20121011Kalich DI.pdf

DAVID J. MEYER
VICE PRESIDENT AND CHIEF COUNSEL FOR
REGULATORY & GOVERNMENTAL AFFAIRS
AVISTA CORPORATION
P.O. BOX 3727
1411 EAST MISSION AVENUE
SPOKANE, WASHINGTON 99220-3727
TELEPHONE: [phone]
FACSIMILE: [phone]
[e-mail]

BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION

IN THE MATTER OF THE APPLICATION ) CASE NO. AVU-E-12-08
OF AVISTA CORPORATION FOR THE )
AUTHORITY TO INCREASE ITS RATES )
AND CHARGES FOR ELECTRIC AND )
NATURAL GAS SERVICE TO ELECTRIC ) DIRECT TESTIMONY
AND NATURAL GAS CUSTOMERS IN THE ) OF
STATE OF IDAHO ) CLINT G. KALICH
)

FOR AVISTA CORPORATION

(ELECTRIC ONLY)

Kalich, Di 1
Avista Corporation
I. INTRODUCTION 1
Q. Please state your name, the name of your 2
employer, and your business address. 3
A. My name is Clint Kalich. I am employed by Avista 4
Corporation at 1411 East Mission Avenue, Spokane, 5
Washington. 6
Q. In what capacity are you employed? 7
A. I am the Manager of Resource Planning & Power 8
Supply Analyses in the Energy Resources Department of 9
Avista Utilities. 10
Q. Please state your educational background and 11
professional experience. 12
A. I graduated from Central Washington University in 13
1991 with a Bachelor of Science Degree in Business 14
Economics. Shortly after graduation, I accepted an analyst 15
position with Economic and Engineering Services, Inc. (now 16
EES Consulting, Inc.), a Northwest management-consulting 17
firm located in Bellevue, Washington. While employed by 18
EES, I worked primarily for municipalities, public utility 19
districts, and cooperatives in the area of electric utility 20
management. My specific areas of focus were economic 21
analyses of new resource development, rate case proceedings 22
involving the Bonneville Power Administration, integrated 23

Kalich, Di 2
Avista Corporation
(least-cost) resource planning, and demand-side management 1
program development. 2
In late 1995, I left Economic and Engineering 3
Services, Inc. to join Tacoma Power in Tacoma, Washington. 4
I provided key analytical and policy support in the areas 5
of resource development, procurement, and optimization, 6
hydroelectric operations and re-licensing, unbundled power 7
supply rate-making, contract negotiations, and system 8
operations. I helped develop, and ultimately managed, 9
Tacoma Power’s industrial market access program serving 10
one-quarter of the company’s retail load. 11
In mid-2000 I joined Avista Utilities and accepted my 12
current position assisting the Company in resource 13
analysis, dispatch modeling, resource procurement, 14
integrated resource planning, and rate case proceedings. 15
Much of my career has involved resource dispatch modeling 16
of the nature described in this testimony. 17
Q. What is the scope of your testimony in this 18
proceeding? 19
A. My testimony will describe the Company’s use of 20
the AURORAXMP dispatch model, or “Dispatch Model”. I will 21
explain the key assumptions driving the Dispatch Model’s 22
market forecast of electricity prices. The discussion 23

Kalich, Di 3
Avista Corporation
includes the variables of natural gas, Western Interconnect 1
loads and resources, and hydroelectric conditions. I will 2
describe how the model dispatches its resources and 3
contracts to maximize customer benefit and tracks their 4
values for use in pro forma calculations. Finally, I will 5
present the modeling results provided to Company witness 6
Mr. Johnson for his power supply pro forma adjustment 7
calculations. 8
Q. Are you sponsoring any exhibits in this 9
proceeding? 10
A. Yes. I am sponsoring one exhibit marked 11
Confidential Exhibit 5, Schedule 1. It provides summary 12
output from the Dispatch Model and data that are used by 13
Company witness Mr. Johnson as input for his work. All 14
information contained in the schedule was prepared under my 15
direction. 16
17
II. THE DISPATCH MODEL 18
Q. What model is the Company using to dispatch its 19
portfolio of resources and obligations? 20
A. The Company uses EPIS, Inc.’s AURORAXMP market 21
forecasting model (“Dispatch Model”) and its associated 22

