20040622Sterling Direct.pdf

HECEJVED
F ! L~E D
r:1tLJ
('"'
L.:i
Yr:
~)
/hi - c::
~,.'"
;1 'w'
! , '...) . .. .. . ..;,. .
UriL j r iES CO ("j ISSiON
BEFORE TH
IDAHO PUBLIC UTiliTIES COMMISSION
IN THE MATTER OF THE APPLICATION
OF A VISTA CORPORATION FOR
AUTHORITY TO INCREASE ITS RATES
AND CHARGES FOR ELECTRIC AND
NA TU RAl GAS SERVICE TO ELECTRIC
AND NATURAL GAS CUSTOMERS IN
THE STATE OF IDAHO.
) CASE NO. AVU-O4-) AVU-O4-
DIRECT TESTIMONY OF RICK STERLING
IDAHO PUBLIC UTiliTIES COMMISSION
JUNE 21 , 2004
Please state your name and business address
for the record.
My name is Rick Sterling.My business
address is 472 West Washington Street, Boise, Idaho.
By whom are you employed and in what
capaci ty?
I am employed by the Idaho Public Utilities
Commission as a Staff engineer.
What is your educational and professional
background?
I received a Bachelor of Science degree in
Civil Engineering from the University of Idaho in 1981 and
a Master of Science degree in Civil Engineering from the
University of Idaho in 1983.I worked for the Idaho
Department of Water Resources from 1983 to 1994.In 1988,
I became licensed in Idaho as a registered professional
Civil Engineer.I began working at the Idaho Public
Utilities Commission in 1994.My duties at the Commission
include analysis of utility applications and customer
petitions.
What is the purpose of your testimony in this
proceeding?
The first purpose of my testimony is to
discuss the Company s weather normalization.Another
purpose is to detail the test year power supply
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 1
adjustments proposed by Avista and describe my
investigation of those adjustments.I will also discuss
Avista s investments in the Coyote Springs 2 , Kettle Falls
CT and Boulder Park proj ects.
Are you sponsorlng any exhibits?
I am sponsoring Staff Exhibit Nos. 128Yes.
through 131.
Please summarlze your testimony.
My review of the Company s weather
normalization consisted of replicating the results
obtained by the Company, in addition to evaluating the
effects of varying the weather data and period of record
used in the Company s analysis.I conclude that the
weather normalization performed by Avista is accurate and
reasonable, and recommend that it be accepted.
The test year power supply adjustments
proposed by the Company in this case consist of
contractual changes due to new or expiring contracts, and
changes due to specific contract rates or terms and power
supply cost adjustments for normalized loads and water
condi t ions.As a result of these adjustments, the Company
has proposed a net, system-wide decrease in test year
expenses of $30.5 million.
My investigation of test year power supply
adjustments included evaluation of known and measurable
CASE NOS. AVU-04-1/AVU-04-06/21/04
STERLING, R.
STAFF
(Di) 2
changes through August 2005 and replication of the
Company s dispatch simulation model and evaluation of its
inputs and assumptions.I specifically focused on short-
term sales and purchases and long-term wholesale sales and
purchase contracts.
I found that the power supply pro forma
adjustments proposed by the Company adequately reflect
known and measurable changes that will occur through
August 2005.I also found that the dispatch simulation
model adequately reflects anticipated dispatch of Company
resources, the availability and price of regional surplus
energy, the normalization of hydro resources, and the
normal cost of fuel for Company-owned thermal resources.
Therefore, as a resul t of my investigation, I recommend
that the Commission accept the power supply adjustments as
proposed by the Company.
Based on my review of the Company s decision
to pursue the Coyote Springs 2 proj ect (CS2), I concluded
that the Company s need for power justified the decision.
My review of the Request for Proposal (RFP) process also
led me to conclude that the process was fair and that the
CS2 proj ect was the best al ternati ve.Because the proj ect
was transferred from Avista Power to Avista Utilities
cost, I believe that it was appropriate to consider the
proj ect as an al ternati ve in the Company s RFP evaluation.
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 3
Despi te the problems caused by the bankruptcy of the
construction contractor, and the numerous problems
experienced with the generator step-up transformer, I
believe Avista did all it reasonably could to minimize the
construction delays and the cost overruns.
The Kettle Falls CT and Boulder Park proj ects
were pursued to obtain some relief from the extremely poor
water conditions and high market prices in 2000 and 2001.
I reviewed the Company s analysis justifying the Kettle
Falls project and conclude that it was reasonable given
the circumstances at the time.In reviewing the Boulder
Park proj ect, however, I found that there were exceptional
cost overruns and delays.While some of the cost overruns
and delays were unavoidable, others could have been
avoided if Avista had better planned and managed the
proj ect Because the cost overruns and delays were so
excessive, I contend that ratepayers should not be stuck
with all of the excess costs and recommend that ten
percent of the proj ect investment not be allowed in rate
base.
WEA THE R NORMAL I ZA T I
What is the purpose of weather normalization?
Customer energy usage in the test year
typically higher or lower than normal due to unusually
warm , cold, wet or dry weather.The purpose of weather
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 4
normalization is to adjust test year customer energy usage
to reflect a level of usage that would reasonably be
expected in a year wi th normal weather condi tions
Normalized customer energy usage is then used to establish
retail sales revenue that can be expected in a normal
year.It is also used to determine the demand that must
be met by the Company s generation or purchased resources,
thus it affects the normalized net power supply expenses.
Have you reviewed the weather normalization
performed by the Company in this case?
Yes , I reviewed it in detail.I replicated
the method used by the Company in order to verify the
accuracy of the Company's resul ts I also varied the
analysis by using weather and customer usage data for
different periods of record than used by the Company.
also examined different weather variables.In addition , I
performed weather normalization analysis for each of the
Company I s customer classes to determine which classes are
sensitive to weather conditions.
Avista made separate weather normalization
adjustments for usage by its electric and its gas
customers.Did you review the Company s weather
normalization for its gas customers?
Yes, I conducted a similar reVlew of the
Company s gas weather normalization as I did for the
CASE NOS. AVU-04-1/AVU-04-
06/21/04 STERLING, R.
STAFF
(Di) 5
electric weather normalization.The techniques and
weather variables used by the Company were nearly
identical for both the electric and gas weather
normalization.
What is your opinion of the Company I s weather
normalization?
I believe the Company I s weather normalization
fairly and accurately adjusts test year energy consumption
and that no further adjustment to the weather
normalization proposed by the Company is necessary.
POWER SUPPLY EXPENSE AND REVENUE ADJUSTMENTS
Why is it necessary to make adjustments to
the test year power supply costs?
The Company s adj ustments to the 2002 test
period power supply revenues and expenses are designed to
reflect the normalized level of revenues and expenses, and
to include known and measurable changes to the revenue and
expense items.The purpose of the adj ustments is to come
up with revenues and expenses that can be reasonably
expected going forward wi th the rates that are established
by the Commission.
What are the primary differences in net power
supply costs since Avista ' s last general rate case in
1997?
Net power supply costs in this case are
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 6
approximately $11 million (Idaho share) higher than in the
last general rate case in 1997.The two primary changes
include a reduction in wholesale sales revenue (PGE
capacity sale) of $6 million , and an increase in net fuel
expense for thermal generation (primarily Coyote Springs
2) of $4.5 million.
Have you reviewed the testimony of Company
wi tness Johnson and the power supply adj ustments shown in
Exhibi t No.1 0, Schedul e I?
I have reviewed Mr. JohnsonYes.
testimony, Exhibit No. 10, Schedule 1 , Company workpapers
that support the exhibit and Company responses to Staff
production requests.
What are the primary reasons for the proposed
power supply adjustments?
There are two prlmary reasons for the
proposed adjustments to the 2002 test year power supply
revenue and expenses.The majority of the adjustments are
associated with contracts.These can be due to the
expiration of an existing contract or the initiation of a
new contract, or due to specific, proj ected or estimated
changes in contract rates or charges.The remaining
changes resul t from the dispatch simulation model, and
mostly include proj ected fuel expenses.
Staff Exhibi t No. 128, enti tled 2002 Test
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 7
Year Power Supply Adj ustments, provides a categorical
breakdown of total Company power supply expense and
revenue adjustments.Expenses have been reduced by $ 8 5 . 9
million and revenues have been reduced by $55.4 million
for a net decrease in revenue requirement of $30.5 million
from the 2002 test year.
Please generally describe the types of power
supply adjustments summarized in Staff Exhibit No. 128.
Avista has made 67 pro forma power supply
adjustments to 2002 test year actuals to reflect power
costs for the twelve-month period beginning September 1
2004 and ending August 31, 2005.Fifty-two of these
adjustments are to test year expenses, while
adjustments are to test year revenues.Many of the
adjustments are associated with changes in wholesale power
contracts from 2002 through August 2005.Some of these
adj ustments reflect new or expiring contracts, while
others reflect contractual rate and cost changes for
services purchased, services rendered and acquisition of
fuel supplies over the same period.In some cases,
adjustments are based on specific contractual rates
applied to historical averages or estimates for such
things as generation or transmission quantities.The
remaining adjustments have been categorized as power
supply, and are the resul t of output from the Company
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 8
dispatch simulation model under normal load and water
condi tions.
What prlmary cri terion did you use to decide
whether a proposed adjustment is reasonable?
The primary criterion is whether the
adj ustment is known and measurable.
Are the power supply adjustments proposed by
the Company and presented by Mr. Johnson reasonable?
I have reviewed the workpapers provided by
the Company for each of the proposed power supply
adj ustments presented by Mr. Johnson and recommend that
they be approved as proposed.There is li ttle question
that the specific changes such as new contracts , expired
contracts, and contract-specific changes in rates or
charges occur at a date certain and are therefore known
and measurable.When expense and revenue adjustments
shown on line 4 of Staff Exhibit No. 128 are combined,
this category of adjustments represents approximately a
$7.09 million increase in power supply revenue requirement
(Net adj ustment in power supply costs = Net adj ustment
expenses - Net adjustment in revenues, or -$11.172 million
(- $18 . 260 mi 11 ion) = $ 7 . 088 mi 11 ion)
When the expense and revenue adjustments
shown on line 8 that represent estimated, proj ected and
miscellaneous contract changes are combined , they
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 9
represent a decrease in power supply expenses of $34.
million.Al though these changes are not all specifically
stated within a contract, I believe they represent
reasonable estimates based on historic averages, proj ected
