20140702Comments.pdf

Teresa A. Hill
ISB No. 6175
K&L Gates LLP
One Columbia St. Suite 1900
Portland, OR 97258
Telephone: [phone]
Fax: [phone]
Attorneyfor Renewable Northwest and American Wind Energt Associotion
BEFORE THE
IDAHO PUBLIC UTILITIES COMMISSION
itIC* t1,i il i:i
2014 JUL -2 Pil lr: 35
l-! I I rr
UT lLlIl IS CC[iiri i;isi iii:
TN THE MATTER OF THE
APPLICATION OF IDAHO POWER
COMPANY TO UPDATE ITS WIND
TNTEGRATION RATES AND
CHARGES.
) Case No. IPC-E-13-22
)
) CoMMENTS OF AWEA AND
) RENEWABLE NORTHWEST
)
INTRODUCTION
Pursuant to Idaho Public Utilities Commission ("IPUC" or "Commission") Rule of
Procedure 203 and the Modified Procedure Schedule agreed to by the parties on May 27,2014,
the American Wind Energy Association ("AWEA") and Renewable Northwestl hereby file these
Comments on Idaho Power Company's ("Idaho Power" or the "Company") November 29,2013
Application to Update Wind lntegration Rates and Charges (the "Application"). Idaho Power's
Application seeks to increase the wind integration rates and charges applicable to qualifying
facilities ("QFs") under the Public Utility Regulatory Policies Act of 1978 ("PURPA").2 As
originally presented to the Commission, the Application would modifu the rates of both QFs
I Following the commencement of this proceeding, in recognition of the organization's twentieth
anniversary, Renewable Northwest Project changed its name to "Renewable Northwest."
2 In the Matter of the Application of ldaho Power Company to Update its l(ind Integration Rates and
Charges, Case No. IPC-E-13-22, Application at Part IV (Nov. 29,2013).
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
with existing contractual legally enforceable obligations and any new wind generator QFs that
may come online in the future.3
Several parties filed Motions to Dismiss the Application on the basis that it requested the
Commission to modiff the rates and terms in existing contractual legally enforceable obligations
without the QFs' consent.4 AWEA and Renewable Northwest filed comments in support
ofdismissal and requested workshops to address the issues raised in the Application and
accompanying Wind Integration Study.s
In Order No. 33030, the Commission denied the Motions to Dismiss, but clarified that
any Commission-approved modifications to Idaho Power's wind integration rates and charges
would only apply prospectively to new contracts.6 The Order also provided that intervenors
seeking to protect the rights of existing contracts could withdraw from the proceeding if they
believed they no longer had a direct and substantial interest in the proceeding.T Although some
intervenors withdrew from the proceeding, our organizations have remained involved to
represent our continuing interest in accurate wind integration studies, discussed further below.
II. COMMENTS
Before addressing the merits of the Application, we wish to clarify our continued interest
in this proceeding. Our organizations have been deeply involved with wind integration studies
and proceedings on wind integration issues for many years. This proceeding is unique in that we
t Id.
a See, e.g., In the Matter of the Apptication of Idaho Power Company to update its Wind Integration
Rates and Charges, Case No. IPC-E-13-22, Motion to Dismiss of Cold Springs Windfarm, LLC, et al.
(Jan. 3 1,2014).
s In the Matter of the Application of ldaho Power Company to (lpdate its Wind Integration Rates and
Charges, Case No. IPC-E-13-22, Comments of AWEA and RNP in Support of Motion to Dismiss (Feb.
7,2014).
6 In the Matter of the Application of ldaho Power Company to Update its Wind Integration Rates and
Charges, Case No. IPC-E-13-22, Order No. 33030 at 8 (Apr. 30,2014).
' Id.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
do not expect the outcome to have a near-term impact on the rates associated with new wind
development on Idaho Power's system, as we do not anticipate ony new wind projects selling
their output to Idaho Power under PURPA in the near-term-largely due to the current published
rate cap and Idaho Power's onerous contractual provisions. Nevertheless, we have remained in
this proceeding because of the fundamental flaws in Idaho Power's most recent Wind Integration
Study (the "2013 Study") that underpins the Company's Application in this proceeding. It is in
the public interest to ensure the technical accuracy of such studies, and failure to do so runs
counter to the interest ofratepayers and shareholders alike.
Our Comments here focus on identifring the major flaws in the Company's 2013 Study
and making recommendations on how to correct these flaws. The primary shortcomings of the
2013 Study are that it does not accurately portray Idaho Power's actual operating procedures or
best practices by (1) using the day-ahead wind forecast error instead of the hour-ahead forecast
error in calculating the reserve requirement for wind; and (2) calculating reserve requirements
based on the outdated assumption that reserves accommodate wind variability on a stand-alone
basis, when in reality, grid operators balance the deviations of net load (load minus wind and
other generation).8 In addition to these methodological flaws, the Company has not accounted
for available operational tools that enable u,ind to be integrated more efficiently. Our analysis
suggests that these methodological shortcomings result in a roughly three-fold exaggeration of
the incremental reserve requirement for wind.
