Mark McGranaghan, Bill Roettger

Economic Evaluation
of Power Quality
E lectric power quality disturbances can have
significant economic consequences for many
different types of facilities. A wide variety of solu-
tion technologies exist for mitigating the conse-
quences of such disturbances, and a methodology
for performing a comparative economic analysis is
featured in this article.
Different technologies are evaluated by estimating
the improved performance that can be expected after
the technology has been applied. The power quality
cost savings are calculated for each technology along
with the costs of applying the technology. The net ben-
efits, expressed in terms of annual costs, are shown as a
means of comparing the various technologies.

Economic Impacts
of Power Quality
The costs associated with power outages can be tre-
mendous. Manufacturing facilities have costs rang-
ing from $10,000 to millions of dollars associated
with a single interruption to the process. The costs to
commercial facilities (banks, data centers, customer
service centers, etc.) can be just as high if not higher.
Unfortunately, these facilities can be sensitive to a
wider range of power quality disturbances than just
outages that are counted in utility reliability statistics.
Momentary interruptions or voltage sags lasting less
than 100 ms can have the same impact as an outage
lasting many minutes.
This has resulted in a wide variety of technologies
for equipment protection and improving power qual-
ity. Utilities must also evaluate the need to improve
the quality of supply if there are large numbers of
customers impacted by power quality variations.
The evaluation of power quality improvement al- ®
© DIGITAL VISION LTD
ternatives is an exercise in economics. Facility man-
agers and utility engineers must evaluate the
economic impacts of the power quality variations
against the costs of improving performance for the different al- Improving facility performance during power quality varia-
ternatives. The best choice will depend on the costs of the tions can result in significant savings and can be a competitive
problem and the total operating costs of the various solutions. advantage. Therefore, it is important for customers and suppli-
Note that the solutions should include options for improving ers to work together in identifying the best alternative for
performance on the utility supply system. achieving the required level of performance.
The economic evaluation methodology described here con-
This article is part of a series of articles on power quality appearing in this issue
sists of the following four steps:
l Characterize the system power quality performance
and the the August, September, and November 2001 issues of IEEE Power Engi-
neering Review. M. McGranaghan and B. Roettger are with Electrotek Con- l Estimate the costs associated with the power quality varia-

cepts, Inc. tions

8 0272-1724/02/$17.00©2002 IEEE IEEE Power Engineering Review, February 2002
 l Characterize the solution alternatives in terms of costs and Standard 1159, Table 2-1, where voltage sags lasting between
effectiveness 0.5 and 30 cycles are identified as “instantaneous,” those lasting
l Perform the comparative economic analysis. between 30 cycles and 3 seconds are identified as “momentary,”
and those lasting between 3 s and 1 min are defined as “tempo-
Characterizing Power Quality Performance rary.” In addition to magnitude and duration, it is often impor-
The first step in the process is to understand the kinds of distur- tant to identify the number of phases involved in the sag, since
bances that occur on the system and the frequency of their oc-
currence. Voltage sags and momentary interruptions will
usually be the most important in terms of their impact on facility
operations, and these will be our focus in this analysis.
When defining performance indices, it is important to under-
stand the characteristics of disturbances that can cause customer
equipment to misoperate. Load susceptibility to rms voltage E D
variations is very dependent on the specific load type, control
settings, and application. Consequently, it is often very difficult
to distinguish which characteristics of a given rms variation are
115 kV C
likely to cause equipment to misoperate. Equipment susceptibil-
ity to rms variations can be divided into three main categories: System B
l Equipment sensitive to only the voltage during an rms
SLGF
variation. This group includes devices such as F A
undervoltage relays, process controls, motor drive con-
trols, and many types of automated machines (e.g., semi- Transmission
conductor manufacturing equipment). Devices in this
group are sensitive to the minimum (or maximum) voltage 13.8 kV Distribution
magnitude experienced during a sag (or swell). The dura-
Circuit
tion of the disturbance is usually of secondary importance Breakers
1 2 3 4
for these devices.
l Equipment sensitive to both the magnitude and duration of

an rms variation. This group includes virtually all equip- 480 V

ment that uses electronic power supplies. Such equipment
misoperates or fails when the power supply output voltage
drops below specified values. Thus, the important charac-
teristic for this type of equipment is the duration that the
rms voltage is below a specified threshold at which the
Fuse
equipment trips.
l Equipment sensitive to characteristics other than magnitude