third party budgets or historic service costs or revenues
under existing contracts.
Power Supply adjustments, the final category
of expense and revenue adjustments, are from the dispatch
simulation model and are shown on lines 10 and 11 of Staff
Exhibi t No. 128.After analysis of the simulation model
examination of Company workpapers and review of production
request responses, I believe that the adjustments for
short-term sales and purchases, and fuel price changes for
thermal resources are reasonable.When added together,
this category of adjustments represents a decrease of
$ 3 . 53 mi 11 ion.I will discuss the dispatch simulation
model and the associated adjustments in more detail later
In my testimony.
How did you evaluate the Company s proposed
adj ustments for contracts?
I reviewed the workpapers provided by the
Company, which in some cases consisted of the contracts
themselves and in other cases consisted of excerpts from
the contracts showing the rates and terms that would
affect power supply costs.The workpapers showed
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 10
beginning and termination dates of the contracts, the
quantities and delivery schedules, and the rates for
purchase or sale.
Are there some contracts for which
adjustments have been made where a precise rate is not
specified?
Yes , there are some.For those contracts the
adjustments were based on estimates made by the
contracting parties.
There appear to be very large power supply
adj ustments in both expenses and revenues in the
miscellaneous " category (line 7) of your Staff Exhibit
No. 128.Please explain why these adj ustments are so
arge
Nearly all of the adjustments in this
category, both on the expense and the revenue side, are
attributable to gas that was purchased, but not consumed,
for generation during the 2002 test year.The pro forma
expense for this gas is zero since it is assumed that all
gas purchased will be used for generation.Similarly, the
pro forma revenue for this gas is also zero since there
would normally be no gas to sell.
The second most noticeable adjustments are In
the short -term purchases/sales " category (line 10) of
your Staff Exhibit No. 128.Please explain why these
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 11
adj ustments are so large.
The short - term market purchases and sales
adjustments are based on output from the dispatch
simulation model (AURORA)The adjustments are the
combined effect of differences from the 2002 test year in
both the quantities of purchases and sales, and the prices
of those purchases and sales.In general, there would be
fewer short-term purchases and more sales in a normal
year.This reflects the fact that the CS2 plant would be
available in a normal year , and the fact that 2002 was
below normal for hydro generation.
The final category of large adjustments is in
fuel expenses (line 11 of Staff Exhibit No. 128)Please
explain this adj ustment
Fuel expense adj ustments are based on the
results of the Company s system dispatch model.The
maj ori ty of the fuel expense increase is associated wi
operation of the CS2 plant.The Boulder Park and Kettle
Falls CT proj ects also contribute to this adjustment.
Note on Staff Exhibit No. 128 that the increase in fuel
expense is more than offset by a net decrease in the cost
of short-term purchases and sales.
Do you believe it is appropriate to pro form
the normalized 2002 test year power supply expenses to the
period of September 1 , 2004 through August 31, 2005?
CASE NOS. AVU-E- 04 -l/AVU-G- 04-
06/21/04
STERLING, R.
STAFF
(Di) 12
Yes, I believe that it is appropriate to
allow adj ustments that reflect power supply cost during
the period proposed for several reasons.First, as
previously discussed, all of the adjustments must be
reasonably known and measurable to be considered
reasonabl e Second , the adjustments must be based
strictly on test year loads and be independent of future
retail load conditions.Finally, by the time the rates go
into effect in this proceeding, we will be at the
beginning of the pro forma period and the test year will
be more than two years old.
Is it unusual in a general rate case to pro
form test year power supply expenses to a period more than
two years later than the test year , in this case from a
2002 test year to a pro forma period of September 1 , 2004
through August 31 , 2005?
No.In Avista s last general rate case, Case
No. WWP-98-, the Company used a 1997 test year and a
pro forma power supply period of July 1, 1999 through June
30, 2000.Thus, the pro forma period followed the test
year by about two and a half years.
By using a pro forma power supply period of
September 1, 2004 through August 31, 2005, do you believe
there is any potential for a mismatch between revenues and
expenses?
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 13
There is always a potential for a mismatch of
revenues and expenses.That is why we typically use a
historical test year and try to limit adjustments as much
as possible.In using a historic test year and making
prospective adjustments, it is very important to make only
those adj ustments that are known and measurable.I have
carefully reviewed each of the power supply adjustments
proposed by the Company and bel ieve all of them are
reasonably known and measurable.
But isn t it possible that the Company
power supply adjustments include known expense increases
and known revenue decreases due to ei ther new or expired
contracts, but not include potential revenue increases due
to unknown future events and prices?
If Avista has contracts that explre and are
not replaced during the pro forma period, the dispatch
simulation model will either buy or sell generation to
replace the effect of the contract.Thus, for example, if
a power sales contract expires before the end of the pro
forma period leaving Avista with surplus generation for
some period of time, the system dispatch model will simply
sell the surplus into the market at whatever prices the
model computes.Thus, the revenue lost when the contract
expires is replaced by revenue determined by the system
dispatch model.Similarly, if a purchase contract by
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 14
Avista expires, the model will purchase replacement
resources from the market at computed prices.Al though
the purchase and sales prices computed by the model are
not precisely known and measurable, they are as accurate
as can be determined, short of having a contract in-hand.
Moreover , they are no less accurate than the normalized
fuel expenses.
According to Mr. Storro s testimony at page
, lines 6 - 9, Avista ' s annual net resource energy position
does not become deficient until 2008 and beyond, and the
Company s capacity position is either surplus or nearly
balanced through 2007.Is it possible that the Company
surplus is too large, resulting in increased costs but not
proportionately increased revenues?
It is important to realize that the Company
surpl us condi t ion is on an annual bas is, and that there
are times during the year when the surplus is either
greater or less than the annual average.Avista operates
its own resources to make economy sales in the market
whenever its resources are not needed to meet its own
load.However, if those resources cannot be economically
operated to make off - system sales, they si t idle.
Nevertheless Avista still may need all of its resources
times, and must always maintain a required reserve margin.
(Avista currently maintains a reserve margin of about 15%
CASE NOS. AVU-04-1/AVU-04-06/21/04
STERLING, R.
STAFF
(Di) 15
based on forecasted peak loads.In addition, Avista is
required by the Western Electricity Coordinating Council
to maintain an operating reserve equal to 5% of its hydro
generation and 7% of its thermal generation capacity)
Having too great of a surplus can indeed cost the Company
and its ratepayers more.However , I do not believe that
Avista has an unacceptably large surplus.Further, I
believe the planning cri teria used by the Company for
deciding whether and when to acquire new resources
appropriate.
Is it unusual to have 67 power supply expense
and revenue adjustments in a general rate case?
No.In Avista ' s last general rate case there
were 97 power supply adjustments.As I stated earlier
the maj ori ty of the adj ustments in this case are
contractually related, and the remaining adjustments are
pro forma fuel cost adjustments.
DISPATCH SIMULATION MODEL
Has Avista done anything differently from its
1997 general rate case in terms of analysis using a
dispatch simulation model?
The primary difference is that theYes.
Company is now using the AURORA model.AURORA dispatches
resources on an hourly basis, unlike the previous model
that used a monthly time step.An hourly dispatch more
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 16
accurately reflects the true system dispatch of Avista
resources and of other generation resources throughout the
reglon.The use of hourly data also more accurately
recognizes hourly load variations and properly evaluates
the costs and benefi ts of peaking resources.In my
opinion, the adoption of an hourly dispatch model is a
substantial improvement over prior system dispatch models
and I am more comfortable wi th the resul ts it produces.
You stated that the power supply adjustments
proposed by Mr. Johnson were reasonable.How did you
evaluate the adjustments that result from running the
dispatch simulation model?
The first step in evaluating the power supply
expense and revenue adjustments was to replicate the
Company s results using the AURORA model.Throug h it s
software licensing agreement, Avista has provided Staff
wi th a copy of the model.Avista has also provided Staff
with a complete copy of all input data that it used in its
analysis.By replicating the Company s resul ts, I was
able to better understand the relationships between energy
demand, supply energy and market conditions throughout the
reglon.I then evaluated the hydro generation and
regional resource input data provided mostly by third
parties, the long-term contract demand obligations as
adj usted in the pro forma test year , the monthly energy as
CASE NOS. AVU-E- 04 -l/AVU-G- 04-06/21/04 STERLING, R.
STAFF
(Di) 17
calculated by the model for short-term purchases and
sales, and the generation and cost for each Company-owned
thermal resource.The final step was to evaluate the
effect of different natural gas prices on the annual fuel
cost for the Company s thermal resources.
How do you know that the hydro conditions
assumed by the model represent normal water conditions?
In the model , hydroelectric generation for
the Northwest was based on the Northwest Power Pool'
2000-2001 Headwater Benefits Study.The study provides
generation estimates for northwest hydroelectric plants
including Avista s plants, utilizing current regulation
and sixty water years (1929-1988) of historical stream
flows.Because AURORA dispatches resources throughout the
WECC, data sets for plants outside of the Northwest (e. g.
Canada and California) were also used.These data sets
were provided by EPIS, the developer of AURORA, and are
based on information from Canadian sources and from the
U. S. Department of Energy.Because the hydro data used in
this rate case has been developed by independent sources
for a variety of uses by many different utilities, I
believe it fairly reflects normal water conditions and
produces unbiased resul ts
It would seem that the resul ts of the
dispatch simulation model would be highly dependent on the
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 18
fuel prlce assumptions used in the model.Did you reVlew
Avista s fuel price assumptions and do you believe they
are reasonabl
It is true that the resul ts of the dispatch
simulation modeling are highly dependent on the fuel price
assumptions used.For its analysis, Avista used actual
contract prices for its coal plants and for its wood-fired
Kettle Falls plant.For its gas-fired plants, the Company
used Henry Hub NYMEX natural gas forward prices on
December 10, 2003 for the power supply pro forma period.