In connection with our analysis, we have performed some calculations on what we would
expect the correct rates to be, and those rates are lower than those proposed in the Application.
However, more information is needed to arrive at fully accurate rate calculations. What our
8 We use the terms "balancing reserves," "reserve requirement," "Reg-I-Jp," and "Reg-Down" all to refer
to the same flexible generating capability necessary to accommodate the within-hour variability of load
and generation (including wind, solar, and conventional generation).
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
calculations demonstrate is that the flaws in Idaho Power's 2013 Study result in excessive,
inaccurate wind integration charges. We recofirmend that the Commission refrain from
approving the Company's proposed wind integration rates and charges that would apply to any
new QFs or be used in any integrated resource plan until such time as Idaho Power revises its
2013 Wind Integration Study to address the flaws identified herein.
1. Comparison of Idaho Power's 2007 and 2013 Wind Integration Studies.
Alarmingly, it appears that the Company's use of best practices in its wind integration
methodology has actually diminished over the past seven years, as many of the errors in the 2013
Study were not made in Idaho Power's 2007 Wind lntegration Study (the"2007 Study").
In the Company's 2007 Study, which includes the original Study and an Addendum,e the
Company employed the services of the well-respected electric engineering and consulting firm,
EnerNex. This earlier analysis was well grounded in sound practices and employed the then-
current methodologies. Specifically, the2007 Study appropriately relied on the hour-ahead
forecast error for wind, and netted the variability of load, wind, and conventional generation to
reflect the reality that balancing reserves are held for net load variability, which is far smaller
than the sum of the parts of the reserve requirements for wind and load.
In preparing the 2013 Study, Idaho Power has abandoned the use of outside consultants
to inform the study methodology and has instead chosen to pursue methodologies that are
inconsistent with current utility practice and wind integration science. Most importantly, Idaho
Power's 2013 Study does not incorporate the use of hour-ahead wind forecasts and fails to net
the reserve requirements of wind and load; these methodological flaws are described in detail in
the following two sections.
'Idaho Power's 2007 Study was conducted in February of 2007, and then updated with the Addendum in
October of that same year.
IPC-E-13-22 -July 2,2014 Comments of AWEA and Renewable Northwest
As a general comment, we struggle to understand why Idaho Power has moved away
from the well-established best practices for wind integration studies it used seven years ago,
particularly given that most utilities have made strides in this area. Indeed, most utilities have
greatly improved their wind integration operations over the past seven years by improving their
wind forecasting and scheduling capabilities and by using dynamic approaches to committing
and holding balancing reserves. In contrast, Idaho Power's 2013 Study suggests that the
Company has become less effrcient at integrating wind.
2. Use of Day-Ahead Versus Hour-Ahead Forecast Error.
One of the primary methodological errors in the 2013 Study, and a major driver of the
costs identified by the Study, is the use of day-ahead forecasts to predict the output of wind
plants. Use of "day-ahead" forecasts in this context means that system operators set their wind
forecast for a given hour the day before the actual hour ofproduction occurs. In contrast, "hour-
ahead" forecasting uses wind forecasts set during the hour before the wind is generated. This is
an important issue because hour-ahead wind forecasts are inherently more accurate than day-
ahead forecasts. Indeed, it is a well-established scientific fact that wind energy forecast error is
greatly reduced as one moves closer to real-time, as one would expect for any forecast.lo
'o For example, data from Germany show that the normalized root mean square error ('NRMSE") is
reduced from 5 .7%o of installed wind capacity for a day-ahead forecast to 2.6%o for a 2-hour ahead
forecast. Hannele Holttinen, et al.,Design and Operation of Power Systems with Large Amounts of
Wind Power at28-29 (2009), ovailable athttp:llwww.vtt.filinflpdfltiedotteetl2}}9lT}493.pdf. Hour-
ahead forecasts show even greater accuracy, with NRMSE falling well below 2o/o at a forecast horizon of
one hour (as shown by the blue and brown lines in the EU chart, indicating the NRMSE for European and
German wind energy forecast models respectively). 1d. Closer to home, data from the Alberta
independent system operator ("AESO') shows a nearly identical relationship, with wind energy forecast
enor falling drastically as one gets closer to real-time, as shown in the AESO chart. See R. Widiss and K.
Porter, A Review of Variable Generation Forecasting in the West, NREL Report SR-6A20-61035 at 4
(July 20 I 3 ), avail able ar http://www.nrel. gov/docs/f,i I 4osti/6 I 03 5.pdf.
