20040220Sterling Direct.pdf

RECEIVED2004 February 20 PM 4:59IDAHO PUBLICUTILITIES COMMISSIONBEFORE THEIDAHO PUBLIC UTiliTIES COMMISSIONIN THE MATTER OF THE APPLICATIONOF IDAHO POWER COMPANY FORAUTHORITY TO INCREASE ITS INTERIMAND BASE RATES AND CHARGES FORELECTRIC SERVICE.) CASE NO. IPC-O3-DIRECT TESTIMONY OF RICK STERLING
IDAHO PUBLIC UTiliTIES COMMISSION
FEBRUARY 20, 2004
Please state your name and business address forthe record.My name is Rick Sterling.My business addressis 472 West Washington Street, Boise, Idaho.By whom are you employed and in wha t capaci ty?I am employed by the Idaho Public utili tiesCommission as a Staff engineer.What is your educational and professionalbackground?I recei ved a Bachelor of Sci ence degree in Ci vi IEngineering from the Uni versi ty of Idaho in 1981 and aMaster of Science degree in Civil Engineering from theUni versi ty of Idaho in 1983.I worked for the IdahoDepartment 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
1 7 Utilities Commission in 1994 My duties at the Commission
include analysis of utili ty applications and customer
peti tions
What is the purpose of your testimony in this
proceeding?
The first purpose of my testimony is to discuss
? ')- ,
the methodology and resul ts of Idaho Power's load
normalization , and to make a recommendation on whether
believe the Company s resul ts should be accepted.Next, I
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
discuss the Company's power supply modeling and discuss anal ternati ve method that I used to evaluate Idaho Powerresul ts Finally, I discuss the Danskin project and makea recommendation on whether I believe the project costsshould be allowed in rate base.Load N orlualizatiollWhat is load normalization?Load normalization is a process to determinewhether actual electrici ty sales were higher or lower thannormal as a resul t of actual weather.Energy use statistically estimated as a function of weather and non-wea ther variablesWhy is load normalization important and how does
it affect the Company's revenue requirement?
Load normalization is important because it
establishes the loads that must be met by Idaho Power in a
1 7 normal year which in turn are used for jurisdictional
separation normalization of power supply costs , and cost
of service.Normalized loads are also used to determine
the revenue that the utili ty would be expected to receive
in a normal year.
Please describe the load normalization performed
? ')- ,
by the Company in this case.
Idaho Power used mul tiple regresslon analysis
normalize loads Normalization was performed separately
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
r;,
uslng eleven different regresslon equations - two thatdescribe Idaho Power s total system residential andcommercial sales , two that describe Oregon s residentialand commercial sales, five that describe irrigation salesfor each of thedescribessal esdescribessal es Company's operating centers , one thatto the Ci ty of Wei ser , and one tha to Raft Ri ver Rural Electric Cooperati ve,Inc.To explain electrici ty use , the regression equa tionsutilize weather concepts such as heating, cooling andgrowlng degree-days and precipi tation , as well as economlCand demographic information such as electrici ty price,electric space heat saturation , and air conditioningsaturation.Once regression equations were developednormal variable values were entered into the equations
compute normalized loads These normal loads were then
used by the Company in i ts power supply modeling,
1 7 jurisdictional allocation and cost of service studies
Do you agree wi th the normalized loads proposed
by the Company?
Yes I do.The regression equations developed
by the Company are very accurate predictors of usage by
various customer groups based on historic condi tions and
? ')- ,
consumption levels The correlation coefficients obtained
by the Company that indicate the accuracy of predictions
in its analysis are very high.I bel i eve tha t the
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
..5
methodology used by the Company is appropriate and thatthe resul ts are reasonable.Power Supply l\tlodelillgHave you revi ewed the power supply model i ngperformed by the Company as part of this case?Yes, I have.Do you agree wi th the normal i zed power suppl costs proposed by the Company?Al though I believe the power supply model theCompany used in this case could be improved I concludethat the normalized power supply costs proposed by IdahoPower appear conserva ti ve and so Staff does not oppose theCompanys proposal The Company compu ted a ne t power
supply cost of $49.6 million for the 2003 test year.With
known and measurable adjustments , the Company is proposlng
that a net power supply cost of $47 7 million be adopted
1 7 in this case.In the Company s last general rate case
(Case No. IPC-E-94-, a normalized net power supply cost
of $48 million was accepted.