Kalich, Di 4
Avista Corporation
database for determining power supply costs.1 The Dispatch 1
Model optimizes Company-owned resource and contract 2
dispatch during each hour of the January 1, 2013 through 3
December 31, 2013 pro forma year. 4
Q. Please briefly describe the Dispatch Model. 5
A. The Dispatch Model was developed by EPIS, Inc. of 6
Sandpoint, Idaho. It is a fundamentals-based tool 7
containing demand and resource data for the entire Western 8
Interconnect. It employs multi-area, transmission-9
constrained dispatch logic to simulate real market 10
conditions. Its true economic dispatch captures the 11
dynamics and economics of electricity markets—both short-12
term (hourly, daily, monthly) and long-term. On an hourly 13
basis the Dispatch Model develops an available resource 14
stack, sorting resources from lowest to highest cost. It 15
then compares this resource stack with load obligations in 16
the same hour to arrive at the least-cost market-clearing 17
price for the hour. Once resources are dispatched and 18
market prices are determined, the Dispatch Model singles 19
out Avista resources and loads and values them against the 20
marketplace. 21

1 The Company is using AURORAXMP version 11.1.1001.

Kalich, Di 5
Avista Corporation
Q. What experience does the Company have using 1
AURORAXMP? 2
A. The Company purchased a license to use the 3
Dispatch Model in April 2002. AURORAXMP has been used for 4
numerous studies, including each of its integrated resource 5
plans and rate filings after 2001. The tool is also used 6
for various resource evaluations, market forecasting, and 7
requests-for-proposal evaluations. 8
Q. Who else uses AURORAXMP? 9
A. AURORAXMP is used across North America and in 10
Europe. In the Northwest specifically, AURORAXMP is used by 11
the Bonneville Power Administration, the Northwest Power 12
and Conservation Council, Puget Sound Energy, Idaho Power, 13
Portland General Electric, Seattle City Light, Grant County 14
PUD, Snohomish County PUD, and Tacoma Power. 15
Q. What benefits does the Dispatch Model offer for 16
this type of analysis? 17
A. The Dispatch Model generates hourly electricity 18
prices across the Western Interconnect, accounting for its 19
specific mix of resources and loads. The Dispatch Model 20
reflects the impact of regions outside the Northwest on 21
Northwest market prices, limited by known transfer 22
(transmission) capabilities. Ultimately, the Dispatch 23

Kalich, Di 6
Avista Corporation
Model allows the Company to generate price forecasts in-1
house instead of relying on exogenous forecasts. 2
The Company owns a number of resources, including 3
hydroelectric plants and natural gas-fired peaking units, 4
which serve customer loads during more valuable on-peak 5
hours. By optimizing resource operation on an hourly 6
basis, the Dispatch Model is able to appropriately value 7
the capabilities of these assets. Forward prices for the 8
proforma 2013 period are 38% higher in the on-peak hours 9
than off-peak hours at the time this case was prepared. 10
The Dispatch Model forecasts on-peak prices for the pro 11
forma period to average 39% higher than off-peak prices. A 12
graphical representation of the differences in on- and off-13
peak prices over the proforma period is shown below in 14
Chart No. 1. 15