IPC-E-13-22 - July 2, 2014 Comments of AWEA and Renewable Northwest
.E 5,0004
-
E 4oooo/ol[t
E H*,oor"ES
*,Ep.oo"r,*EHo
H t I .oo"/o
!EE2, o,osyo
2 3 4 5 6 7
ForecaS horlzon [hl
AESO Shortterm Foracast Mean Absolute Error August 2012
Forecast
Fersistence
6
Forecast Horizon Ihours]
t4
t2
10
I
6
4
2
0
1?10
'{-Sgsmo-EU-# persistence{r- Cosmo-DE
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
Idaho Power's 2007 Study uses hour-ahead forecasts to calculate the integration costs-
not day-ahead forecasts. Section 7 of the 2007 Study explains how the incremental reserve
requirement for wind is input into the Vista DSS model: it assumes that the next-hour wind
forecast is delivered to operations at 65 minutes before the next hour. The 2007 Study explains,
"this allows the operations group time to assess, plan and execute transactions to meet next hour
needs in a manner consistent with current trading practices."ll In other words, the2007 Study
does not use the day-ahead forecast error to calculate the incremental reserve requirement need.
Using day-ahead forecasts, as Idaho Power does in the 2013 Study, is inconsistent with
other wind integration studies and with corlmon utility scheduling and dispatching practices.
The effect of this methodological error is to greatly inflate the amount and cost of balancing
reserves needed to integrate wind on Idaho Power's system. The impact of this methodological
change in the 2013 Study can be observed by comparing the results of the two studies. For
example, whereas the2007 Study finds that 900 MW of wind would require an incremental
reserve requirement of 98 MW of Reg-Up and 98 MW of Reg-Down, the 2013 Study finds that
800 MW of wind would require between 199-274 MW of Reg-Up and between 219-353 MW of
Reg-Down. In short, the changes Idaho Power has made to the study methodology cause a
roughly three-fold increase in the amount of balancing reserves needed for wind.
When asked about the rationale for basing the reserve requirement for wind on the day-
ahead schedule error in the 2013 Study, Idaho Power responded:
Idaho Power views the simulation of day-ahead scheduling as
appropriate due to system scheduling practices. Day-ahead
scheduling is reflective of the time frame in which Idaho Power
makes dispatching decisions... [if not] ... the amount of balancing
reserves is smaller because it is based on the hour-ahead errors in
forecast wind... [and so].. . the dispatchable generators would not
tt 2oo7 Study at 46.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
be scheduled to allow them to respond to day-ahead forecast
errors, meaning that the response to these larger errors is only
achieved by some other means, which in Idaho Power's view
would too often translate to a risky reliance on the wholesale
electric market.tl2l
We disagree with Idaho Power's characterization of system scheduling practices, as it
exaggerates the importance of day-ahead "scheduling." Standard utility practice is to "dispatch"
generation close to real-time, while the day-ahead process of selecting which plants are likely to
operate the next day is called "unit commitment." Unit commitment is very different from
dispatch, as unit commitment typically only entails evaluating supply and demand to determine
which units are likely to be needed the next day so that any units that require a long lead time to
start up can begin to do so. Unit commitment decisions frequently deviate greatly from how
power plants are dispatched the next day, as actual dispatch decisions change with evolving
supply and demand conditions.
We understand that Idaho Power starts its system management with a day-ahead forecast
for load, wind, other generation, and off-system transactions. However, Idaho Power is well
aware that its day-ahead load forecast will be wrong and that it needs to have enough generation
available to meet its load under all reasonabie conditions. As the day unfolds, Idaho Power must
adjust to real-time conditions and make hourly dispatch decisions to meet its load and all other
contractual arrangements in a least-cost manner.
Indeed, any utility system is constantly changing its hour-to-hour dispatch to serve load
reliably and at least-cost. Dispatchable generation has always been changed on an hourly basis
to follow the net of hourly changes in load and deviations in the output of conventional
generators, and operators use the same approach for following net load changes on systems with
" Idaho Power Response to Data Request No. 4.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
wind. As the wind comes up, a utility will back off more expensive generation previously made
available in the day-ahead timeframe to meet load, which saves the utility operating costs. As
the wind drops off, a utility will increase the use of its more expensive generation, hour-by-hour,
adjusting for the wind's output and mindful to not use any more higher-cost generation than
necessary.
Because Idaho Power can use hourly adjustments from its other generators or market
transactions to cover the hour-to-hour changes in net load, Idaho Power only needs to carry
enough balancing reserves to cover the hour-ahead forecast error. This is why Idaho Power used
hour-ahead forecasts to estimate balancing reserve needs in its 2007 Study.