Avista then adjusted the Henry Hub prices using basis
differentials intended to capture ancillary costs such
transportation and taxes.A different set of gas prlces
was derived for Coyote Springs 2 , Rathdrum, and the
combination of Boulder Park, Northeast and the Kettle
Falls CT.The source used by Avista for these prlces was
the same system the Company uses to make gas fired
resource dispatch decisions.
Because the modeling resul ts are so highly
dependent on gas prlces, I investigated gas price changes
and their effect on annual expenses.I first examined a
historical record of NYMEX forward prices for delivery in
each month of the pro forma period.I reviewed historical
daily NYMEX forward prices from April 2003 - April 2004 to
determine whether the December 10, 2003 prices used by
CASE NOS. AVU-04-1/AVU-04-
06/21/04
STERLING, R.
STAFF
(Di) 19
Avista were unreasonably high or low.In my judgment,
Avista did not choose a particularly high or low priced
day.Generally, gas prices have steadily increased Slnce
December 10, 2003 when Avista chose prices for its
analysis.
Nevertheless, to analyze the effect of gas
prlces on net power supply costs I estimated gas prices
that were lower and higher than the prices used by Avista.
In the low price scenario, I selected prices on May
2003 because they were nearly the lowest of any day in the
past twelve months.For the pro forma period, the prices
averaged about $4.77 per MMBtu.For the high gas prlce
scenarlo, I selected futures prices on May 5, 2004 because
they were close to the highest on any day in the past
twelve months.The average price in the pro forma period
under the high price scenario was approximately $6.09 per
MMBtu.Using these high and low gas prlce scenarios, I
determined a corresponding range of thermal fuel costs to
be $46.32 million to $63.49 million.The thermal fuel
cost computed by Avista using its December 10, 2003 fuel
prices is $50.0 million.Based on the range I computed
for high and low gas prices, I concluded that the gas
prices Avista used in its modeling are reasonable.
How critical is it that Avista use accurate
gas prlces in determining its net power supply costs?
CASE NOS. AVU-04-1/AVU-04-
06/21/04
STERLING, R.
STAFF
(Di) 20
Of course, it is desirable to use gas prices
that are close as possible to what the Company will
actually encounter.It is impossible to know these prlces
in advance, however.Nevertheless, if gas prices are
estimated too high or too low , deviations in actual net
power supply costs will be captured in the Company
annual power cost adj ustment (PCA)Under the PCA , Avista
is entitled to recover or refund to customers up to
percent of deviations from normal.This sharing between
the Company and its customers helps to minimize the built-
in incentive for Avista to establish its base net power
supply costs too high.Again, I do not believe Avista
chose to use December 10, 2003 gas prlces in an effort to
set its base net power supply costs high.Instead, I
believe the gas prices chosen by Avista are reasonable.
Do you recommend any changes in the thermal
fuel adj ustments proposed by the Company?
I believe that the dispatch simulationNo.
model adequately estimates the amount of energy that will
be generated at each resource under normal water
conditions.I also believe that the fuel price changes
proposed by the Company are reasonable based on my reVlew
of Company workpapers.
Does the dispatch simulation model include
speculative sales or purchases?
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 21
The dispatch simulation model includesNo.
only Avista s hourly native loads, so resources are
dispatched to meet only those loads.However , whenever
Avista has resources of its own that can be operated
economically to meet other loads in the region, they will
be operated and the revenues will accrue to Avista and its
customers.Similarly, Avista regularly makes off -system
purchases whenever its own resources are insufficient to
meet load.These off -system purchases and sales are not
speculative and therefore are appropriately included in
power supply modeling.
COYOTE SPRINGS 2
When did Avista first identify a need for the
Coyote Springs 2 proj ect?
In July 2000, Avista submitted an update to
its 1997 Integrated Resource Plan (IRP)The updated 1997
IRP served as the basis for a Request for Proposals that
the Company intended to release in August 2000.In the
1997 IRP update, Avista s load-resource balance showed
that the Company was deficit, both for energy capacity,
beginning immediately and extending throughout the entire
planning horizon.Deficits in 2000 were 395 MW of peak
capacity and 237 aMW of energy.One of the primary
reasons for the deficits was the sale of the Company
share of the Centralia plant.Avista had a contract to
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 22
purchase output from Centralia after the sale, but that
contract expired at the end of 2003.A second reason for
the expected deficits was a decreased reliance on long and
short-term contracts, in part due to their risk and the
recent volatili ty in market prices.I believed that the
Company s need for new resources was sufficiently
demonstrated in the 1997 IRP update and I supported the
Company s decision to issue a Request for Proposals.
Do you believe the RFP issued by Avista was
fair?
Yes, I believe the RFP was fair.Staff
reviewed preliminary drafts of the RFP prior to its
release and provided comments to Avista.All of Staff'
comment s, both wri t ten and verbal, were addressed by
Avista in the preparation of the final draft RFP.Avista
then submitted the draft RFP and its 1997 IRP Update to
the Commission for comment.Commission Staff commented
noting that it believed that issuing the RFP was
appropriate.The Commission issued Order No. 28542 noting
that the Company s filings of its 1997 IRP Update and the
RFP were informational and were not required by statute or
Commission Order.The Company solicited only comment
therefore, Commission approval was not necessary.The
Commission commended Avista for soliciting public input
into its RFP process.
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 23
Avista s RFP was an ~all source " competitive
bid based on the Company s identified need for 300 MW of
new electric power starting in 2004.The 1997 IRP Update
described the Company s loads and resources, provided an
overview of technically available resource options, and
demonstrated need for resources.
In its filing with the Commission, the
Company stated that it would consider any offer of
resources including but not limited to, energy and
capacity, energy efficiency, turnkey plans, construction-
for Avista-of a generating plant on a site provided by the
bidder, and construction by a bidder on a site furnished
by Avista.
I believe that the RFP was fair in all
respects, and not intended to favor specific proposals,
locations, technologies or bidders.
Briefly describe the response Avista received
In response to the RFP.
Thirty-two proposals were received from
bidders for a total of 2 , 700 MW of resources in response
to the all-resource RFP.The proposals included 24 offers
for new generation , six of which were for renewables, one
customer-owned emergency generation proposal, and seven
energy efficiency proj ects.
Do you believe that the evaluation criteria
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 24
developed and used by Avista were fair to all proposals?
Avista went to great lengths to insureYes.
that the evaluation criteria it developed were fair and
impartial.Besides seeking input from the Idaho and
Washington Commission Staffs, it retained R.W. Beck, an
engineering consul ting company, to also review the
evaluation criteria.R . W. Beck made recommendations on
the evaluation criteria and on the assumptions to be used
in analyzing proposals, and on the dispatch modeling and
economlc analysis used by Avista.
Do you believe it was appropriate to consider
the Coyote Springs 2 proj ect as an al ternati ve, Slnce
rights to develop the proj ect were owned at the time by
Avista Power, an unregulated Avista Corp. subsidiary?
Yes, I do believe it was appropriate.
participated in meetings with Avista and with a
representative from the Washington Commission Staff in
which this issue was specifically discussed.My opinion
and the opinion of the Washington staff member was that
CS2 should be considered as an al ternati ve as long as the
project assets at the time (permits, site , turbine
contract, rights to develop, etc.) would be transferred at
cost to Avista Utilities.Early on in the proposal
evaluation phase, it was apparent that the CS2 project
could be a very competi ti ve proposal.It was fel t that
CASE NOS. AVU-E- 04 -l/AVU-G- 04-06/21/04 STERLING, R.
STAFF
(Di) 25
excluding it might eliminate what could ultimately be
Avista s best and least cost option.
Do you believe there was any impropriety in
the transfer of rights to the CS2 proj ect from Avista
power to Avista Utilities?
No, because the transfer was made at cost.
Staff auditors have reviewed the transaction and have
assured me that the transfer was indeed at cost.Nei ther
Avista Power nor the shareholders of Avista Corp. made any
profit from the transfer.
What was Staff's involvement in the RFP
process?
I participated on behalf of the Idaho
Commission Staff.I reviewed and helped develop
evaluation criteria , and reviewed the results of Avista
analysis of proposals.I participated in several meetings
with Avista and a representative of the Washington
Commission staff to review Avista ' s evaluation and ranking
of the proposal s .We reviewed the Company s first round
screening resul ts and provided input into the decision
about which proj ects should move on to the second round of
screenlng.We also identified things we believed needed
further investigation before further evaluation and
ranking could take place.During the final screening
process, we reviewed in detail Avista ' s economic analysis
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 26
as well as all the other factors that were used in
assessing the proposal s .
just days before Avista
the Board of Directors.
I al so at tended a final meet ing
staff made their recommendation to
Are you convinced that Avista chose the best,
least cost proposal?
The Company ' s selec~ion of CS2 asYes, I am
a resource from its 2000 all-resource Request for
Proposals process was reasonable.
Do you believe it was reasonable to sell half
of CS2 to Mirant?
Yes, I do believe it was reasonable, glven
the financial challenges facing the Company at the time.
I reviewed the analysis done by the Company of the options
available at the time.Al though it would have been
desirable to have more interested bidders in the plant, I
believe that the Company s analysis supports the decision
to sell half of the plant to Mirant.
Avista witness Lafferty s testimony includes
extensive discussion of the litany of problems experienced
during the construction and start-up of CS2 , along with
the costs associated wi th those problems.Do you be 1 i eve
that the cost overruns that resul ted from these problems
should be allowed in rate base?
The problems and associated cost overruns
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 27
seemed to be associated primarily wi th two factors, the
bankruptcy of Enron and ultimately of NEPCO, its
construction subsidiary, and failures of the generator
step-up (GSU) transformer.
I do not believe the bankruptcy of Enron and
NEPCO could have ever been envisioned at the time
construction on the proj ect began.There was virtually
nothing Avista could do other than try to mi tigate the
effects on the CS2 construction costs and schedule.
believe Avista made a good effort keep costs under
control and to construction delays following themlnlmlze
bankruptcies therefore, I do not believe Avista or its
shareholders should be held accountable for any cost
overruns and delays caused by the bankruptcies.
Wi th regard to the repeated GSU transformer
failures, I believe that these too were beyond the control
of Avista.Decisions about the transformer design and