Kalich, Di 7
Avista Corporation
Chart No. 1 – Monthly AURORA modeled versus forward Mid-C 1
Prices 2
3
Forward prices month to month are tracked very closely 4
in the Dispatch Model, given that the AURORA model is using 5
normalized hydro, load, and resource outages. In summary, 6
the Dispatch Model appropriately values the energy from 7
Avista’s resources during on-peak periods in a manner 8
similar to that recently experienced in the Northwest 9
region for the 2013 proforma period. 10
Q. On a broader scale, what calculations are being 11
performed by the Dispatch Model? 12
A. The Dispatch Model’s goal is to minimize overall 13
system operating costs across the Western Interconnect, 14
including Avista’s portfolio of loads and resources. The 15
-
5
10
15
20
25
30
35
40
45
50
20
1
3
_
0
1
20
1
3
_
0
2
20
1
3
_
0
3
20
1
3
_
0
4
20
1
3
_
0
5
20
1
3
_
0
6
20
1
3
_
0
7
20
1
3
_
0
8
20
1
3
_
0
9
20
1
3
_
1
0
20
1
3
_
1
1
20
1
3
_
1
2
$/
M
W
h
Off-Peak
On-Peak
Forward Off Peak
Forward On Peak

Kalich, Di 8
Avista Corporation
Dispatch Model generates a wholesale electric market price 1
forecast by evaluating all Western Interconnect resources 2
simultaneously in a least-cost equation to meet regional 3
loads. As the Dispatch Model progresses from hour to hour, 4
it “operates” those least-cost resources necessary to meet 5
load. With respect to the Company’s portfolio, the 6
Dispatch Model tracks the hourly output and fuel costs 7
associated with portfolio generation. It also calculates 8
hourly energy quantities and values for the Company’s 9
contractual rights and obligations. In every hour the 10
Company’s loads and obligations are compared to available 11
resources to determine a net position. This net position 12
is balanced using the simulated wholesale electricity 13
market. The cost of energy purchased from or sold into the 14
market is determined based on the electric market-clearing 15
price for the specified hour and the amount of energy 16
necessary to balance loads and resources. 17
Q. How does the Dispatch Model determine electricity 18
market prices, and how are the prices used to calculate 19
market purchases and sales? 20
A. The Dispatch Model calculates electricity prices 21
for the entire Western Interconnect, separated into various 22
geographical areas such as the Northwest and Northern and 23

Kalich, Di 9
Avista Corporation
Southern California. The load in each area is compared to 1
available resources, including resources available from 2
other areas that are linked by transmission connections, to 3
determine the electricity price in each hour. Ultimately, 4
the market price for an hour is set based on the last 5
resource in the stack to be dispatched. This resource is 6
referred to as the “marginal resource”. Given the 7
prominence of natural gas-fired resources on the margin, 8
this fuel is a key variable in the determination of 9
wholesale electricity prices. 10
Q. How does the Dispatch Model operate regional 11
hydroelectric projects? 12
A. The model begins by “peak shaving” loads using 13
system hydro resources. When peak shaving, the Dispatch 14
Model determines which hours contain the highest loads and 15
allocates to them as much hydroelectric energy as possible. 16
Remaining loads are then met with other available 17
resources. 18
Q. Has the Company made any modifications to the 19
EPIS database for this case? 20
A. Yes. The EPIS database was modified to include 21
various assumptions used in the Company’s 2011 Integrated 22
Resource Plan. It uses a similar database as the Company’s 23

Kalich, Di 10
Avista Corporation
last rate case filing, with updates to prices, resources, 1
and loads. For example, Avista’s resource portfolio is 2
modified to reflect actual project operating 3
characteristics. Natural gas prices are modified to match 4
forward prices over the pro-forma period, regional 5
resources and loads are modified where better information 6
is made available, and northwest hydro data are replaced 7
with Bonneville Power Administration data. 8
9
III. HYDRO MODELING ASSUMPTIONS 10
Q. Please provide additional detail on how the 11
Company has modeled hydroelectric generation for this case. 12
A. Avista is modeling the Clark Fork, the Mid-13
Columbia (Mid-C) projects, and the upper four Spokane River 14
projects (Post Falls, Upper Falls, Monroe Street, and Nine 15
Mile) identically as we did in the last rate case. For the 16
two lower (Long Lake and Little Falls) Spokane River 17
projects, the Company is now using its Avista Hydro 18
Optimization model; the same model was adopted for use in 19
the last case for the Clark Fork and Upper Spokane River 20
systems. 21
Avista uses historical stream flow data from the 22
Bonneville Power Administration (“BPA”) for the Clark Fork 23

