1. Economic Planning Study (EPS) Updates & Study Results ColumbiaGrid Planning Meeting December 8, 2016

2. Outline Background Study Assumptions Study Results To Do
WECC 2026 Common Case Assumptions
Modeling Changes by ColumbiaGrid
Study Results
Flows within the Northwest
West of Cascades
West of Hatwai
Northwest Dispatchable Generation
CAISO Duck Curve
Flows to California
Observation
Market Issue
To Do

3. Background
ColumbiaGrid EPS is intended to provide insight into the changing Western Market and potential impacts to NW power system
Annual study, based on the latest information
This presentation is based on preliminary work using WECC 2026 Common Case v1.30 as the starting point
Modeling objective: A day-ahead market where areas commit and dispatch control supply for their own needs prior to market sales

4. EIM Market
Participants in the Energy Imbalance Market (EIM) have no impact on a day-ahead PCM run
EIM rules require each participant to supply a day- ahead forecast of contracted supply to meet hourly load
The value of an EIM is in optimizing intra-hour operation which is outside the scope of this analyst
EIM study requires multiple PCM run: day-ahead (DA) run → real-time (e.g. hour-ahead (HA)) run which also includes 5-min EIM Market run (EIM)
The HA and EIM run rely on inputs from the previous run

5. WECC 2026 Common Case Assumptions
Started with WECC 2026CC v1.30 dataset
Key changes from WECC 2024 dataset:
Behind the Meter (BTM) PV in CA increased from 7,650 to 12,120 MW
This accounts for 86% of the increased solar generation in California
BTM-PV outside of CA increased from 1,240 to 4,065 MW
Net modeled wind/solar in CA BA 8,240/24,710 MW (total of 32,950 MW)
Net modeled wind/solar of 29,590/33,600 MW (total of 63,190 MW) for the entire WECC
Net increase of wind/solar of 11,185 MW with 79% of the increase located in CA

6. Modeling Changes by ColumbiaGrid#1
Applied applicable changes issued by NTTG as part of coordinated efforts
Correct unusual load shapes to align with historic shape
Corrected modeled California load to properly account for BTM-PV
Lower BTM-PV in California by 1,700 MW to match CEC forecast
Station service was deleted from the PCM to reflect the net generation modeled in the PCM,
Non-Dispatchable supply other than wind and solar was changed from a dispatchable units to a flat monthly shape based on historical operation

7. Modeling Changes by ColumbiaGrid#2
Modeling of dispatchable Hydro in California was changed
From an hourly shape to proportional load following pointing to load minus solar generation
This shifts peak Hydro generation to align with the observed daily peak
AB32 Import Fee:
CA has imposed an import fee to account for the CO2 emission of the imports. The generic value is equivalent to an 8.06 HR unit burning NG
Low CO2 emitters can apply for a lower CO2 import fee based on historic CO2 emissions called Asset Control Supply (ACS)
Modeling of this was change from a flat monthly profile to a nomogram with BPA load – select Hydro generation – CGS

8. Modeling Changes by ColumbiaGrid#3
Split the modeling of combine cycles from plant to 1x1 configuration, i.e. by gas turbine
CC units outside of CA control were cycling mid- day in response to the CC Duck curve. To reduce mid-day cycling, the start cost of CC and GTs in CA was lowered to zero
Manually shifted maintenance in the Northwest to correspond with spring run-off: Apr, May & Jun
Manually shifted maintenance in CA to correspond with peak solar generation: Mar, Apr & May

9. Sensitivity Cases
With uncertainty surrounding the installation of new thermal capacity in the Northwest West three cases are run
For purposes of this analysis, combine cycles are assumed as generic new capacity
The Cases are:
Case Ro: No new capacity is installed
Case R5: Five generic CC are added
Case R7: Seven generic CC are added

10. West of Cascade Monthly
With the retirement of Centralia and Colstrip 1&2 flow shifts from North to South path
Net North/South flow is within historic range
Avg monthly flow for all WECC paths is available upon request