which manufacturer to select were not unreasonable.
Whenever problems were encountered, it appears Avista did
everything it could to make repairs or acqulre a
replacement.The Company also appears to have diligently
exercised warranties and pursued insurance claims.
The cost overruns associated wi th these
problems have been estimated by Avista to be approximately
$15 million.This amount represents 16 percent of the
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING , R.
STAFF
(Di) 28
total original project cost estimate of $93.9 million.
Staff does not oppose inclusion of these costs in rate
base for the CS2 plant.
KETTLE FALLS CT
Why did Avista build the Kettle Falls gas-
fired combustion turbine (CT) proj ect?
The Kettle Falls CT proj ect was one of at
least five potential generation proj ects identified as
possible solutions to help mitigate the effect of very low
water condi tions and extremely high and volatile electric
prices that occurred during the June 2000 through December
2001 period.Eventually the Company decided to pursue the
Kettle Falls CT proj ect and the Boulder Park proj ect, but
not pursue three small proj ects involving installation of
natural gas or diesel- fueled generators at other
locations.Two gas-fired engine generators like those
installed at Boulder Park were purchased by Avista for
installation at the Spokane Industrial Park, but were
never installed after power prices receded in late 2001.
Recovery of the cost of these generators is not being
requested in this case.
Have you reviewed the final cost of the
Kettle Falls CT proj ect?
The final cost of the Kettle Falls CTYes.
proj ect as verified by Staff auditors is $9.2 million, or
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 29
approximately 8.2 percent above the estimated proj ect
cost of $8.5 million.
It appears the proj ect exceeded its cost
estimate by nearly $700,000.What does Avista attribute
the cost overruns to?
There are two primary reasons identified by
Avista.First, $543,000 in additional costs were incurred
because of additional work that had to be completed by the
proj ect contractor.Most of this work was associated with
the construction cost of the turbine building.Second, an
additional $153,000 was incurred directly by Avista for
work outside of the scope of the contractor
responsibili ty.Of this amount, $133,000 was paid to the
contractor in accordance wi th contract requirements for
exceeding the performance requirements of the turbine.
Do you recommend that the full final cost of
the Kettle Falls CT proj ect be allowed in rate base?
Yes, I do.Despite the fact that the final
proj ect costs exceeded its original estimate and took a
little longer to complete than expected, I believe the
cost overruns were within a reasonable range and not
unusual for a proj ect of this type.
BOULDER PARK
Was Boulder Park or an equivalent plant
included in Avista s 1997 or 2000 IRPs before the Company
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 30
made its decision to pursue the proj ect?
The need for such a plant was notNo.
identified In any of the Company s previous IRPs.Avista
decided to pursue the proj ect primarily in response to the
extreme low water conditions and market prices in
2000-2001.
Do you believe it was reasonable for Avista
to develop the Boulder Park proj ect?
Yes, I do.Market prices at the time were
extremely high and no one knew if or when such high prices
might subside.Most utilities in the Northwest were
pursuing a variety of options for relief from the high
prices including diesel generation, gas-fired generation,
customer buy-backs and demand management programs.Avista
also considered many of these options, and the Boulder
Park proj ect appeared to be one of the Company s most cost
effective al ternati ves.I thoroughly reviewed the
Company s analysis that it completed at the time a
decision was made to pursue the proj ect.At that time, I
believe a decision to proceed was reasonable.
What was the Company s estimated cost for
Boulder Park?When did the Company expect to complete
construction?
When the proj ect was first proposed, Avista
estimated the construction cost to be $21.0 million.
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 31
June 17 , 2001 , Avista revised its estimate upward to
$23.65 million.The original estimated completion date
was September 1, 2001.
It appears that there were considerable cost
overruns and delays on the proj ect Have you reviewed the
information provided by the Company in response to Staff'
production requests concerning cost overruns and delays?
Yes, I have.The final cost of Boulder Park
was approximately $32.1 million.This is $11 million more
than ini tially proj ected, and represents a greater than
50% cost overrun.Completion of construction was delayed
by eight months until May 2002.
What reasons does Avista gl ve for the cost
overruns and delay in completion?
In response to production requests,
Avista states that:
The excess costs for the Boulder Parkproj ect generally stemmed from the fast
track design-build approach that the
Company chose in order to bring small
generation on line as quickly as
practical in order to mitigate the high
prices and volatility in the electric
power market during the energy crisis.Al though not new technology for the
power industry, the natural gas fired
reciprocating engine generators were the
first project of its kind for Avista,
which contributed in part to actual
construction costs being higher than the
original estimates.
Avista provided a summary by cost category of the amounts
CASE NOS. AVU-04 -l/AVU-G- 04-06/21/04 STERLING, R.
STAFF
(Di) 32
of the cost overruns, along wi th a brief description of
the reasons for the cost variations in each category.
have included this summary as Staff Exhibi t No. 129.
Do you believe the explanations cited by
Avista for the cost overruns are reasonable?
I believe that some of the explanations are
reasonable.Avista clearly did not anticipate many of the
problems encountered in the proj ect' s construction or many
of the requirements imposed on the proj ect by other
agencles.For example, the Company cites incomplete
construction plans being provided by the engine generator
manufacturer, handicapped building access requirements,
road width requirements, paved instead of graveled si te
grounds , building soundproofing requirements and
construction plan approval delays as among the many
unexpected factors.I agree that many of these delays and
requirements could not have been anticipated.
Nevertheless, it is simply impossible to
19nore that the final proj ect cost exceeded the ini tial
estimate by nearly 53 percent.While many of the causes
of cost overruns could not be anticipated, I believe some
of them could have been if Avista had better planned and
managed the proj ect Blaming a fast track construction
process for cost overruns might make sense if the proj ect
had actually been completed on a fast track schedule, but
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 33
the fact is that construction took eight months longer
than expected.The higher costs due to the fast track
schedule apparently cost the Company qui te a lot but
gained nothing.
It is common to include a contingency amount
in the cost estimate for large construction proj ects
lnsure that funds are available in the event of unplanned
problems, circumstances or condi tions.The amount of the
contingency can vary considerably for construction
proj ects depending on many things such as material and
equipment costs, installation complications and unknown
si te condi tions.Contingency amounts for proj ects similar
to this one are typically in the range of 5 -15 percent.
In fact, CS2 and Kettle Falls contingencies totaled 16 and
8 percent, respectively.Avista may not have any
experlence in building this particular type of plant, but
it should have some experience with building practices and
requirements in Spokane County, a place where it has buil
many things.
The explanations put forth by Avista may be
understandable, but the excessive cost overruns should
primarily be the responsibili ty of Avista.I believe
ratepayers should be able to expect the utility to have
the ability to construct proj ects at least cost.
Construction of new proj ects cannot simply be a blank
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 34
check signed by ratepayers.It is reasonable to expect
the utility to have the expertise and experience to
construct and manage any proj ect it undertakes at a
reasonable cost.
Do you recommend that all of the cost of the
Boulder Park plant be allowed in rate base?
No, I do not.I recommend that ten percent
of the final proj ect cost be disallowed.
What is the basis for recommendipg ten
percent disallowance?
In reviewing Staff Exhibi t No. 129, three
particular cost categories stand out.First, the final
construction management cost of $2 159,000 was 2.25 times
the revised proj ect estimate.This addi tional cost was
primarily due to the contractor being required to spend
twice the amount of time working on the proj ect.The
second cost category that stands out is $1, 110,000 for
Avista s proj ect management, engineering and proj ect
commissioning.There was no amount included for these
costs in the revised estimate.Finally, an addi tional
$912 714 was incurred because of the additional time
required to complete the proj ect The total' cost overrun
in just these three cost categories comes to $3,221,714,
approximately ten percent of the total final proj ect cost
Undoubtedly, some of the cost overruns in these categories
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLINS, R.
STAFF
(Di) 35
would have occurred due to reasonable construction delays
and probl ems.However, it is also likely that there are
some unreasonable cost overruns spread throughout nearly
every cost category.Consequently, I believe a ten
percent disallowance from rate base is a fair amount.The
effect of a ten percent disallowance from rate base is a
reduction in annual revenue requirement of approximately
$205,000 on an Idaho jurisdictional basis.Staff witness
Patricia Harms further discusses this adjustment in her
testimony.
I might also add that uslng the ini tial
construction cost estimate as the basis for judging the
reasonableness of the final construction cost is not
necessarily always fair.The ini tial estimate could be
low or inaccurate.
Have you examined any other evidence to
determine a reasonable cost for gas fired reciprocating
engines similar to Boulder Park?
Yes, al though cost information for these
types of englnes is somewhat difficul t to obtain because
there are few utilities or public entities that have
recently installed these types of units.Normally, uni ts
like these are installed by non-public entities such
hospitals, institutions and industries for cogeneration or
backup purposes.Nevertheless, I was able to obtain some
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 36
information for comparison purposes.Six different recent
reports all reference the same source for cost figures.
Thus, I have included excerpts from only one report as
Staff Exhibi t No. 130.As second source ci ting a cost
range of $350 to $600 per kW is included as Staff Exhibit
No. 131.As shown by Staff Exhibit No. 130, total plant
costs range from $695 per kW for the largest units to
$1030 per kW for the smallest units.Boulder Park
consists of six units similar in size to the largest unit
shown in the exhibi Boulder Park's total plant cost
came to $1303 per kW.The initial estimate of the plant
cost was approximately $850 per kW.It is absolutely true
that actual costs for a specific plant could vary quite
significantly from the estimates shown in the exhibit
however , Boulder Park's cost seems exceptionally high by
compar l son.Even with the ten percent disallowance
recommended by Staff, Boulder Park's cost would still far
exceed the estimates from other sources.
Does this conclude your direct testimony in
this proceeding?
Yes, it does.
CASE NOS. AVU-04-1/AVU-04-06/21/04 STERLING, R.
STAFF
(Di) 37
~:
;
d