Kalich, Di 11
Avista Corporation
and the Spokane River projects. For the Mid-C, where the 1
Company does not have adequate data to model them, NWPP 2
generation values are used just as in previous rate cases. 3
As in previous cases, the NWPP data are modified slightly 4
to address the NWPP model’s tendency to overstate 5
generation in high-flow periods, to maintain year-to-year 6
consistency in project operations, and to account for 7
encroachment on our Mid-C project shares. 8
Q. What hydroelectric record is being used in this 9
case? 10
A. 1929-1998. 11
Q. How is the generation then used for ratemaking 12
purposes? 13
A. The monthly generation levels for each project 14
(Mid-C, Spokane River, and Clark Fork) are input into the 15
dispatch model (AURORAXMP) where Avista’s portfolio value is 16
quantified for ratemaking purposes. 17
Q. Please describe the Avista Hydro Optimization 18
Package. 19
A. The Avista Hydro Optimization Package is a mixed-20
integer linear programming-based system emulating the 21
operation of the Company’s projects. It was developed in 22
support of system operations, financial forecasting, and 23

Kalich, Di 12
Avista Corporation
hydro upgrade efforts. Operating on an hourly time-step, 1
they accurately represent individual turbine and reservoir 2
operations. License constraints (e.g., minimum flows, 3
elevation limits) are honored in all periods. The 4
optimization package is comprised of four components. 5
Q. What is the first component of the Avista Hydro 6
Optimization Package? 7
A. The first component is the Avista Hydro Water 8
Budget Model. This model looks over the longer record of 9
water flow optimization to ensure storage water is released 10
during the most valuable times of the year. Outputs are 11
weekly beginning and ending project elevations for each 12
storage project. These elevations are exported to the 13
system optimization model Output table. 14
Q. What is the source for hydroelectric flows in the 15
Avista Hydro Water Budget Model? 16
A. The source data for the Water Budget Model is BPA 17
daily flow data derived from the U.S. Army Corp of 18
Engineers monthly flow study. This work re-creates 19
historical flows on Avista hydro projects back to 1929 20
based on today’s river system.2 This data is housed in the 21

2 Accounting for additional irrigation depletion, new in-river
developments, and present regulation requirements due to environmental
requirements.

Kalich, Di 13
Avista Corporation
Avista Hydro Optimization Model Input Database, and is the 1
second element of the Avista Hydro Optimization package. 2
Q. What is the third element of the Avista Hydro 3
Optimization Package? 4
A. The third element is the Avista Hydro 5
Optimization Model itself. This hourly model uses a mixed-6
integer optimization routine to maximize the value of the 7
hydroelectric projects over time. Each project is 8
represented in detail, including individual turbine 9
efficiency curves, physical and license-constrained 10
reservoir elevations, tailrace elevations, and minimum and 11
maximum flow constraints. 12
The Avista Hydro Optimization Model shapes generation 13
into the most beneficial (i.e., most economic) time periods 14
using the projects’ storage reservoirs. It also maximizes 15
the value of the generation by flowing water through the 16
turbines at their most economically efficient points on the 17
power curves. 18
Q. What is the fourth element of the Avista Hydro 19
Optimization Package? 20
A. The fourth element is the Avista Hydro 21
Optimization Model Output Database. This database contains 22
the results from the Avista Hydro Optimization Model, 23