When asked to provide all wind integration studies that Idaho Power relied upon in
making the decision to base the study of reserve requirements for wind on the day-ahead
scheduling/forecasting error, Idaho Power produced one paper from 2009.13 Specifically, Idaho
Power quotes from page 4 of that study, noting that,'oautility will make unit commitment
decisions well in advance of real-time, based on wind forecasts and load forecasts, and will make
additional decisions when correcting dispatch in real-time given the original commitment set."l4
Idaho Power's data response goes on to claim that, "consistent with the IEEE publication, Idaho
Power makes unit commitment decisions in the day-ahead market... [and]... based the reserve
requirement decisions for wind on the wind forecast day-ahead schedule error."l5
As stated above, we do not disagree that utilities must make unit commitment decisions a
day in advance of real-time, but the important aspect of system scheduling that Idaho Power
leaves out-and which is clearly identified in the same sentence they quote from-is that a
" Idaho Power Response to Data Request No. 5 (citing Erik Ela, et al.,TheEvolution of Wind Power
Integration Studies: Past, Present, and Future at l-8 (2009IEEE)), ovailable at
http://ieeexpiore.ieee.org/xpi/loqin-isp?tp-&arnunnber:5275981&url:hUp%3A%2F%2Fieeexp!or.trli-
ts Id.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
utility "will make additional decisions when correcting dispatch in real-time." It is this real-time,
hour-by-hour management of variable loads and wind that Idaho Power's curent study
completely ignores. While Idaho Power admits that it adjusts generation hour-by-hour for
load,l6 the 2013 Study completely ignores this routine aspect of utility operations.
Furthermore, the IEEE paper relied on by Idaho Power makes no mention of basing the
reserve requirement for wind on day-ahead schedule error. It says that the utility must make
"unit commitment" decisions based on day-ahead forecasts of load and wind, but does not say
that the commitment decision must be to hold a fixed amount of balancing reserves based on a
day-ahead forecast error. Doing so would be hugely inefficient and would ignore all of the lower
cost options available to the utility through the hourly dispatch of least-cost resources.
We agree that Idaho Power must set up the system in advance, a day-ahead, to be able to
provide sufficient balancing reserves to manage the hourly forecasting error from wind and load,
but that is not the same thing as needing to determine the hour-by-hour wind schedules a day in
advance of real time and hold reserves based on that iarger forecast error. Instead, the common
practice in utility wind integration studies (and tbr most Northwest utilities) is to determine the
maximum (90th percentile) amount of hour-ahead balancing reserves needed (based on a 45-, 30-
, l5-, or l0-minute-ahead persistence fcrrecast) and hold that amount every hour of the year. This
approach recognizes that the hourly dispatch of other resources and market purchases will play a
significant role in following the fluctuations in load and wind throughout the day. This approach
would reduce costs significantly and is consistent with what Idaho Power did in its 2007 Study.
Standard best practice in many regions of the country include (1) setting the wind schedule 10
minutes ahead of the operating interval basecl on current wind output, which drastically reduces
tu See, e.g.,2007 Studyat 17,21,45-46;2013 Studyar 18.
IPC-E-13-22 - July 2, 2014 Comments of AWEA and Renewable Northwest
10
the wind forecasting error; and (2) dynamically adjusting the quantity of reserves based on
information provided in the wind and load forecast and statistical analysis.
3. Netting the Variability and Scheduling Error of Generation and Load.
Another primary source of error in Idaho Poweros 2013 Study is the failure to account for
the netting between the forecasting/scheduling errors of load and wind, primarily, but also with
the deviations of other generators. The approach used in Idaho Power's 2013 Study is to first
separately calculate the amount of balancing reserves required for load, have the system
optimizer hold that amount of reserves, and calculate the production costs associated with this
portfolio. Next, Idaho Power separately calculates the amount of balancing reserves needed for
wind, constrains the system optimizer to carry this additional amount of reserves, and then again
calculates the production cost associated with the "wind scenario." Comparing the difference
between these two production cost runs provides the incremental cost associated with balancing
wind and directly feeds into Idaho Power's proposed wind integration rates.
The problem with this approach is that in reality, Idaho Power will not use the arithmetic
sum of the individually calculated balancing reserve needs for wind and load because (1)
balancing reserves for load and wind are rarely needed at the same time; and (2) scheduling
errors of the load and wind will often at least partially cancel each other out. This concept is
called "netting."
To explain the netting concept further, consider that Idaho Power is carrying balancing
reserves for load every day and every hour of the year. Idaho Power states that it carries
balancing reserves equal to three percent ofload, and that these reserves are used for in-hour
load following requirements, regulation requirements, and load forecast error.lT The reason that
wind and load scheduling/forecasting error should be netted is because sometimes the load
r Idaho Power Response to Data Request No. 15.