(j
t
I
i
~
~
,.
.
.
.
.
.

i
f
;
.

e
n
::
r
--
-

r
l
"

(
t
)
...
.
.
.

(t
)

cr
"
+:
:
-

'"
1

Z

...
.
.
.
.
::
:
:
;
'
.

0

rl
"
::
!

.
Z
gq

~~
~
if
;
.
~
~
"
"
'
"
ee
N
.
~

Cl
.
0
+::
-
t-

,.
.
.
.
.
.
~
2
.
(
)
'
02

.' \
m
.

mS
i
l
l
i
.

~.
m
l
l

R

.
..
"~
.
(
g
v
!
/
;
.

11
1
.
.
.
.

..
S
W

..
.
~
.
.
~
.
.

L
l
l
i
~
.
"
.
.
.
l
I
1
8
i
j
J
t
i
Q
.
mi
l
.
.
Ei
l
m
$

,
,

. .
m.
m
~
l
.
1
;
~
i
M
~
m
E
E
M

,
,

'
JI
~
I
I
I
)
NE
T
EX
P
E
N
S
E
S
RE
V
E
N
U
E
S
AD
J
U
S
T
M
E
N
T
Li
n
e
No
.
TY
P
E

O
F

C
H
A
N
G
E
S
SP
E
C
I
F
I
C

C
O
N
T
R
A
C
T

C
H
A
N
G
E
S
NE
W

A
N
D

E
X
P
I
R
E
D

C
O
N
T
R
A
C
T
S
$1
2
01
6
$1
8
54
6
$6
,
53
0
CO
N
T
R
A
C
T

S
P
E
C
I
F
I
C

R
A
T
E
$8
4
4
$2
8
6
$5
5
8
SU
B
T
O
T
A
L
$1
1

,

1
7
2
$1
8
,
26
0
08
8
CO
N
T
R
A
C
T

R
A

TE
/
R
E
V

C
H
A
N
G
E
ES
T

I

M
A

TE
D
/
P
R
O
J

E
C
T
E
D
07
0
52
3
$3
,
54
7
MI
S
C
81
0
41

18
4
62
6
SU
B
T
O
T
A
L
$7
3
74
0
$3
9
66
1
$3
4
07
9
PO
W
E
R

S
U
P
P
L
Y
SH
O
R
T
-
TE
R
M

P
U
R
C
H
A
S
E
S
/
S
A
L
E
S
$3
6
20
3
53
0
$3
8
73
3
FU
E
L
20
1
20
1
SU
B
T
O
T
A
L
00
2
53
0
53
2
TO
T
A
L

N
E
T

A
D
J
U
S
T
M
E
N
T
$8
5
91
4
$5
5
39
1
$3
0
52
3
Summary of Costs
Boulder Park Generating Station
Part I - Wartsila Costs 17 est.actual difference
Wartsila Recipricating Engine/Generators (Units 1 - 6)$ 13,300,000 $ 13,300,000Change orders 208,000 208,000
Wartsila Subtotal 300 000 508,000 208,000
Part II - Contractor Construction Costs
Construction Management (KBI)960 000 159,000 199 000Buildings and Sound Enclosures (Furnish and Install)250 000 228,000 978,000Ventilation/Exhaust/Duct System (fabricate & install)170,000 299,000 129,000Mechanical equipment Installation and commissioning 130,000 712 000 582 000Electrical equipment Installation and commissioning 720,000 546,000 826,000
Contract Construction Subtota 230 000 944 000 714 000
Part III - Avista Construction Costs
Site Work 220,000 410 000 190 000
Gas System 160,000 103.000 (57 000)
S u bsta tionfT ransmiss i 0 niDi stribution/Commu n ication 136 000 1 ,488,000 352 000
Permits/Property Acquisition/Legal Fees 450,000 280,000 (170,000)
Miscellaneous Items
Fire Detection & Suppression Systems 237 000 237 000
Electrical and mechanical systems 415 000 415 000
Emission Testing 000 35,000
Spare Parts and Tools 100 000 100 000
Avista Commissioning/Management/Engineering 110 000 110 000
vista Subtotal 966 000 178 000 212,000
Subtotal (Wartsila, Contractor, and Avista)$ 21,496 000 $ 28,630,000 134,000
Washington State Sales Tax (8.1%)772 546 080,000 307,454
8&0 Tax 000 000
AFUDC 387 286 300,000 912 714
TOTAL (Units 1 to 6)$ 23,655 832 064 000 408 168
Exhibit No. 129
Case No. A VU-04-
A VU-04-
R. Sterling, Staff
6/21/04 Page 1 of 4
Boulder Park Generating Station
Cost Summery Variance Details
Part I - Wartsila
The project had 13 Change orders issued for a total of $208 000. The major cost increase was
$123 000 to cover the additional time Wartsila had to spend on the site over and above that
which they contracted for.
Part II - Contractor Construction Costs
The total contractor construction cost over run was $4 714 000. This was primarily the extra cost
associated with the following:
a. Construction Management. The project took much longer than anticipated to complete
thereby increasing the construction management costs by approximately $600 000 for
supervision labor and $400,000 for additional purchasing and construction markups on the
oveITuns on materials and subcontractors. Change orders for engineering changes totaled
approximately $200 000. Total oveITun from estimate is $1 199 000.b. Buildings and Sound Enclosures. The original estimate did not include the consumables
building ($150 000), special inspections ($80,000), nor control room building ($500 000).
The original building estimate from the consultant was lower than the actual cost by
$400 000. The sound enclosures oveITan $60 000 due to design changes. The total overrun
on buildings was $978 000.c. VentilationlExhaustlDuct systems. Change orders to add ventilation air louvers and
piping/sheeting changes added $129,000 total.
d. Mechanical equipment installation and commissioning. This was the single largest overrun
on the project. The mechanical piping work ran $977,000 over due to the complexity of the
piping as required versus the simple piping runs as bid from the minimal design prints. The
exhaust stack was not in the original design and added $200 000. The exhaust duct
insulation was not known in the original design and added $195,000. The foundation work
associated with the auxiliary work outside the main building was not in the original estimate
due to unknowns and underestimates of what was actually needed thereby adding $275,000.
Commissioning costs were less here than estimated but resulted in increased A vista
commissioning costs in Part III. Total cost ovelTUn here was $1,582,000.e. Electrical equipment installation and commissioning. The total overrun was $826,000. This
was due to additions to the scope of work (ie. fire detection system) as well as the lack of
electrical design especially in the control wiring.
Exhibit No. 129
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 2 of 4
Part III - A vista Construction Costs
The total A vista construction cost over run was $2 212 000. This was primarily the extra costassociated with the following:
a. Site work. Road work was larger and more difficult than expected because Spokane County
required a 24' road instead of a 20' road ($35,000). Site work was larger and more difficult
than expected due to rocks, larger footprint of buildings and auxiliaries, as well as' fire and
water system increases ($130 000). The fence work was overlooked in original estimate
($25,000). Total overrun here was $190,000.
b. Gas system. Relocating the station further east shortened the gas run and was $57,000 less
than estimated.c. SubstationtrransmissionlDistribution/Communication systems. The substation transformer
was more expensive than expected, the substation work was more extensive, but the
transmission/distribution work was not as extensive as predicted for a total overrun of
$220 000. The communication system was far more extensive and complicated than
originally anticipated due to microwave not feasible and fiberoptic being required to handle
the load thereby costing an additional $132 000. d. PermitslPropertylLegal. The land was $150,000 less than expected and the legal was
$20 000 less than expected for a cost underrun of $170 000.e. Miscellaneous. These were not included in the original estimate. The fire detection and
suppression systems were $237,000; electrical and mechanical system work was $415,000
(broken down to control systems (g) $160,000; larger power cables and terminations
(g)
$35 000; extra grounding inside station (g) $20,000; work platforms (g) $150,000; handicap
access ramp (g)$50 000); emission testing was $35,000; spare parts and tools was $100 000;
and the Avista commissioning/management/engineering was $1 110 000. The extra labor
costs were due to the fact that to get this project completed, A vista essentially took over from
the construction management firm the commissioning and final engineering.
Taxes - The extra sales tax was from the increase in the cost of the project. The B&O taxes
were not included in the original estimate. The extra AFUDC was accrued due to the extra time
the project took to complete.
Exhibit No. 129
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 3 of 4
Boulder Park Generation Station
CAR data-backup 2-
Major Changes from original handwritten CAR form:
Original estimate = $23.
Est. 1-30-02 $31.5M
$ 8.0M required to complete project.
Major changes in scope of work:
. ..
Extra time on project for Wartsila, KEI, contractors, and Avista construction personnel
Extra AFUDC accumulated due to increase in length of construction process
Control building size increased 25%
Handicapped access required by Spokane County
Complete cooling system containment and oil system containment required by Spokane
County
Air Handling system added to achieve cooling and charge air requirements
Extra catalyst required to achieve acrilyn and formaldehyde limits for SCAP
Quieter radiator fans and silencers from Wartsila to meet sound limits
Additional piping required to handle unforeseen complexity of mechanical systems
Additional electrical work to handle unforeseen complexity of electrical systems(especially
control systems)
Road building changed from 14' driveway to 24' road complete with paving to satisfy
Spokane County requirements !plus extra rock problems encountered
Site grading size increased 20%/ extra rock problems encountered
Added 115 Kv transmission line work
Increases in Washington State Sales Tax and B&O tax
Estimated total increase for above section = $5.7 M
:...
Major portions of work not included in original estimate;
Communication system to tie plant into remote operating facility
Work platforms and cell hoists
Fire & gas detection system
Fire suppression system
10" fire line and hydrants! " water line
Remote and air handling computer control systems
Security system
Annunciator system
Interior painting and insulation
4!0 power cable & terminations
emergency shutdown generator and connections
interior building grounding system
emission testing
Commissioning (A vista labor)
Operations training for A vista personnel
. A vista Management and Engineering time
Estimated total increase for above section = $2.3 M Exhibit No. 129
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 4 of 4
Ga s.. Fired D i st rib ute d
Energy Resource
Technology Characterizations
Bringing you a
prosperous future where
. ,
energy is dean, abundant,
reliable, and affordable
Prepared for the Office of Energy Efficiency and Renewable Energy
November 2003. NRELlTP-620-34783
Exhibit No. 130
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 1 of 6
November 2003 N REL/TP-620-34 783
Gas-Fired Distributed
Energy Resource
Technology Characterizations
Larry Goldstein
National Renewable Energy Laboratory
Bruce Hedman
Energy and Environmental Analysis, Inc.
Dave Knowles
Antares Group, Inc.
Steven I. Freedman
Technical Consultant
Richard Woods
Technical Consultant
Tom Schweizer
Princeton Energy Resources International
Prepared under Task No. AS73.2002
. ".."..".,,"""...
1. r-1'iL
+..
National Renewable Energy Laboratory
1617 Cole Boulevard
Golden, Colorado 80401-3393
NREL is a U.S. Department of Energy Laboratory
Operated by Midwest Research Institute. Battelle
Contract No. DE-AC36-99-GO10337
Exhibit No. 130
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 2 of 6
3 Performance and Efficiency Enhancements
Brake Mean Effective Pressure (BMEP) and Engine Speed
Engine power is related to engine speed and the Brake Mean Effective Pressure (BMEP) during
the power stroke. Reciprocating engines can produce more power from a given displacement
volume (cubic inches or liters) by increasing engine speed and/or the pressure inside the engine
cylinders. BMEP can be regarded as an "average" cylinder pressure on the piston during engine
operation, and is an indication of the specific load on an engine. Engine manufacturers often
include BMEP values in their product specifications. Typical BMEP values are as high as 230
psig for large natural gas engines and 350 psig for diesel engines. Corresponding peak
combustion pressures are about 1 750 psig and 2 600 psig, respectively. High BMEP levels
indicate high specific power output, and generally result in improved efficiency and lower
specific capital costs and maintenance costs.
BMEP can be increased by introducing larger volumes of combustion air and fuel into the engine