Kalich, Di 14
Avista Corporation
including hourly turbine discharge and spill flows, hourly 1
generation levels, and hourly reservoir elevations. 2
Q. How did the Company ensure that the Avista Hydro 3
Optimization Package accurately reflects the operations and 4
value of Company-owned projects? 5
A. The Avista Hydro Optimization Package is 6
benchmarked against the Company’s 2000-2009 actual results 7
at the projects to ensure its accuracy. 8
Q. How did the initial results compare, and how was 9
the package adjusted to match with the 10-year record? 10
A. The Avista Hydro Optimization Package initially 11
over-estimated generation relative to the 2000-2009 periods 12
by approximately 5.5 percent for the Noxon project. It 13
understated generation by 0.6 percent for the Cabinet Gorge 14
project. For the four upper Spokane River projects, 15
generation was overstated by between 5% and 18%. These 16
results were expected, as Avista does not operate its 17
projects in isolation. Instead the Company uses its hydro 18
projects to meet all of its needs, including operating 19
reserves. There are also times where units are out on 20
maintenance or forced outage. To synch the Avista Hydro 21
Optimization Package to history the power curves for each 22
project were therefore adjusted by the differences 23

Kalich, Di 15
Avista Corporation
described above. After the benchmarking process, the model 1
generated levels equal to actual generation during the 2
2000-2009 period. The adjustments are presented below in 3
Table No. 1. 4
5
Table No. 1 – Avista Hydro Optimization Benchmarking 6
Adjustments 7
8
Q. Are the hydro models included in the Company’s 9
filing? 10
A. Yes. All four components of the Avista Hydro 11
Optimization Package for each major Company hydro system 12
(Spokane River and Clark Fork River) are included in my 13
work papers, including all input and output data. 14
Q. Does the Avista Hydro Optimization Package 15
account for recent upgrades at the Noxon Rapids project? 16
A. Yes. Once the original model was benchmarked 17
against recent generation years that did not benefit from 18
upgrades at Noxon, the newly upgraded units (1, 2, 3, and 19
Projects
Model
Overestimating
Percentage (%)
Model
Underestimating
Percentage (%)
Applied
Benchmark
Adjustment
Percentage (%)
Noxon Rapid 5.5 105.5
Cabinet Gorge 0.6 99.4
Post Falls 16.8 116.8
Upper Falls 12.2 112.2
Monroe Street 4.7 104.7
Nine Mile 18.3 118.3
Long Lake 7.8 107.8
Little Falls 2.5 102.5

Kalich, Di 16
Avista Corporation
4) were input into the model to reflect the higher 1
anticipated generation levels. 2
Q. How is the Company using the new Avista Hydro 3
Optimization Package in its business operations? 4
A. The Avista Hydro Optimization Package is an 5
essential tool to assist the Company with optimizing its 6
system operations, both in short- and long-term planning. 7
Its results are also used for Company budgeting and hydro 8
project market valuation studies. It has been used to 9
support various upgrade option studies. Given its speed it 10
is possible to run large hydro-flow records through it, as 11
is necessary for rate filings such as the one before you 12
today. It was used by the Company in its last rate case 13
before the Commission. 14
Q. How does the AURORAXMP Dispatch Model Operate 15
Company-controlled hydroelectric generation resources? 16
A. The Dispatch Model treats all hydroelectric 17
generation plants within each river system as a single 18
large plant. To account for the actual flexibility of 19
Company hydroelectric resources, the Company develops 20
individual hydro operations logic for each of its 21
facilities. This separation ensures that the flexibility 22

Kalich, Di 17
Avista Corporation
inherent in these resources is credited to customers in the 1
pro forma exercise. 2
Q. Please compare the operating statistics from the 3
Dispatch Model to recent historical hydroelectric plant 4
operations. 5
A. Over the pro forma period the Dispatch Model 6
generates 69% of Clark Fork hydro generation during on-peak 7
hours (based on average water). Since on-peak hours 8
represent only 57% of the year, this demonstrates a 9
substantial shift of hydro resources to the more expensive 10
on-peak hours. This is identical to the five-year average 11
of on-peak hydroelectric generation at the Clark Fork 12
through 2011. Similar relative performance is achieved for 13
the Spokane and Mid-Columbia projects 14
15
IV. OTHER KEY MODELING ASSUMPTIONS 16
Q. Please describe your update to pro forma period 17
natural gas prices. 18
A. Natural gas prices for this filing are based on a 19
1-month average from June 28, 2012 to July 27, 2012 of 20
calendar-year 2013 monthly forward prices. Natural gas 21
prices used in the Dispatch Model are presented below in 22
Table No 2. 23