Why is the Company s normalized net power supply
cost nearly the same as it was in Idaho Power's last
general ra te case?
? ')- ,
As discussed in Company wi tness Said'
testimony, several factors have caused upward pressure on
power supply expenses while others have caused downward
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
pres sure.The net effect of these factors has caused amodest $1 . 9 million lncrease in normalized net powersupply costs before known and measurable changes Afterknown and measurable changes, the difference lS a $0million decrease from the last rate case.As described in Mr. Said's direct testimony,factors that have caused upward pressure on power supplycosts include higher market prices along wi th higherseasonal and peak hour loads tha t must often be met usinghigher cos t resources Factors tha t have caused downwardpressure on power supply costs include a slight netdecrease in annual system load expiration of FERCjurisdictional contracts , and overall decreases in coalcontract prlces
Did you explore or devise an al ternative method
to evaluate the normalized power supply expenses proposed
1 7 by the Company.
Besides reviewing the Company s determination of
normalized power supply expenses using AURORA I also
performed a regression analysis to estimate a range of
normal power supply expenses In the analysis, I chose
the following eight independent variables that affect
? ')- ,
power supply costs
(a) Brownlee inflow
(b)Installed generation capaci ty
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
...J
(c)Electric market pr 1 (d)Unit cost of fuel at Bridger(e)Unit cost of fuel at Boardman(f)Unit cost of fuel at Valmy(g) System firm load(h) PURPA purchasesI used net power supply cost as the dependentvariable in the regression analysis I used twenty-fouryears of historical data in the analysisWha t did you hope to accompl i sh wi th yourregression technique?My goal was simply to generally compare theval ue proposed by Idaho Power to es tima ted net powersupply cos t using other methods
What did you conclude from your regresslon
analysis?
1 7 I concluded that the normalized net power supply
expenses proposed by Idaho Power are reasonable and are
probably low.
Do you recommend that the Commission accept the
normalized net power supply costs as proposed by Idaho
Power?
? ')- ,
Yes I do.However I al so recommend tha t the
Company and Staff moni tor the actual net power supply
costs in the coming few years to assure actual net power
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
supply expenses properly track water condi tionsDallsliillPlease summarlze Commission Order No. 28773(Case No. IPC-E-01-12) concernlng the Danskin plant.In Order No. 28773 , the Commission authorizedIdaho Power to proceed wi th the construction of theDanskin plant.In doing so , however , the Commissionstated:We note that the procedure followed in thiscase has limi ted the type and extent ofreview that would otherwise occur in acertificate filing.The information provided however insufficient to determine the reasonablenessof the related costs As reflected in Staffcomments, it is unknown whether the MountainHome Station was the least cost al ternati ve.
Because the Mountain Home Station was not
selected pursuant a RFP process , we are unable
to conclude based on the information provided
that the commi tment estimate is reasonable. The
Company in its Application , we note, also
provides no compari son of al terna ti ves
(al ternati ves available but not chosen)
...
There
is no record as to whether other al ternatives
were also considered and rejected. We are
unconvinced tha t the bes t measure of the cos t
al ternative resources is market price estimates
in effect at the time the decision to proceed
was made. The record supporting such a finding
remai n s to be developed.
1 7
We find that there is insufficient record
assess and determine the reasonableness of the
Company s commi tmen t e s tima te and canno therefore provide the Company wi th a doll
amount of rate base assurance. As we indicated
in our prior Milner decision Order No. 23520
when the Commission authorizes construction of
new generation
,...
it informs the Company, its
? ')- ,
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
ratepayers and its investors that , in theordinary course of events prudently incurredcosts of construction in bringing the authorizedplant on line will later be recognized in theCompanys revenue requi remen too." a t page 20then went on to discuss examples of what type ofrecovery is not guaranteed. That being said, wenevertheless note that implici t in our decisionin this case to approve a certificate forconstruction of the Mountain Home Station recovery of some reasonable amount as rate baseaddi ti on The Company needs to provide theCommission wi th more information. What otheral ternatives were considered? What was theCompany's f oreca s ted need? The Companyexpressed concern that we will assess itsdecision to build based on hindsight and from perspecti ve of changed market condi tions assure the Company that the review standardemployed by the Commission will be what Companyknew or should have known a t the time i t madeits decision to build.Did Idaho Power provide addi tional justificationfor Danskin in its testimony in this case?