IPC-E-13-22-July 2,2014 Comments of AWEA and Renewable Northwest
I1
forecast will be in error in a positive direction (e.g. + l0 MW) and at the same time, the wind
forecast error will be in a negative direction (e.g.- 13 MW). In this example, the wind and load
error cancel each other out for all but 3 MW and the system operator only needs to dispatch 3
MW of balancing reserves. Three megawatts, in this simple example, would be the incremental
balancing reserve need attributable to wind. Idaho Power's 2013 Study proposes to charge wind
for all l3 MW of balancing reserves-a significant difference.
When asked if Idaho Power nets the scheduling/forecasting errors for wind with these
errors for other generation and load, the Company responded that it does not because "wind and
load are not correlated ... [thereby] validat[ing] Idaho Power's decision not to net the scheduling
effors for wind with the scheduling errors for other generation and load."l8
We agree that load and wind are not correlated but disagree completely that this
"validates" Idaho Power's decision not to net the scheduling error for wind and load. Actually, it
is precisely because load and wind are not correlated that Idaho Power should net the scheduling
errors of these two sources of variability on the system." O.rly if wind output were always
moving in the exact opposite direction of load, by the exact opposite magnitude, would wind's
variability not be partially canceled out. Indeed, Idaho Power's own 2007 Study recognized this
arithmetic fact: "[b]ecause of the similarity between load and wind with respect to real-time
r8 Idaho Power Response to Data Request No. 3.
re Statistically speaking, only if wind and load variations were perfectly negatively correlated with a
correlation coefficient of -1.0 would Idaho Power be correct that it is not necessary to net their offsetting
variabilities. Well-established statistical principles dictate that the combined variability of uncorrelated
sources ofvariability is equal to the square root ofthe sum ofthe squares ofthe individual sources'
variability. As an example, given a fictitious power system with wind variability of 30 MW per hour,
load variability of 100 MW per hour, and conventional generator variability of 40 MW per hour, the
method for accurately calculating total power system variability is as follows:
Sum of squares variabiliry = (302 + 1002 + q1z\:900+10,000+1,600: l2,5OO
Total Power System Variability : Square Root of (302 + 1002 + 402)= I I 1.80 MW
This method is essential to accurately capture the statistical fact that the combined variability of several
uncorrelated factors is less than the sum of their parts, hence why in the example above the combined
variability of 30 MW, 40 MW, and 100 MW is only 1l1.80 MW and not 170 MW.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
t2
operations, it is useful to couple their separate regulation components into a single total
regulating reserve level. It is understood that because of interaction between load and wind, a
straight arithmetic sum of the separate components results in reserve levels that are
inappropriately high."2o Instead, in2007,Idaho Power calculated the total reserve requirement
by using a root-sum-square methodology, which is much more appropriate.
Indeed, all wind integration studies that reflect best practices in the field calculate
reserves based on net load (load minus wind), as it is widely rurderstood that calculating reserves
for wind alone results in an incorrect answer. For example, the list of "Common Errors in
Integration Analyses" ina20l1 National Renewable Energy Laboratory ("NREL") report by the
leading wind integration experts explains why it is critical to analyze net load variability, not
wind-alone variability:
The concept of balancing the net load with conventional generation
is well-understood in the integration literature and power system
operations. In fact, the NERC Area Control Error (ACE), Control
Performance Standards (CP S I &2) standa"rds, Di sturbance Control
Standard (DCS), and balancing requirements are based upon it.
However, within the past year we have seen two integration
analyses that have attempted to balance wind and solar in isolation
from the remaining load. This means that when wind/solar and
load are both increasing, a conventional generator must decrease
output to hold the wind and solar constant, but at the same time,
generation must increase to meet the increasing load. This does
not reflect how power systems are operated and greatly overstates
the balancing costs of wind and solar.t2ll
Similarly, the IEEE paper, which Idaho Power cites as the authority for much of its 2013 Study
methodology, also explains why netting load and wind variability is so important:
20 2oo7 Study Addendum at 19.
2r Michael Milligan, et al., Cost-Causation and Integration Cost Analysis for Variable Generation at 27
(June 20 1 l\, availab le ar http://www.nrel. gov/docs/& I I osti/S I 860.pdf.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
t3
To realistically simulate power system operation, the uncertainties associated with load
forecast erors and wind forecast errors are important. Because wind and load forecast
errors are generally statistically independent, they do not add arithmetically, and should
be developed for the simulation in as realistic a manner as possible...This is because the
unit commitment process is done to target the combined loads and wind forecast, whereas
the forecast errors will not become apparent until the operating hour.22
The Federal Energy Regulatory Commission ("FERC") has also weighed in to explain
that a failure to account for net load results in an incorrect calculation oftotal reserve needs. For
example, in August 2009, FERC issued a letter finding that the utility Westar's proposal to
calculate wind reserve needs and integration costs in isolation without accounting for the
diversity benefits of different resources' variability would incorrectly result in an over-collection
of reserve costs.23
Westar's own data provided in that proceeding compellingly illustrates that deviations in
load and other generation cancel out much of the variability of wind energy. The stand-alone
variance of load (variance is proportional to the incremental reserve need caused by each source
of variability) was found to be 5 times larger than that of wind, and the variance of conventional
generation was found to be 2l times larger than that of wind.2a As a result, 695 MW of wind
" Erik Ela, et al.,The Evolution of Wind Power Integration Studies: Past, Present, and Future at} (2009
IEEE)), available at
A%zF%2Fieeexnlor.
