitchen around it to raise the temperature of room air that then has to be
cooled. A computer can be built to "shut down" all but memory-maintenance activities in a
computer when not being used, reducing its energy consumption by 80% when idle. A heat
pump can be designed to start its compressors and heat them, several minutes before
starting its blowers, marginally improving its performance.
Usually, efficient appliances employ measures from both categories to gain their
improved performance. However, appliance retrofit programs, which improve the
efficiency of existing appliances in a home or business without replacing them, usually
include onl> the later category of improvement. The best example is "water heater
wrapping," which involves putting a blanket of insulation around the water heater tank and
hot water pipes in a home or business. This cuts heat losses of a water heater by a small but
noticeable amount (about 5-10%). It is economically justifiable because it is easy to do and
inexpensive. Another example is "set back" thermostats, which save energy by scheduling
home heating and air conditioners in a more efficient manner.
Gradual trend of improvement. A very real appliance upgrade program in the residential
sector has been implemented in the U.S. by action of the federal government. The Energy
Efficiency Act (1993) set gradually increasing targets for efficiency of household
appliances such as air conditioners and refrigerators. Over time, the minimum allowable
efficiency of units remitted for sale in the U.S. is mandated to increase. Thus, as older
appliances wear out, they will be replaced by more energy efficient devices.
Technological advance provides something of the same trend, of a lower rate, through
improvements made in appliances by manufacturers. For example, the efficiency of
household kitchen refrigerators more than doubled from 1940 to 1990 (i.e., energy required
for any amount of cooling fell by half). Since then government regulations have stipulated a
slightly higher rate of improvement than this.
Utility perspective and programs for appliance upgrades. Generally, a utility can and
should count on a certain amount of steady, gradual energy efficiency improvements due to
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 337
appliance upgrades as a factor in its load forecasts. It can promote and encourage efficiency
where it makes sense through programs of public awareness, financing (utilities in the
author's home state offer financing for major appliances that meet certain standards), and
actual incentives (for example a $100 rebate when one turns in an old refrigerator and
submits proof of purchase of a new, efficient one, where the need is great). But generally,
appliance upgrades happen as programs external to the utility, driven by market forces
(owners want efficient appliances because they save them money), technological factors
(manufacturers improve efficiency as a competitive advantage), and government mandate.
Distributed (backup) generation
Backup generation, as opposed to distributed generation, can be viewed as a distinct
category of customer-site systems. Contrary to the impression created by many modem DG
advocates, DG has been widely used in the industry for decades, but almost exclusively as
backup generation - aimed at providing a type of improved reliability. Energy consumers,
mostly commercial and industrial users with high costs of sustained interruptions, have
created a very large and healthy market for backup generation systems. In the past ten
years, a number of utilities have started to offer backup generation as a service. Distributed
generation, in all forms, will be discussed in section 10.3.
Building shell improvements
Improvements in building shell insulation, sealing of doors, windows, and vents, and air
handling systems (duct work, etc.) can render a large improvement in the heating and air
conditioning energy consumption of a home or office building. Energy requirement
reductions of up to 66% are possible in most buildings in the United States, although
economically justifiable reduction levels are generally only about a third of that.
Building shell energy efficient technology. Adding more and better insulation to older
homes, and insulating ductwork exposed to ambient temperatures in attics and crawl spaces,
is perhaps the most popular form of energy management in this category. Other popular
measures are intense weather-sealing of doors and windows, replacement of windows with
double- or triple-paned glass and casements that are heavily insulated and sealed, glazed
glass or reflecting film put on windows, and awnings.
Like appliance upgrades, a certain amount of this type of improvement occurs as a
"natural trend," driven by a combination of market and technological forces. Older homes
are often replaced by newer homes built to much higher energy efficiency standards. When
combined with upgrades of the cooling and heating units themselves, reduction in energy
usage of up to 50% can occur between an old structure and a modern replacement.
3
Improved energy, peak, and user satisfaction. Building shell improvements are one
category of energy efficiency improvement that almost always improve consumer
satisfaction with end-use performance. Well-insulated homes are famously "less drafty,"
having fewer unwanted warm and cold corners, and far less temperature variation during
the day. This improvement is often a selling point stressed in any program pushing these
measures.
3
As an example, in 1990 the author left a charming but old 1,500 square foot home near Pittsburgh for
a new 3,300 square foot home in Gary, NC. Built in 1939, the Pittsburgh residence had been
upgraded in 1964 and 1987. The 1990 home, of "high energy efficiency standards," used noticeably
less energy for cooling each summer, despite having more than double the floor space and having to
deal with 20% greater cooling degree days in North Carolina as opposed to Pittsburgh.
Copyright 2004 by Marcel Dekker, Inc.
338 Chapter 10
Building shell energy programs almost always makes an improvement in both annual
energy usage and peak demand. The impact on energy usage is obvious and well described
above, and is usually the motivation for their implementation, at least by consumers.
Usually, they reduce coincident peak demand but not non-coincident (individual consumer)
peak demand. Since the utility really cares only about the former, most of these programs
are at least somewhat a type of peak reduction approach.
The peak reductions come about because the building shell improvements reduce the
burden on existing heaters, heat pumps, and air conditioners. These units cycle on and off to
operate. For example, during peak summer temperature periods, the central air conditioner
in a home might operate with a 90% duty cycle (cycling on and off for a total 54 minutes
operation out of the peak hour). Insulation and other building shell improvements that
reduce energy losses in the home by 10% will reduce the need for cooling during this peak
hour. Very likely the duty cycle of this AC unit will fall to about 81%, so that the
household's contribution to the utility's coincident peak hourly demand would also drop by
about 10%.
4
Customer and utility implementation issues. Building shell energy improvements are
not simple to implement nor without a noticeable, often significant, cost. They may also
involve a considerable amount of inconvenience (removing and rebuilding the outer walls
of a home). In addition, effective economic improvement of a residence's or office
building's shell requires considerable attention to detail and some basic engineering
analysis and design.
Some electric utilities provide services to consumers to promote these types of
upgrades; a few finance them, often a holdover program from the regulatory-driven
programs of the 1990s. Energy service companies (ESCos, the energy providers in a
completely competitive, de-regulated electric market) often make energy efficiency
improvements of this type a part of marketing programs aimed at attracting consumers with
a "better service, lower overall cost" value package.
End-use storage
End-use storage includes a number of appliance designs that de-couple, or separate, the
time of a consumer's use of the appliance's product from the time of the appliance's use of
energy to produce the product. Thus, end-uses that are heavily skewed toward peak times
can be satisfied while the usage is pushed to off-peak.
A widely, nearly universally applied example is a storage water heater - a "water
heater" as far as most homeowners are concerned. Residential water heaters, either
electric or gas, usually have a tank that holds from 40 to 80 gallons of hot water, ready
for use. This means that their heating elements do not have to, and sometimes do not,
operate during the time the homeowner draws water from them.
Storage water heaters are widely used not for energy storage reasons, but because the
storage of hot water permits instant use of a good deal of hot water, without having to
wait for it to be heated. However, they are often retro-fitted with controls that shut down
their elements during peak periods, as a peak reduction measure, in company with
perhaps larger tanks so they have more product stored for use.
End-use storage technologies. Sound and effective storage space heaters and coolers
using proven technologies are widely available, at least in some parts of the world. Storage
4
The reduction would not be exactly 10% for a variety of secondary effects and ifs, ands, or buts. For
more detail in the behavior of such appliances, see Spatial Electric Load Forecasting, Second Edition,
H. Lee Willis, Marcel Dekker, 2002, Chapter 3.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 339
home heaters are widely used in Europe. They consist of an insulated box inside of which is
placed an array of ceramic "bricks" of very high energy absorption density which surround
resistive heating elements. To reduce peak demand, the elements are run at night, heating
the bricks to high temperatures. During the day, they are shut down and a small fan blows
air over and through the ceramic brick lattice to produce heated air to keep the home warm.
These units are relatively simple and fairly reliable and durable.
Similarly, "cool storage" units store "cool" in the form of chilled brine (salt water can
be cooled to below freezing while still remaining a liquid) which is then run through
radiators and baseboard pipes in a building to cool it during the heat of a summer day.
The chiller is run only at night, and the only air conditioning load during the day is the
small pump, and perhaps fans, to circulate air. Cool storage units utilize only proven
technology, but generally they are fairly large, not home sized, and thus applicable
mostly to commercial applications.
End-use storage units produce peak load reductions, but usually increase energy use.
Both room-size and central thermal storage heating units are widely used in Europe, with
residential and small-business market penetration in some regions approaching 50% of all
homes using electric heating. Cool storage units have been used in some public schools in
Texas. Both are very effective at reducing heating and cooling demand of the affected
consumers during utility peak times, by as much as 90%.
But these technologies increase both the energy usage and actual demand of the unit,
compared to non-storage units. There are several reasons. First, both store heat (or lack of
it) and encounter slight losses as that heat escapes during the day. More important,
though, thermal storage heaters use resistive heating elements, roughly only half as
efficiency of a heat pump.
5
Similarly cool storage units cannot utilize "every cooling trick
in the book" as can non-storage AC units. As a result, they use more energy during a 24-
hour period, and when they are "recharging" (at night, off-peak) they may use more
energy than a non-storage unit would when running at peak during the day. (A thermal
storage unit equivalent to an 8 kW heat pump might use 10-15 kW of heating elements.)
End-use storage generally appeals only to the utility. It reduces utility peak demand in a
very noticeable way. But it increases total energy used by the consumer and thus has no
market appeal unless supported by significant rebates or incentives. With very rare
exceptions end-use storage is only successful when implemented by the utility. In Europe,
the widespread use of thermal storage home heating systems in parts of Germany is due to
the utilities there having both encouraged their use and rewarded it through particularly
attractive off-peak rates to those users.
Fuel switching
One option that will reduce electric energy usage and peak demand on nearly any T&D
power system is fuel switching - moving major energy uses like space and water heating,
cooking, and even cooling from an electric base to natural gas or oil. Few electric utilities
consider or encourage this because it cuts deeply into revenues. However, many energy
consumers consider and effect these measures, particularly when they face high electric
rates.
5
A heat pump cannot be used to drive a compact thermal storage unit, because it does not generate
intense heat. Thermal storage units store a lot of heat in a relatively small compact amount of matter
(several hundred pounds is all) by using resistive elements to raise its temperature by several hundred
degrees Centigrade. By comparison, a heat pump can only generate a differential of perhaps 50
degrees. The units would be too bulky to fit in apartments and small homes, or to retrofit to even large
homes.
Copyright 2004 by Marcel Dekker, Inc.
340 Chapter 10
During the 1980s and 1990s fuel-switching was an option required for utility study by
some state regulatory commissions. It was felt that fuel-switching had the potential to
permit utilities to avoid high generation and delivery costs under some circumstances.
There was also an opinion among some energy conservation advocates that gas and oil
usage for water and space heating was inherently more efficient and should be pursued
for that reason.
Technology. For applications where gas and oil are viable alternatives to electric power,
proven gas appliances are widely available to consumers. Gas heaters, washer-dryers,
ranges, and ovens are all "stable" technology and have been available for decades. Gas air
conditioners work well, and have been available since the 1970s, but are a constantly
evolving and improving technology, with newer units providing more reliability and
efficiency than those of even a decade before.
Electric energy and peak load reduction. Fuel substitution is an effective way of
reducing the use of electric energy and cutting peak load. Shifting of some energy intensive
end-uses such as water and space heating and air conditioning to gas can make a big
difference in demand and energy needs in an area, if applied to a significant portion of the
consumer base there.
In some cases, the use of gas rather than electricity results in a net reduction in total
energy usage, too. For example, in the case of most residential water heaters, it takes less
energy to heat water in a home or business by using gas there than to make electricity with
it at a generating plant and move that electric power to the home for use there to heat water.
In other cases the situation is not so clear, as with heat pumps in areas where gas delivery
and distribution costs, and losses, are relatively high. However, overall, using gas at
distributed sites for some of these "heavy" applications is more efficient from an overall
energy standpoint than using electricity. For this reason, "fuel switching" enjoyed some
popularity with regulators during the 1980s and early 1990s, and was a required option for
study for some utilities for that reason.
However, the situation is not quite that simple. Generally, emissions and environmental
impact are lessened if electric power is burned under the superior control and emissions
management of a large central generating plant, utilizing new technology, than at many
small, perhaps not as well maintained sites. Then, too, there are the distribution systems to
be considered, with a host of issues that often dominate decision-making in a region. For
this reason, many of the "fuel switching" studies done in the 1980s and 1990s were
simplistic and of little lasting value. Probably the best societal strategy to address the "gas
versus electric" question is the one the U.S. has taken: let the market decide.
Fuel availability and distribution systems. Availability of distribution systems plays an
important part in both consumer and utility perspectives on fuel substitution. Electric
distribution networks are nearly ubiquitous, whereas many areas of the nation do not have,
and probably will never have, gas delivery systems. The marginal costs of construction and
system constraints in different areas vary so much within each category (gas and electric)
that it is impossible to generalize which is less expensive to distribute.
Consumer perspective. Surveys and market results indicate that consumers make
decisions among electric and gas options almost entirely on the basis of cost. But overall,
electricity has several advantages with respect to consumer appeal. First, electric
distribution systems reach more consumers: there is a significant portion of households in
the U.S. and elsewhere in the world that have electric but not gas distribution available.
Second, most households could get by without natural gas, but not electricity: gas-powered
televisions, stereos, computers, and microwave ovens don't exist, and gas lighting is
suitable only for ornamental applications. Finally, a segment of consumers perceives that
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 341
electricity is safer than gas.
6
On the other hand, for a majority of consumers where gas is
available it is the less costly energy source for "energy intensive" household applications
like water and space heating.
Utility perspective. Gas versus electric choices for a utility are either very simple or
incredibly complicated. They are simple for an electric utility that views its business as
competing in a free marketplace for energy business of consumers. Gas is the competitor,
and programs to shift consumers to gas seldom make sense (at least if the utility has been
well managed). But utilities that distribute and sell both energy sources face a much more
complicated business problem, //"the utility has complete freedom to decide what and how
it invests in electric and gas systems, it can decide where to extend or limit gas and
electricity distribution, and when and where to promote gas or electric consumption, and
how to price one energy source versus the other. It can optimize any of a number of
business goals: maximized profit, overall (least) cost, minimized business risk, or maximum
consumer choice, etc. Generally, regulatory rules severely limit the utility's choices here
and clearly identify what and how it will optimize the use of both systems. Generally, fuel
switching is not an approach considered by utilities for tactical reasons
Distributed generation. For the standpoint of this discussion, distributed power
generators can be viewed as devices that effect a type of fuel switching. They will turn gas
into electricity on a site-specific basis, and permit gas to power televisions, lighting, and
other purely electric applications. DG will be considered later in this chapter.
Interlocking
Interlocking is a very simple peak load control mechanism that has an impact on a utility
system's load that is incredibly difficult to analyze and predict with accuracy. It involves
wiring the thermostats or control mechanisms of two or more appliances at a location in
series, so that only one of them can operate at any one rime. The most common
application of interlocking is the connection of a home's space conditioner (e.g., heat
pump) and water heater in a master-slave arrangement, so that the electric water heater
cannot operate whenever the air conditioner/heater is operating.
Normally, interlocking is intended to reduce only peak demand while having no impact
on energy usage. In practice, if it works well as a peak load reduction mechanism, it will
render a very slight, usually insignificant reduction in energy usage, too.
Technology. Interlocking involves nothing more than re-wiring of the low-voltage (12-
24 volts DC) control wiring in a home or business. This requires a few dozen feet of 24-volt
rated control wire and perhaps replacement of single-pole control relays in the appliances
with new double-pole relays. All of this is simple, commodity equipment of very low cost
and wide availability. Re-wiring is straightforward. For this reason, it is among the least
expensive and simplest peak load reduction mechanisms available.
Simple to implement but complicated to model. Interlocking's impact on a utility's load
is complicated because its operation interacts with appliance diversity and operations
schedules. Accurately modeling and predicting its effects requires detailed assessment of
appliance-level load behavior and non-coincident versus coincident load curve shapes, as is
discussed in Chapter 3, Section 3.3.
6
Several electric utilities discovered through market surveys and focus groups that they could achieve
noticeable market share in electric heating against gas by stressing its "safe and warm" aspect. These
marketing programs did not imply gas was unsafe (and neither is the author) but rather simply
stressed that electricity was safe, essentially playing on concerns some consumers may have already
had.
Copyright 2004 by Marcel Dekker, Inc.
342 Chapter 10
NORMAL.
Ti me >
INTERLOCKED
Figure 10.2 At the top, an air conditioner and a water heater each act according to the control of their
thermostats. Both use all the energy they want and occasionally their usage overlaps. At the bottom,
when interlocked they alternate usage and the "slave" appliance (the water heater), obtains less energy
because it is prevented from operating as much as it will if not interlocked. This results in a load
reduction during peak times.
How it works. Suppose that a particular household has a 5 kW electric water heater
interlocked to its 10 kW central electric air conditioner. Further, assume that during the
peak summer temperatures, the AC unit will operate at a 95% duty cycle, while the
household water heater might want to operate at only a 15% duty cycle. Both of these duty
cycles are fairly typical for peak summer hours.