No.
Why is Staff providing testimony in support of
1 7 Danskin cost recovery when the Company did not?
Danskin s plant cost recovery represents a large
portion of increased revenue requirement requested in this
case.Staff believes it is important to address the issue
and provide the Commission wi th the Staff posi tion.
Has the Commission Staff audi ted the construction
? ')- ,
costs for the Danskin plant?
Yes The total plant cost including the
substation , step-up equipment, and structures and
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
improvements is $52 484 209 as of year-end 2003.Do you believe all of the costs incurred forconstruction of the Danskin plant are reasonable andshould be allowed in rate base?Yes I do.The plant's capi tal costs wereprojected to be $46 million upon completion in 2001 Withan addi tional 20% for contingencies, Idaho PowerCommi tment Estimate" for the capi tal cost portion of theplant was $55 2 million.The Staff-audi ted cost of $52mi II ion is cl ear I y below the Company s commi tmen testimate.The Danskin plant was nearly as costly to buildas the Bennett Mountain plant is expected to be yet theBennett Mountain plant will have a capaci ty of 162 MW
compared to Danskin's 90 MW.Why was Danskin so expensi
compared to Bennett Mountain?
1 7 The commi tment estimate for construction of the
Bennett Mountain plant is $54 million while the cost of
Danskin was $52 5 million.Bennett Mountain's uni t cost
therefore , is expected to be $336 per kW while Danskin
was about $583 per kW - more than 1 7 times the cost of
Benne t t Moun tai n .
? ')- ,
One reasonable measuring stick for Danskin'
plant cost is generating plant cost estimates prepared by
the Northwest Power and Conservation Council for use in
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
i ts Fifth Power Plan.The estimates were prepared onApril 5 , 2002 , therefore , they are likely veryrepresentative of costs at the time Danskin was built.Al though the Fifth Power Plan has yet to be released, itspower plant cost assumptions have not changed.TheCouncil's capi tal cost estimate for gas-fired simple cycleplants ranges from $540 to $660 per kW , wi th $600 per kWbeing the base case estimate.Danskin's cost of $583 perkW is very close to the Council's base case estimate.Bennett Mountain s expected cost of $336 per kWis very low compared to simple cycle plant costs of justtwo years ago.The demand for gas turbines surged in the1998-2001 time frame peaking in 2000 During this timeperiod, turbine manufacturers could not keep pace wi
orders for new equipmen t and buyers bargained wi th each
other for higher slots on manufacturer's wai ting lists
1 7 Since that time , however electric market prices have
modera ted and demand for new gas turbi nes has pI ummeted
At the time Idaho Power commi tted to Bennett Mountain
turbines could be obtained at a highly discounted prlce.
That is the primary reason Bennett Mountain is so much
cheaper than Danskin on a cost per kW basis
? ')- ,
Wha t has been the actual cos t of energy from
Danskin?
The Company s Application in the Danskin Case
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
(Case No. IPC-E-01-12) indicated that the preliminaryestimate of the levelized cost per MWh would range from upper level of $223 per MWh based on a capi tal cost forthe plant of $55 2 million , 500 hours of annualgeneration, and levelized fuel costs of $5 05 per MMBtuover the 30-year life of the plant , to a lower range costof $77 per MWh based on a plant cost of $46 million , 5140hours of annual dispatch , and average fuel costs of $5per MMB tu The actual cos t of the plan t ended up beingcloser to the high estimate but the actual hours ofoperation has been close to the low estimate.Gas priceshave varied substantially throughout the past two yearsand the estimated gas price may still be reasonable overthe 30-year plant life.Consequently, Danskin s actual
energy cos ts have so far been much closer to $223 per MWh
than to $ 7 7 per MWh.Future changes in gas prices and
1 7 operating hours will , of course , change the cost of energy
from the plant.
If the cost of energy from Danskin is so
expensi ve , why did Idaho Power build the plant?
First, it is important to recognize that the
Danskin plant is a peaking plant , not a base-load plant.
? ')- ,
As a peaking plant , it is intended to be operated for only
brief periods during peak hours in the summer and winter.
