20150601Kalich Direct.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-15-05
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)

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 4
Avista 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 13
in 1991 with a Bachelor of Science Degree in Business 14
Economics. Shortly after graduation, I accepted an 15
analyst position with Economic and Engineering Services, 16
Inc. (now EES Consulting, Inc.), a Northwest management-17
consulting firm located in Bellevue, Washington. While 18
employed by EES, I worked primarily for municipalities, 19
public utility districts, and cooperatives in the area of 20
electric utility management. My specific areas of focus 21
were economic analyses of new resource development, rate 22
case proceedings involving the Bonneville Power 23
Kalich, Di Page 1 Avista Corporation

Administration, integrated (least-cost) resource planning, 1
and demand-side management 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
includes the variables of natural gas, Western 24
Kalich, Di Page 2 Avista Corporation

Interconnect loads and resources, and hydroelectric 1
conditions. I will describe how the model dispatches its 2
resources and contracts to maximize customer benefit and 3
tracks their values for use in pro forma calculations. 4
Finally, I will present the modeling results provided to 5
Company witness Mr. Johnson for his power supply pro forma 6
adjustment calculations. 7
Q. Are you sponsoring any exhibits in this 8
proceeding? 9
A. Yes. I am sponsoring one exhibit marked 10
Confidential Exhibit 5, Schedule 1. It provides summary 11
output from the Dispatch Model and data that are used by 12
Mr. Johnson as input for his work. All information 13
contained in the exhibit was prepared under my direction. 14
15
II. THE DISPATCH MODEL 16
Q. What model is the Company using to dispatch its 17
portfolio of resources and obligations? 18
A. The Company uses EPIS, Inc.’s AURORAXMP market 19
forecasting model (“Dispatch Model”) and its associated 20
database for determining power supply costs.1 The Dispatch 21
Model optimizes Company-owned resource and contract 22
1 The Company uses AURORAXMP version 11.5.1083 with a Windows 7
operating system.
Kalich, Di Page 3 Avista Corporation

dispatch during each hour of the January 1, 2016 through 1
December 31, 2017 pro forma periods. 2
Q. Please briefly describe the Dispatch Model. 3
A. The Dispatch Model was developed by EPIS, Inc. 4
of Sandpoint, Idaho. It is a fundamentals-based tool 5
containing demand and resource data for the entire Western 6
Interconnect. It employs multi-area, transmission-7
constrained dispatch logic to simulate real market 8
conditions. Its true economic dispatch captures the 9
dynamics and economics of electricity markets both short-10
term (hourly, daily, monthly) and long-term. On an hourly 11
basis the Dispatch Model develops an available resource 12
stack, sorting resources from lowest to highest cost. It 13
then compares this resource stack with load obligations in 14
the same hour to arrive at the least-cost market-clearing 15
price for the hour. Once resources are dispatched and 16
market prices are determined, the Dispatch Model singles 17
out Avista resources and loads and values them against the 18
marketplace. 19
Q. What experience does the Company have using 20
AURORAXMP? 21
A. The Company purchased a license to use the 22
Dispatch Model in April 2002. AURORAXMP has been used for 23
numerous studies, including each of its integrated 24
Kalich, Di Page 4 Avista Corporation

resource plans and rate filings after 2001. The tool is 1
also used for various resource evaluations, market 2
forecasting, and requests-for-proposal evaluations. 3
Q. Who else uses AURORAXMP? 4
A. AURORAXMP is used all across North America, 5
Europe, and the Middle East. In the Northwest 6
specifically, AURORAXMP is used by the Bonneville Power 7
Administration, the Northwest Power and Conservation 8
Council, Puget Sound Energy, Idaho Power, Portland General 9
Electric, PacifiCorp, Seattle City Light, Grant County 10
PUD, and Snohomish County PUD. 11
Q. What benefits does the Dispatch Model offer for 12
this type of analysis? 13
A. The Dispatch Model generates hourly electricity 14
prices across the Western Interconnect, accounting for its 15
specific mix of resources and loads. The Dispatch Model 16
reflects the impact of regions outside the Northwest on 17
Northwest market prices, limited by known transfer 18
(transmission) capabilities. Ultimately, the Dispatch 19
Model allows the Company to generate price forecasts in-20
house instead of relying on exogenous forecasts. 21
The Company owns a number of resources, including 22
hydroelectric plants and natural gas-fired peaking units 23
serving customer loads during more valuable on-peak hours. 24
Kalich, Di Page 5 Avista Corporation

By optimizing resource operation on an hourly basis, the 1
Dispatch Model is able to appropriately value the 2
capabilities of these assets. Forward prices for the 3
proforma 2016 were 30% higher in the on-peak hours than 4
off-peak hours at the time this case was prepared; 2017 5
on-peak prices were 31% higher than off-peak prices. The 6
Dispatch Model forecasts on-peak prices for the pro forma 7
period to average 33% higher than off-peak prices in 2016 8
and 31% higher in 2017. Both are close to the forward 9
prices. A graphical representation of the differences in 10
on- and off-peak prices over the proforma periods is shown 11
below in Illustration Nos. 1 and No. 2. 12
Kalich, Di Page 6 Avista Corporation