Westar Energt, Inc.,Docket No. ER09-1273-000, Deficiency Letter from FERC re Proposed Balancing
Area Services Agreement and Schedule 3,{, Generator Regulation and Frequency Response Service (Aug.
3,2009), available ar http://elibrary.ferc.gov/idmwsicommon/opennat.asp?filelD:12099675. In the
letter, FERC states "your proposed regulation requirement for intermittent resources reflects the
regulation capacity that would be required to balance the output of three intermittent resources in
isolation. [Failing to] take into account the diversity among deviations of all system resources and load . .
. could result in regulation charges for resources and load that overstate and over-recover total system
regulation costs. For example . . . during a l0-minute interval, the output of certain intermittent
generation may be declining, while during the same interval, the output of other intermittent or
dispatchable generation may be increasing, or load may be decreasing, effectively cancelling out the
regulating requirements to some extent on a system-wide basis."
2a Westar Energt, Inc.,Docket No. ER09-Lz73-O}O,Informational Filing, OATT, Schedule 3,A' (Mar. 14,
2014), available ar hftp://elibrary.ferc.gov/idmws/common/opennat.asp?fileID=13484564. The box
under Step 2 in the attachment to the filing repofts the variances of wind, load, and conventional
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
14
energy only incrementally increased the total regulation reserve need for the Westar system by 3
MW.25
4. Grid Operating Reforms Would Substantially Reduce Integration Costs.
Wind integration costs are largely caused by obsolete grid operating practices. For
example, analysis conducted as part of the Puget Sound Energy wind integration case before
FERC confirmed that moving from hourly scheduling to 30-minute scheduling intervals with a
l0-minute forecast lead-time would have reduced the wind reserve need by more than 60%o.26 In
addition, regions with efficient grid operating practices see much smaller integration costs. Data
from ERCOT indicate that the cost of integrating more than 10,000 MW of wind generation is
less than $0.504{Wh of wind energy, thanks to its large grid operating area and practice of
dispatching generation at five-minute intervals and setting wind schedules based on persistence
forecasts from l0 minutes before real-time.27 As shown in the chart below, regions with fast
sub-hourly scheduling have much lower integration costs than regions with hourly scheduling.28
generation (RBASE Generation), and for the calculations above the variances of 'oinside wind" and
"outside wind" are summed.25.Id. Subtracting the wind variance from the calculation in the box labeled "Regulation Requirement"
and taking the square root of that sum without the wind variance included yields a total reserve need of
129.6 MW, versus 132.5 MW with the wind resources included.
26 Puget Sound Energt, Docket No. ERl l-3735-000, Protest of American Wind Energy Association and
Renewable Northwest Project at 13 (July 5,2011), available at
Mark Alhstrom, A Market Perspective on F'orecast Value at 5-8 (Feb. 26,2073), available at
rp://\ ^vw.uwlg.org/slcroreworK/Anlsrom-Desslon I .pol.
Ryan Wiser and Mark Bolinger, 2012 Wind Technologies Market Report at 64 (USDOE Aug. 2013),
available ar https://www Leere.energy.&ov/wind/pdfs/2O12-winci_technologies_market-report.pdf.
l5
IPC-E-13-22 - July 2,2014 Comments of AWEA and Reneu'able Northwest
EEg!El*.p;s
'E$stkdGtlEEiEEE:to-eXi*8#
c
4#o
fis5&
*0&
efi{*
2$q&
f$1&
fst&
$9&
' **ry{-*'**
g
|'*" "' -' ""a"'"'"" * -{''10ft 20qu Ns% {0!& foqe
Y{ir"id FEn*eton {CaBxcity OadB)
.$$Brft$trts sdr{Ed : <16+ntrt
r.Ap8 (2007i-+'Avidr €007!- BPA{2M}. BPAi2or1i* SAtAPS0OTI{*}r. ctuso {2{07}* CAISO(?ClDlt I: EHCOrn00Si-r ERC0Y{201:}.' El,YtTS (?010)+Hat6 po,mr {r0$?je ld*$FEdtrr{?Ut2}
r$s-i,rE (201d)*tr"M!,t{ilHi0 f?OOStz itsbf&$€ttoloitrt.t !la* York {100$l-irltor$$estilrn (r0! I lr !.ao{h$lsr$rn (20121-.-I,tHIlO t?0t01 .