7
Interlocking the two appliances prevents
the water heater from operating at more than a 5% duty cycle, that being all that is left to it
by the air conditioner. Thus, a reduction in 10% of the water heater's duty cycle is affected,
ar>d
f
he household's energy usage during the peak hour is reduced by
Reduction in demand at peak = 10% x 5 kW = .5 kW (10.1)
Thus, the household's energy use during this peak hour is reduced by a noticeable
amount. Water available for use in its tank might drop in temperature somewhat, however,
storage water heaters generally can go for an hour or more without energy at all and still
satisfy most household needs, and the one in this example is getting a bit of power during
peak, anyway. Energy use is not greatly affected: after the peak period is over, the water
heater will "be hungry" and operate at a higher than normal duty cycle until it returns the
water stored in its tank to full operating temperature. Figure 10.2 illustrates this.
By contrast, during an off-peak summer hour, the AC might operate at an 80% duty
cycle. In that case there is more than enough time for both appliances to obtain all the
energy they need. Thus, interlocking limits usage only during periods of intense (peak)
energy usage, and makes no impact otherwise.
7
Water heaters seldom operate at really high duty cycles. Typically, during periods of peak water heater
usage, the average duty cycle for water heaters over a utility system is only about 35%.
Copyright 2004 by Marcel Dekker, Inc.
8
Distributed Resources 343
Reduction in weather sensitivity. A recognized capability of interlocking, which was
never completely included in utility evaluations of it carried out in the 1980s and 1990s,
was its reduction in weather sensitivity. Really extreme high temperatures once-in-ten-
year extreme summer weather - can cause air conditioners to reach 100% duty cycle,
leading to unexpectedly high peak loads that can overstress electrical equipment.
8
With
interlocking, the demand weather sensitivity of a household like that used in the example is
cut almost in half. As temperature rises, the AC unit is called upon to remove more heat
from the home and its duty cycle increases thereby reducing by a like percentage the duty
cycle of the home's interlocked water heater. Thus, the increase in demand seen by the
system, as temperature rises above a threshold, is reduced to:
(AC connected load - Water Heater connected load) (10.2)
(AC connected load)
of what it was without interlocking. The threshold is the temperature at which the duty
cycles of the two interlocked devices sum to 100%.
Reduction in non-coincident demand. Interlocking also reduces the peak non-coincident
demand of a household, not only during peak, but all 8760 hours of the year. This makes no
impact on household use of energy or convenience of the appliances (homeowners would
rarely if ever notice any impact), but it does reduce losses and stress on the utility's service
transformers and secondary circuits somewhat.
When interlocked, a water heater and an air conditioner/heater can never operate at the
same time. As is discussed in Chapter 3, if left to their own, random operation, both devices
will cycle on and off as their thermostats dictate, and occasionally their usage will overlap,
creating "needle peaks" of the type illustrated in Figure 3.2. Such sharp peak loads usually
last only a few minutes, until one or the other of the two shuts off. They occur more
frequently on peak days (the appliances operate more often so they have more chance to
overlap. But they occur "randomly" during the entire year.
Interlocking effectively synchronizes the electric usage of the two devices so that they
operate one after the other. Non-coincident demand of the household is cut permanently.
But coincident load is not affected except during peak conditions. During off-peak periods
the interlocking makes no impact. At those times, even rho-igh the non-coincident load
curves of households with interlocked appliances have been altered by "rescheduling" of
the "on" periods for one of their major appliances, the household non-coincident curves will
"add up" (see Figure 3.4) to the same aggregate coincident load curve as they did when not
interlocked.
The net effect of this 8760-hour-a-year reduction in non-coincident load are reductions
in the peak demands seen by the utility's secondary circuits and service transformers and
the losses that occur in the consumer's service drops. Total losses reduction at this level
may approach 15%. Voltage regulation to the consumer is also improved, although by only
a small margin. The author is aware of no situation where interlocking has been applied
because of its impact at this level, but the losses savings and improved voltage regulation
are a benefit that can be counted in assessments of its benefit.
Poor anticipation of the impact of extreme weather caused several widespread outages for utilities in
the U.S. when a one-in-ten summer hit the midwestern and northeastern United States in 1999. Many
utility T&D systems had been designed handle typical summer weather, but could not handle extreme
weather. Large (25,000+ customers), long (8+ hours), and significant (downtown, City Hall)
interruptions resulted. For more, see Chapters 5 and 6 in Willis, 2002.
Copyright 2004 by Marcel Dekker, Inc.
344 Chapter 10
Interlocking's history. Interlocking of AC and water heater units was tried by a number
of utilities in the southern and southwestern United States, and for electric space heating
and water heaters by a small number of rural cooperatives in the northern states, in the
1980s and 1990s. Most of these utilities had relatively high rates of load growth and thus
had an incentive to cut peak load.
Interlocking developed a checkered reputation, working well in some places and not so
well, or not at all, in others. It worked very well in one or two summer peaking systems and
well enough to be considered a success in several winter-peaking utilities in the Midwest.
However, it failed to provide any peak load reduction at one southwestern utility because
planners had estimated duty cycles for AC and water heaters of 95% and 15% respectively,
as in the example earlier. In actuality they averaged 90% and 8%.
9
Since their sum was less
than 100%, there was no reduction in peak demand.
10
These results were widely
disseminated within the industry, in part by manufacturers of load controllers who saw
interlocking as a competitor to their products and gave it a poor reputation.
Interlocking's potential. Interlocking works well in cases where the utility can verify
that the sum of the interlocked devices' duty cycles exceeds 100% at time of coincident
peak. It is robust, simple to implement, and effective in these cases. It has the added benefit
of reducing stress and losses on the secondary voltage system. It is simple to explain to
consumers and to install, makes only a minimal impact on consumer usage, and is quite
durable and long-lasting.
Only consumers with space conditioning and water heating, or similar loads of that type,
are candidates for interlocking. In some utility systems this is a small percentage of the
utility's customer base. But in other systems, particularly in rural and suburban areas, a
significant portion of the residential customer base falls into this category.
Consumers have no incentive to interlock their appliances - it provides no benefit to
them. Thus, a utility will have to offer an incentive to gain acceptable and market
penetration. In some cases, the most expedient and simple way to gain participation might
be for the utility to offer a free replacement water heater, only available as an interlocked
device, to anyone who needs a replacement or wishes to add one (new construction,
conversions). The utility still gets the energy sales, but sees no or very little peak load
increase.
Lighting efficiency
Approximately one-fifth to one-third of the electrical energy sold by the average electric
utility in the United States is used by its customers for lighting applications. Nearly 70
percent of this lighting energy is consumed in the commercial, public, and industrial
sectors. Among the numerous types of lighting devices (lamps) available, efficiency varies
greatly. However, efficiency is only one of a number of important factors to consumers,
9
Residential AC units are usually sized to operate at 90% duty factor during the hottest normally
expected weather in a summer. Water heaters typically have a duty cycle during the hour of peak
water heater demand of less than 30% and off-peak duty cycles as low as 10%. But a majority of units
in this utility system were located in un-air-conditioned garage spaces, where ambient temperatures
around them were typically 125F during the hottest part of the day. As a result, most water heaters
operated but a few minutes an hour during summer afternoons.
10
In actuality there was a very small impact, barely measurable, because not all homes were average. In
a few, less than 15%, duty cycles did exceed 100% during peak loads, and a slight reduction was
effected. However, the action of this program was far below expectation, and the utility abandoned it
as "ineffective." In this case, it was.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 345
including initial cost, color of light, lifetime, and operating characteristics (some lamps do
not provide light the instant they are turned on, but take several minutes to work up to full
output). The efficiencies of the principal types of electric lamps vary considerably.
Efficiency is measured by lumens per watt, where a lumen is the measure of light flux (on a
per energy basis, one lumen is equal to 0.00413 watt), and varies among types.
Energy and peak reduction. Lighting efficiency programs reduce both energy usage and
peak load.
Lighting efficiency technology. Lighting efficiency improvement programs focused on
interior applications, most often on replacing incandescent lighting with fluorescent tubes or
compact fluorescent bulbs, which use only one-half to one-third the energy for equivalent
amounts of light. A 23-watt compact fluorescent produces equivalent illumination to a 75-
watt incandescent bulb. In summer-peaking utilities the bulb produces an additional energy
savings. A majority of the energy used by an incandescent bulb is turned directly into heat
(the bulb radiates more heat than visible light energy). That heat has to be removed from
inside the home by an air conditioner. A compact fluorescent bulb thus reduces air
conditioner operation and load, too. From the consumer standpoint, an additional appeal is
that compact fluorescent devices have a much longer lifetime than incandescent (lasting an
average of six to eight years in residential service).
Often interior light fixtures, particularly those with reflectors, are replaced. Better
(shinier metal and better shaped to disperse light where needed) reflectors can improve the
efficiency of an existing fluorescent light fixture by up to 15%.
Exterior lighting efficiency programs, for. exterior building illumination, night-time
security, and parking lots and storage yards, often see changes in all aspects of the lighting
system as a result of a performance/efficiency upgrade. They may include changes in the
number of lighting fixtures, their type and orientation, and their scheduling.
Residential lighting upgrade programs usually concentrate on replacing incandescent
bulbs with compact fluorescent "bulbs." The degree of improvement in lighting energy
efficiency seen in the residential sector is usually greater than in commercial applications.
Up to a five-fold increase is possible. In the 1980s, utilities often had to provide
considerable incentives to consumers to try compact fluorescent lighting. Wider availability
of these devices has made these largely unnecessary.
Although the economics work out very well (compact fluorescent "bulbs" more than
pay for themselves with reduced energy and replacement costs), consumer acceptance has
been slow. Many do not like the starting latency of these devices, a roughly one-second
delay between the time the light switch is activated and when the "bulb" begins producing
light. There is not much that the bulb's designers can do about that unless the device's
energy efficiency is substantially reduced. Compact fluorescent "bulbs" have three other
disadvantages with regard to incandescent replacement. Their initial cost is about six times
that of incandescent bulbs of equivalent output. Second, compact as they may be, many
cannot fit into older overhead light fixtures or table and desk lamps. (Design progress is
gradually reducing the size of the smallest compact fluorescent bulbs available, improving
this liability.) Finally, fluorescent produces a white light, not yellow, and this represents a
change for customers (the white light is sometimes perceived as harsher).
Utilities promoted compact fluorescent light bulbs to residential customers during the
1990s, as DSM. Most such programs ended a number of years ago. Usage continues to
increase, driven by the consumer savings the devices produce and due to wider
availability (stores like Home Depot and Ace Hardware carry them, and they can be
found in the light bulb section of some food markets). In addition, newer designs are
more compact and include some models with a glass shell so the unit looks very much
like a "standard" bulb.
Copyright 2004 by Marcel Dekker, Inc.
346 Chapter 10
Commercial and Industrial lighting efficiency often includes replacement of
incandescent with compact fluorescent, upgrades of interior fluorescent light trays with
improved reflectors, and revised exterior illumination.
Finally, many utilities worked with architects and home builders to both design new
structures so they used natural illumination well and have efficient lighting throughout (e.g.,
fixtures that will accept compact fluorescent bulbs).
Modern application of lighting efficiency programs. All lighting programs reduce
energy sales and thus erode revenues. Thus they do not have a strong appeal to utilities. In
particular, programs that reduce night-time illumination load (building security, parking lot
and storage yard lighting) have very little appeal. They do not bring a peak load reduction,
and cut off-peak sales.
But they do appeal to consumers, because they lower cost and potentially provide better
illumination. In the residential sector, utilities can depend on trends in increasing use of
compact fluorescent lamps to continue. Most utilities have not and probably will not
aggressively push residential programs, partly because much residential usage is off-peak
and thus does not return a big peak load reduction.
Commercial and industrial programs are a different matter. A lot of commercial
lighting, particularly interior lighting, is coincident with peak demand. The savings in cost
due to a good lighting upgrade can be sufficient to attract businessmen without incentives
from the utility. In a de-regulated industry, someone - usually an energy service company
can very likely make a profitable business out of offering lighting upgrade survey, design,
and implementation services.
Load control
Load control can be implemented in many ways. But in all cases, it means "utility" control
of the operating schedule of appliances like air conditioners and water heaters through some
remote means. "Utility" is given in quotations here, because in a de-regulated industry it
may very well be an energy service company that implements load control, not the delivery
utility.
Load control of electric water heaters was a classic DSM program of the 1980s and
early 1990s. A utility would offer a customer an incentive, perhaps $5 per month rebate,
in return for which it would connect a remote-control relay to the electric water heater at
their site. During peak periods, it would open the relay, shutting off the water heater, for a
period of one, two, and, in some utility cases, four hours.
Many other household and business loads were, and are, candidates for load control,
including air conditioners and heat pumps and central electric heaters. Here, the units are
not shut down for anything more than a brief period, but this is done frequently. For
example, an air conditioner might be shut off for 7.5 minutes during each half hour. This
effectively limits its duty cycle to 75%. If during peak times the unit would run at a 90%
duty cycle, this effects a reduction in coincident peak load equal to 15% of its connected
load.
Peak load reductions and rebound. Load control is a peak load reduction measure,
which may, during peak days, also reduce energy usage slightly, However, theory and
experience showed that it had little effect on energy usage, due to an effect called rebound.
As an example, the reader can consider the operation of a water heater that has been denied
electric power for two hours after that control is terminated. It will immediately switch on
and draws power until it raises the temperature of water stored in its tank back up to its
thermostat's setting, operating at a higher duty cycle than it would during the post-peak
period, if it had not be "controlled."
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 347
Technology and past applications. Many load control pilot programs were implemented
during the "DSM era" and a few grew into stable mainstream programs. Florida Power
Corporation (today, Progress Energy, Florida) at one time had several hundred thousand
sites under control. Similarly Detroit Edison had enough water heaters under control that it
could shave peak load by the equivalent of an entire generating unit. Equipment and control
strategies varied greatly. A myriad of different control and communications technologies
were developed in the 1980s and 1990s to communicate with load controllers, using several
types of power line carrier, phone, and numerous different radio communication methods.
Some systems were one-way communication only: the utility sent signals out to the
controllers, but got no verification back that they had received the control signal or operated
correctly. Other systems, generally more costly, included two-way communication which
would verify operation.
Utility and consumer appeal Load control has no appeal to energy consumers, at least
those not on real-time or demand-sensitive rates. It offers them no advantages. In fact, it can
offer a disadvantage: slightly cool "hot" water and slightly warm "cool" air during peak
periods.
But to a utility, load control can effect significant reductions in peak load in a manner
that permits it to "program" the timing, the amount, and the location of the reductions. This
has great appeal. However, load control is expensive and complicated. First, it requires
incentives. Second, it requires control modules at every controlled site, a communications
system to control there, and a central load control dispatch center to coordinate it all. Third,
this equipment will need service, increasing maintenance costs.
Motor and equipment efficiency upgrades
This category of energy efficiency program can best be described as "the industrial
equivalent of appliance upgrades." It consists of a variety of programs aimed at
improving the efficiency of industrial electric energy use. Among these are the direct
analog of residential class appliance upgrades - motor efficiency upgrade programs - in
which electric motors are replaced with high-efficiency units that do the same job using
less electricity. But in addition there are many other measures that fall into this category
of improving energy efficiency of industrial processes, such as programs to replace
relatively inefficient boilers, pumps, rolling machines, stamping machines, etc., with
equipment that is, ultimately, much less demanding of energy.
A considerable portion of the electric energy sold by most utilities is converted into
mechanical rotation using electric motors of either induction or synchronous type. Many of
these motors are included in air conditioners, heat pumps, chillers, rollers, and other heavy
industrial equipment. But beyond those uses, motors are needed for a diverse range of
applications, including conveyer belt drives, water well pumps, dryer fans, garage door
openers, swimming pool pumps, and hundreds of other applications, in residential,
commercial, and particularly industrial classes. Motor sizes, applications, and efficiencies
vary tremendously in these classes, but in nearly every commercial and industrial
facility, there are numerous motors, of all sizes, and thus a potential for energy savings.
Motor efficiency technologies for industrial applications. As was the case in residential
appliance upgrades, there are two ways than an industrial facility's energy efficiency can be
improved. First, equipment, including motors, but also magnets, heating elements, boilers,
and flash processors, can be made more efficient in terms of its use of electricity to produce
rotation torque, magnetic attraction, heat, and intense radiation, respectively. Second, the
device's efficiency with respect to the mechanical function it performs can be improved. A
pump that is turned by a motor can be made 5% more efficient in performing the pumping
function. Energy efficiency is improved by 5%.
Copyright 2004 by Marcel Dekker, Inc.
348 Chapter 10
The efficiency of an electric motor in converting electric power into rotation power is a
function of the type of motor (induction, synchronous, single-phase, three-phase), details in
its design and assembly (air gaps, tolerances, etc.) and the choice of materials (better
magnetic materials in the stator core, for example) can significantly improve efficiency. As
a result, high efficiency motors are quite easy to design and build for any application. They
simply cost a good deal more than inefficient motors because they employ more expensive
materials and require more parts and closer tolerance assembly.