cylinders through improved turbocharging, improved after-cooling, and reduced pressure losses
through improved air-passage design. These factors all increase air charge density and raise
peak combustion pressures, translating into higher BMEP levels. However, higher BMEP
increases thermal and mechanical stresses within the engine combustion chamber and drive-train
components, along with a potential increase in the tendency for detonation, depending on fuel
type. Proper design and testing is required to ensure continued engine durability and reliability.
Turbocharging
Essentially, all modem industrial engines above 300 kW are turbocharged to achieve higher
power densities. A turbocharger is basically a turbine-driven intake air compressor. The hot
high-velocity exhaust gases leaving the engine cylinders power the turbine. Very large engines
typically are equipped with two large or four small turbochargers. On a carbureted engine
turbo charging forces more air and fuel into the cylinders, increasing engine output. On a fuel-
injected engine, the mass of fuel injected must be increased in proportion to the increased air
input. Cylinder pressure and temperature normally increase as a result of turbocharging,
increasing the tendency for detonation for both spark ignition and dual-fuel engines and
requiring a careful balance between compression ratio and turbocharger boost level.
Turbochargers normally boost inlet air pressure by a factor of 3 to 4. A wide range of
turbocharger designs and models is used. Heat exchangers (called after-coolers or inter-coolers)
are often used to cool the combustion air exiting the turbocharger compressor to keep the
temperature of the air to the engine under a specified limit and to increase the air density.~ 4.4 Capital Cost
This section provides estimates for the installed cost of natural gas spark-ignited, reciprocating
engine-driven generators. Two configurations are presented: power-only and CHP. Capital
costs (equipment and installation) are estimated for the five typical engine genset systems
ranging from 100 kW to 5 MW for each configuration. These are "typical" budgetary price
levels to the end user. Installed costs can vary significantly depending on the scope of the plant
~quipment, geographical area, competitive market conditions, special site requirements
Gas-Fired Distributed Energy Resource Technology Characterizations
Reciprocating Engines Page
Exhibit No. 130
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 3 of 6
emissions control requirements, prevailing labor rates, and whether the installation is a new or
retrofit application.
In general, engine gensets do not show the economies of scale that are typical when costing.
industrial equipment of different sizes. Smaller genset packages are often less costly on a
specific cost basis ($/kW) than larger gensets. Smaller engines typically run at a higher speed
(rpm) than larger engines and often are adaptations of high-production-volume automotive or
truck engines. These two factors combine to make the small engines cost less than larger
slower-speed engines.
The basic genset package consists of an engine connected directly to a generator without a
gearbox. In countries where 60 Hz power is required, the gensets run at speeds that are multiples
of 60 - typically 1 800 rpm for smaller engines and 900 or 720 rpm for large engines. In areas
where 50 Hz power is used, such as Europe and parts of Japan, the engines run at speeds that are
multiples of 50 - typically 1 500 rpm for smaller high-speed engines. The smaller engines are
skid-mounted with a basic genset control system, fuel system, radiator, radiator fan, and starting
system. Some smaller packages come with an enclosure, integrated heat-recovery system, and
basic electric-paralleling equipment. The cost of the basic engine genset package plus the cost
for added systems needed for the particular application or site comprise the total equipment cost.
The total installed cost includes total equipment cost, plus installation labor and materials
(including site work), engineering, project management (including licensing, insurance
commissioning, and startup), and contingency.
Table 3 provides cost estimates for current power-only systems. The estimates are based on a
simple installation with minimal site preparation required. These cost estimates are for base-load
or extended peaking operation and include provisions for grid interconnection and paralleling.
The package costs are intended to reflect a generic representation of popular engines in each size
category. The engines all have low emission, lean-bum technology (with the exception of the
100 kW system, which is a rich bum engine that would require a three-way catalyst in most
urban installations). The interconnect/electrical costs reflect the costs of paralleling a
synchronous generator, although many 100 kW packa~es available today use induction
generators that are simpler and less costly to paralle1.l However, induction generators cannot
operate isolated from the grid and will not provide power to the site when the grid is down.
Labor/materials represent the labor cost for the civil, mechanical, and electrical work - as well as
materials such as ductwork, piping, and wiring - and is estimated to range from 35% of the total
equipment cost for smaller engines to 20% for the largest. Project and construction management
also includes general contractor markup and bonding, as well as performance guarantees, and is
estimated to range from 10% of the total equipment cost for small engines to 8% for the largest
engines. Engineering and permitting fees are estimated to range from 5% to 8% of the total
equipment cost depending on engine size. Contingency is assumed to be 5% of the total
equipment cost in all cases.
19 Reciprocating Engines for Stationary Power Generation: Technology, Products, Players, and Business Issues
GR!, Chicago, IL and EPRIGEN, Palo Alto, CA: 1999. GRI-99/0271 , EPRI TR-113894.
Gas-Fired Distributed Energy Resource Technology Characterizations
Reciprocating Engines Page
Exhibit No. 130
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 4 of 6
Table 3. Estimated Capital Cost for Typical Reciprocating Engine-Generators in
Grid-Interconnected Power-Only Applications (2003)
Nominal Capacity (kW)
Cost ($/kW)
Equipment
Genset Package
In terco TIll ec tIE I e ctri c al
Total Equipment
LaborlMaterials
Total Process Capital
Proj ect and Construction
and Management
Engineering and Fees
Proj ect Contingency
Total Plant Cost (2003 $/kW)
100
400
250
650
228
878
030
300 000 000 000
350 370 440 450
150 100
500 470 515 515
175 141 103 103
675 611 618 618
$790 $720 $710 $695
Source: Energy and Environmental Analysis, Inc., estimates
Table 4 shows the cost estimates on the same basis for combined heat and power applications.
The CHP systems are assumed to produce hot water, although the multi-megawatt size engines
are capable of producing low-pressure steam. The heat recovery equipment consists of an
exhaust heat exchanger that extracts heat from the exhaust system, a process heat exchanger that
extracts heat from the engine jacket coolant, a circulation pump, a control system, and piping.
The CHP system also requires additional engineering to integrate the system with the on-site
process. Installation costs are generally higher than power-only installations due to increased
project complexity and the higher perfonnance risks associated with system and process
integration. Labor/materials, representing the labor cost for the civil, mechanical, and electrical
work - as well as materials such. as ductwork, piping, and wiring - is estimated to range from
55% of the total equipment cost for smaller engines to 35% for the largest CHP installations.
Project and construction management is estimated to be 10% of the total equipment cost for all
engines. Engineering and pennitting fees are estimated to range from 10% to 8% of the total
equipment cost depending on engine size. Contingency is assumed to be 5% of the total'
equipment cost in all cases.
Gas-Fired Distributed Energy Resource Technology Characterizations
Reciprocating Engines Page 2-Exhibit No. 130
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 5 of 6
Table 4. Estimated Capital Cost for Typical Reciprocating Engine-Generators in
Grid-Interconnected CHP Applications (2003)
/ ,'.' ",.,;( ..", ..,',.."..~....
sfem. $'" f ." S ""'t' ,
.c'
" .
~m:erP'
... ......
Nominal Capacity (kW)100
Cost ($/kW)
Equipment
Genset Package
Heat Recovery
Interconnect/Electrical
Total Equipment
500
incl.
250
750
Labor/Materials
Total Process Capital
413
163
Proj ect and Construction
and Management
Engmeering and Fees
Proj ect Contingency
Total Plant Cost (2003 $/kW)350
300
350
180
150
680
306
986
$1 ,160
Source: Energy and Environmental Analysis, Inc., estimates
000
370
100
560
240
800
$945
.:$YS
~~,
000
440
580
220
800
$935
000
450
555
210
765
$890
5 Maintenance
Maintenance costs vary with engine type, speed, size, and number of cylinders, and typically
include:
Maintenance labor
Engine parts and materials, such as oil filters, air filters, spark plugs, gaskets, valves
piston rings, electronic components, and consumables (such as oil).
Minor and major overhauls.
Maintenance can be done either by in-house personnel or contracted out to manufacturers
distributors, or dealers under service contracts. Full maintenance contracts (covering all
recommended service) generally cost 0.7 to 2.0 cents/kWh, depending on engine size, speed
and service, as well as customer location. Many service contracts now include remote
monitoring of engine perfonnance and condition and allow predictive maintenance. Service
contract rates typically are all-inclusive, including the travel time of technicians on service calls.
Recommended service is comprised of routine short-interval inspections/adjustments and
periodic replacement of engine oil and filter, coolant, and spark plugs (typically at 500 to 2 000
Exhibit No. 130
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 6 of 6
Gas-Fired Distributed Energy Resource Technology Characterizations
Reciprocating Engines Page
" ,, '""" .:~"",, "
f,."
",
:1"""t..;
~ ", ,
t;t.. . ..;)'
' ", ..,.."",, ' ', , , , , " ," .""'" , .., ," "' ", ,' , . ', "'" ," '
ean-i..
,.., , ,..,-.. ' , ', '
oo
, ., , , ', ," "" ," ", .. '" .. , "
Natura - Gas
hr;; lfecitJtt!:(! ti"jv'fEttlfiJetJr f!J$.):(;)fiS()fft1!;i
a\fifi;W311Jprnt:tntvf tl t1tJYlPl LJt J($t~i:J,.burrfirlfl
g~'A's rfJ(;;.t~!lT)C,fl~/ngfJnginC3;3! .ibut 1,J~tJ llit~h
tl7€w lflJnIWJtd0i;~f/;;:) tU!Lit'JJ!J?
Exhaust-Gas Recirculating
Aftertreatment
")UBSCRiBE
Convinced that reciprocating engines fired by natural gas will
playa major role in the future of distributed energy but that key
technology challenges remain to be addressed, the United States
Department of Energy has set the goal of a more efficient and
cost-effective lean-bum gas engine within the next five to seven
years. The goal for this new era is a fuel-to-electricity conversion
efficiency of at least 50% (30% higher than what's cUITently
available), NOx emissions of 0.1 g/bhp/hr. (a 95% reduction
which still will need aftertreatment to meet tough air-quality standards in such places as
Califomia s South Coast Air Quality Management District), installed capital costs of $400-
$S40/kWe and significant reduction in maintenance costs. The program is called Advanced
Rec.iprocating Energy Systems (ARES) and so far has the support of the major engine
manufacturers working in concert with the national laboratories and selected universities to
expand the use of reciprocating engines for distributed-generation (DG) applications.
Cost of Electricity
Comparison
COMMENT
ON THIS