Kalich, Di 18
Avista Corporation
Table No. 2 – Pro Forma Natural Gas Prices 1
2
Q. What is the Company’s assumption for rate period 3
loads? 4
A. Pro forma loads used in this case are weather-5
adjusted loads between July 1, 2011 and June 30, 2012. 6
Table No. 3 below details actual, weather-adjusted load. 7
Table No. 3 – Pro Forma Loads 8

Q. Please discuss your outage assumptions for the 10
Colstrip units. 11
A. As with our assumptions for other plants, we use 12
a 5-year average through 2011 to estimate long-run 13
performance at the Colstrip plant. The 9.6% forced outage 14
rate is based on this average and is slightly higher than 15
the 8.7% level used in the rate case last year. 16
Basis Price ($/dth)Basin Price ($/dth)
AECO 3.18 Stanfield 3.45
Malin 3.53 Sumas 3.61
Spokane 3.63 Henry Hub 3.61
Rockies 3.45 S. Calif.3.70
Month Actual
Weather
Adjusted Month Actual
Weather
Adjusted
Jan-12 1,237.9 1,246.2 Jul-11 1,019.1 1,056.9
Feb-12 1,187.9 1,190.8 Aug-11 1,097.4 1,081.5
Mar-12 1,097.3 1,088.2 Sep-11 1,005.9 970.4
Apr-12 1,006.4 1,018.0 Nov-11 1,000.5 1,006.4
May-12 963.6 960.0 Nov-11 1,156.9 1,155.9
Jun-12 957.2 965.1 Dec-11 1,251.0 1,265.5
Average 1,081.7 1,083.5

Kalich, Di 19
Avista Corporation
V. RESULTS 1
Q. Please summarize the results from the Dispatch. 2
A. The Dispatch Model tracks the Company’s portfolio 3
during each hour of the pro forma study. Fuel costs and 4
generation for each resource are summarized by month. 5
Total market sales and purchases, and their revenues and 6
costs, are also determined and summarized by month. These 7
values are contained in Confidential Exhibit 5 Schedule 1 8
and were provided to Mr. Johnson for use in his 9
calculations. Mr. Johnson adds resource and contract 10
revenues and expenses not accounted for in the Dispatch 11
Model (e.g., fixed costs) to determine net power supply 12
expense. 13
Q. Does this conclude your pre-filed direct 14
testimony? 15
A. Yes, it does. 16

DAVID J. MEYER
VICE PRESIDENT AND CHIEF COUNSEL FOR
REGULATORY & GOVERNMENTAL AFFAIRS
AVISTA CORPORATION
P.O. BOX 3727
1411 EAST MISSION AVENUE
SPOKANE, WASHINGTON 99220-3727
TELEPHONE: [phone]
FACSIMILE: [phone]
[e-mail]

BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION

IN THE MATTER OF THE APPLICATION ) CASE NO. AVU-E-12-08
OF AVISTA CORPORATION FOR THE )
AUTHORITY TO INCREASE ITS RATES )
AND CHARGES FOR ELECTRIC AND )
NATURAL GAS SERVICE TO ELECTRIC ) EXHIBIT NO. 5
AND NATURAL GAS CUSTOMERS IN THE )
STATE OF IDAHO ) CLINT G. KALICH
)

FOR AVISTA CORPORATION

(ELECTRIC ONLY)

Exhibit No. 5
Case No. AVU-E-12-08
C. Kalich, Avista
Schedule 1, p. 1 of 3

CONFIDENTIAL

Dispatch Model Summary Output

Pages 1 through 3

THESE PAGES ALLEGEDLY CONTAIN TRADE SECRETS OR
CONFIDENTIAL MATERIALS AND ARE SEPARATELY FILED.