In general, smaller motors of any type are less efficient than larger ones. This is partly
due to the fact that small electric motors are built to inexpensive, commodity designs, while
larger ones are generally built with some regard to energy efficiency and long lifetime, even
if those are not predominant thoughts on the part of the designer. But there are
physical reasons why a large motor can be more efficient. In some types of motor design
magnetic leakage in the end of the stator winding (which is unavoidable unless the motor is
to have infinite length) only doubles even if the motor's size and power is increased by a
factor often. Thus, whether built to efficient design or not, motors have a definite efficiency
of scale.
Figure 10.3 shows the range of efficiencies in percent (HP out versus electrical energy
in, with HP rated at 746 watts/HP) for the typical range of small industrial
applications (5-50 HP). Replacing a "normal" motor in the smaller size categories with a
high efficiency motor can improve conversion efficiency by about 5%. The potential
improvement drops to almost negligible amounts for larger sizes. Thus, motor efficiency
programs often are best aimed at medium and small motor applications.
Tailoring and design are the key. Unlike its residential equivalent appliance upgrades
- an industrial energy efficiency improvement program cannot be generalized. A standard
type of energy efficient refrigerator will fit in and improve the energy usage of most every
household's kitchen, but there is no standard type of "high efficiency motor" and if there
were, it would fit in but a small portion of applications.
Almost every industrial plant is different for other plants in the details of the
exact process it employs, its layout, and the models and types of machines it
employs. To achieve energy efficiency, motor and other energy upgrades have
to be carefully selected and engineered to fit that particular plant. It is impossible to
generalize about motor applications or load patterns. Some motors, such as. those
powering conveyer belts, run nearly continuously for long periods and are candidates for
replacement with efficient motors. But other motors, like those used to open gates at the
plant entrance, etc., run infrequently and only for a short period each time. It would make
little economic sense to replace them with efficient motors - the savings could hardly be
justified.
Gradual trend of improvement. Again, as was the case with residential appliances,
industrial equipment and plants in nearly all industries are gradually becoming more energy
efficient over time. This is due almost entirely to cost considerations: industrial plant
owners and operators are very cost conscious and will seek and continuously drive for cost
reductions wherever they can be found. Utilities can count on long term continuation of two
trends in industrial energy usage. First, all existing industries will gradually become more
energy efficient. Second, new industries will spring up which are more energy intensive
than ever, as previously unaffordable, or un-thought-of industries are made profitable by-
technological progress. No utility effort is needed to drive these trends. This is "progress" in
industry.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 349
95
90
LJ
z
UJ
LJ
80
75
HIGH
EFFI CI EMCV
10
MOTOR SIZE - HP
100
Figure 10.3 Standard (bottom) and high-efficiency (top) motor efficiencies. Efficiency generally
increases with increasing motor size.
Example end-use motor efficiency program. In the 1980s a rural utility in the central
United States determined that nearly 40% of its residential customers and about half of
commercial and industrial customers had at least one water well pump. These motors
operate pumps to pull water out of the ground and push it into a pressurized tank. In this
utility's case, the water table was deep in most parts of its service territory, so that the
required work for most pumps was considerable. A survey indicated the average water well
pump in the system was a .75 HP single-phase motor with an efficiency of 62%,
mechanically loaded to about 80% of rated horsepower and operating about 1,000
hours/year, with a duty cycle of 40% during the peak hour. This means that the average
motor consumed about 850 kWh during a year, and contributed about .4 kW to system
peak.
Such motors have a lifetime in service of about 8-12 years. The utility offered a $50
rebate
11
to customers who replaced a failed pump motor with one whose efficiency was
75% or better. (Most fractional horsepower motors are only of about 50-66% efficiency, so
75% is "high efficiency" for a motor of this size range). Such a motor cuts energy usage by
100 kWh per year and reduces contribution to peak by about 50 watts. Further energy
savings in this end-use category were probably possible by improving pump mechanisms
themselves (many are no doubt inefficient in using the motor's mechanical energy to raise
and pressurize water), but this was not pursued.
Utility perspective and programs for appliance upgrades. Generally, a utility can and
should count on a certain amount of steady, gradual energy efficiency improvements due to
improvements in existing industry. Nationwide, and globally, the century long trend of
11
High efficiency motors cost about $70 to $100 per HP more than standard motors. Thus, this
incentive covered only about half the cost differential.
Copyright 2004 by Marcel Dekker, Inc.
350 Chapter 10
gradual increases in total industry energy use should also be assumed. However, what and
how much growth occurs in any particular utility service area is a question whose answer
requires careful study.
Despite their focus on costs, most industrial companies and plant managers do not want
to have to manage their own energy usage. They want to focus on their core business.
Substantial opportunity exists for utilities or energy service companies (ESCos) to offer
energy improvement and cost reduction services in this sector of utility consumer bases.
But while there is tremendous potential in industrial motor and equipment upgrades for
large industrial energy consumers, it is not easy to engineer or implement. Good, effective,
and economically optimized energy efficiency often requires a partial re-design of each
plant, its cooling, waste water, ventilation, process conveyors, and a myriad of other
equipment. This can reduce energy usage considerably without having an adverse effect on
production. But often a good deal of professional analysis and design work is necessary to
do this well. On the plus side, the energy usage of some industrial plants, which costs
millions of dollars per year, clearly justifies such effort. Overall, most utilities and ESCos
conclude that energy efficiency programs for large industrial energy users are best
implemented as a series of customized projects, perhaps marketed under an umbrella
program for identification and PR purposes, but individually tailored to each opportunity.
A motor efficiency upgrade program offered to a utility's or energy service company's
customer base as a whole needs to concentrate on ferreting out and marketing to
applications that use smaller motors and where efficiency will be worthwhile. Identification
with particular end-uses and consumer segments makes identification of customers and
marketing much easier. The water well pump example cited above is an excellent program
in this regard. It focused on a wide usage category that customers would find easy to
identify ("Do you have a water well?"), the end-use represented a lot of energy and on-peak
demand, and it used relatively small sizes of motors, so considerable room existed for
improvement. Utilities seeking programs to apply widely across their service area need to
look for these same characteristics. A combination of a near-continuous usage, so there is a
real savings to be had, and smaller size, for which efficiency improvements provide a larger
percentage improvement, offer the best potential for generalized programs.
Renewable energy
Distributed renewable electric energy generation has an appeal to many consumers, both
because they may obtain "free" energy from wind or solar generation at their site, and
because most energy consumers understand that renewable generation is more
environmentally benign than most central plant technologies. There is about renewable
energy an image of living more in harmony with nature, letting it do the work, which also
seems to increase its appeal. And, potentially, small renewable power units located at
customer sites do have the potential to lower both the energy and peak demand that must be
delivered over the T&D system.
Technology. There are three types of small, renewable energy power generation sources
that represent 99% of all renewable applications. These are solar, wind, and small hydro.
These technologies are discussed in section 10.4.
Effective, proven solar and wind generators exist in a variety of sizes and types, from
very small to megawatt-plus sizes (Willis and Scott, 2000). But none of these units can
compete head-to-head with electric utility power on the basis of cost. At present, under
ideal ambient conditions, when implemented in sizes that achieve large economy of scale,
and when professionally engineered and managed, solar and wind power are only barely
able to compete on the fringe of the electric power industry. When implemented in smaller
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 351
sizes and sited at customer locations rather than where they have optimal access to natural
energy, solar and wind generation perform less cost effectively.
However, this does not stop a significant portion of consumers from trying these
technologies anyway. In 1998 the author and colleagues performed a quick survey of rural
electric energy consumers in western Kansas and eastern Colorado. Approximately 3% of
ranch and farm houses had wind generators, not a large market penetration, but remarkable
considering that the local utilities provided no direct support for these programs. However,
only 20% of these units were actually being used to generate power. Several had been given
up as a lost cause by their owners, maintenance and operation simply being more hassle
than the ranchers expected.
Wind and solar have disadvantages, but they have some advantages. Solar is firmly
entrenched, even dominant, as a source of power for small, isolated, low-energy needs. An
example is the emergency telephones located on the side of many rural roads. These use a
small solar panel to trickle-charge a battery, that will power the phone when it is needed.
Wind generation is used to produce power at isolated facilities and communities where
T&D lines do not run and where bringing in fossil fuel for power generation is expensive.
Non-dispatchable power. One problem with solar and wind generation is that the users
have to take the power when the sun is shining, or the wind is blowing, as the case may be.
Solar power is not available at night. Wind generator power fluctuates with wind speed.
The power from solar and wind generation is not dispatchable one cannot count on
certain amounts of power being there on demand. Providing dispatchable power requires
using renewable energy in combination with any of the various storage technologies,
including battery, flywheel, and other, that can store the renewable electric power for later
use.
Non-electric renewable energy applications. Often, electric energy needs can be
reduced by using solar or wind power to directly address the end-use function. Two
residential applications are drying clothes with wind power (a clothesline) and using solar
power to heat water directly. Clothesline drying of washed clothing has certain
disadvantages, which is why electric and gas clothes dryers were invented in the first place.
However, it works well and uses little electric energy (but more human energy). In some
parts of the U.S., outside of cities where grime and pollution are present, it is popular. It is,
however, perhaps too archaic an application to use as an example for most utilities.
Direct solar water heating is perhaps the most widespread and most effective residential
and small business use of renewable energy for a normal electric energy application. The
technology is simple, involving only plastic pipes, solar radiators, and a storage tank
(needed in electric applications, too). Properly designed systems heat water even on cloudy
days and store enough hot water to make it through the longest nights. Small systems can
augment, and larger systems completely replace, a heavy household energy usage. In many
areas of the U.S., particularly where sunlight is plentiful and electric rates are high, these
units make economic sense. Their most serious drawbacks are esthetics (the units are too
large to easily hide, and not attractive) and hassle. Maintenance requirements are much
more than for a normal electric water heater.
Consumer interest. Renewable energy is simply not cost competitive in small, customer-
size packages, so it does not have a wide and lasting appeal to most consumers. It also is
something of a hassle - units require more maintenance and give more operating problems
that most homeowners expect. However, a small percentage of consumers are attracted by
the novelty, environmental, and/or individualistic aspects of having their own distributed
renewable generation. This segment will probably continue to experiment with renewable
generation, but it cannot be expected to grow until the technology makes considerable
(50%) improvement in cost effectiveness.
Copyright 2004 by Marcel Dekker, Inc.
352 Chapter 10
Utility interest. Few utilities have any incentive to implement renewable energy on a
distributed basis. It is not cost competitive. Keeping systems operating will require
increases in and new types of maintenance activities. Renewable energy looks best when
implemented at the system level (e.g., wind generator parks) with its power used to defer
peak and energy needs that would otherwise be satisfied by purchase from the wholesale
grid.
Site-specific electric-use automation
One very workable strategy to reduce the cost of energy use is to use it more intelligently.
For households and small businesses, home and business automation systems provide this
function, scheduling appliances in a manner that obtains savings due to peak reduction and
use of lower-cost off-peak energy. Home and small business automation systems can be
regarded as intelligent scheduling and interlocking systems, or microprocessor-controlled
demand limiters, or appliance usage schedulers, or, in the case where they communicate
with the utility, real-time price control or similar systems. They are all of these. Residential
systems are "home versions" of industrial plant control and factory automation systems
computerized control systems for the home.
Technology. A wide variety of home and small business automation systems are
available. Although they differ in details, all have several features. First, they have some
way of communicating with major appliances. Many use power line carrier in the
household wiring. Others use a very low power radio. A few use infra-red signals, like
those used by remote controls for televisions.
Second, the automation system has a way of controlling the devices and perhaps of
sensing if it has been turned on or wants to turn on. This often usually consists of small
"controller" boxes which a homeowner can plug into an electric outlet, into which the
appliance is then plugged. Thus, a refrigerator is linked into the home automation system by
plugging it into a control box, which is controlled by the system.
Heavy appliances, such as heaters, AC units, water heaters, and dishwashers, and some
others such as outside lights, etc., have to be hard-wired with controllers, work that often
may (should) involve an electrician. But overall, the system needs to be able to control most
of the non-interior lighting energy uses in the household or business.
The third feature is a central control capability, the "smarts" of the system which
monitors usage and controls the appliances. This permits the homeowner to set
preferences, schedule usage, and program efficiency goals ("Wash the dishes whenever it
is most off-peak during the night, but I will always have hot water at 7 AM in the
morning, period."). The homeowner or business owner programs the system with his or
her individual preferences and priorities. Set-back thermostat operation, delay of
dishwashers and clothes washers to off-peak, and other similar schedule decisions can be
made once and then implemented on a continuing basis by the system. Conditional
requirements can also be set with regard to time, demand level, and end-use ("Keep the
AC set back to 80 degrees and heat to only 50 degrees respectively on weekends.") as
desired. The system will routinely schedule those functions as programmed and effect the
goals given it.
The most effective home and small business automation systems are those that
communicate with the utility or energy service company system. Perhaps the soundest
approach is real-time pricing, or at least identification of when rates are in various rate
tiers. Homeowners can program their systems with a load reduction priority. One
homeowner could request that water heating be the device cut first, if load must be
limited. Another could identify the swimming pool pump, or air conditioning and
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 353
heating, or the refrigerator and freezers. Regardless, the utility can see substantial
response to its price signals and interruption requests from systems like these.
Cost of home automation and programmable appliance controllers is currently higher
than other load control or interlocking options. Systems cost from $500 to over $1,000 for
the master control station plus an additional cost of $50 for each appliance controller. A
utility would still have to pay incentives for control of appliances. While this cost is
considerably more than other load control options, costs of these systems can be expected to
drop in the future. Beyond this, home automation performs numerous other functions
besides load control, many of which will attract homeowners exclusive of any load
reduction benefit.
Peak and energy reduction. Home and business automation systems have the potential
to render significant reductions in energy usage and peak demand. However, utility
planners need to realize that that is not the major purpose of these systems, as viewed by
their owners. The systems are intended to improve value: to lower cost and increase
efficiency. Peak and energy reductions come about as part of this process.
Modern utility industry perspective. At the present time, home automation systems are
in their infancy. Frankly, many do not work well, being somewhat undependable
(controllers sometimes don't) and, according to user surveys, too difficult to use.
12
Over
time, home and business automation systems will become both more capable and hopefully
easier to use and more widely available. They will be built into new homes. Utilities can
therefore count on a slow but steady increase in these systems.
Automation systems offer utilities and energy service companies a "smart consumer
end" to real-time and scheduled-rate pricing systems, something necessary if those systems
are going to work well. A company that wants to accelerate these trends and take advantage
of these systems and the capabilities they provide should probably begin by effecting some
standards within its area of operations. It should identify and perhaps make available to
customers a preferred system, compatible with its real-time pricing systems and
communications. The systems can be sold on the convenience and lower cost for power that
they provide. Incentives generally are necessary only to greatly accelerate the rate of
acceptance.
Uninterruptible and high-quality power supplies
A wide variety of "reliability improvement" and "power quality improvement"
equipment is available for installation at an energy consumer's home and business. Many
of these devices are appliance specific, intended to be connected to only one device (e.g.,
a computer). Some are installed on circuits, or groups of circuits, affecting only some
loads and some parts of the facility. Others are installed at or near the service entrance
and affect the service to the entire site. All share one thing in common. They improve the
value the consumer obtains from the use of electricity, and his appliances, by controlling
or mitigating unwanted characteristics of the delivered power. These can include voltage
sags and surges, interruptions of service, and high harmonics content.
Technology and types of devices. Uninterruptible power supplies (UPS) consist of an
energy storage device (batteries or, in a few newer systems, flywheels), a charger
connected to the utility power system to keep the stored energy resource charged, and a
power controller to output power to appliances and equipment as needed. They are
12
A pilot program in the late 1990s, with several hundred homeowners in France resulted in only about
5% of homeowners using the systems on a long-term basis. The rest felt the systems were too
complicated to operate.
Copyright 2004 by Marcel Dekker, Inc.
354 Chapter 10
available from "commodity" units sold at hardware and home improvement stores for
only about $100 to large "site sized" units as big as cargo containers that can serve an
entire commercial building.
Characteristics that are deliberately varied in the design of various types and
qualities of UPS design are energy, power, and power quality. Energy is the total amount
of kWh that the unit can store and supply. Power is the rate at which it can supply that
energy: a .5 kWh, 250 W UPS can serve up to a 250 watt load (i.e., a home PC) for about
two hours.
Power quality characteristics - frequency stability, voltage regulation, and lack of
harmonics - vary greatly among available UPS. Most low-cost commodity units offer far
less in this regard than "utility grade" power quality. Some low cost units output what is
basically a chopped square wave, a waveform full of harmonics. Frequency often varies
by several Hz either randomly or as a function of load.
13
Voltage may vary by as much as
10% from no to full load. On the other hand, the author's computer doesn't seem to care,
working well on such a UPS. A small FM radio, however, buzzes loudly when powered
by the same UPS.
The higher the power quality the "cleaner" the power that is produced - the higher
the cost. Truly "perfect" power quality, voltage that varies very little from an exact
sinusoidal waveform regardless of source or load behavior or events, is quite expensive to
produce. High quality filters and power quality cleanup in the UPS converter, which
provides the output power when needed, can double, triple, or even further increase the
cost of the entire UPS over a standard unit. But high quality units are available in a range
of types and capabilities, so that potentially a consumer can have whatever he or she
wants if they are willing to pay for the quality.
However, not all UPS systems provide anything above "mediocre" power quality.