ARTiCLE
CREATE A LINK
TO THIS ARTICLE
ON YOUR SITE
According to fanner ARES Program Manager Joe Mavec, the project was launched in
September 200 I and will proceed over three phases with research on advanced materials
fuel- and air-handhng systems, advanced ignition and combustion systems, catalysts, and
lubricants. Phase I is scheduled for completion during 2004-200S , while the deadline for final
Phase III is 2009-2010. Cummins Power Generation, Caterpillar Inc., and Waukesha Engine
Dresser Inc. have received Phase I grants and are "following individual research paths " as
John Hoeft, director of marketing for Waukesha, puts it, based on each company s marketing
target. "At Waukesha we re working on the I-megawatt-size product " says Hoeft
, "
and we
looking at a redesign of our VGF (engine), our V16 platform to get there.
A non-nonsense, long-established, and extensively used power-generating technology that
requires fuel , air, compression , and a combustion source, reciprocating engines fall into two
categories: spark-ignited engines fueled by natural gas and compression-ignited engines that
run on diesel fuel. Gas engines are culTcntly available in two versions: rich-bum and lean-
burn, the latter made commercially viable when microprocessors made it possible to
efficiently control critical fuel How and fuel-air gas mixture plus ignition timing. In a lean-
burn engine, excess air is introduced into the engine with the fuel, which reduces the
temperature of the combustion process, which in turn reduces by almost half the amount of
nitrogen oxide produced comparcd to rich-bum engines. And because excess oxygen is
available, combustion is more efficient, producing more power with the same amount of fuel. ExhIbIt No. 131
Case No. A VU-04-
Distributed-power appJications favor A VU-04-
R. Sterling, Staff
6/21/04 Page 1 of 6
$600 per kilowatt.
i)HOrO~ CA l't'JrPILLAR
natural-gas technologies first and
foremost because they deli ver lo\v air
emissions," says Caterpillar s Gas
Product Marketing Manager Michael
Devine. "Diesel-fueled systems still
dominate in standby and shOJi-run
installations, but right now gas is better
at combining availability, price, and
environmental compliance. Gas-fueled
generator sets can be on-line and
producing power within three to six
months of when theyre ordered at a
cost that varies from about $350 to
Devine says Caterpillar has already hit the market with ARES-style improvements. "The
G3500C engine program and its advanced gas-engine control module is an offshoot of ARES.
The new control system solves some of the challenges that have typically affected the
efficiency of lean-burn engines, including maintaining air-fuel ratio and constant emissions
control. "
Technological advances aside, choosing a natural-gas learn-burn generator set from what'
now available requires a thorough assessment of the amount and duration of power to be
generated, which must in turn be balanced against installed cost, engine efficiency, and
emissions control. While large-scale DG applications have sometimes favored 24/7
cogeneration systems, Devine reports that smaller industrial users and some utilities are
opting for selective usage, sometimes running as fe"v as 500 hr./yr.
But Stan Price, project manager for
Northern Power Systems Inc. in San
Francisco, CA, wonders about such
short-hour applications. "We try to
select equipment so that it runs at least
000 to 4 500 hours a year as close to
its full rating as possible. If the
capacity factor is below 60%, I begin to
wonder whether the economics are
going to make sense for the customer.
What's got to drive the decision to put
in a genset for, say, 1 200 hours a year
is the fact that loss of power during an
inten'uptible period is very expensive
in terms onost product. The company
is not just saving money on electricity,
they re saving on product costs.
PHO to:CATERPt Lt..
AtWaukesha, Hoeft thinks the choice of an engine begins with emissions requirements.
Once you look at kilowatt size, you make your decisions based on the product mix and
meeting the emissions requirements, then on how much efficiency you want. It's a tradeoff
between emissions and efficiency and first (installation) costs.
Chach Curtis, vice president of onsite generation for Waitsfield , VT-based Northern Power
Systems, notes that while Jean-burn engines have become the industry standard - particularly
in Europe because they are typically anywhere from 3 to as much as 10(10 more efficient in
converting fuel to electricity - there also is a market for rich-burn engines. "In states like
California and New Jersey and New York and now Massachusetts, both systems are going to
need some kind of aftertreatment. For the rich-burn engines, it's a cheaper, simpler process.
, in these states, you have to look at the higher cost of aftertreatment to meet emissions
standards on a lean-burn engine versus how much additional savings youre going to generate
from the highcr electrical efficiency a lean-burn system is going to give you. Then you have
to determine if that's going to pay for itself in a reasonable timeframe. Tfnot, the customer ExhIbIt No. 131
might be better off with a rich-bum engine and saving some money up-front on the emissions Case No. A VU-04-equipment. A VU-04-
R. Sterling, Staff
6/21/04 Page 2 of 6
A year ago you could install a lean-burn engine in Massachusetts without the tougher area-
based SCR (selective catalytic reduction). And, in California, although they ve extended the
incentive program to the end of 2007, they ve lowered the emission requirements in order to
qualify.
As Curtis points out, the only aftertreatment technology currently on the market to bring lean-
burn engines into compliance where NOx standards are tight is SCR, which some end users
are uncomfortable about utilizing for cost and safety reasons. But because the major
manufacturers are solidly behind lean-bum technology, they are quick to play down states
'vvhere higher emission standards can make compliance costly, and the industry itself is
looking for new aftertreatment technologies to come on-line that will eliminate the perceived
risk of storing and using the ammonia that's added to a lean-bum engine exhaust stream.
Within the next two or three years, you re going to see exhaust gas-circulation technologies
emerging for lean-bum (engines) that will bring them down into compliance " says John
Kelly, director of distributed energy for the Gas Technology Institute (GTI) in Chicago, IL.
But Ritchie Priddy of Attainment Technologies LLC in New Iberia, LA, says that time is
already here (see sidebar).
At Caterpillar, Devine agrees that meeting local emissions standards is one of the factors that
needs to be considered in what he calls "the economic equation " to determine whether
generating your own electricity is competitive against purchasing power from a utility. "When
a user is trying to determine the cost of operation for a gas engine, they usually think of the
installed first cost of the system, the fuel and maintenance costs, but they also need to figure
the cost of meeting the local emissions regulations, which can be met either inside the engine
or outside the engine. With rich-bum engines, there is just enough air to mix with the right
amount of required fuel to make the power required. Given that nitrous oxide is created in the
exhaust stream in the presence of heat, the higher the temperature and the longer the exposure
to that heat, the more NOx will be created. To minimize exhaust emissions, a three-way
catalyst is then used to convert the exhaust gas into essentially water and nitrogen. This type
of system is similar to automotive systems used today - you end up with very high exhaust-
gas temperatures, and because of the way this type of engine consumes fuel, your efficiency is
typically in the 33% to 35% range. A le~m-burn engine deals with most emissions in the
engine. You still have the same amount of fuel introduced into the cylinder to make the
required power, but you re putting excess air into the cylinder with the fuel. You
dis1Tibuting the same amount: of heat over a larger volume, so your exhaust-gas temperatures
are lower, greatly reducing the formation ofNOx. In areas where very low exhaust emissions
are required, a simple oxidation catalyst or SCR may be used to meet the local standards. An
added benefit of lean-burn engines is that the lower exhaust-gas temperatures translate into
higher power density, longer maintenance intervals, and lower owning and operating costs.
After installation, a 1.75-MW cogeneration system at the
Chicago Museum of Science and Industry will provide up to
80% of the museum s heat, hot water, and electricity.
Herman Van Niekerk, vice president of
engineering at Cummins, agrees that a
fundamental difference between rich-
burn and lean-burn engines is that the
lean-burn is more fuel-efficient, but he
adds a qualifier. "As the engine gets
bigger, the gap in performance and
efficiency gets wider. The newer lean-
burns are 39% efficient or better, while
the rich-burns are about 32%. With that
soli of eftlciency gap, you can afford to
do all sorts of aftertreatments to meet
emissions requirements. But if you get
down to 300 kilowatts or less, then the
advantage of having lean-bum over
rich-burn is not that great. You may
(gain) two percentage points of
efficiency with lean-burn, but you have
the cost of the aftel1reatment. I've done
several feasibility studies on lean-bum
projects in which a small unit just
doesn t cut it.Exhibit No. 131
Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 3 of 6
PHotO: I::tJMMIW$
Otherwise it s a purely economical
situation. 'vVe run a feasibil1ty study
with the data we get from the utility
company - every 15 minutes of use -
and from the customer about his site
including his thermal load profile and if
it's a cogeneration project. Then we
model an engine on the resulting load
curve and simulate real-life conditions
for an entire year so we will know
exactly what will happen iJwe try to
generate power on the customer s site.
This makes it easy for us to then
compare rich-burn and lean-bum
engines of different sizes and from
different manufacturers.
The Cummins lean-burn generator set produces up to 1.
MW/hr. of electricity and 4 000 Ib.lhr. of steam.
This process also gives me a financial model, which allO\vs me to give the customer a full
financial-impact study on what it will take to do the job. Some customers want a simple
payback in two to three years. Others want to bOlTOW the money. Our program will take the
cash flow from construction to ten years and calculate the return on investment. Customers
must be clear on these questions before any of the modeling work can be done.
A case in point is a large automobile manufacturer headquartered in Ton-ance, CA , that
elected a simple payback, Van Neikerk says. The company installed a combined heat and
power system that uses a Cummins I.MW natural gas-fired generator with a 250-ton Trane
absorption chiller. Modeling convinced decision-makers that a CHP unit was environmentally
and economically responsible, says Garth Sellers, manager of national facilities services. "
knew that we wanted to generate power, especially with the cost of energy in California. We
also lalew we wanted to use the byproduct of heat. Eventually we determined that we could
use the heat in an absorption chiller to produce air conditioning, which we needed, We
generate enough elect'icity to fully supply our central plant in ToITance during the summer
months. During the winter months and in the evenings and on weekends, we supply several
other buildings on campus. Our goal is to run the generator at 100% load, 98% of the time.