r Odu{Slllh,clr peci$corp (AO7l-r:,-Faa{iCorp (20rO}*r*FacrltCe6 {20i?}""irf-Epp{m{ole *pp"$Hftc{r$1El' $lb Erlirde3 t?0S3)s w\rrsls (2010): Xcot-lrHOOC (?0041
-*-)Orl"P$Co t2006)t; Xsaf-lJlrd0B$O$l
Sre 1O-1{, ffi% S0}& jl$}& SC% oqt
rred Pa$eafi dfr {C*s€dty B8*ib}
$0!6
Moreover, the chart indicates that the wind reserve levels indicated in Idaho Power's
2013 Study (light blue line near the top of the left panel) are al extreme outlier, even among
other utilities with inefficient operating practices. Idaho Power is holding around 35 MW of
reserves for every 100 MW of wind capacity, while all other studies for utilities with inefficient
operating practices have found a reserve need of less than 25 MW for every 100 MW of wind
capacity atthat level of wind penetration. In fact, most studies (including Idaho Power's 2007
Study) are under 12 MW, or one-third of the amount calculated in Idaho Power's 2013 Study.
For utilities with efficient operating practices, all studies found a reserve need of less than 12
MW for every 100 MW of installed wind capacity at those wind penetrations. This highlights
the impact of Idaho Power's incorrect assumptions on its calculated reserve need and therefore
its calculated wind integration cost.
A number of studies have documented that reserve needs and costs are drastically
reduced when balancing area operations are coordinated and dispatch intervals are reduced, as
they would be under an Energy Imbalance Market ("EIM"). For example, the Northwest Power
Pool Market Assessment and Coordination Committee has studied this issue and found that
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
t6
participating in an EIM would reduce Idaho Power's reserve requirement by 29-65o/o,'n *d
would reduce the Company's operating costs by $2.418 million per year.3'
When asked if the Company has analyzedthe costs and benefits ofjoining the CAISO-
PacifiCorp EIM in zDls,Idaho Power responded that it has not, primarily because it "does not
have firm transmission rights to access the California market."3' Ho*"r.r, Idaho Power is well
connected with PacifiCorp's transmission system, and PacifiCorp is a leading participant in the
CAISO-PacifiCorp EIM. This transmission access to PacifiCorp's system would be sufficient to
join this EIM effort and reap the associated benefits.
Because integration costs are largely caused by outdated grid operating practices, it is
unreasonable to allocate these costs to wind generators. Instead, the entity responsible for setting
its grid operating procedures should be required to take steps to reduce its integration costs, since
it is best positioned to mitigate these costs by changing its operational practices. As a
transmission provider, Idaho Power is required to comply with FERC Order 764,which requires
transmission providers to implement grid operating reforms before being allowed to impose
integration charges on wind generators. Although Order 764 is not directly applicable to the
integration charges assessed against wind QFs in the PURPA context, the reforms that Idaho
Power is already required to implement under Order 764 should be applied to the Company's
integration of wind QFs to bring down those costs.
When asked about operational tools, Idaho Power explained in its data response that it
has "adopted a wind forecasting tool which has improved forecasting accuracy of wind
2e Northwest Power Pool Members'
at Attachment 3, Table 1, p.28 (Oct.
Market Assessment and Coordination Initiative, Final Phase I Report
21, 2013), ovailable ar hltplwty$a.n-lyptrorg/documents/MC-
3t Idaho Power Response to Data Request No. 7.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
generation on both the day-ahead and hourly forecast basis."32 We support the development of
this tool and understand that it has increased the efficiency of Idaho Power's ability to integrate
wind considerably. We are perplexed, however, as to why the capabilities of this tool do not
factor into Idaho Power's 2013 Study methodology-nor do they seem to factor into the wind
integration rates and charges proposed in the Company's Application. Wind integration rates
and charges should be going down to reflect the efficiency improvements resulting from the
forecasting tool and other operational tools available to the Company. Instead, they are moving
in the opposite direction.
5. Adjustments to the Company's Rate Calculations to Account for Study Flaws
and Changing Market Prices.
Due to the methodological elrors in Idaho Power's 2013 Study, it makes more sense to
use the 2007 Study as guidance in this docket, especially for determining the incremental reserve
requirement associated with future wind generation. Although the2007 Study is generally
superior to the more recent study, one useful component of the 2013 Study is the updated market
price data. The market price indices Idaho Power used in the 2013 Study average at $40.98, a
42Yo decrease from what was used in the 2007 Study. The decrease in market prices suggest that
the Idaho Power's wind integration costs should be lower today than they were in 2007 because
the opportunity cost in the market is the primary basis for how Idaho Power values the
incremental balancing reserves it allocates to integrating wind.