and duration. Some devices are affected by other rms varia-

tion characteristics such as the phase unbalance during the SLGF
disturbance, the point-on-wave at which the variation is ini- Figure 1. Example of fault locations that could cause a voltage sag that would
tiated, or any transient oscillations occurring during the dis- affect customer equipment
turbance. These character-
istics are more subtle than 13.8 kV Voltage Sags (Line-Line)- 60 Second Aggregation
magnitude and duration, 38 Events (Vbase = 13.8 kV) - Circled Data Points Are Plant Outages
and their impacts are much 1.2
more difficult to general-
ize. As a result, the rms
variation performance in- 1
dices defined here focus on
the more common magni- 0.8 Semi F47 Curve
tude and duration charac-
Volts (PU)

teristics.
We will use the industry 0.6
standard method of characteriz-
ing voltage sags using the mini-
mum voltage magnitude and the 0.4
duration of the sag (time that the
voltage is below specified 0.2
thresholds). The sag duration is
determined by the length of
time required for the protective 0
0.00 0.01 0.10 1.00 10.00 100.00 1000.00
devices to detect the fault and
Seconds
open. One useful duration char-
acterization is defined in IEEE Figure 2. Year 2000 rms voltage variations and process disruptions

IEEE Power Engineering Review, February 2002 9

 Table 1. Example of weighting factors for different
voltage sag magnitudes
Event Category Weighting for
Economic Analysis
Interruption 1.0
Sag with min. voltage below 50% 0.8
Sag with min. voltage between 50 and 70% 0.4
Sag with min. voltage between 70 and 90% 0.1
this can affect both equipment sensitivity and solution technol-
ogy capabilities.
The magnitude of the voltage at a given facility during a volt-
age sag event will depend on the location of the facility with re-
spect to the fault location on the system. Unfortunately, faults
over a wide part of the system can often affect the operation of a
facility. This is often called the “area of vulnerability” for the fa-
cility sensitivity to voltage sags. Figure 1 illustrates examples of
fault locations that could cause misoperation of equipment at a
facility.
Once the area of vulnerability is determined for a specific
evaluation, the expected number of sags of a given severity can
be calculated based on expected fault performance of the trans-
mission and distribution circuits within the area of vulnerability.
Of course, the easiest way to characterize performance is
with power quality monitoring. The voltage sags can be charac-
terized over time. Figure 2 illustrates measured voltage sags that
have been characterized by their magnitudes and durations and
then plotted along with the equipment ride through standard for
semiconductor manufacturing equipment (SEMI F47 Stan-
dard). The circled data points represent those events that re-
sulted in a process interruption. As expected, the deeper the
voltage sag and/or the longer the sag lasts, the more likely it is
that the sag results in a process interruption. It is interesting that
sensitivity of the process for the example system (candy wrap-
per manufacturer) seems to correspond closely to the sensitivity
curve specified for semiconductor manufacturing equipment.
There are several uncircled sags on the plot in Figure 2 that
are at first surprising. These uncircled sags occurred while the
plant was not operating either because it was already down due
to a previously experienced sag or it was down for some other
reason like preventative maintenance.
Estimating the Costs for
Power Quality Variations
The costs associated with sag events can vary significantly from
nearly zero to several million dollars per event. The cost will
vary not only among different industry types and individual fa-
cilities but also with market conditions. Higher costs are typi-
cally experienced if the end product is in short supply and there
is limited ability to make up for the lost production. Not all costs
are easily quantified or truly reflect the urgency of avoiding the
consequences of a voltage sag event. This concept was punctu-
ated for the author during a recent visit to a waste water treat-
ment facility where the consequences of tripping a 100 hp dc
motor as a result of a voltage sag were indeed a messy affair.
The cost of a power quality disturbance can be captured pri-
marily through three major categories:
l Product-related losses, such as loss of product/materials,
lost production capacity, disposal charges, increased in-
ventory requirements, etc.
10
Table 2. Combining weighting factors with expected voltage sag
performance to determine total costs of power quality variations
Event Weighting for No. Events per Total Equivalent
Category Economic Year Interruptions
Analysis
Interruption 1.0 5 5
Sag with min.
voltage below
50% 0.8 3 2.4
Sag with min.
voltage
between 50 and
70% 0.4 15 6
Sag with min.
voltage
between 70 and
90% 0.1 35 3.5
Total 16.9
l Labor-related losses, such as idled employees, overtime,
cleanup, repair, etc.
l Ancillary costs such as damaged equipment, lost opportu-
nity cost, and penalties due to shipping delays.
Focusing on these three categories will facilitate the devel-
opment of a detailed list of all costs and savings associated
with a power quality disturbance. One can also refer to IEEE
Standard 1346, Appendix A, for a more detailed explanation of
the factors to be considered in determining the cost of power
quality disturbances.
Costs will typically vary with the severity (both magnitude
and duration) of the power quality disturbance. This relationship
can often be defined by a matrix of weighting factors. The
weighting factors are developed using the cost of a momentary
interruption as the base. Usually, a momentary interruption will
cause a disruption to any load or process that is not specifically
protected with some type of energy storage technology. Voltage
sags and other power quality variations will always have an im-
pact that is some portion of this total shutdown. The base costs
associated with a momentary interruption can be designated as
Ci. If a voltage sag to 40% causes 80% of the economic impact
that a momentary interruption causes, then the weighting factor
for a 40% sag would be 0.8. Similarly, if a sag to 75% only re-
sults in 10% of the costs that an interruption causes, then the
weighting factor is 0.1.
After the weighting factors are applied to an event, the costs
of the event are expressed in per unit of the cost of a momentary
interruption. The weighted events can then be summed, and the
total is the total cost of all the events expressed in the number of
equivalent momentary interruptions.
Table 1 provides an example of weighting factors that were
used for one investigation. The weighting factors can be further
expanded to differentiate between sags that affect all three
phases and sags that only affect one or two phases. Table 2 com-
bines the weightings with expected performance to determine a
total annual cost associated with voltage sags and interruptions.
The cost is 16.9 times the cost of an interruption. If an interrup-
tion costs $40,000 the total costs associated with voltage sags
and interruptions would be $676,000 per year.
IEEE Power Engineering Review, February 2002
 Table 4. Effectiveness of the power quality improvement
Increasing Cost
options for a particular example case
4 - Utility 3 - Overall 2 - Controls 1 - Equipment
Solutions Protection Protection Specifications Interruption Sags with Sags with Sags with
Inside Plant 1 (%) Min. Min. Min.
Voltage < Voltage Voltage
50% (%) 50-70% 70-90%
2 Controls (%)
3 CVT (controls) 0 20 70 100
4 Motors
Group of Dynamic sag