Some low-cost UPS appear to have been designed to maximize "store appeal," with the
highest possible power and energy levels that can be advertised on the box ("Powers 250
watts for up to 2 hours") and a very low cost. No details on voltage regulation, frequency
stability, or harmonics content are given. Waveforms are usually closer to a square wave
than a sinusoidal. There is nothing inherently wrong with this, since the units will power
most equipment, even computers. And frankly, most consumers would not think to look
for such statistics and would not necessarily know what the data meant if they did. But
this does emphasize two important points. First, if power quality is important, a buyer
needs to be very careful to study the details of specifications, and will need to be very
selective when buying a UPS. Second, it would be best to make certain just what power
quality is needed: one has to pay a premium for higher quality and it makes economic
sense to make certain the capability is needed.
Surge and sag protection devices are another popular power quality device with
widespread appeal to consumers. The goal of the lowest cost surge/sag protectors is
simply to protect an appliance from damage from high or low voltage (transient or
otherwise, in most cases). The simplest and least expensive of these devices open a
breaker if a surge or sag is detected and expect the owner to close the breaker again to re-
activate the device. The appliance (e.g., computer) ceases to operate. These are
essentially appliance-specific voltage- rather than current-sensitive breakers. More
expensive surge suppressors are available that "shave" incoming surges, allowing the
appliance to ride through surges and voltage spikes.
13
A UPS purchased at a local hardware store and used by the author for a desktop PC at his home has
an output that varies from 57 to 62 Hz, depending on load.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 355
Similarly, there are devices, including some mid- and high-quality UPS systems, that
will shave voltage surges and spikes and fill in voltage sags, maintaining power to the
protected appliance regardless. The author's home computer UPS is not one of them. It
was advertised as an UPS with a "surge protection circuit." Unfortunately, that is a
breaker located on the appliance side of the UPS. A voltage spike is not shaved by this
device but instead travels through the UPS to the output side causing the breaker to open,
dropping the appliance load. Thus, this UPS, which will keep a computer operating
through a service interruption, will drop it instantly if there is a voltage spike.
14
Harmonics mitigation equipment. Harmonics filters and blocks, which eliminate in
one way or another off-primary frequency power from circuits, are widely available but
not nearly as much a "retail" technology as UPS and surge/spike protection circuits.
Harmonics are generated by nearly any non-linear impedance, such as a saturated
magnetic core in a transformer or motor, a switched power supply, a light dimmer circuit,
or any of numerous other household and industrial processes. But harmonics problems
are not nearly as pervasive as interruptions and voltage surges and spikes.
15
Almost all
electric energy consumers experience interruptions, and are potentially exposed to sags
and surges. By contrast, few will ever experience harmonics to the degree that it will
affect the operation of their appliances.
Peak and energy impacts. None of these devices make any noticeable impact on peak
load and energy usage.
Reliability and power quality impacts. The whole purpose of these devices is to
improve reliability of operation of consumer appliances.
Utility and consumer perspective. Consumers have created a high demand for
"commodity" UPS systems and surge/sag protection for computers, big screen televisions,
and other home electronics. Stores like Home Depot sell a lot of them. Commercial and
industrial businesses also buy a lot of these (often just sending employees out to Home
Depot), but they also seek whole-site and more capable systems, particularly for critical
loads. A number of companies specialize in offering large, customized UPS systems for
entire sites, often in company with backup generation that will start within a few minutes.
Traditionally, most electric utilities stayed far away from offering UPS, surge and sag
protection, or harmonics abatement equipment, perhaps because to do so would have
been to admit that power quality was an issue. However, in the past five years several
utilities have begun to sell UPSs, surge/sag protection services, and fee-paid harmonics or
power quality improvement services.
Utility-Sponsored Load Reduction Programs
A 21
st
century electric distribution utility may find itself implementing something similar to
the DSM programs of the late 20
th
century, for several reasons. First, they may be ordered
to apply conservation and demand reduction measures where they are cost effective, or at
least evaluate them and show that those they are not using are not cost effective. Second, a
14
The author treasures this device, and the box it came in, as a demonstration of "getting exactly what
you pay for." The unit's brightly colored box promised to support "up to 250 watts" (it powers a 220
watt computer with no problem), for "up to two hours" (it has kept the computer going for nearly that
long), and promised that it "protects computers and similar equipment from voltage spikes in the
utility supply." It seems to do this, but doesn't keep the computer energized after the spike.
15
Harmonics, particularly if measured with respect to current, are ubiquitous in most household and
business wiring systems. But high harmonic current content seldom causes problems. High harmonic
voltage content may (usually will) lead to problems. When harmonics do cause a problem, solutions
are not "one size fits all" but need to be carefully tailored to solve the particular problem at the site.
Copyright 2004 by Marcel Dekker, Inc.
356 Chapter 10
utility may decide to consider these methods as a way of reducing its costs, without
regulatory prompting. Either way, planners will need planning methods that (1) assess
energy and demand reductions, (2) can address coincidence and location issues of those
reductions, (3) provide reasonable cost estimates of the DSM methods for cost effectiveness
evaluation, (4) balance these methods and their benefits and costs against T&D needs and
the benefits and costs of traditional system additions.
10.3 CONSERVATION VOLTAGE REDUCTION (CVR)
Conservation Voltage Reduction (CVR) involves lowering the load on the distribution
system by reducing the voltage on the distribution system, as shown in Figure 10.4. It is a
"DSM" program that can be implemented only by the utility.
The concept behind CVR is that a reduction in voltage will reduce the load of every
appliance used by every customer, and hence reduce peak and energy usage. For pure
impedance loads, such as incandescent lighting or resistive water heaters, instantaneous
power consumed is proportional to the square of the voltage (Power=V
2
R). Hence a 5%
reduction in voltage results in a 10% reduction in its demand. However, some loads such as
induction motors draw more current as voltage is reduced -- their load remains constant, or
even increases slightly (if losses are considered), as voltage is reduced. The composition of
load in a system needs to be taken into account in determining if it is a good candidate for
CVR. But overall, on many systems, tests have established beyond any doubt that an
instantaneous 3% drop in voltage at the substation instantly reduces the load on a
distribution feeder by something approaching 3%, and one of 5% produces about 5%.
Practical Matters Associated with Reducing Voltage on the System
CVR is implemented on a permanent basis by adjusting transformer turns ratio taps at the
substation or adjusting the settings of LTCs or voltage regulators at the substation and out
on feeder trunks. Most utilities have little problem reducing voltage by 2-5% for CVR using
these means.
But making it work in conjunction with meeting all other, and prior, distribution
operating goals, particularly that of meeting consumer voltage delivery standards, requires
some additional effort. Overall, the impact on the operation and design of the distribution
system due to CVR is a lequirement that it operate with less voltage drop overall, and
ultimately a need for much better focus on voltage drop and control of voltage during
operation of the system. Figure 10.4 shows the most simple, ideal case, one that does not
happen very often. There, the feeder was operating with a voltage drop much less than
permitted by standard, and CVR could be effected by simply lowering voltage at the
substation.
Figure 10.5 shows a more common situation. The feeder is close to the voltage drop
limit at the end, something that was identified in Chapter 9 as good engineering - use all of
the resources available to the planner before spending money on additional capacity or
voltage performance. In this case, reducing voltage means that the utility must add voltage
control equipment in order to keep voltage above the minimum required by its service
quality standard at the end of the feeder and keep it low over the rest of the range of the
feeder. Thus, CVR requires more and more carefully placed voltage regulators and
switched capacitors. But these changes can be implemented successfully. Whether they are
cost-effective depends on the utility's perspective and its energy and peak reduction needs,
as will be discussed below.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 357
minimum primary level voltage limit
110
1
Distance - miles
Figure 10.4 Conservation voltage reduction lowers voltage on a feeder to the minimum possible.
Under the right conditions this will reduce peak load and energy usage.
Distance - miles
Figure 10.5 Conservation voltage reduction can result in too much voltage drop for customers at the
ends of the feeders. Here, customers for the last .5 miles on this lateral branch see voltage below
minimum guideline (dotted line) with CVR. Note that the slope of the voltage profile with CVR is
steeper than without (total voltage drop to the end is about .45 volt more with CVR). The loads on this
feeder are predominantly constant-power loads. As voltage goes down, they compensate somewhat by
drawing more current, increasing voltage drop.
Copyright 2004 by Marcel Dekker, Inc.
358 Chapter 10
C9
e
>
110
minimum primary level voltage limit
0 1 2
Distance - miles
Figure 10.6 Two different plans for rather extreme levels of CVR re-engineering of the feeder
produce different voltage profiles. Solid line involves reconductoring the feeder for less voltage drop.
Dotted line uses the existing conductor and installs a regulator near the feeder midpoint. It applies
some CVR across the entire feeder. The reconductoring produces far fewer electric feeder losses than
the regulator scheme, but the regulator scheme produces a lower average voltage up and down the
feeder.
How much can voltage be dropped in a realistic CVR program?
The amount of permissible CVR program voltage reduction permissible on utility
systems varies a great deal depending on design, load characteristics, etc. However,
generally:
Rarely can one implement a "theoretical" across the board constant
reduction (Figure 10.4). Most often, reductions affect only parts of each
feeder, as illustrated by the re-engineered feeder profiles in Figure 10.6.
Most feeders can absorb reductions of averaging about 1-1.5% without
too much re-engineering. There will be exceptions, feeders where no
CVR can be applied without major re-engineering and additions.
Reductions of about 3% can be effected through considerable refinement
of voltage engineering on feeders and the installation of properly tailored
line regulators and switched capacitors.
Reductions of up to 5% and in some cases a bit more than 6% can be
effected during emergencies, if the utility has previously prepared to
handle a 3% reduction without problems, as described immediately
above as an emergency peak shaving measure.
On many utility systems, for the emergency situation listed above, the voltage delivered
to consumers near the end of feeders will fall noticeably below the minimum dictated by the
utility's standards for normal voltage service. Such situations are tolerable for brief periods
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 359
because the "emergency conditions" permit application of emergency voltage standards.
For example, utilities using ANSI standard C84.1-1989 as their voltage standard can get
over three additional volts for emergency CVR. The standard lists 105% to 95% of nominal
of utilization voltage as the standard for normal conditions but a range of 105.8% to 91.7%
as allowable for certain temporary or infrequent operating conditions, i.e., emergencies.
CVR Works Well in Short-Term Tests
Without doubt reducing voltage reduces load. This is easy to see with what is called a
"notch test": drop the voltage at the substation low side bus by adjusting LTC controls or
regulation by a total reduction of 5% while monitoring current. The reduction in power
usage is obvious. Averaged over more than 20 utilities the author has worked with on CVR
programs, systems generally respond slightly better than linearly: a 5% CVR gives about a
5-6% reduction in load. Raise the voltage back to its original level fifteen minutes later and
the load instantly goes up by a like amount. It is hard to argue with this type of "notch test"
(short-term drop and rise back in voltage). Every one of these the author has observed has
indicated that voltage reduction reduces load by about or a little better than a 1:1 ratio.
But this does not mean CVR works this well as an energy reduction measure. What is
often not recognized is that such "notch tests" do not mean that that reduction lasts
permanently. In fact, CVR interacts with diversity of appliance loads, so that certain
appliances "fight" the voltage reduction after a period of time, in a process called "duty
cycle rebound."
Duty Cycle Rebound: A Complicating Factor
Consider an electric water heater rated at 4,000 watts at a nominal 230 volts. What
happens when its voltage is reduced by 5% to 207 volts? Its load instantaneously drops
by very nearly 10%, to 3,610 watts. However, this water heater is still operating under
thermostatic control, and that thermostat still intends to operate the water heater elements
so that it maintains the water in its tank at the same temperature as it was prior to the
CVR occurring. The water heater thus compensates for its lower heating output by
running at a 10% higher duty cycle. For example, where it might normally have operated
for fifteen minutes in an hour, a duty cycle of 25% at 4,000 watts, it now stays on for an
extra 1.5 minutes. Operating now for 16.5 minutes in the hour (a duty cycle of 27.5%, ten
percent more than before), it obtains the same 1 kWh of energy that it would have
without CVR:
Energy use prior to CVR =15 minutes x 4 kW = 1 kWh
Energy use after CVR = 16.5 minutes x 3.61 kW = 1 kWh
In the long term daily kWh usage and its contribution to system peak in every hour where
it runs less than 100% of the time remains exactly the same as before. Since under normal
conditions, even of extreme peak conditions, water heaters never run more than about
60% of the time in any hour, this means that over the course of a year, no net reduction in
energy usage by water heaters occurs due to CVR. All draw power at a slightly slower
rate, but each draws it for a correspondingly longer period.
Also, no peak reduction occurs due to CVR impact on water heaters. Prior to CVR,
perhaps there were 100,000 electric water heaters, each of average connected load of 4
kW, operating on the system at 25% duty cycle during the peak hour. That means an
average of 25% were activated at any one time during peak hour, so water heater
contribution to peak demand was
Copyright 2004 by Marcel Dekker, Inc.
360 Chapter 10
25% x 100,000 x 4 kW = 100 MW
After CVR has been implemented, the connected demand of each water heater is now
only 3.61 kW instead of 4 kW. But duty cycles have rebounded to 27.5% to make up for
the slower rate of energy delivery to each water heater. Thus, contribution to peak
demand is
27.5% x 100,000 x 3.61 kW = 100 MW
Similarly, air conditioners, heaters, electric ovens, stoves, water well pumps pressurizing
storage tanks, and most of the other "heavy" loads in most homes and businesses operate
on some sort of thermostat of "performance" control, and no long-term reduction in
energy usage occurs among these loads. Only lighting, particularly incandescent lighting,
and non-controlled impedance and constant current loads do not rebound in this manner.
CVR Does Provide Energy Reductions, But of
Lower Amounts than "Notch Tests" Indicate
But CVR does provide permanent reductions in overall energy usage and in peak
demand. They are just a bit lower than might be expected from "notch tests." While a
majority of home and business appliances (as measured by demand and energy usage) on
most utility systems are of the type that rebound, as explained above, a good portion are
not. But roughly 25-30% of electric load is lighting. Another 10% typically consists of
inexpensive "regulated" AC-DC power supplies (in low-end electric and consumer
equipment) which often looks something like a constant current load.
Assuming all lighting responds to CVR like an impedance load would, when voltage is
reduced by 5%, the load of this 25-30% portion of demand would drop by 10%, for a
reduction amounting to 2.5% to 3% of the total. The constant current load portion would
drop by 5%, for a reduction of .5% of the total. Thus, a 5% CVR would obtain a permanent
reduction in energy usage and peak loads of about 3-3.5% from a 5% reduction. Of course,
not all lighting is constant impedance load, but there a small portion of other non-controlled
impedance loads on most utility systems. Overall, the author believes that this rule-of-
thumb based model of CVR impact provides the best general estimate of its capability.
Long-lasting demand reductions of from 50% to 75% of the amount of voltsge reduction
can be obtained. CVR does work, well, as an energy reduction means. It does not work
nearly as well as often expected based on short-term tests, which show a reduction of about
twice that much, but it can be an effective energy reduction measure.
Less rebound during some extreme winter peak conditions
Duty cycle rebound cannot occur if the duty cycle of the affected appliances is at or very
near 100%. For example, if the duty cycle of the water heaters used in the example above
had been at 100% during the peak hour, reduction in demand that hour would have been
10%, not zero. Had it been above 91%, some reduction would have been seen, an amount
equal to 100% minus the actual duty cycle.
But few thermostatic or pressure-controlled appliances operate at 100% duty cycle, even
during peak. Water heaters typically operate, at the time of water heater peak demand, at
only between 30% and 15% duty cycle, depending on the characteristics of both usage, and
water supply temperature in the utility service area (see Chapter 3). Air conditioners and
heat pumps are designed to operate at 80 to 90% during peak conditions. (And in addition,
they are mostly motor - i.e., constant power - loads so CVR makes only a small reduction
on their demand anyway).
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 361
In order for a major appliance to yield a large reduction during peak, it would have to be
an impedance load and be at or near 100% duty cycle. The only appliance in wide usage
among households and small businesses that meets this categorization is resistive space
heating, either of the direct or radiator and baseboard (hot water flow) type. Also included
here are the resistive assist heating elements in many heat pumps designed for intense
winter usage. All of these appliances are mostly constant-impedance loads, and can be very
near 100% duty cycle during extreme weather winter conditions (worst winter in 20 years,
etc.).
Thus, during really extreme winter peaks, when heating demand is very high, the load of
all types of resistive space heaters is affected by CVR, perhaps greatly. It thus acts as a
brake on winter extreme weather peaks. Summer peaks, however, will not react in this way
because AC and heat pump loads act mostly as constant-power loads, so even if they are at
100% duty cycle, CVR does not affect them (there is neither a significant load reduction or
any rebound).
Using CVR as a Peak Shaving Method
CVR, implemented as described above (on a permanent basis, lowering voltage to the
lowest possible value that meets minimum standards) will reduce peak demand, too, since
it reduces demand every hour of the year. However, it can be used as a peak reduction
tool only in a way that provides temporary reductions in peak demand that are much
greater than this "permanent CVR" amount. To do so, the utility must apply CVR only
during peak periods.