Illustration No. 1 – 2016 Monthly AURORA modeled versus 1
forward Mid-C prices 2
3
4
5
6
7
8
9
10
11
12
Illustration No. 2 – 2017 Monthly AURORA modeled versus 13
forward Mid-C prices 14
15
16
17
18
19
20
21
22
23
-
5
10
15
20
25
30
35
40
1 2 3 4 5 6 7 8 9 10 11 12
2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016 2016
$/
M
W
h
Aurora Off-Peak
Off Peak
Aurora On-Peak
On Peak
-
5
10
15
20
25
30
35
40
45
1 2 3 4 5 6 7 8 9 10 11 12
2017 2017 2017 2017 2017 2017 2017 2017 2017 2017 2017 2017
$/
M
W
h
Aurora Off-Peak
Off Peak
Aurora On-Peak
On Peak
Kalich, Di Page 7 Avista Corporation

The forward Mid-Columbia prices in the graphs are the 1
latest one month average (Feb 20, 2015 through Mar 19, 2
2015) of Intercontinental Exchange (ICE) quarterly prices 3
at the time the study was prepared. 4
Dispatch Model and forward prices can and sometimes 5
will differ, as forward prices are based on market 6
expectations whereas the data used in the Dispatch Model 7
are normalized for hydro, loads, and resource outages. 8
Where the market expects a low hydro year forthcoming, 9
forward market prices could be higher than Dispatch Model 10
prices. Referring back to Illustration No. 1, the average 11
price for the 2016 forward period is $25.21 per MWh; the 12
Dispatch model result is $25.61 per MWh. Referring back 13
to Illustration No. 2, the average forward price for 2017 14
is $27.92 per MWh; the Dispatch model result is $28.36 per 15
MWh. These results explain that the market is not 16
forecasting a bias in future conditions (e.g., a low hydro 17
year). The Dispatch Model therefore provides a very close 18
approximation to what the actual market is predicting, and 19
provides a good data set for the pro forma. 20
Q. On a broader scale, what calculations are being 21
performed by the Dispatch Model? 22
A. The Dispatch Model’s goal is to minimize overall 23
system operating costs across the Western Interconnect, 24
Kalich, Di Page 8 Avista Corporation

including Avista’s portfolio of loads and resources. The 1
Dispatch Model generates a wholesale electricity market 2
price forecast by evaluating all Western Interconnect 3
resources simultaneously in a least-cost equation to meet 4
regional loads. As the Dispatch Model progresses from 5
hour to hour, it “operates” those least-cost resources 6
necessary to meet load. With respect to the Company’s 7
portfolio, the Dispatch Model tracks the hourly output and 8
fuel costs associated with Avista’s portfolio generation. 9
It also calculates hourly energy quantities and values for 10
the Company’s contractual rights and obligations. In 11
every hour, the Company’s loads and obligations are 12
compared to available resources to determine a net 13
position. This net position is balanced using the 14
simulated wholesale electricity market. The cost of 15
energy purchased from or sold into the market is 16
determined based on the electric market-clearing price for 17
the specified hour and the amount of energy necessary to 18
balance loads and resources. 19
Q. How does the Dispatch Model determine 20
electricity market prices, and how are the prices used to 21
calculate market purchases and sales? 22
A. The Dispatch Model calculates electricity prices 23
for the entire Western Interconnect, separated into 24
Kalich, Di Page 9 Avista Corporation

various geographical areas such as the Northwest and 1
Northern and Southern California. The load in each area 2
is compared to available resources, including resources 3
available from other areas that are linked by transmission 4
connections, to determine the electricity price in each 5
hour. Ultimately, the market price for an hour is set 6
based on the last resource in the stack to be dispatched. 7
This resource is referred to as the “marginal resource.” 8
Given the prominence of natural gas-fired resources on the 9
margin, this fuel is a key variable in the determination 10
of wholesale electricity prices. 11
Q. How does the Dispatch Model operate regional 12
hydroelectric projects? 13
A. The model begins by “peak shaving” loads using 14
hydro resources with storage. When peak shaving, the 15
Dispatch Model determines the hours with the highest loads 16
and allocates to them as much hydroelectric energy within 17
the constraints of the hydro system. Remaining loads are 18
then met with other available resources. 19
Q. Has the Company made any modifications to the 20
EPIS database for this case? 21
A. Yes. As we have in the past, Avista’s resource 22
portfolio is modified from EPIS’s default database to 23
reflect actual project operating characteristics. Natural 24
Kalich, Di Page 10 Avista Corporation