Utility Machines corrector/DVR 0 20 90 100
Source Other Loads
Flywheel ride
through
Sensitive Process Machine technologies 70 100 100 100
Figure 3. Different protection options for improving performance during UPS (battery
power quality variations ride through
technologies) 100 100 100 100
Static switch 100 80 70 50
Table 3. Example costs for different types of power quality
improvement technologies Fast transfer
switch 80 70 60 40
Alternative Typical Cost ($) Operating and
Category Maintenance Costs (% of
initial costs per year)
In general, the costs of these solutions increases as the power
Controls protection (< 5 kVA) level of the load that must be protected increases (Figure 3). This
CVTs 1,000/kVA 10 means that economies usually can be achieved if sensitive
equipment or controls can be isolated and protected individually
UPS 500/kVA 25
from equipment that does not need protection.
Dynamic sag Ideally, the appropriate ride through will be part of the equip-
corrector 250/kVA 5
ment design (thus the motivation for the new semiconductor
Machine protection (10-300 kVA) manufacturing equipment standard). However it is often not a
practical option when trying to improve the operation of an ex-
UPS 500/kVA 15
isting facility. Original equipment manufacturers may also be
Flywheel 500/kVA 7 reluctant to incorporate “voltage sag ride through” capabilities
Dynmic sag in their equipment because the added costs may not translate
corrector 200/kVA 5 into an appropriate perceived value for many of their customers.
Manufacturers are more inclined to offer a “voltage sag ride
Facility protection (2-10 MVA)
through” option that could be purchased by those customers
UPS 500/kVA 15 with the need.
Flywheel 500/kVA 5 Each solution technology needs to be characterized in terms
of cost and effectiveness. In broad terms, the solution cost
DVR (50% voltage should include initial procurement and installation expenses,
boost) 300/kVA 5
operating and maintenance expenses, and any disposal and/or
Static switch (10 salvage value considerations. A thorough evaluation would in-
MVA) 600,000 5 clude less obvious costs such as real-estate or space-related ex-
Fast transfer penses and tax considerations. The cost of the extra space
switch (10 MVA) 150,000 5 requirements can be incorporated as a space rental charge and
included with other annual operating expenses. Tax consider-
ations may have several components, and the net benefit or cost
Characterizing the Cost and can also be included in with other annual operating expenses.
Effectiveness for Solution Alternatives Table 3 provides an example of initial costs and annual operat-
A wide range of potential solutions, with varying degrees of cost ing costs for some general technologies used to improve perfor-
and effectiveness, are available to mitigate the consequences as- mance for voltage sags and interruptions. These costs are
sociated with poor power quality. Power quality solutions can be provided for use in the example and should not be considered in-
applied at different levels or locations within the electrical sys- dicative of any particular product.
tem. The four major options are: Besides the costs, the solution effectiveness of each alterna-
l Supply system modifications and equipment that affect tive must be quantified in terms of the performance improve-
multiple customers ment that can be achieved. Solution effectiveness, like power
l Service entrance technologies that affect a single targeted quality costs, typically will vary with the severity of the power
customer quality disturbance. This relationship can be defined by a matrix
l Power conditioning at equipment locations within a facil- of “% sags avoided” values. Table 4 illustrates this concept for
ity the example technologies from Table 3 as they might apply to
l Equipment specifications and design. the specific example of the candy wrapper manufacturer.
IEEE Power Engineering Review, February 2002 11
Performing Comparative Economic Analysis
The process of comparing the different alternatives for improv-
ing performance involves determining the total annual cost for
each alternative, including both the costs associated with the
power quality variations (remember that the solutions do not
typically eliminate these costs completely) and the annualized
costs of implementing the solution. The objective is to minimize
these annual costs (PQ costs + solution costs).
Comparing the different power quality solution alternatives
in terms of their total annual costs (annual power quality costs +
annual power quality solution costs) identifies those solutions
with lower costs that warrant more detailed investigations. The
“do nothing” solution is generally included in the comparative
analysis and is typically identified as the base case. The “do
nothing” solution has a zero annual power quality solution cost
but has the highest annual power quality costs.
Many of the costs (power quality and O&M) are by their na-
ture annual costs. The costs associated with purchasing and in-
stalling various solution technologies are one time up-front
costs that can be annualized using an appropriate interest rate
and assumed lifetime or evaluation period.
Figure 4 gives an example of this type of analysis for a candy
wrapper manufacturer. The facility has a total load of 5 MW, but
only about 2 MW of load must be protected to avoid production
Figure 4. Example of comparing solution alternatives with the base case using
disruptions. The voltage sag performance was given in Table 2. total annualized costs
The costs for an interruption are $40,000 per event, and the costs
for voltage sags are based on the weighting factors given previ- A thorough economic analysis would also include a parame-
ously. The six options given in Table 4 are analyzed, and the an- ter sensitivity evaluation where the uncertain parameters could
nual costs are presented. The annualized costs are calculated be characterized by minimum, maximum and average values.
based on a 15-year life and an interest rate of 10%. The probabilistic nature of power quality events coupled with
It is interesting to note that all of the options reduce the total market conditions that can greatly affect power quality costs
annual costs (in other words, any of these options would have a would typically justify the need for the sensitivity evaluations.
net benefit to the facility with the assumed interest rate and life-
time when compared to the existing conditions). It is also inter-
esting that the best solution in this case involves applying References
equipment on the utility side (fast transfer switch). However, R. Gilleskie, et al., “Economic evaluation procedure for assessing
this has a major assumption that a backup feeder would be avail- power quality improvement alternatives,” presented at the 1997 PQA
able and that there would be no charge from the utility for pro- North America Conference, Columbus, Ohio.
viding a connection to this backup feeder except the equipment
IEEE Recommended Practice for Monitoring Electric Power Quality,
and operating costs.
IEEE Standard 1159, 1995.
More commonly, the solution would be implemented in the
facility and either a dynamic sag corrector or flywheel-based Specification for Semiconductor Processing Equipment Voltage Sag
standby power supply might make sense for protecting the 2 Immunity, SEMI F47-0200.
MW of sensitive loads. In this case, protecting just the controls IEEE Recommended Practice for Evaluating Electric Power System
with CVTs does not provide the best solution because the ma- Compatibility with Electronic Process Equipment, IEEE Standard
chines themselves are sensitive to voltage sags. 1346, 1998.