11. West of Cascade Flow Duration
With the retirement of Centralia and Colstrip 1&2 flow shifts from North to South path
Net North/South flow is within historic range

12. Net West of Cascade Hourly Flow
Difficult to determine if change in flow is significant
Flow from BC is impacting hourly flow on West of Cascades (Clean up item)
Need multi year average and operating range is see potential impact on flow (green area in previous charts)
Including previous year of operation would give insight into current market behavior
Avg hourly flow by month for all WECC paths is available upon request

13. Flow on Hatwai Historic flow: 960 MW Avg flow from cases: 25 MW
Reduction in flow 935 MW
Retirement of Colstrip 1 & 2: 614 MW
Mile City AC-DC-AC intertie exports 200 MW in all but summer months (import /5 – 8/4)
Previously modeled as 200 import year round
NWMT has a higher load forecast

14. Northwest Dispatchable Generation
Dispatchable generation remain relatively constant between the cases
In Case R0 existing NG supply makes up 68% of the reduced coal generation
Most of the change in generation between the three case is seen in Feb, Mar, Aug, Sep and Oct
To Do: Run with 2001 Hydro year (Critical year)
Charts for all cases are available upon request

15. California Duck Curve 24,710 MW of solar in CA matters
Duck Curve or observed load: Load minus solar generation
The Duck Curve changes traditional relationship:
Daily min occurs mid-day
Daily min load is significantly lower than traditional min
Daily peak occurs during the evening ramp only
Mid-day peak shift to evening
This results in a significant impact on ramp rate

16. California Duck Curve Ramp Rate
Load w/o solar the avg daily ramp of 12,000 MW ramp in 7 hrs is exceeded during three months
Load minus solar a 12,000 MW ramp in 3 hrs is seen in all but one month

17. Avg Daily NW Flow to California - Apr
Morning ramp: +2,000 MW
Mid-Day ramp: –5,500 MW
Evening ramp: +4,600 MW
COI: 7.5 hrs of imports
PDCI: 5.5 hrs of imports
Net: 7 hrs of imports

18. Avg Daily NW Flow to California - Aug
You can see flow changes to CA based on sensitivity case
COI: Focus on evening ramp
PDCI: Mid-Day imports
Net: Peak evening ramp
Net evening ramp: 2,000 MW

19. Avg Daily NW Flow to California - Dec
COI: double hump
PDCI: Off-peak and shoulder is high with mid- day imports
Net: Double hump with potential mid-day imports
Net evening ramp: 3,250 MW

20. Observations From the study results:
When developing a power flow case, selecting the time of day will be as import as the season
Uncertainty surrounding operational flexibility on the Columbia River generation
Traditional On-Peak 16 hour block sales to California no longer fits their needs
This will impact all sales in the West
To adequately describe interaction in the Western market, the On-Peak time period needs to be divided into three new periods: Morning Ramp, Mid-Day, Evening Ramp

21. Observations
During light load months in California, intra-day sales could see exports/imports/export
During some hours, imports on PDCI and exports COI were observed

22. Market Issues
Balance Areas (BA) are obligated to serve load and have a balanced day-ahead schedule
Exchange between BAs is driven by several factors
Contracts, economic sale, sales do to over commitment, and sales to prevent de-commitment
In general, for DA, current market clearing price do not support daily cycling of CC external to CA
Potential options for California to support its daily ramp:
Contracts: Daily exchange, mid-day sales,..
Higher clearing price to support daily unit cycling
As issues become apparent, more products may be developed

23. To Do Update BC and review Hydro operations
Additional review of Hydro operation
Any Hydro plant on HTC will respond to CA Market (LMP)
What’s reasonable operation to expect?
Review PAR operations
Reduce starts for CC outside of California control
Plants outside of California should only cycle mid-day if they are reasonable compensate for to start
Increase starts for CC and GTs in California to respond to the duck curve
2001 Hydro will help identify supply issues when run with the three cases: R0, R5 and R7

24. Question: Kevin Harris (503) 943-4932 harris@columbiagrid.org