Review Draft
MeetingSummary
NorthWestern Energy Electricity Technical Advisory Committee
November 13, 2013
Attendance
Those participating in or attending the Energy Electricity Technical Advisory Committee (ETAC) meeting in person or via the web and by teleconference included:
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
NameOrganization
Jaime StamatsonMontana Consumer Council (MCC)
Kyla MakiMontana Environmental InformationCenter (MEIC)
Diego RivasNW Energy Coalition
Jeff BlendMontana Department of Environmental Quality (DEQ)
Thomas M. PowerDistrict XI Human Resource Council
Brian DeKiepNorthwest Power and Conservation Council
Neil TempletonMontana Public Service Commission (MPSC)
Will RosquistMPSC
Mike DaltonMPSC
Frank BennettNorthWestern Energy (NWE)
Dave FineNWE
John BushnellNWE
Al BroganNWE
Chuck MagrawNatural Resource Defense Council
Steve FisherLands Energy
Todd GuldsethNWE
Joe StimatzNWE
Joe SchwartzenbergerNWE
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
Patrick BarkeyUniversity of Montana - Bureau of Business and Economic Research (via telephone)
Gerald MuellerConsensus Associates
Agenda
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
The agenda for this meeting focused on a presentation by representatives of Ascend Analytics of its PowerSimm resource planning model and its application to NWE resource planning and acquisition. The specific topics to be addressed include:
•Portfolio characteristics and valuation results;
•Quantifying the value of risk;
•Market inputs;
•Simulation validation;
•Operating dynamics and detailed results; and
•Quantification and monetization.
Introduction
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
Dave Fine introduced the meeting. He noted that modeling discussed at previous ETAC meetings did not incorporate NWE’s proposed acquisition of the PPL Montana hydroelectric facilities in Montana because of NWE’s non-disclosure agreement with PPL Montana during the negotiation. NWE planning staff work has shifted from a planning focus to analyzing the pending hydro facility acquisition. Today’s meeting will continue the process of providing education about Ascend Analytics’ model and its application to NWE’s planning and hydro acquisition. The results presented today are preliminary. Final results will be presented to the ETAC when NWE files its application with the MPSC in December for pre-approval of the hydro acquisition.
Gary Doris, Sean Burrows, and Mark Dyson with Ascend Analytics used a PowerPoint presentation to cover the specific topics listed above in under the agenda heading. A version of the PowerPoint with modeling results redacted will be posted on the ETAC web site at the following address:
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
Gary Doris stated that PowerSimm provides a Monte Carlo framework to construct resource portfolios using simulations of forward markets and during delivery conditions of weather, load, wind, hydropower, and spot market conditions to conduct a risk analysis. Uncertainty is introduced following a highly structured process producing “meaningful” risk analysis that probabilistically envelopes future outcomes. Risk is quantified and monetized using an actuarial approach to help level the economic playing field between different portfolio choices. The monetization approach is more transparent than the efficiency frontier approach to comparing the risk and cost of portfolio alternatives. Ascend’s modeling incorporates information about NWE’s existing generation assets and loads. Load data is combined for all customer sectors;individual load sectors are not broken out.
Highlights of the ETAC discussion during the presentation follows.
ETAC Discussion
Comment - The Ascend approach is not necessarily superior to displaying the modeling analysis in an efficiency frontier, but aids the decision maker in resolving the value of risk. Decision makers will continue to use their judgment in selecting the resource portfolio and making resource acquisitions.
Question - How is the Dave Gates Generation Station treated in the modeling?
Answer - It is modeled as a must-run plant with a constant 7 megawatt (MW) output. It may be possible to address the ancillary services potential of this plant in the future.
Question - Is the combined cycle gas-fired plant assumed to be water or air cooled?
Answer - Air cooled. We recognize that air cooling is a conservative assumption but seems appropriate for MT given the limited availability of water rights.
Question - One of the parameters you use to compare resources is maximized revenue. Why did you choose this parameter rather than minimizing cost?
Answer - Maximizing revenues reflects the fact that NWE can sell into the market. However, maximizing revenues minimizes costs by default.
Question -What time period do you use looking back?
Answer - We start with a monthly period to allow for adjustments to resource operation, e.g. heat rates, with time.
Question - What portfolios did you examine to analyze the hydro acquisition?