Peaking plants will always have high energy costs due
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
thei r I imi ted opera ti ng hour s .Second , i t is impor tan t to remember thecircumstances at the time the decision was made cons truct the Danskin plan Idaho Power made i tsdecision to pursue construction in early 2001 , at theheight of the electric market price run-up. Idaho Powermarketing and trading analysts were predicting that heavyload period market prices for the next few years wouldlikely be in the range of $50 to $350 per MWh , and thathourly prices could exceed $1000 per MWh in the near term.A severe drought also persisted throughout the Northwestat that time which was part of the reason for such highmarket prices This combination of exceptionally lowstream flows and extremely high market prices forced
utili ties to scramble for al ternati ves to meet load.
Beginning in mid-2000 , Idaho Power found it
1 7 necessary to go to the electric market and make large
purchases at extremely high prices Consequently, the
Company began deferring massive power supply costs unlike
any tha t had been made before.The upper graph of Exhibi
No. 124 shows the Company s PCA deferrals between 1999 and
2003 In sinqle months from late 2000 to mid 2001 , total
? ')- ,
deferrals frequently exceeded $20 million and sometimes
approached $50 million.In early 2001 , no one knew how
much longer extremely high market prices would persist.
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
We did know , however , that drought condi tions could notend until at least the following winter.In response to the dire circumstances , inJanuary 2001 , Idaho Power began identifying al terna ti vesto market purchases . In addi tion to building a simple-cycle peaking plant , the Company planned buy-backs fromirrigatorsASTARIS and Simplot.The Company al so plannedto lease mobile diesel generators and to purchase hedgesto guard against price volatili ty.Later , on May 1 , 2001anticipating continued high prlces and poor stream flowsthe Commission issued Order No. 28722 in Case Nos . IPC-E-01-7 and IPC-E-01-11 directing Idaho Power to prepare andfile a report which would identify and outline plans formeeting loads during the summer and winter of 2001
The Danskin proj ect , wi th i ts short cons truction
I ead time , wa s in tended to be on -I i ne in time to provi de a
1 7 resource that could mi tigate exposure to extremely high
near-term market prices
Did Idaho Power lssue a request for proposals or
solici t bids for the Danskin project?
No, Idaho Power did not issue a request for
proposals , nor did it formally bid the equipment contract
? ')- ,
or the cons truction con tract.While conceding in Case No.
IPC-E-01-12 that an ideal way to determine the cost of
available al ternative resources would be to ini tiate
CAS E NO. I PC - E 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
..5
request for proposals , the Company contended that pursulngthe RFP route would likely have delayed the resourceacquisi tion until 2002 , thereby exposing the Company toincreased levels of market purchases through fall and intothe win ter season.Bef ore the ex treme pr ice run -up began , howeverIdaho Power did issue a Request for Proposals as a resul of its 2000 IRP.The Company received proposals for gas-fired combustion turbines and coal-fired generation.addi tion , the Company evaluated self-build al ternativesusing gas-fired combustion turbines The Garnet proposalwas eventually selected , al though the project was laterabandoned.The proposals received during this processgave Idaho Power at least some indication of the costs of
new gas-fired generation.However because the RFP was
seeking 250 MW of capaci ty during a limi ted number of days
1 7 in only fi ve mon ths I do not believe the bids provided a
fair approximation of the cost that could be expected for
90 MW simple cycle plant.Al though the RFP was broad
enough that smaller projects could be proposed , only a
handful of proposal s were recei ved in response to the RFP
and of the propo al s re cei ved only two we re for Ie s s than
? ')- ,
the reques ted amoun t of capaci ty and energy.
In the Company s 2000 IRP , a number of other
technologies for generation were evaluated, including
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
coal combined cycle gas wind and other renewables Theevaluations were non-si te-specific , however , and most werenot realistic al ternati ves to building a simple cycleplant due to the urgency wi th which new generation wasneeded.How did the Danskin plant compare to the otheral ternatives available to Idaho Power at the time?Obviously, one of the al ternatives constructing Danskin would have been to continue to makeenergy purchases from the market.However given theexceptionally high prices, poor stream flow condi tions,and the extremely high PCA deferrals , it was believed thatcontinued reliance on the market would only exacerbate theproblem.