Consider a summer peaking utility that operates at "normal" voltage levels for most of
the year, so that at the beginning of its peak hour, late in the afternoon of a July day, it
can drop voltage by 3%. All loads that will rebound, including those like water heaters
that will "rebound" by operating at longer duty cycles, instantly drop in load, so the
utility sees a system wide reduction of something between 3% and 6%. For the sake of
argument here, it will be assumed this reduction is 4.5%, or about three times the long-
term impact that a permanent reduction would provide.
How long will this reduction last? Tests carried out by the author in the mid 1980s
indicated it lasts longer than might be expected. Almost an hour and half passes before
there is a noticeable indication the reduction is dropping, and it takes more than two
hours before the decrease in loading degrades to the long-term figure of only 1/3 as
much. The reason is the duty cycles of water heaters and other appliances that are not
operating at the time the reduction is implemented. Consider a water heater that just
switched on at the time the voltage reduction is applied. It is at a typical water heater
cycle, 25%, so it wants to run for 15 minutes, and will now operate for 16.5 minutes
under CVR. Thus it will be about 15 minutes before the "rebound effect" of this water
heater's additional duty cycle operation will be seen on the system.
In the normal course of events, during the next few minutes, other water heaters, off
when the voltage drop took place, cycle back on, exactly as they would have had the
voltage stayed normal (voltage reduction makes no difference in the duration of the "off
cycle" of any device). They come back on at a time when not quite as many water heaters
as one would have expected have shut off, and thus there are a few more water heaters
running than there would have been if the voltage had been left the same. All are running
at a lower load because of the lower voltage, but over time, due to the slightly longer
operating cycle of each, more are operating at any one instant (in fact exactly 10% more)
and the total coincident demand due to water heating coincident demand is back to where
it would have been had voltage never been reduced. Eventually, 10% more water heaters
will be operating at any one time, each at 10% less demand. There will be no reduction.
Copyright 2004 by Marcel Dekker, Inc.
362 Chapter 10
Now consider a water heater that switched off just a second before the voltage
reduction occurred. Again, operating at 25% duty cycle (and assuming that its cycles take
about an hour), it will be 45 minutes before it switches on, at a time it would have even
without CVR. Its demand will then by lower than it normally would have been, and it
will be a further 15 minutes before its rebound occurs, too. Thus, it will be an hour and
fifteen minutes before the system sees its rebound. In actuality, during summer, the duty
cycle length of many water heaters is well over an hour, so rebound can take as long as
two hours to occur.
This correction in the number-of-units-operating-at-any-one-moment occurs only after
one complete off-on cycle of all water heaters in the system has taken place, and because
water heaters operate infrequently, this can take up to two hours. Thus, gradually during the
hour or two after voltage is reduced, duty cycle rebound slowly erodes a portion of the load
reduction, at least that of water heaters and similar devices.
Other controlled impedance appliances, including heaters and ovens, also react in this
way: they only recognize they need to operate longer when they switch on. Thus, rebound
takes time to occur, being a function of the duty cycle. After this period of adjustment, the
actual peak load reduction will be somewhat less than a simplistic approach suggests.
CVR Cost
CVR's major cost impact of implementing CVR is at the distribution level, where it can
increase cost of feeders in some cases. In other feeder cases, it can be applied at no cost,
occasionally with little more effort than the resetting of transformer taps and LDC settings.
However, CVR reduces the voltage drop allowed to distribution planners to distribute
power, a consideration in systems that are well designed from a load reach standpoint and
have no margin in voltage.
In some cases, CVR can be implemented only by modifying the feeder system (adding
regulators or re-conductoring) to overcome voltage drop problems caused by CVR's
effective change in voltage standards.
16
Beyond this, by limiting the voltage drop available,
feeders will have less margin to accept new load growth before the (now revised) voltage
drop criteria are exceeded - they will need to be reinforced more often and perhaps in a
more costly manner.
Additionally, the utility needs to consider the impact of CVR on load reach and the
loss in value of that. In some cases, the loss in capability of the distribution system is
considerable, and the overall cost of future additions and expansion is considerable. This
is not to say that CVR is not economical or that it should not be examined as an option,
merely to point out that the present and future costs on distribution should be taken into
account in developing evaluations of benefit/cost.
Utility Perspective
Traditionally, most vertically integrated electric utilities preferred to use CVR as a "last
ditch" peak shaving method. Using it as an energy reduction method reduces revenues, so
they generally do not view it favorably. For this reason, some utilities resisted CVR, and
others merely did not respond with the same enthusiasm they would have had if CVR
would increase revenues by 3%. Lower voltages also meant less margin for unknowns or
voltage regulation problems, and planners, engineers, and operators all had concerns
about that. Finally, the appeal of a significant peak reduction capability if used only for
16
Some state utility commissions require CVR implementation at the distribution level but permit
utilities to not apply it in cases where they can demonstrate an adverse cost impact.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 363
peak shaving was great. CVR does work, it can be implemented where and when needed
and it delivers relatively big peak reductions, if only for brief times. Thus many utilities
reserved it for their peak shaving mechanism and did not implement it under normal
conditions or as an 8760-hour measure.
In fact, most utilities operated, and continue to operate, their systems nearer the high
than the low end of the permissible voltage spectrum. This is not due entirely, or perhaps
even in the majority, to the fact that this reverses the CVR effects and increases revenues
by a small amount. The author has grown convinced in over 25 years of working with
utilities that this is mainly a culture attitude. Utility planners, engineers, and operators
believe that providing slightly higher than nominal voltage is simply better for their
customers. They also know that higher voltage gives them a margin that will help them
maintain service during unexpected operating contingencies. Certainly, however, most
electric utilities tolerated this situation because it had no adverse, and perhaps a slight
positive, impact on revenues.
In a de-regulated power industry, many utilities will no doubt continue to have the
same perspective about CVR. These companies are essentially still vertically integrated,
at least to the extent of being local delivery (T&D and retail sales) companies and still see
the revenue reduction as something to be avoided. In particular, utilities under rate
freezes have to be very concerned about anything that would erode revenues. They have
no mechanism to adjust rates to compensate for a change in the coverage of their fixed
costs (see the Rate Impact Measure Test discussion of cost evaluation, later in this
chapter).
However, utilities that meet the strict definition of wire companies - electric delivery
utilities who operate the T&D system but do not actually sell the power transported over
them - might take a different view of CVR, depending on the regulatory framework
within which they work. Depending on how they are compensated for their delivery
services, these utilities may be indifferent to whether the energy delivered over their
system is reduced or not. Or they may have a different financial incentive that makes
them less reluctant to consider reductions. Finally, many will service at least a part of
their income, and some a major portion, by billing for the peak demand they transport
over the system (that being what applies against their capital charges for capacity and
allocation of a large part of their fixed costs). Therefore, CVR's reduction in peak
reduction demand will not appeal to them except as a very "last ditch" emergency
measure.
10.4 DISTRIBUTED GENERATION
Three Ways to "Burn" Fuel to Provide Electric Power
Distributed generation most often uses some form of conventional fossil fuel, like gasoline,
diesel or fuel oil, natural gas, propane, methane, or gasified coal, to produce electric power.
In every case, regardless of fuel, through very careful design and often intricate timing
of events, measured amounts of the fossil fuel are oxidized - purposely combined with
oxygen - to produce heat, and perhaps pressure, and, ultimately, electricity. There are three
major ways in which oxidizing fossil fuels are used to produce electric power.
In two of these three approaches, the fuel is burned and a portion of the heat produced is
transformed into mechanical rotation, which spins an electric generator, producing
electricity. The first of these approaches, the reciprocating piston engine, uses the heat and
pressure from combustion to move a piston inside a cylinder, converts that linear motion to
rotation of a crankshaft, and uses that rotation to spin an AC electric generator.
Copyright 2004 by Marcel Dekker, Inc.
364 Chapter 10
The heat and pressure created by combustion can also be used to spin a turbine wheel.
The fossil fuel is burned very close to the turbine vanes, producing a rapidly moving gas
that passes over the turbine wheel's vanes, causing it to rotate. Again, this rotation is used to
turn an electric generator.
The third way to obtain electric power from fossil fuel is with a fuel cell. Fuel cells do
not use combustion -- oxidation is performed by chemical catalyst. Fuel cells are essentially
fuel-powered batteries, producing low-voltage DC current via the microscopic catalyst-
enhanced oxidation of hydrogen from the fossil fuel, in the presence of electrolytes. This
DC power is converted to AC power using electronics.
Other, less proven ways to produce electric power from fossil fuel have been developed
(i.e., thermophotovoltaic), but these three methods appear to be the only economically
viable approaches.
From Less than 1 kW to More than 25,000 kW
Distributed generators are available in sizes from less than 5 kW to 25,000 kVA (even
larger turbine and diesel units are available, but 25,000 kW is the top of the "DG range"
being considered in this book). Generally, large reciprocating and combustion turbine
generators are designed for heavy, long-term use, and are available in sizes from 1,000
kW on upwards. Large fuel cell systems are available in capacities from 1,000 kW up to
10,000 kVA. These larger size DGs are usually installed at the primary distribution
voltage, e.g., on portions of the electric system between 2 and 25 kV phase to ground,
and are restricted to applications at large industrial sites, or on the electric utility system
itself. Some are applied as base load, used 8,760 hours per year, while others serve as
peak-reduction units used only during periods of high power demand.
Smaller DG units are available in sizes from 1,500 kW down to as small as 5 kW. These
units are intended for very dispersed applications, as generators for individual homes and
small businesses or as portable power units for construction sites, etc. Reciprocating piston,
fuel cell, and a type of turbine (micro-turbine) are all available in this range.
Such "mini" and "micro" generators are almost always installed on utilization voltage
level (120/240, 480, or 600 volt) circuits, often on the customer side of the electric utility
meter. Applications for these types of units can include providing power for all of the
eiectiical demand at a residence or small commercial site, or just providing power for peak
shaving. They can also be devoted solely to improving availability of power, including
usage in UPS (uninterruptible power supply) and standby or emergency power systems.
The Carnot Cycle
Most fossil fuel DG units are powered by heat engines. They convert heat into mechanical
motion by allowing that heat to move from a place of high temperature to a place of lower
temperature in a manner that causes mechanical motion. Fuel cells can also be viewed as
electrical heat engines: they create electrical rather than mechanical motion while
channeling heat flow from high to low temperature points.
The fuel efficiency of a heat engine has an upper bound that is defined by the basic
Carnot cycle equation:
*high ~ 1 low
Upper limit on engine efficiency = (10.3)
Thigh
where T^is the temperature on the high side of the energy transformation machinery, for
example, that inside a turbine engine, and TI
OW
is the temperature on the low side, for
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 365
example, on the outlet side of the turbine exhaust, all measured in absolute terms (degrees
Kelvin).
While the Carnot cycle and equation 10.3 above have many implications for DG
engineering, the most important implications are:
1. The higher Thigh, the higher the efficiency. It can be raised by using a
hotter-burning fuel, by burning the fuel under higher pressures, by
burning a leaner fuel mixture (higher air/fuel ratio), or by insulating the
combustion areas so the internal areas of the engine run hotter.
2. The cooler TI
OW
, the higher the efficiency. Heat engines work on a
difference in temperature. Regardless of Thigh, the efficiency can be
improved if heat can be "dumped" to a lower low temperature.
Qualitatively, the relative efficiency of almost all types of DG, and the basic concepts
behind various efficiency-improvement approaches, can be understood by applying this
simple pair of "Carnot cycle rules." In almost all cases, whether one is talking about diesels,
turbines, fuel cells, or even solar power for that matter, units that run at higher temperatures
and "dump" their heat to lower temperatures produce more power from the same amount of
fuel (or sunlight).
Fuel Efficiency and Heat Rate Curves
All fossil-powered generating units have an efficiency that varies as a function of output
level. The heat rate curve, or efficiency-vs.-output curve, for a generator shows how many
BTUs of fuel it requires in order to produce a kWh of electric power, as a function of the
amount of power it is producing. Figure 10.7 shows heat rate curves for three DG units, in
per unit form, to stress how the efficiency curve shape varies from one unit to another. Note
that while curve shapes are quite different, each type - fuel cell, turbine, and reciprocating
piston - has some point or range in which it is most efficient, where it requires slightly less
fuel per kW than at any other output level. Heat rate curve shapes vary widely among DG
units. A DG designer can alter the shape of a unit's heat rate curve by changing various
elements in the machine's design, to make the same DG unit have different "optimal"
operating points for it? intake, exhaust, fuel injection, generator, and others parts.
Piston Engine Driven Distributed Generators
Reciprocating piston engines produce rotating mechanical power by the use of pistons -
essentially round seals that slide back and forth in cylinders within which a fuel is burned.
The burning fuel forces the pistons out. A crankshaft converts that linear motion to rotation
and, through its inertia, forces the piston to slide back and forth in the cylinder so the
process can be repeated cyclically. Figure 10.8 illustrates the basic operation of a piston
engine.
The reciprocating piston engine provides the mechanical motion needed to drive a
rotating electric generator. Reciprocating piston engines turn at roughly the speed of steam
turbines (1,200-6,000 RPM), which is relatively slow compared to some gas turbines,
particularly smaller micro-turbines (up to 100,000 RPM). This rather "traditional" speed of
rotation for piston engines (similar to that of steam turbines) means that traditional types of
generators are most typically used. Four types of generator can be attached to reciprocating
Copyright 2004 by Marcel Dekker, Inc.
366 Chapter 10
Example Heat-Rate Curve Shapes
25 50 75
Output: Percent of Maximum Rating
100
Figure 10.7 Heat rate curves for three different DG units, in per unit scale of their maximum output,
illustrate the different way in which efficiency varies as a function of output depending on the type
and design of a DG unit.
engines to produce electric power, synchronous AC, inductive AC, DC with DC/AC
conversion, or written-pole synchronous AC generation. By far the most widely used is
the synchronous AC generator - the same type used on "real" (large central station)
generators. Alternating-current frequency is therefore controlled by controlling engine
speed.
Piston-engine generators use a gasomis, diesel, natural gas, or propane/methane
powered piston engine to spin an electric generator, with the engine's crankshaft and the
generator's rotor usually spinning at the same rate on the same or directly coupled shafts.
Piston-engine DG is both the most popular DG unit at the present time and the
technology that sets the performance/cost benchmark that other types of DG must meet to
see any significant market success. Interestingly, it currently accounts for 95+% of all
distributed generation on the planet, a figure that has remained stable, or slightly
increased, in the past five years. Pistons may be an older technology compared to some
other DG types, but they are still a viable, competitive technology. They will likely
remain the power source of choice for very small (< 250 kW) electric generators in the
foreseeable future.
Reciprocating piston engines are a proven, mature, but still improving method to
provide power for distributed generation systems. They have potential fuel economies as
high as 45%, but their greatest advantages are a low-cost manufacturing base and simple
maintenance needs. This last is often an overwhelming advantage: Among all the various
types of DG, only a reciprocating piston engine - in particular a diesel engine - is so
universally familiar that one can find someone to repair it virtually anywhere on the planet.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 367
Figure 10.8 The basic Otto cycle of most reciprocating piston engines consists of four strokes, each
during the movement of the piston through a full back or forth motion that corresponds to one-half
revolution of the crankshaft. This particular engine uses flapper valves to control intake and exhaust.
Disadvantages of piston-driven generators are a general lack of good waste heat for
co-generation applications compared to turbines, exhaust emissions, noise, vibration, and
weight. All of these can be mitigated by various means and probably improved greatly in
future designs.
Most people think of piston engines for DG in terms of the engines that propel
automobiles and trucks. Some DG engines are derived from automotive applications. The
piston engines in automobiles typically have between four and eight cylinders, in which the
pistons are about 2 to 4 inches in diameter, and reciprocate at up to 8,OOO RPM. But the
piston engines used to drive larger piston-engine generators (diesel generators up to 50,000
kW have been built) have much bigger pistons, often more than 4 feet in diameter and
weighing several hundred pounds each. Engines with such heavy reciprocating masses turn
at speeds as low as 50 RPM, far below the RPM rate of smaller engines, partly because it
requires too much mechanical strength to reciprocate pistons char heavy at hlgh speed B3ui
more important, the exploding gas (fuel-air mixture) inside a cylinder will expand with the
same characteristics whether in a small or large combustion chamber. A 5-foot-wide piston
with a 5-foot stroke, turning at 150 RPM, moves at the same speed as a piston with a two-
inch stroke, moving at 4,500 RPM, and is just as compatible with the burn characteristics of
the fossil fuel. Thus, large generators turn at lower speeds, small ones at higher speeds.
Low-speed motor-generators operate at 50 to 150 RPM. Medium-speed generators run
at up to 800 RPM, and high-speed generators at 1,200 to 1,800 RPM. Note that all these
speeds are far below the normal rate of rotation for automobile piston engines. Automotive
derived DG units usually run at 1,200-2,400 RPM. Regardless, although there are
exceptions, almost all piston-engine DG units use a constant-speed alternating-current
generator and run at as close to a constant speed as possible.
While proven by decades of use, and not as exotic as some other forms of DG,
reciprocating engines have tremendous potential for further future improvement. The
impending demise of the internal combustion piston engine has been forecasted by
various proponents of newer designs for most of the last one hundred years. Its
replacement has successively been predicted to be the turbine, various new designs or
cycles involving rotors, impellers, vibrating fluids, or other approaches for heat-to-
Copyright 2004 by Marcel Dekker, Inc.