gas prices are modified to match the latest one month 1
average of forward prices over the pro-forma period. 2
Regional resources and loads are modified where better 3
information is available. And Northwest hydro data are 4
replaced with data from the Bonneville Power 5
Administration. The EPIS database is further modified to 6
include various assumptions used in the Company’s 2015 7
Integrated Resource Planning process and other new 8
resource information where available. 9
Q. Has the Company made any changes to the way it 10
models hydro in this case? 11
A. Methodologically, no. We did update the hydro 12
record to include ten additional years of hydrology that 13
have become available since our last general rate case. 14
We now model 80 years, from 1929 through 2008. Further, 15
BPA data now is being used for the Mid-C projects. This 16
change provides a consistent data set across all Avista 17
hydroelectric projects. 18
Q. How does BPA data for the Mid-C projects compare 19
to the NWPP data? 20
A. The BPA 80-year record provides 2.5% more 21
generation in the pro forma period than the NWPP 70-year 22
record historically used by the Company. This difference 23
Kalich, Di Page 11 Avista Corporation

decreases revenue requirement relative to continued use of 1
NWPP data. 2
Q. Why did Avista modify its analysis to use an 80-3
year record? 4
A. Consistent with precedent, Avista uses the full 5
hydro record for its rate filings. 6
Q. How does the AURORAXMP Dispatch Model operate 7
Company-controlled hydroelectric generation resources? 8
A. The Dispatch Model treats all hydroelectric 9
generation plants within each river system as a single 10
large plant. To account for the actual flexibility of 11
Company hydroelectric resources, Avista develops 12
individual hydro operations logic for each of its 13
facilities. This separation ensures that the flexibility 14
inherent in these resources is credited to customers in 15
the pro forma exercise. 16
Q. Please compare the operating statistics from the 17
Dispatch Model to recent historical hydroelectric plant 18
operations. 19
A. Over the pro forma period the Dispatch Model 20
generates 68% of Clark Fork hydro generation during on-21
peak hours (based on average water). Since on-peak hours 22
represent only 57% of the year, this demonstrates a 23
substantial shift of hydro resources to the more expensive 24
Kalich, Di Page 12 Avista Corporation

on-peak hours. This is identical to the five-year average 1
of on-peak hydroelectric generation at the Clark Fork 2
through October 2014. Similar relative performance is 3
achieved for the Spokane and Mid-Columbia projects. 4
5
III. OTHER KEY MODELING ASSUMPTIONS 6
Q. Please describe your update to pro forma period 7
natural gas prices. 8
A. Consistent with past general rate case filings, 9
natural gas prices are based on a one-month average from 10
February 20, 2015 through March 19, 2015 of calendar-year 11
2016 & 2017 monthly forward prices. Natural gas prices 12
used in the Dispatch Model are presented below in Table No 13
1. 14
Table No. 1 – Pro Forma Natural Gas Prices 15
16
17
18
19
20
21
Basis Price ($2016/dth) Price ($2017/dth)

Kalich, Di Page 13 Avista Corporation

Q. What is the Company’s assumption for rate period 1
loads? 2
A. Consistent with prior general rate case 3
proceedings, historical loads are weather-adjusted. For 4
this filing weather normalized 2014 load is 1,057.1 5
average megawatts compared to actual loads of 1,061.5. 6
Table No. 2 below details data included in this 7
proceeding. Further information on the weather 8
normalization is within witness Ms. Knox’s testimony. 9
Table No. 2 – Weather Normalized Loads 10
11
12
13
14
15 Q. Please discuss your outage assumptions for the 16
Colstrip units. 17
A. As with our assumptions for other plants, and 18
consistent with prior cases, Avista uses the most recent 19
available five-year average forced outages to estimate 20
long-run performance at the Colstrip plant. The 11.42% 21
forced outage rate is based on the average outages between 22
2010 and 2014. Maintenance outages use the six-year 23
average of planned outages. Six years is used because the 24
Month Load Month Load

Kalich, Di Page 14 Avista Corporation

plants are on a three-year maintenance schedule meaning 1
that a five-year average would over- or under-estimate 2
average maintenance for these plants. 3
Q. Are there any other modeling changes from the 4
last rate filing? 5
A. In the past Avista has not reflected the costs 6
of station service in its proforma power supply expenses 7
because AURORA was unable to account for it. Station 8
service is now tracked in AURORA. Station service is 9
calculated using average station service load between 2010 10
and 2014 for each plant. The cost is determined by 11
multiplying station service consumption by the hourly 12
simulated Mid-Columbia market price. 13
14
IV. RESULTS 15
Q. Please summarize the results from the Dispatch 16
Model. 17
A. The Dispatch Model tracks the Company’s 18
portfolio during each hour of the pro forma study. Fuel 19
costs and generation for each resource are summarized by 20
month. Total market sales and purchases, and their 21
revenues and costs, are also determined and summarized by 22
month. These values are contained in Confidential 23
Schedule 1C and were provided to Mr. Johnson for use in 24
Kalich, Di Page 15 Avista Corporation

his calculations. Mr. Johnson adds resource and contract 1
revenues and expenses not accounted for in the Dispatch 2
Model (e.g., fixed costs) to determine net power supply 3
expense. 4
Q. Does this conclude your pre-filed direct 5
testimony? 6
A. Yes, it does. 7
Kalich, Di Page 16 Avista Corporation