The 2007 Study Addendum also provides useful guidance and data points. A sensitivity
analysis in the Addendum estimates that for 600 MW of wind, and assuming average market
prices of $44.44-prices that are T.TVohigher than what Idaho Power is projecting today-the
wind integration cost would be $6.33AvIWh. Adjusting the $6.33 rate for the lower market prices
32 Id.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
l8
seen today yields arate of $5.84/ItzIWh ($6.33 - ($6.33 *.077)). This rate of $5.84/lt4Wh should
be considered the ceiling for Idaho Power's wind integration costs going forward. There are
several methodological improvements, operational tools, and market reforms that would further
decrease this cost of $5.84:
o Neither the 2013 Study nor the 2007 Study accounted for the scheduling error of
conventional generators, orthe expected interconnection ofsolar generation on
Idaho Power's system.
o The 2013 Study does not include the benefits of Idaho Power's new wind
forecasting tool.
o FERC Order 764 reqtires all transmission providers, including Idaho Power, to
offer l5-minute scheduling, and liquidity in this short-term market is increasing.
o Energy Imbalance Markets are being widely discussed in the West, and Idaho
Power may be able join an EIM as soon as October 1,2015.
Taking into account these other factors, we think a reasonable integration cost for Idaho
Power is $5.30/\uIWh, which is consistent with what other transmission providers in the region
have calculated. However, our calculations are based in part on limited data from Idaho Power,
and more analysis may be required to arrive at accurate wind integration rates. We recommend
that at a minimum, the Commission refrain from approving Idaho Power's proposed wind
integration rates and charges until the Company revises its 2013 Study to address the significant
methodological flaws andanalyze costs and benefits of the operational reforms we have
identified.
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
l9
IV. CONCLUSION
WHEREFORE, Renewable Northwest and AWEA respectfully request that the
Commission decline to approve the Company's proposed wind integration rates and charges that
would apply to any new QFs or integrated resource plans until such time as Idaho Power revises
its Wind Integration Study to address the flaws identified herein.
DATED this 2nd day of July,20l4
K&L Gates, LLP
,,
Teresa Hill
Attorney for AWEA and Renewable
Northwest
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
20
CERTIFICATE OF SERVICE
I hereby certiff that on the 2nd day of July, 2014, atrue and correct copy ofthe foregoing
COMMENTS OF RENEWABLE NORTHWEST AND AMERICAN WIND ENERGY
ASSOCIATION, Case No. IPC-E-13-22,was served by electronic mail to:
ID,A,HO POWER COMPANY:
COMIVtrSSION STAFF:
IDAHO WINDS LLC:
COLD SPRINGS WINDFARM, LLC;
DESERT MEADOW WINDFARM, LLC;
HAMMETT HILL WT\IDFARM, LLC;
Donovan E. Walker
Idaho Power Company
r22t W. Idaho St. (83702)
P.O. Box 70
Boise,ID 83707-0070Email: [e-mail]
dockets@ idahopower. com
Michael J. Youngblood
Greg Said
Idaho Power Company
l22lW. Idaho St. (83702)
P.O. Box 70
Boise,ID 83707-0070Email: [e-mail]
[e-mail]
Kris Sasser
Deputy Attorney General
Idaho Public Utilities Commission
472 W . Washington (837 02)
P.O. Box 83720
Boise, lD 83720-0074Email: [e-mail]
Dean J. Miller
McDevitt & Miller LLP
420W. Bannock Street
Boise,ID 83702Email: [e-mail]
Rick Koebbe, President
Idaho Winds, LLC
5420 W. Wicher Road
Glenns Ferry, lD 83623Email: [e-mail]
Bob Eggers, Legal Counsel
Idaho Winds, LLC
Email: [e-mail]
Peter J. Richardson
Richardson Adams, PLLC
515 N.27th Street
IPC-E-13-22 - July 2,2014 Comments of AWEA and Renewable Northwest
2l
MAINLII\IE WINDFARM, LLCi RYEGRASS
WhIDFARM,LLCI AI\ID TWO PONDS
WINDFARM,LLCZ
CASSIA WINDFARM LLC;
HOT SPRINGS WINDFARIVI LLC;
BEI\INETT CREEK WINDFARM LLC;
CASSIA GULCH WII\D PARK, LLC; TUANA
SPRINGS ENERGY, LLC; AITTD IIIGH MESA
EI\-ERGY, LLC:
Boise,ID 83702
peter@.richardsonadams. com
Benjamin G. Huang, Manager
c/o Mountain Air Projects
6000 N. Foxtail Way
Glenns F..ry, ID 83623
Gregory M. Adams
Richardson Adams, PLLC
515 N.27th Street
Boise,lD 83702
gre [e-mail]
Paul Ackerman
Assistant General Counsel
Exelon Business Services Corporation
100 Constellation Way
Baltimore, MD21202
DATED this 2th day of July,2014.
IPC-E-13-22 - July 2, 2014 Comments of AWEA and Renewable Northwest
22