Another option was to ini tiate buybacks wi th
some of i ts I arges t cus tomer groups Idaho Power agreed
1 7 to purchase 50 MW from ASTARIS for a two-year period at a
cost of $159 per MWh.Thirty megawatts were also
purchased from Simplot at $75 per MWh in the first year
$90 per MWh in the second year and 85% of market price
the thi rd year.An addi tional block of 8 MW was purchased
from Simplot at two-thirds of market prlce.A buy- back
? ')- ,
program for large commercial and industrial customers was
also ini tiated but no customers participated.
A buy-back program for irrigators was also
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
impl emen ted The Company purchased 262 MW of loadreduction at a cost of $150 per MWh.Two large QF con tracts , one wi th Simplot and onewi th Amalgamated Sugar , were also re-negotiated duringthi s time frame.Finally, the Company leased mobile dieselgenera tors The generators were capable of providing MW at an estimated cost of $124 per MWh.Exhibi t No. 125providessumma rycon trac ts pur suedthe price run -up.Over the of the short - term programs andduring this time period in response course of time during which they werein effect most of the programs proved quite expensive.The ASTARIS buy-back cost a total of nearly $128 million.
The irrigation buy-back cost $86 million.The mobile
diesel generators , despi te never being used to satisfy
1 7 load , cost almost $5 5 million.The lower graph on
Exhibit
each
cost
No. 124 shows PCA deferrals by month as a resul
of these three measures Compared to the total
these alternatives , Danskin s $52 5 million
capi tal cos t doesn t seem so large.In analyzing the
Danskin project , Idaho Power estimated the present value
? ')- ,
of the revenue requirement over the 30-year expected plant
life to be approximately $180 million.
Didn t Idaho Power receive an unsolici ted
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
competing proposal for the Danskin plant?Yes Power Development Associates , LLC of Boisesubmi tted a proposal to Idaho Power to install two 45 MWsimple cycle gas turbines near Mountain Home at a si tedifferent than the Danskin si te.The proposed turbinesbelieved, were more efficient in a simple cycle mode thanthe turbines Idaho Power planned to install but were lessefficient in a combined cycle mode.I daho Powereventually rejected the proposal primarily because ofuncertainty about whether the project could come on-linesoon enough to meet the Company s immediate need to berelieved of purchasing from the market.As it turned out , Power Development AssociatesLLC was the predecessor to Mountain View Power , Inc., the
successful bidder to construct the Bennett Mountain plant.
The si te of the Bennett Mountain plant is the same as the
1 7 si te proposed as an al ternati ve to Danskin.Bennett
Mountain s plant capaci ty and equipment package
different than what was proposed ini tially, however.
Power Developmen Associates proposal had been selected
an alternative to Danskin , the Bennett Mountain plant
would not have recently been available as an option.
? ')- ,
Do you believe Idaho Power adequately considered
other al ternatives to construction of the Danskin plant?
Yes, I do, gi ven the ci rcums tances tha t exi ted
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
1 7
at the time the decision to build Danskin was made.What has been the history of operation of theDanskin plant so far?Since the plant went on-line at the end ofSeptember 2001 , the pI an t has opera ted on average abou 500 hours per year.The plant has been operated most the summer months, al though it has operated at least somein every month of the year.Exhibi t No. 126 shows thegeneration of the plant by month since it went on-line inSeptember 2001Will construction of the Bennett Mountain plantmake the Danskin plant no longer useful?I don t believe so.Operation of theDanskin plant could change after Bennett Mountain becomes
available but I believe Danskin will continue to be used
to meet peak loads primarily in the summer and winter.
1 7 Bennett Mountain will be a more efficient plant than
Danskin , thus it will have a lower dispatch cost.
However Bennett Mountain will not always be able to meet
the Company s peak load requirements by itself making
Danskin necessary.In addi ti on , I thi nk there coul d
times when Danskin would be dispatched before Bennett
? ')- ,
Mountain because Danskin's two 45 MW turbines can be
dispatched independently, whereas Bennett Mountain will
have a single 162 MW uni Small peak load needs might be
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF
more economically met uslng Danskin despi te i ts higherdispatch cost.Does this conclude your direct testimony in thisproceeding?Yes i t does
1 7
? ')- ,
CAS E NO. I PC - E - 0 3 - 1 3
02/20/04
STERLII-TG, R. (Di)
STAFF