368 Chapter 10
motion conversion, and the fuel cell. Every such prognostication overestimated the pace
of development and advantages of the technology predicted to take pistons place, and
underestimated the improvement that would continue in piston engines.
No doubt, eventually, something will displace the internal combustion piston engine
as the most widely used mechanical power source for small and medium applications on
this planet. But frustrating as it has been for the proponents of rotating piston (e.g.,
Wankel) engines, small turbines, fuel cells, and other types of newer concepts, piston-
engine technology always seems to stay just far enough in front to justify its continued
dominance. And even at the beginning of the 21 century, there is no sign that this
replacement will occur any time in the near future.
Turbine Driven Distributed Generators
A gas turbine generator uses a turbine spun by the expanding high-pressure gases of
combustion to rotate an electric generator, as shown in Figure 10.9. Advantages of the
gas turbine over all other direct-fueled engines include continuous combustion without
reciprocating motion, low vibration, and very high power-to-weight ratio - lightweight
turbine generators weigh only a sixth as much as reciprocating piston units of comparable
output. Gas turbine driven generation has achieved significant market acceptance in the
utility sector, particularly as central station peaking units, and for some larger
distributed utility applications in the 10-25 MVA range. Some power producers have
committed a significant portion, up to 1008, of their central generating capability to
turbines. Gas turbines are also widely used in industrial power applications, in large
petrochemical plants, paper mills, and other facilities that have significant power needs.
Gas turbines are simple, compact, robust, but not outstandingly efficient devices which
can be applied to turn electric generators in distributed generation systems. By varying the
design, and using physics in different ways, relatively satisfactory fuel economy and
durability can be obtained from turbines over a range of more than four orders of magnitude
in size, from about 15 kVA to more than 150,000 kVA.
Gas turbine generators fall into three categories of distinctly different design and
operating characteristics. The easiest way to distinguish these categories is by size (physical
or electrical output) as illustrated in Table 10.2, but utility planners must realize that size is
not the only, or even the most sigmficant, difference among turbines in these different
categories.
Figure 10.9 Turbines use a continuous flow with pressure differences due to intake and output
turbine wheels that segment the flow into four distinct regions. The first corresponds to the intake and
compression cycles of a piston engine (Figure 10.8), the second to the power cycle, and the third to the
exhaust cycle. The turbine produces a high-flow, low pressure exhaust flow (4) that is very suitable for
CHP applications.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 369
All three categories of turbine work on the same principle, but use markedly different
designs and different types of generators depending on their size. As was the case with
piston engines, they vary in design according to size because the flame and gas-expansion
characteristics of the gas energy are the same regardless of size. A two-inch turbine in a 20
kW micro-turbine generator must deal with the same expanding gas as a 100-inch turbine
wheel in a 150 MW turbine. As a result, the former has to spin at 90,000 RPM to match the
larger turbine's blade tip speed at only 1,800 RPM.
The largest category is the traditional utility gas turbine generators that range from
about 10,000 kVA peak output to more than 150,000 kVA, with a typical unit being about
70,000 kVA. This is the only one of the three categories for which the turbines were
designed specifically for electric power applications. Turbines in the two smaller categories
are invariably modified versions of turbines originally designed for other applications.
In the middle of the size scale are mini gas turbine generators, units in the range of
800 kVA up to about 10,000 kVA. Most of the turbines in this category are based on
designs first produced for small aircraft, coastal and military patrol ships, helicopters, or
battle tanks. As such, when used for DG, they are not quite as efficient and do not have
quite as favorable maintenance and operating costs as the larger turbines designed solely
for utility use. On the other hand, many are derived from military designs, which mean
they are robust, durable, and capable of operation under very adverse ambient conditions.
The smallest category of gas turbine generator, and the one that has gained
tremendous cachet as a "revolutionary concept" that will alter the electric industry
entirely, is the micro-turbine generator. Units in this category range from less than 20
kVA up to about 750 kVA. The turbines used in these units were mostly designed
originally for vehicular application, and many were not originally designed as turbines,
but as the turbine in a turbo-charger for a large piston engine. Table 10.2 compares these
three categories of turbine generator. More discussion of their design, operation, and
similarities and differences can be found in Willis and Scott (2000).
Table 10.2 Comparison of Gas turbine Generator Categories
Characteristic
Available range - kVA
About the size of ...
Original design based on ...
Most typical fuels are . . .
Makes about as much noise as
Out of service once every . . .
Turbine generator usually is ...
Turbine spins at about . . .
Generator type used is ...
Generator turns at about . . .
Turbine & generator run at ...
Best fuel efficiency is about . . .
Can be bought and installed in .
Typical cost/kW is ...
Micro
20-500
a refrigerator
bus, tmck engines
nat. gas, diesel
a car at 40 mph
two years
single shaft
70,000 RPM
DC with AC conv.
70,000 RPM
variable speed
32%
a week
$950/kW
Mini
650-10,000
a large truck
aircraft engines
nat. gas, diesel
3-4 leaf blowers
eight months
two shaft
15,000 RPM
AC sync.
3,600 RPM
constant speed
30%
two months
$450/kW
Utility
12,500-265,000
a building
utility needs
nat. gas, fuel oil
a jet plane
year and a half
two or three shaft
1,800 RPM
AC sync.
1,800 RPM
constant speed
37%
a year or two
$265/kW
Copyright 2004 by Marcel Dekker, Inc.
370 Chapter 10
Fuel Cell Powered Distributed Generation
Fuel cells take a unique approach to using fossil fuel for producing electricity. Instead of
burning fossil fuel to produce motion to drive a generator, they oxidize the hydrogen in a
fossil fuel in a chemically controlled (catalyst-driven) process that causes ion migration
through an electrolyte, creating a direct electric current and, hence, electric power. If
designed and operated appropriately, they are nearly silent, very efficient, and produce
virtually no polluting emissions.
Fuel cells fall into five categories depending on the chemical basis for their operation.
Ranked in ascending order of internal temperature (and thus potential Carnot-cycle
efficiency) they are: proton exchange membrane fuel cells (PEMFC), alkaline fuel cells
(AFC), phosphoric acid fuel cells (PAFC), molten carbonate fuel cells (MCFC), and solid
oxide fuel cells (SOFC). AFCs are unsuitable for electric power application and used
almost exclusively only in spacecraft. Alone among fuel cell types they require pure oxygen
and cannot not "run" on air. The four of the remaining fuel cells types are potentially viable
for at least certain niches in the DG field. In particular, PAFC, MCFC, and SOFC seem
suitable for stationary power generation applications.
Fuel cells are simple in concept, but very complicated in actual execution. High fuel-to-
electrical efficiency comes from high internal temperatures, on par with or higher than
inside combustion turbines. This means they must be robust designs made of high-
temperature materials. In all of them, the fossil fuel needs to be "reformed" - stripped of its
hydrogen, which the fuel cell needs, while the rest of the fuel's contents is discarded
(exhaust). Reformers use a type of catalytic converter to do this. Both reformer and fuel cell
operate best under high pressure, requiring (high temperature) sealing and pumps for both
fuel and air. The power is produced at high, direct current and low voltage and needs to be
converted to AC at utilization voltage. All of this results in a machine that is actually quite
complicated. Maintenance needs are perhaps no more than those needed by piston-engine
DG, but they are very different: One can find a diesel engine mechanic just about anyplace
on the planet, but fuel cell service persons are extremely rare.
For over three decades, fuel cell technology has been developing and improving, always
promising widespread commercial success "just around the corner." Despite considerable
work in the last decade, the three barriers to widespread fuel cell application are still high
cost, lack of proven durability, and a need for a large, different service infrastructure (the
aforementioned "fuel cell mechanics," among other things). Solutions to the second and
third will follow automatically upon the first, so it all boils down to whether durable and
reliable fuel cells can be produced at a competitive price. This is being addressed by
government, industry, and manufacturers, but remains a challenge that no one seems able to
meet.
For some applications - those having particularly sensitive environments in which
noise, vibration, or emissions are a major concern - today's fuel cells are perhaps best for
DG applications. But for most distributed applications they are too expensive and, frankly,
not yet proven sufficiently for widespread application.
Exhaust Heat for Co-Generation and CHP Purposes
All fossil-fueled generators produce hot exhaust gases which contain a good deal of
remaining energy which can, under some circumstances, be harnessed to provide additional
value. Turbines and high-temperature fuel cells (SOFC) are best in this regard, producing
hot exhaust gases that are useable for both additional power applications or other purposes.
However, piston-engine units are sometimes fitted for this purpose.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 371
Generators that produce additional power from the exhaust gases are called combined
cycle units. The exhaust of some turbines is used to boil water to produce steam to power
an auxiliary steam turbine in what is called a combined cycle turbine. A Cheng cycle
turbine routes the steam back into the gas turbine, where it mingles with the fossil fuel gas.
Both types can approach 45% fuel-to-electric efficiency in annual operation.
Another option is to use the exhaust heat to boil water for hot-water purposes or for an
industrial process. Paper mills often use gas turbine generator exhaust for their pulp boiling.
In fact, often paper-mill turbines are designed as a balance between electrical and exhaust
water heating applications tailored to the specific plant. DG units are often used to heat
water for hot-water applications at smaller industrial plants and in commercial and multi-
family residential applications.
It is also possible to use absorption cooling equipment to use the waste heat for cooling.
Specific piston and turbine "CHP" (cooling-heating-power) units have been designed
specifically for office and apartment building applications. A good CHP application - one
where the power, hot water and heating, and cooling outputs match the need of the site -
can have an overall fuel to end-use efficiency approaching 80%. This makes CHP-based
DG so cost-effective that very little can compete against it on the basis of overall cost.
Renewable Energy: Wind and Solar
Renewable energy resources offer a potential for power production using solar, wind, and
a host of other power sources derived from ongoing natural processes of the earth's
environment. By their very nature renewable power sources are small, modular, and
geographically distributed. For these reasons, and because they connect to the lower
voltage parts of the grid, they are often classified as DG.
Renewable power systems require no source of fossil or nuclear fuel, as in the case of
gas, steam turbine, reciprocating piston, and fuel cell generation units. Some do require
burning of a fuel: garbage, animal waste (methane), and bio-mass among them. Often, the
motivation for building renewable generation is not to add local peaking support or
reliability backup, but to obtain "green" energy production. Regardless an important
advantage for remote location applications is that "fuel" is delivered to the site at no
charge. For the significant portion of small communities worldwide that use diesel
generation for electric generation, the delivery cost of fuel exceeds its actual cost. In
addition, environmental impact is deemed low in most cases, and the fact that these
power sources are sustainable without using any natural resources is appealing to many
persons.
Most renewable generators make far less environmental impacts than fossil fuel and
nuclear power generation, but are less cost-effective as well. Many are subject to some
degree of unpredictability in their energy availability and hence the net power output, and
using them to provide dispatchable power depends on combining them with some form of
energy storage. Many types of renewable energy also have site requirements that
constrain them to locations with the right combination of natural factors.
Hydro-generation is by far the most widely used, proven, and cost-effective
renewable energy source. Where one has access to continuously moving water, some
degree of power generation can be produced. Even small (250 kW) hydro units can be
efficient and reliability if well designed, and can produce power that is cost-competitive
with power form the grid. Of course, few consumers have access to the requisite river,
stream, or waterfall, and there are often legal restrictions on the use of such sources for
any purpose, generation included. But for farmers, ranchers, and rural homeowners who
do have such sources available to them, small hydro power is an option, particularly if the
water source has a rather continuous flow year around and if a small reservoir or pond is
Copyright 2004 by Marcel Dekker, Inc.
372 Chapter 10
available to provide "energy storage" in the form of an upstream reservoir.
17
Solar power. Among all possible renewable energy sources, the most flexible and
applicable in many respects is solar power. The fuel costs nothing, but is available only
during the day. However, unlike wind generation, solar power's "fuel availability" is quite
predictable (well-designed solar units will produce power even when the sky is overcast).
But daily sunlight cycles remain a major constraint on design and duty cycle, as will be
discussed in this section.
Photovoltaic generation converts light energy directly into electric power, using any of
several types of flat semiconductor diodes built with as large an area as possible and with
their p-n junction located very close to the exposed surface. Solar electric cells produce DC
electric power at low voltage, typically around 0.5 volt, whenever exposed to sufficient
light. A single cell may be less than a square centimeter in size, and produce only a small
amount of power, often much less than a watt. Usually many cells are connected in series to
provide higher voltage, and in parallel to produce higher current, in what is called a
photovoltaic array. As was the case with fuel cells, the DC output of PV solar cells must be
converted to AC power by inverter/filter equipment if they are to be connected to the power
grid or used to power AC appliances. For very small PV systems, the cost of this
conversion equipment can be a significant portion of its cost. PV is a proven technology
with more than 200,000,000 PV arrays in operation worldwide - the vast majority of them
in calculators and similar small appliances.
Solar thermal conversion systems use mirrors to concentrate reflected sunlight to
produce intense heat energy which is used to transform water into steam, which in turn
drives a steam-turbine generator. Larger (5 MVA and up) systems store great amounts of
heat overnight so that they are dispatchable power sources.
Smaller solar thermal systems use a Stirling cycle reciprocating engine, and have no
storage. Either way, solar thermal units usually employ an AC generator, which means they
have electrical output characteristics identical to traditional generation sources. In this
respect, they represent a proven and risk-free technology, well understood and for which
maintenance and operating experience is considerable. Overall cost is generally two to three
times that of fossil fuel generation, but for remote applications where fuel delivery is costly,
these simple, robust, and easy-to-repair systems are worth consideration.
Wind power generation harnesses the energy in wind to drive electric power generators,
using some form of "wind turbine" - essentially an optimized windmill driving an electric
generator. Individual wind turbines fall into the DG size range, being between 75 and 3,000
kVA output. However, most commercial applications include groups of these turbines
arranged in "wind parks" of between 25 and 200 MVA total, and are treated as "green"
central station power plants. Something on the order of 8,000 MW of installed wind
turbines were in commercial operation worldwide, with over 1,000 MW of additions being
made every year, nearly all of it in wind parks.
The advantages of wind generation are no fuel cost, a total lack of exhaust emissions,
modularity with a fairly linear power vs. cost relationship for large-scale (wind park)
installation, a very robust, simple, proven technology, and potentially a 24-hour per day
supply of energy. Disadvantages are somewhat higher cost than for fossil fuel generation,
temporal unpredictability of energy production, as well as environmental impacts generally
considered to be greater than solar power's.
17
Even a small reservoir, storing a day's worth of water upstream, converts a non-dispatchable "run of
river" generator to something that can be relied on to track load well, improving its value as a
generation source immensely.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 373
Production cost of the resulting electric power depends greatly on location, local labor
costs, and other factors. Most likely, small wind turbines will not be competitive with small
fossil-fueled gas turbine generators and other similar distributed generation for sites located
on distribution systems of major utilities, but large (3 MVA and up) units will become cost
competitive when built in large wind parks in optimal wind areas. That said, the reader
should recognize that there is a large market for wind and solar energy in developing
countries with rapidly growing loads, because fossil fuel carries additional costs along with
it, not often included in direct comparisons of price. Many developing countries simply do
not want to import fossil fuel even if affordable wind energy not only defers expenses, but
improves their balance of payments. Finally, in developed nations such as the United States,
electric services providers are discovering a considerable market for "green power." A good
portion of the residential market, perhaps up to 10 percent, will pay a premium for power
which has been produced by environmentally benign, renewable power generation
methods. Wind turbine power is one viable means of producing this green power in
"utility" amounts.
Hydro-electric power is a proven and mature technology for electric power production.
While most hydro plants have a substantial head (difference between high and low water
levels on opposite sides of the water turbine), it is possible to generate electric power in DG
amounts with a difference of only a dozen feet and a flow equal to that of many small rivers
and large streams. Low-head hydro has been evaluated as feasible and economically
competitive at many locations throughout North America, and no doubt large amounts are
similarly available worldwide.
Hydropower plants, low-head or otherwise, provide good 24-hour-a-day power
production. Most low-head installations are "run of river" plants. They do not have a
substantial reservoir of water storage behind them to vary output, but instead work with
whatever flow the river provides.
18
Thus, their output varies on a seasonal basis and from
year to year according to the amount of rainfall upstream. Making a low-head hydro site
dispatchable means building a reservoir behind it and making at least minor, and perhaps
very major, changes in the water levels and coverage of lakes and rivers. This raises
concerns about their environmental impacts on aquatic life and the changes wrought on the
natural "cleansing" of the waterways by spring floods. These concerns have limited
application of low-head hydro in many areas and in fact have led to initiatives by
environmental groups to remove existing small hydro plants in states like Maine.
Regardless, low-head hydro is a very proven, robust, and feasible renewable alternative for
renewable DG in many areas of the world.