2002 International Power Quality Conference

21-25 October 2002, Singapore
The first International Power Quality Conference (IPQC 2002) will be held 21-25 October 2002 in Singapore. It is jointly organized by the In-
stitution of Engineers Singapore (IES) and Singapore Power Ltd. and co-organized by the Institution of Electrical Engineers (IEE) Singapore
Centre, the IEEE PES Singapore Chapter, the National University of Singapore (NUS) and the Nanyang Technological University (NTU). The
conference provides a forum for experts, practitioners, researchers and others concerned with power quality to discuss the latest technologies
and management approach for improving power quality and system reliability in the electricity supply.
Conference topics include power quality management and planning, power quality audit and economics, power quality measurement and
monitoring, power quality standards, equipment design in respect of power quality, electrical distribution network, power system harmonics,
lightning protection and surge suppression technologies, distributed generation, power quality enhancement system and devices (e.g. UPS,
DVR, etc.), power supply reliability, power quality requirements of mission critical facilities, grounding, shielding and other EMC issues, case
studies on power quality related issues, and other topics related to power quality.
For more information, contact the conference secretariat, Institution of Engineers Singapore, 70 Bukit Tinggi Road, Singapore 289758, +65
469 5000, fax +65 467 1108, e-mail iesnet@singnet.com.sg.

12 IEEE Power Engineering Review, February 2002