At Northern Power, Pace points out
that there arc advantages to
cogeneration besides what's obvious.
Being an onicial cogenerator based on
the Public Utility Code (means) that
you can apply for incentives, and most
utilities have a special gas tariff rate for
cogeneration, which in some cases is
significantly less than the tariff for
nOlmal boiler heating gas. But one
thing you have to be careful of is the
quality of waste heat you need. Some
processes use 150 psi of steam, and
recip engines are not good matches for
waste heat at ISO-pound steam because
they don t have the required amount of
waste heat at a high enough
temperature. Some manufacturers are
more restrictive than others as to how hot they allow certain waste heat streams to be. Some
\villlimit water-jacket heat to 185i, others will let it go up to 210i, and some (will let it go) as
high as 240i. So understanding the basic energy balance of the engine and the quality of the
heat is important in understanding how you match that specific engine to the process.
PHorO:Al'M(J$ PO\'
From our perspective at GTI " says Kelly, "although heat recovery helps, the really big
impact on decision-making is the electTicity cost in the region. That's the number-one driver.
With utilities having peak and off-peak rates, if you manage the situation coITectly, you can ExhIbIt No. 131
be very economical. At GT!, for example, we run 9 a.m. to 6 p.m. every day, and the payback Case No. A VU-04- l-
on our system is maybe four and a half years. We believe this is the optimum solution because A VU-04-
R. Sterling, Staff
6/21/04 Page 4 of 6
it also takes care of the electrical utility. When we re not running at night, they get to sell their
base, but we re shaving their peale
Whether you re only going to run at peak periods depends on what your nighttime rates are
and what your fuel costs are " says Van Niekerk. "lfyou can generate cheaper than what you
would othen-vise pay for electricity - if you compare both thermal and electric - you always
run the genset 24/7 , and it pays every time. Because even if you only save a penny per
kilowatt-hour, on a megawatt unit, that's almost $100 000 a year. Because deciding when to
run or not is a really tight calculation, at Cummins we also provide a real-time monitoring and
analysis system that will actually look at fuel costs and at electrical rates and then advise the
customer during off peri ods to stop the generator until fuel prices come down.
Except for waste heat, all of these factors\vere figured into decisi on-making when the
research and development operation of a major global manufacturing company based outside
of Chicago decided on self-generation. According to its facilities manager, the company was
experiencing major problems with quality and reliability in the power it received from its
local utility. During summer hot spells, the load could be down by as much as 15%. The
company already had instaUed its own internal distribution network for power it bought off
the grid and its own double-redundant diesel-po\vered system for backup at its corporate data
center. Once the decision was made to generate power on-site, the company brought in Nicor
Solutions, which helped develop the onsite power plant, eventually built the facility, and then
leased it to the client, who runs it on a typical peak-shaving profile, 9 a.m. to 6 p.m. The
company chose two Waukesha VHP 5904-L TD 1- to 25-kW gensets but left enough room in
the building that houses them to add a third unit. "We chose W aukesha " says the facilities
manager
, "
primmily because of their availability in the market, because of their operating
history, and (because of) the fact that they re a relatively simple and straightforward engine. In
my mind, other new technology being offered hadn t been proven. We also liked the fact that
the company is relatively close in case anything happens." Keeping track of fuel costs is
critical to efficient operation. "I'm always looking two years ahead , and when I see that the
price of gas in 2006 is reasonable, I buy a contract and lock in the price. A lot of people do
this, but they don t constantly monitor the market. We have settled into a procedure, which
takes me a minute each moming to look at where our elec1Ticity prices are and then at what
our natural-gas prices are, and then we make a detClmination: Does it make sense for me to
buy energy, leave my plant idle, and sell my natural gas, or does it make sense to generate
electricity on-site?"
Devine agrees that equipment and
operating costs have to be balanced
against what he calls "power reliability
and power quality," and any bottom-
line economic assessment must
consider added costs, such as standby
PHOTO:ATMOSPOWEftSYSUMS charges, exit fees, and additional
incremental costs, for interconnection.
He points to industrial operations, such as Kuntz Electroplating Inc. in Kitchener, ON, where
seconds-long intenuptions in utility-supplied power stopped production for as long as an
hour. The company al-so was experiencing voltage disruptions during periods when high-
demand equipment came on-line, and the resulting damage in solid state processing control
could cause repairs that could shut down production lines for as long as 45 minutes. To solve
these problems, Kuntz installed five Cat G3516 generator sets for a 4.075-MW capacity.
When the system is operating at the rated load, it carries roughly 65% of the plant's total
electrical load; control switchgear sheds noncritical loads in case of utility power
interruptions. The company also recovers heat from engine exhaust and jacket waterloil cooler
circuits to help satisfy a process heat load of 18 million Btu Ihr. for parts cleaning and
electroplating tanks.
Caterpillar also is working with utilities, such as Herber Light and Power (HL&P), a
municipal electric utility in Herber City, UT, to install its own DG systems rather than rely on
customers to pick up peak-time power demands. Devine explains, "\Vhen power shortages hit
Califomia in the summer 01'2000 , HL&P was prepared. By increasing run time on its
distributed-generation resources, which consisted of natural-gas- and diesel-engine-driven ExhIbIt No. 131
generator sets, HL&P avoided purchasing wholesale power at prices that rose from the typical Case No. A VU-04-
$20 per megawatt-hour to as high as $200 per megawatt-hour at peak-demand hours. After the A VU -04-
R. Sterling, Staff
6/21/04 Page 5 of 6
crisis passed, HL&P took further steps to protect reliability and stabilize prices, investing in
three new advanced gas-fueled generator sets rated at a combined 5,52 megawatts. \Vith those
new units on-line as of July 2002, the distributed-generation facility has nine gas and two
diesel units deJivering 11.97 megawatts of capacity. It provides economical load following
year-round and shields HL&P customers against future swings in wholesale power prices, In
case of a major wholesale supply interruption, the facility could cany a substantial share of
HL&P's load, keeping the majority of its customers in service.
Houston, TX-based Atmos Power Systems (APS) designs and installs plants for peak shaving,
shoulder, and interruptible load applications. "Historically," says APS Vice President LaITY
Moore
, "
utiJity-provided power during peak- and shoulder-load operations has always been
the most expensive due to demand charges. APS builds the power-generating facility and
offers its customers long-term leases that allow them to build an equity position in the
generation plant during the term of the contract" One of APS's clients is a food-processing
operation in the Southeast where a large portion of the facility s electricity portfolio was on an
interruptible basis, which meant that the utility had the right, given notice, to reduce power
demand by a certain amount In the face of increasing demands on the utility that supplied its
power, the company wanted to firm up its power delivery and reduce high demand charges.
The decision we had to make " says the company s energy manager
, "
was (this): Do we
continue to take intenuptible power, or do we take the inteITuptible part of our portfolio and
make it fim1? But under most utilities, the real benefit of interruptible power versus firm
pO\ver is that you don t pay the high demand charges. So in effect the demand portion is much
cheaper. So we weighed the increased cost of firming up our interruptible service against the
cost of turning those generators. In effect we were filming up our power because we had
generation on-site.
APS installed a 20-MW plant using 12 Cummins QSV lean-bum generator sets, which
environmentally were permitted to operate 1 200 hr./yr., and then leased the plant to the
customer. Power is generated at 13 800 V and is connected directly to the customer
substation. The company s energy manager acknowledges that leasing the facility rather than
bearing the capital cost of building the plant was attractive but that the company hasn
completely ruled out buying the lease.
With these kinds of numbers, Moore says APS is enthusiastic about the DG market, which he
also predicts will include a combination of utilities and end users. "Utilities benefit from DG
power plants installed in areas of system weakness " says Moore
, "
by being able to defer
capital budget items to upgrade their transmission infrastructure.
Besides emissions, Moore thinks that noise management and equipment maintenance are two
factors that have to be considered from the get-go. "In these kinds of lightly loaded
applications, the life expectancy of a system like we put in with the 12 Cummins gensets is 40
years, after which the engines will be overhauled and allowed to operate for another 40 years.
The key is proper maintenance, which Cummins supplies. The only thing we require of our
customers is that someone walk through and do a periodic check once a day to make sure
everything is running smoothly, that there s no oil on the floor, no antifTeeze. This has the
added benefit that, if five years down the road the customer decides they want to purchase the
power plant, they have people who are qualified and know how it works and are familiar with
its operating history.
Journalist PENELOPE GRENOBLE O'l\IALLEY is afrequent contributor to
environmental publications.
GTI recommends that anyone considering distributed energy develop maintenance
specifications and put them out to hid at the same time they hid the project. Van Niekerk
describes Cummins s "bumper-to-bumper" guarantee as "a fixed feed per kilowatt-houLThe
customer knows exactly \vhat it's costing him to generate electricity. For a penny or a quarter
of whatever that number is per kilowatt-hour, we provide full waITanted maintenance and a
monitoring system, which automatically calls out so everybody lmows what's going on and if 131there arc any problems.1 It
. Case No. A VU-04-
A VU -04-
R. Sterling, Staff
6/21/04 Page 6 of 6
DE - March/April 2004
CERTIFICATE OF SERVICE
HEREBY CERTIFY THAT I HAVE THIS 21ST DAY OF JUNE 2004
SERVED THE FOREGOING DIRECT TESTIMONY OF RICK STERLING, IN
CASE NO. AVU-04-l/AVU-04-, BY MAILING A COpy THEREOF POSTAGE
PREPAID TO THE FOLLOWING:
DAVID J. MEYER
SR VP AND GENERAL COUNSEL
VISTA CORPORATION
PO BOX 3727
SPOKANE WA 99220-3727
KELLY NORWOOD
VICE PRESIDENT STATE & FED. REG.
VISTA UTILITIES
PO BOX 3727
SPOKANE WA 99220-3727
CONLEY E WARD
GIVENS PURSLEY LLP
PO BOX 2720
BOISE ID 83701-2720
DENNIS E PESEAU, PH. D.
UTILITY RESOURCES INC
1500 LIBERTY ST SE, SUITE 250
SALEM OR 97302
CHARLES L A COX
EV ANS KEANE
111 MAIN STREET
PO BOX 659
KELLOGG ID 83837
BRAD M PURDY
ATTORNEY AT LAW
2019 N 17TH ST
BOISE ID 83702
CERTIFICATE OF SERVICE