10.5 ELECTRIC ENERGY STORAGE SYSTEMS
Energy storage can often augment DG in three ways. First, energy storage can be used for
stabilization purposes, permitting the DG to run at a constant, stable output level, even if the
load fluctuates greatly and rapidly. Second, proper amounts of storage can provide energy
to ride through periods when the DG unit is unavailable, for example, during the nighttime
for solar power, or when the DG unit of any type is being maintained or repaired. Third,
energy storage can permit a non-dispatchable DG unit to operate as a dispatchable unit by
permitting its output at any moment to differ from the power being released to the demand
or into the grid. Table 10.3 lists these three purposes of energy storage and the key
18
A large hydro plant may have a substantial water reservoir behind its dam, storing water equivalent to
several hundred MW-years of net electrical power. This permits it to be dispatched and the ability to
largely disregard short-term shortfalls in river flow.
Copyright 2004 by Marcel Dekker, Inc.
374 Chapter 10
characteristics that distinguish them from one another. Energy storage has other uses when
applied without DG, either as peak shaving systems for utilities or as UPS backup for
customer applications, but those are not emphasized in this chapter.
This section begins by examining the application of energy storage to support DG, what
it provides and why and when that can be of benefit. It then examines the various storage
and operating qualities needed in an energy storage unit in order for it to perform those
functions, and compares the various storage methods available.
Energy Stabilization
Figure 10.10, top, shows the typical load behavior encountered by a DG unit serving a
single household load. This non-coincident load curve (see Chapter 3 for a detailed
discussion) varies rapidly from over 10 kW to less than 1 kW, many times during the
course of a day, almost instantly shifting from high to low state and back again shortly
thereafter. The large shifts are caused by major appliances such as space heaters, air
conditioners, water heaters, electric dryers, and the like, which switch themselves on and
off automatically under their thermostatic control. The short duration of the "needles" is due
to the nature of thermostatic control. Seldom do these appliances run for a long period of
time, as if on a contiguous basis: temperature is kept within a narrow band by running the
unit (e.g., a water heater) for only short periods of time, but frequently, never letting the
temperature rise or fall too far.
Table 103 The Three Major Applications of Energy Storage with DG
Planning Reason Why Energy Storage Is Being Applied in Conjunction with DG System
Aspect Energy Stabilization Ride-Through Dispatchability
Reason Shave needle peaks in the
non-coincident load curve
due to large appliances, etc.
Benefit Lowers peak DG capacity
needed. Improves voltage
regulation.
Storage Must be enough to "shave"
appliance peaks and meet
then- short-term needs.
Peak Relatively great: all the
energy stored must be re-
leased in just a few minutes.
Method Based on detailed assess-
ment of daily load curve,
on a minute-to-minute basis.
Design Typically high-energy, low-
storage design with enough
capacity to avoid deep cycle.
Provide energy to serve
load during periods when
DG output is unavailable.
Service from PV, etc., can
now be maintained during
nighttime, etc.
Dictated by load during
"DG unavailable" times.
Usually Vi day's energy.
Relatively small, only one-
eighth to one-tenth of stored
energy.
Based on hourly analysis
of load needs over a year
and DG availability stats.
Must achieve size balance
between storage size and
andDG.
Provide energy stored to
stabilize DG availability
to meet various schedules.
DG owner can now bid and
sell power contracts for
arbitrary schedules.
Must be enough to transform
the DG schedule into the de-
sired sales schedule.
Requires more than for ride-
though but much less, rela-
tively, than for energy stab.
Based on hourly analysis of
desired schedules, DG avail-
ability stats, business cases.
Must achieve an overall
balance among DG unit size,
storage size, and total cost.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 375
Single Household
Mid. 6 Noon 6 Mid.
Time of Day
Coincident Load Curve
10
Mid. 6 Noon 6 Mid.
Time of Day
10 Single Household "Filtered" by Storage
Mid Noon
Time of Day
Mid.
Figure 10.10 Top, non-coincident load of a single household, showing needle peaks due to appliance
activity. Middle - coincident "individual household" load curve, l/100th of 100 such homes, is
smoother because random needle peaks of the various customers in the group "cancel out." Bottom,
non-coincident curve "filtered" by being put in parallel with a battery that fills in the needle peaks,
averaging usage over 30 minute periods.
Copyright 2004 by Marcel Dekker, Inc.
376 Chapter 10
Nearly all small electric consumer sites display some such behavior, with needle peaks
and rapid shifts in demand level (see Chapter 3). A DG unit that will successfully serve
such a load curve must have capacity greater than the high needle peaks. Power quality
problems may develop anyway simply because the unit may not be able to react quickly
enough to shift its output as rapidly as the demand shifts.
Needle peaks occur during all seasons of the year and times of day: they happen more
frequently during peak demand periods, but they occur less frequently, but potentially with
just as high a load, during off peak periods, too. Equipment serving large groups of homes
does not see this type of load curve, as shown in Figure 10.10, middle. The "needles" and
"valleys" of individual customers in the large group cancel out, leaving a coincident load
curve.
The DG unit will see something akin to a coincident load curve as its load, if the load is
connected in parallel with energy storage, so that the energy for the needle peaks is drawn
from the storage unit, not the DG unit, with the energy "paid back" by the DG during the
next valley. This permits the DG unit to run on a smooth, steady schedule, as shown in
Figure 10.10, bottom. The energy storage filters, or "smoothes out," the non-coincident load
curve. It permits a generator to have a smaller capacity than the needle's magnitude and yet
still serve the load with good power quality.
The reader who has not gone through this book thoroughly and finds the needle-peak
behavior or appliances in Figure 10.10 puzzling can look at Chapter 3's comprehensive
discussion of coincident load behavior and why and how it occurs.
Ride-Through Capability
Solar power cannot provide power when the sun is not shining. Wind turbines cannot
provide power when they are becalmed. Even a small hydro may "run dry" in the seasons
between rains. No DG unit can provide power when disassembled for maintenance. The
DG owner who needs power has two choices during these periods:
Use a backup source of power, such as another DG source or the utility
grid
Provide for one's needs with energy stored when the source was
available
The second option is often the most appealing, because another DG unit might not be
any better (more solar power will not help at night, a fossil unit is unacceptable), and the
grid may not be an available or economical option.
The energy storage required to address these needs is quite different from that for energy
stabilization. To begin with, the total amount of energy needed to be stored is much greater.
For example, a system storing 4 kWh (equivalent to what is stored in two heavy-duty "car"-
type lead-acid batteries) will stabilize the DG unit serving the individual household load
shown at the top of Figure 10.10, so that it sees essentially coincident load behavior. But to
make it through a long winter night without power, the DG unit may need to store over 50
kWh.
Beyond this, during those periods when it is the sole source of power, the battery will
have to be able to meet all the needle peak demands, not just help augment a DG unit. It
will need a greater peak capacity, too. Generally, the DG planner will evaluate these energy
storage needs using a load duration curve or an hourly simulation analysis, which would
examine how often over the year, for how long each time, how much energy would have to
be stored in the storage unit for it to do its job. This analysis would also make certain that
there is enough DG capacity to "charge" the energy storage unit during off-peak periods.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 377
Dispatchability
Energy storage is also needed if a basically non-dispatchable energy source, such as PV,
wind, or solar thermal, is to be applied in a dispatchable manner. Viewed purely from the
standpoint of the value of the energy, dispatchable energy is worth more that non-
dispatchable energy. Dispatchability permits the DG owner to commit in advance to certain
production schedules, to provide a promised amount of power at a certain time, regardless
of how much power the DG unit is producing at that time.
The energy storage needs, as well as planning and analysis concepts for achieving
dispatchability, are conceptually similar to those discussed above for ride-through
capability. First, there is a load curve, or production schedule that must be met. Then, a
sufficient amount of storage capacity, and peak capability, must be arranged to reliably
serve this, given the expected production schedule of the DG unit.
The differences in planning storage to achieve dispatchability, versus only ride-through
capability, are related to more complex and demanding definition of "success" and
exacerbated by the planner's knowledge (or lack of it) of the expected production schedule
which the dispatchable generation must meet. In planning ride-through, the planner
generally has a good idea of the load to be served: one household load. Often detailed data
on that load and its behavior over time are available, but even when they are not the target
is clearly identified. By contrast, dispatchability means freedom to vary production
schedule to meet a variety of schedules by changing output times and amounts. The DG-
storage planner must answer the question "how much freedom is needed in being able to
control and shape the net output power schedule?" Once that has been determined, a load
duration curve or hourly load curve analysis of the extremes of the range desired can be
carried out.
Second, the planner studying dispatchability has a more complicated "definition of
success." Ride-through planning had a simple goal: serve the load with no more than some
small probability of failing to meet it. The storage need is dictated by a requirement to serve
the load schedule. By contrast, in planning dispatchability the final decision usually
becomes an economic balancing act: the more the planner spends on more storage, the more
dispatch freedom the resulting system will have, and the more it can earn selling its power.
The question is, where is the best payoif? At this point, the planning becomes a business
planning situation, beyond the scope of this discussion.
Performance Tradeoffs
Energy storage units have eight major areas of performance which can, to a certain extent,
be traded against one another:
Energy density Power density Electrical efficiency Re-charge rate
Control system Service lifetime Physical dimensions Cost
Each type of storage technology carries with it a different interaction in how
compromises must be made among these performance categories. But all have some
interrelationship: obtaining more of one means something must be given up in the others.
Energy density is a measure of the basic capability of the system: How much energy
(kWh) can it store in a specific space - 5 kWh, 50 kWh, or 500 kWh?
Power density is the amount of power the unit can give up, per its unit size. A
particular storage unit that stores 100 kWh might be able to provide power at 10 kW
per hour, meaning it would take ten hours to discharge. Another with the same
capacity might be able to provide power at 200 kW, twenty times that rate, fully
Copyright 2004 by Marcel Dekker, Inc.
378 Chapter 10
expending its stored energy in half an hour. This is usually measured in kilowatts, or
sometimes megajoules.
Electrical efficiency is the percent of power pushed into the unit that is available to
be withdrawn. A unit with 90% efficiency returns 9 kWh of energy for every 10
kWh banked in its storage. Efficiency of most electrical energy units is a function of
both the length of time energy is stored (as a result efficiency is always greater
overnight than, say, over a week).
Re-charge rate is the rate at which power can be pushed into the unit for storage. A
particular 100 kWh storage unit might be able to provide power out at 10 kW, but
accept power in at only 7 kW. This means that while it takes the unit only 10 hours
to fully discharge at its maximum rate, it takes it a bit over 14 hours to be completely
refilled.
Control system design determines the degree to which the energy storage system can
control the voltage and power quality of the AC power it is providing. Storage
systems such as batteries vary their DC output as a function of both the amount of
stored energy they have remaining and the rate of power flow they are providing at
the moment. But this variable impedance of the battery can be compensated for by
proper design of the AC power converter, which can also be designed so it provides
harmonic-free, high-quality power.
Service lifetime varies greatly among available technologies and depends very much
on how the storage units are used. Lead-acid batteries are particularly notorious for
having short lifetimes in applications where they are repeatedly charged and
discharged completely.
Physical dimensions and physical efficiency. The size and weight of a storage
system are often important, particularly in automotive or shipboard applications, but
also in electric power applications where the issue is whether a sufficient storage
capacity will fit at a site. Weight is sometimes a factor in transportation and
structural requirements.
Cost includes both initial and continuing operating costs. Every storage technology
has an initial cost covering the system itself and its installation. Most require
periodic inspection, maintenance, and perhaps re-calibration and replacement of key
parts (bearings and seals) at periodic intervals. Some storage systems simply need to
be replaced every so often.
10.6 DISTRIBUTED RESOURCE COST EVALUATION
Traditional Regulatory DSM Evaluation Frameworks
Traditionally, DSM was viewed in all cases as an optional resource that would be added if it
provided more benefit than its cost. Therefore, regulatory-driven DSM programs were
typically evaluated by utilities and commissions alike using a benefit to cost (B/C) ratio
analysis. Programs that had a higher benefit than cost were identified as feasible. Among
similar DSM options, those with the higher B/C ratio would be the best choice. In selecting
a few DSM options from a large "shopping list," the B/C ratios can be used to rank-order
the DSM programs.
Similarly, consumers, both homeowners and business owners, make decisions about
their energy management options based on comparing the benefit and the cost. There are
substantial differences in the details of these value systems. Utilities and commissions in the
Copyright 2004 by Marcel Dekker, Inc.
Total Resource Cost (TRC) Test
Distributed Resources 379
"DSM era" - the 1980s and 1990s - usually took a long-term (equipment lifetime, or 30-
year investment depreciation) perspective. Most consumers use a very short, "payback
period" perspective: "Does it pay for itself in two to three years?" Modern utilities vary, but
most plan with a shorter horizon than in the 1980s - often only ten years.
19
The benefits of DSM are the avoided costs of supply (generation, transmission, and
distribution capital costs and the costs of energy) and end-use (consumer costs). In a B/C
assessment, the avoided costs represent a savings or benefit. Avoided costs are typically
expressed in $/kW for capacity-related costs only and in 0/kWh for total costs (capacity
and/or energy components). On the other side of the B/C ratio, costs are the expenses
associated with a DSM program and may include the costs borne directly by the
participants as well as the utility. These costs include such items as the technology
(materials and installation), maintenance, and administrative expenses.
But details of just what benefits and costs are included varied in the DSM era, and
may be important today, to some utilities. There are two major types of cost-effectiveness
tests which are utilized to assess DSM options:
Rate Impact Measure (RIM) Test
The TRC test focuses more on minimizing the cost of electricity services while the RIM test
focuses more on minimizing electricity prices to consumers. Each approach represents a
different approach.
Total resource cost methods label a resource option as cost-effective when the present
worth of all benefits over the planning period, no matter who receives them, exceed the
present worth of all costs no matter who pays them. The allocation of costs and benefits
between the utility and energy consumers - who pays and who gains is not a
consideration when applying the TRC test. Neither are questions about whether all, or only
a few, consumers win. If overall more win than lose an energy management program is
considered be "good."
In some regulatory jurisdictions, the TRC test is referred to as the "All Ratepayers Test."
The benefits included in the B/C evaluation are all incremental direct savings accruing from
the program including consumer savings and the electric utility's avoided costs. Costs
include the electric utility's costs related to the program and the total of all consumer costs,
if any. Any loss of revenue by the utility is not included.
Rate Impact Measure looks at how an energy management program affects the utility
rates or the customers total bills (the two are not the same, some states used one
perspective, others the other). Functionally, it differs from TRC mostly in that it considers
changes in the utility revenue as well as costs. For example, suppose that a utility spends
$5,000,000 to implement a program, thereby producing avoided costs for the utility of
$6,000,000. Then it has a positive B/C ratio as evaluated by the TRC test. But suppose that
the program cuts energy usage of consumers so that the utility loses $7,000,000 in revenue
per year. This means that it will eventually have to raise its rates - there are fewer kWh of
sales to cover the total of its costs.
19
The reason is uncertainty about future regulation and technology, the future regulatory backdrop is
not yet completely defined, utilities are naturally unwilling to make long-term commitments until they
know they will be allowed long-term earnings. With respect to technology, utilities learned in the
1980s and 1990s that technological obsolescence is a very real concern with respect to control and
automation systems (as in many energy management systems). Quick payback of three to five years -
within one or two "technology half-lives" - is the basis for their evaluation.
Copyright 2004 by Marcel Dekker, Inc.
380 Chapter 10
The rate impact measure would label such a program as "ineffective" even though it had
a B/C ratio greater than one. The basis for this perspective is the program's impact on
consumers who didn't or couldn't participate. Suppose the program was "swimming pool
pump" control. A homeowner who does not have a swimming pool cannot participate, yet
as a result of this program sees his electric rates increase slightly.
Modern utilities generally prefer to use a business-case approach for evaluation of
distributed resource programs. There, all utility costs, including loss of revenues, are
weighed against all utility savings. If B/C ratio is greater than one, and risk from regulatory,
technological, and other sectors is considered appropriately small, the program will be
approved. Regulators may have concerns about this, particularly if there would be any rate
inequities of the type that the RIM test identifies or if the program produces additional costs
for consumers not taken into account by the utility.
Consumers, who define the market-driven nature of energy management, evaluate
energy management on the basis of how it benefits them alone. Their savings versus their
costs. This is fine, and something everyone has to accept, but it is interesting that this could
have negative rate impacts: if many consumers implemented measures that made sense to
them, which reduced their utility bills, it might have enough revenue-erosion impact on the
utility that it would require it to raises its rates. Market-driven energy management
programs often fail the RIM test. Chapter 6, section 6.4, presents some more detail on TRC
and RIM tests.
The various methods of conducting cost-effectiveness testing should be considered as
different perspectives or various means of assessing a DSM program. Each such measure
provides some alternative insight to the feasibility of a program. The final selection of
any distributed resource programs depends on who is making the decision, how wide they
cast their net on defining costs and savings, and whether they look at only their benefits
and costs or those of some of the possible players or those seen by all of society.
c
o. il
X
HI
(0 Q>
Timing of Expense
initial Continuing
One-time "first"
costs that do
not depend on
usage pattern
One-time "first"
costs that do
depend on
usage pattern
Periodic, on-
going costs that
do not depend
on usage pattern
Periodic, on-
going costs that
do depend
on usage pattern
Figure 10.11 Two characterizations, one based on when costs occur and the other on how they vary
with usage, result in four categories of costs.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 381
Cost and Benefit Evaluation of Distributed Resources
The most important goal of cost evaluation is to include a comprehensive and consistent
(comparable) assessment of all costs involved, both for every DR alternative and for any
T&D options to which they are being compared. Inclusion of all costs is much more
important than in traditional T&D planning studies, because DR options often differ
significantly among themselves, and from traditional T&D options, in categories that are
ignored or "factored out" of traditional T&D planning studies. One example is O&M costs,
which are rarely included in typical T&D studies, because O&M being about the same
among all options is usually left out of consideration. However, O&M differs substantially
among various DG, DR, and T&D options, and thus must be included in comparisons that
cross the DR-T&D boundary. Similarly, traditional T&D options almost never include the
"cost of energy" because, regardless of choice, it is usually the same for all options. In DG
and DR studies, energy (fuel cost) often varies a great deal, and avoided electric purchases
(advantages to consumers, lost revenues to utilities) must be included.
Four Cost Categories
Costs are characterized in two ways depending on whether they are initial or continuing
costs and depending on if they are fixed or variable (Figure 10.11).
Initial costs must be dealt with before a DG unit or energy resource can be used.
For example, the DG unit itself must be obtained, a site must be prepared for it
(foundation, sound abatement walls, etc.), fuel delivery systems installed, electrical
connections made, etc. Costs associated with accomplishing all these tasks are the
initial costs.
Continuing costs are those associated with keeping the unit available and in service.
These include taxes, annual inspections or certifications, fuel, repairs, labor costs
for service personnel, etc. Usually, these are periodic - monthly, annual, bi-annual,
every five years, etc. - expenses that continue as long as the unit is left in service.
Fixed costs do not vary as a function of the amount or patterns of usage. The cost of
the basic DG unit itself is usually a fixed cost: One must buy or lease the unit
regardless of whether it is used every hour of every day or not at all.
20
Likewise,
annually required inspections or certifications, as well as annual property taxes, are
based on the value of the property and are "fixed," even if they vary year by year,
because the unit depreciates in value, etc.
21
Variable costs are those that vary as a function of the amount or type of usage. Fuel
is a big component of variable cost: the more power produced, the more fuel used.
Other variable costs include certain O&M - those that increase or change
depending on the amount of usage.
20
Of course, a lease could be based on usage - in a manner similar to leasing a car on a "per mile" basis
- which would make the DG unit cost a variable cost.
21
This is a subtle but important distinction. In some municipalities, local property tax on machinery is
based upon assessed value, which decreases with equipment age, and thus the annual taxes due on a
DG or other machine decrease each year. Although these costs do change from year to year, they are
not variable costs, because they do not alter as a function of usage. Once the user decides to buy and
install the DG unit, he has committed to paying all those future taxes, a "fixed" amount regardless of
his usage.
Copyright 2004 by Marcel Dekker, Inc.
382 Chapter 10
Table 10.4 Example DG Cost Analysis - 1,100 kW Reciprocating Natural-Gas Fired AC
Generator Unit Used 3,000 Hours per Year
Initial Fixed Costs
Cost of DG unit itself
Tools and diagnostic equipment
Shipping & insurance
Site prep., foundation, control house
Fuel delivery lines, equipment, meters
Backup tank and equipment (10,000 gal.
propane storage and propane rollover,
supply good for 100 hr. operation)
Electric controls, protection (non-grid
operation)
Design, construction permit, survey,
inspection fees, etc.
Backup fuel, 10,000 gal at 800
Total "Cost of the Unit"
Initial Variable Cost
Legal and fees for emissions permit
for "more than 1,001 hr./yr." operation
TOTAL INITIAL COST
Annual Fixed Costs
Property taxes, .70 per dollar on
$550,000 assessed value
Annual mechanical & electrical
inspection (req. for warranty)
Annual Variable Costs
Fuel: natural gas at $3.05 per MBTU
12,000 BTU/ kWh heat rate, for
3,000 hours at full output
Maintenance, estimated at .30 per kWh
Tear-do wn/rebuild every 10,000 hours
(every three years in this case) $72,000
ANNUAL COSTS 3,300,000 kWh/yr.
($431.10/kW) $474,200
$7,800
$8,200
$15,000
$13,500
$50,000
$18,500
$8,400
$8,000
$603,600
$3,200
$606,800 ($551.64/kW)
$3,474
$1,370
$120,780
$9,900
$24,000 per year
$154,680 (4.68# kWh)
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 383
Table 10.4 shows the costs associated with a hypothetical, but very realistic, V-16
natural gas/propane reciprocating piston driven AC generator capable of producing 1,100
kW output (net) serving up to 1,100 kW of peak load. Although this particular DG
technology lacks the cachet of micro-turbines and fuel cells, it is used here to emphasize
that pistons should not be "counted out" as an equal, or perhaps even superior, DG
technology for application in many cases. The table gives costs for this unit, identified as
fixed and variable, as well as initial and continuing, for an application involving an
expected 3,000 hours per year of full (1,100 kW) output, in an isolated (stand-alone)
application - i.e., not connected to the utility system. This is based on an actual case for a
ranch in an isolated area of the pampas in South America which needed power for a series
of 75 kW irrigation pumps to be run only about 3,000 hours per year during the crop
growing season. The examples in this and subsequent chapters used a gas price of $3.05 per
million BTU. This is a lower price (by about 20%) than the annual market price at the time
of this writing, but thus represents a slightly optimistic scenario.
Assessing Initial and Continuing Costs
How costs are to be considered depends on the financial decision-making context of the
evaluation (see Chapters 6 and 29 and the earlier part of this section). Table 10.5 shows 20-
year operating cost for the DG unit from Table 10.4 when run 3,000 hours per the irrigation
schedule expected at the isolated ranch where this unit is being planned to power a series of
electric pumps used only during the growing season. Total outlay over 20 years is
$3,797,280, production 66,000,000 kWh, or 5.750 per kilowatt hour if undiscounted cost
and undiscounted production are used to compute cost/kWh. The cost per
Table 10.5 Twenty-Year Evaluation of Table 10.4's DG Unit for 3,000 Hours per Year
Operation at Full 1,100 kW Output
Study
Year
Unit
&Site
0 $610,274
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
$3,474
MainL
&Insp.
$11,270
$11,270
$83,270
$11,270
$11,270
$83,270
$11,270
$11,270
$83,270
$11,270
$11,270
$83,270
$11,270
$11,270
$83,270
$11,270
$11,270
$83,270
$11,270
$59,270
Fuel
Cost
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
$120,780
TOTAL
Evaluated cost = 5.75 cents per kilowatt
Annual
Cost
$742.324
$135,524
$207,524
$135,524
$135,524
$207,524
$135,524
$135,524
$207^24
$135,524
$135,524
$207,524
$135,524
$135,524
$207,524
$135,524
$135,524
$207,524
$135,524
$183.524
$3,797,280
PW Disc.
Factor Cost
1.000
0.900
0.810
0.729
0.656
0.590
0.531
0.478
0.430
0.387
0.349
0.314
0.282
0.254
0.229
0.206
0.185
0.167
0.150
0.135
$742,324
$121,972
$168,094
$98,797
$88,917
$122,541
$72,023
$64,821
$89,332
$52,505
$47,254
$65,123
$38,276
$34,448
$47,475
$27,903
$25,113
$34,609
$20,341
$24.791
$1,986,661
NetkWh
Production
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3.300.000
Discntd.
Production
3,300,000
2,970,000
2,673,000
2,405,700
2,165,130
1,948,617
1,753,755
1,578,380
1,420,542
1,278,488
1,150,639
1,035,575
932,017
838,816
754,934
679,441
611,497
550,347
495,312
445.781
66,000,00028,987,970
hour (undiscounted) or 6.85 cents (discounted)
Copyright 2004 by Marcel Dekker, Inc.
384 Chapter 10
Table 10.6 Twenty-Year Evaluation of Buying 3,300,000 kWh/Year at 6.85 Cents, with
No Discounted Production, Leading to Evaluated Cost of Only 3.01 Cents
Study Unit
Year & Site
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
TOTALS
SO
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
Evaluated cost =
Evaluated cost =
Maint.
&tasp.
$0
$0
$0
$0
$0
$0
$0
so
$0
so
so
so
so
so
so
so
so
so
so
$0
3,300 MWh Annual
X $68.53 Cost
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
$226,220
PW Disc. Net
Factor Cost
1
0.9
0.81
0.729
0.656
0.59
0.531
0.478
0.43
0.387
0.349
0.314
0.282
0.254
0.229
0.206
0.185
0.167
0.15
0.135
discounted cost/undiscounted production =
$226,220
$203,598
$183,238
$164,914
$148,400
$133,470
$120,123
$108,133
$97,274
$87,547
$78,951
$71,033
$63,794
$57,460
$51,804
$46,601
$41,851
$37,779
$33,933
$30.540
$1,986,661
kWh
Production
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3,300,000
3.300.000
66,000,000
$1,986,661/66,000,000 =
Discntd
Production
3,300,000
2,970,000
2,673,000
2,405,700
2,165,130
1,948,617
1,753,755
1,578,380
1,420,542
1,278,488
1,150,639
1,035,575
932,017
838,816
754,934
679,441
611,497
550,347
495,312
445.781
28,987,971
3.01 cents
discounted cost/discounted production = $1,986,661/28,987,971 = 6.85 cents
kilowatt, when evaluated with discounted costs and discounted kilowatts, is
higher than when evaluated with both undiscounted, indicating that a greater proportion of
the costs lies in the short term than does the output. The outp.it is spread evenly over the 20-
year period, while the costs tend to cluster near the beginning, particularly in the first year
in this case.
22
This is typical in DG studies but not in all DR. Usually the cost analysis of a
large capital asset is done in a discounted manner, because some of the costs and much of
the production will occur far into the future (see Chapter 5's and 29's discussions of
discounting factors used in economic analysis). Thus, 6.85 cents is the more realistic value
to interpret from Table 10.5.
Always Use Discounted Costs and Energies
Table 10.6 shows a bogus analysis of the same DG unit, using a "bias trick" discussed in
Chapter 29 (and briefly in Chapter 5) in which undiscounted production over the period is
used along with discounted costs for the period to determine the per-kilowatt-hour cost.
This results in a computed cost per kilowatt of only 3.01 cents, an impressively low value,
one reason, perhaps, why that this "mistake" is sometimes seen in DG and DR studies.
22
Note that the 11.70 value obtained is not the same value one would get if one calculated an annual
cost per kilowatt, i.e., undiscounted annual cost over undiscounted annual output, for each year, then
discounted those ratios depending on their year's PW factor, and then averaged them to obtain one
value. That approach gives the same 8.40 per kilowatt hour as doing no discounting at all.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 385
Many planners understand and accept the discounting of future costs in a long-term
planning study, but are uncomfortable with the discounting of future production as done to
get the 6.85 cent figure in the example above. Discounting of future costs is an almost
universal feature of least-cost regulated utility planning, to the extent that planners
experienced with only utility planning may believe it is used in all planning (it isn't,
although the time-value of money is always addressed in some manner).
Most traditional T&D planning studies do not involve assessment of energy or
production or long-term benefit, but instead work to minimize the cost to satisfy an
obligation-to-serve criterion. As such, they deal with a discountable quantity only on the
cost side of the ledger. Thus, T&D planners have not seen situations that demonstration
why it is absolutely necessary to discount future production if future costs are discounted.
By contrast, DG and DR studies are nearly always looking at optional or discretionary
resources. Neither is needed to meet a minimal obligation to serve. Both must justify
themselves on the basis of benefits that outweigh costs. Even when those benefits are
measured as "nega-watts" or "mega-watts" rather than dollars, they must be discounted in
company with the costs for any benefits versus cost analysis to be valid. The easiest way to
see that production kilowatt hours must be discounted like cost is by use of a
counterexample, shown in Table 10.6. There, instead of distributed generation power, the
3,300,000 kWh needed by the rancher is purchased at a flat rate of 6.85 cents per kilowatt
over the 20 year period. Annual cost is simply .0685250 x 3,300,000 kWh, or $226,220 per
year. If evaluated on the basis of discounted cost for the period divided by undiscounted
kilowatt hours, the cost/kWh works out to 3.010, an absurdity: buy power for twenty years
at 6.850/kWh and the average cost works out to 3.010/kWh.
In summary, there are four ways one can compute "cost per kilowatt hour."
1. Undiscounted cost over undiscounted production. This is completely legitimate
but neglects "present worth value of money" considerations in planning - the
reason that discounting is done. It evaluates cost of the DG case discussed above
as 5.750/kWh. That is correct in one sense - over the period total cost divided by
total production is 5.750/kWh. However, most of the money has to be spent up
front, whereas energy production is constant over the period, a fact missed
because discounting is not used at all in this particular case.
2. Discounted cost over discounted production. This provides a legitimate
evaluation of what cost/kWh is, taking the time-value-money into account all
around. He^e, the evaluated cost is 6.850/kWh, a value that fairly represents the
time values of money spent and product gained. Usually, this approach is the
"most valid" way to perform such an analysis.
3. Discounted cost over undiscounted production underestimates the cost/unit of
any scenario with up-front costs but continuing benefits. This approach should
never be used in DG and DR planning - it is always a spurious evaluation.
However, planners should look for it because it is often used: such an approach
makes its way into many proponent studies of DG and DR, either by ignorance or
deliberate intent to confuse.
4. Undiscounted cost over discounted production is equally wrong, but seldom used.
Its use would artificially raise the per unit cost, making DR look less attractive.
Chapter 29 discusses "bogus" studies, including this and other mistakes often made in
planning, as well as various "tricks" that are sometimes deliberately applied to distort the
results of planning studies.
Copyright 2004 by Marcel Dekker, Inc.
386 Chapter 10
Table 10.7 One-Page Summary of Chapter 10
Distributed resources are customer-side or customer-site mechanisms for either
reducing need for or increasing efficiency of energy usage.
Distributed resources are often a cost-effective way of solving an energy
consumer's needs for either of the Two-Qs (energy or reliability).
DR often improves customer reliability and consequently has additional
value that should be taken into account.
UPS systems are a legitimate DR measure, one that exemplifies DR's
reliability impacts.
Electric utilities are often viewed as responsible for implementation and
administration of DR programs because their monopoly franchise makes them
"steward" of energy management practices in their service territory.
Distribution generation (DG) and distributed storage (DS) are subsets of DR
which are implementable either on the customer or utility side of the meter.
DSM will defer transmission additions, but will not defer most distribution
projects. Load reductions at the transmission level basically accumulate to look
like a reduction in load growth rate. As a result, DSM can delay when
transmission projects are needed. At the distribution level, DSM looks different
and has a different effect (see below). It will have no impact on the need for
projects to extend service to new areas and new customers.
DR reduces load density. DSM reduces the load, but not the area that the
distribution system must cover, the number of customers that must be connected
to the system, or the date when any new customer will move in and demand
power. As a result, DR's impact on distribution is best capitalized upon by
reducing capacity density of the distribution.
Maximizing T&D savings from DR requires re-design of the T&D system.
That can require a lot of re-planning, but it usually doubles or triples the
distribution-level savings, increasing the overall T&D savings by 50% or more.
DR measures that individually do not pass "screening tests" for cost-
effectiveness may be cost-effective in an overall plan. If the DSM program is
just shy of having enough load reduction to produce a deferral or cancellation, it
may be cost-effective to include "marginal" programs to reach the reduction
target.
DR peak reduction measures like load control and re-scheduling are usually
more cost-effective at reducing T&D needs than energy conservation alone. The
reason: T&D requirements are linked almost entirely to peak load level. Reducing
energy off-peak, which is the additional benefit conservation brings, may cut
annual losses costs, but these are a minor aspect of T&D costs.
DR and feeder switching interact Properly planned, they can be used to
support one another. In particular, feeder re-switching can be planned in
conjunction with DSM targeted to augment and multiply the effectiveness of the
DSM load reductions.
The best areas for DR targeting are not necessarily where T&D costs are
highest. DSM should be targeted where the combination of DR cost effectiveness
and T&D avoided costs are best.
Copyright 2004 by Marcel Dekker, Inc.
Distributed Resources 387
10.7 SUMMARY
A utility can affect measures on the consumer side of the electric meter that can reduce
energy usage, cut peak load, and improve the reliability of service and usability of its
product as seen by the consumer. Some of these programs are implemented "by market
forces" in spite of what the utility may desire or actions it might take. However,
distributed resources have a potential benefit that should not be ignored. Impact on all
aspects of the utility's performance, including revenues and consumer satisfaction, must
be taken into account. One important point to keep in mind is that a T&D system is the
ultimate distributed resource. At the customer end, the T&D system reaches every
customer. That is about as distributed as a resource can get. The point is that DR planning
should always include the T&D system. Table 10.7 gives a one-page summary of this
chapter's key points.
BIBLIOGRAPHY
J. M. Studebaker, The Electricity Purchasing Handbook, Penwell, Tulsa, 1997.
L. J. Vogt and D. A. Conner, Electrical Energy Management, Lexington Books, Toronto, 1977.
H. L. Willis, Spatial Electric Load Forecasting Second Edition, Marcel Dekker, New York, 2002.
H. L. Willis and G. B. Rackliffe, Introduction to Integrated Resource T&D Planning, ABB Systems
Control, Gary, NC, 1994
H. L. Willis and W. G. Scott, Distributed Power Generation - Planning and Evaluation, Marcel
Dekker, New York, 2000.
Copyright 2004 by Marcel Dekker, Inc.