Answer - A base or market case, a combined cycle plant case, and the hydro case.
Question - What data was used for the hydro plants?
Answer - We used historic hydro plant data and replicated it through time. We have five years of monthly data augmented with five years of hourly data. We did a statistical analysis of the Rainbow plant to see how it would have operated given its upgrade from 30 MW to 60 MW.
Question - So the historic data were extrapolated into the future?
Answer - We used historic relationships between generation, prices and load to do so.
Question - How did you calculate the 439 MW capacity for the hydro plants?
Answer - We summed the total of the individual plants and subtracted the Kerr capacity because its ownership will shift to the Confederated Salish and Kootenai Tribes in 2015.
Question - Will the modeling for the advanced approval application to the MPSC dispatch the hydro?
Answer - The hydro plants will not be dispatched directly; their capability will be offset.
Question -Does each of the stochastic runs draw from a distribution of water years?
Answer - Yes.
Question - Given the fact that precipitation changes and the hydro plants do not last for ever, how do planners typically value hydro assets?
Answer - Necessary maintenance is included in the operation budget. Water rights are assumed in perpetuity.
Comment -Hydro plant operation is subject to Federal Energy Regulatory Commission (FERC) licenses which are of finite duration.
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
Response - Economic analysis assumes that the water right is perpetual. Because of the hydro plants’ large residual value at the end of the planning period, one should change their discount rate. We did not do so, which makes the analysis of the hydro value conservative, i.e., it under values it.
November 13, 2013 NWE ETAC Meeting SummaryPage 1
Review Draft
Comment - You quantify the value of risk by integrating under the cost distribution curve determined from compiling the results of all of the individual stochastic PowerSimm runs. For past plans, NWE assessed risk by calculating the 5th and 95th percentile of the curve. We need to understand which variables you perturb in the individual model runs.
Question - Is the mid-point of the distribution near the quantified risk value?
Answer - Close. The risk premium corresponds approximately to the difference between the mean cost and one standard deviation up in costs. We add the risk premiumto the total cost of the portfolio.
Question - There is a change in the distribution of costs at 2021. What happened then?
Answer - Carbon costs are added in 2021.
Question - Have you considered examining the impact on the portfolio if carbon costs are not added?
Answer - No. Not adding carbon costs would be imprudent as other utilities in the region include carbon costs in their planning. The value we added, $20 per ton of CO2, is at the lower of the range of values used by the other regional utilities.
Question - How much does the electricity price change when carbon costs are incorporated?
Answer - It increases by about $12 per ton of CO2 emitted, calculated by multiplying the carbon cost by 0.6.
Question - Does this mean that electricity customers will pay 60% of the carbon costs?
Answer - The calculation is 0.6 X $20 per ton of CO2. The 0.6 is the conversion factor.
Question - Shouldn’t the conversion factor decrease with time as electricity generation becomes less carbon intensive?
Answer - The conversion factor is based on gas-fired plants as the marginal generation unit. This is a conservative assumption, i.e. it lessons the value of the hydro plants over time. A higher carbon charge would be required to push existing plants fueled with Powder River Basin coal to the margin and cause them to be retired because of costs.
Question - Does the cost of carbon assume that utilities receive carbon emission allowances?
Answer - No, it is based on opportunity costs.
Question - How is the 0.6 conversion factor derived?
Answer by Todd Guldseth - The Energy Information Administration gives a carbon cost of $15 per ton of CO2 in 2015. A combined cycle gas-fired plant produces 0.4 tons of CO2 per megawatt hour (MWh) of electricity production. A simple cycle plant produces 0.6 tons of CO2 per MWh. A coal plant produces a little over one ton of CO2 per MWh.
Question - What is the experience with other emission costs?
Answer - Ascend has been modeling carbon costs in Europe for 18 years. Carbon costs have come down somewhat. For criteria pollutants such as SO2 and NOx, the costs were initially high but dropped as control technology cost dropped.
Question - How did you determine a distribution of gas costs?
Answer - The distribution was created using historic future price curves and option price quotes. We do not simply make random draws of gas prices because it would not result in meaningful price changes over time. Instead we use a mean reversion component of 0.998 to capture price changes in succeeding modeling time steps. We also maintain the cost relationship between electricity and gas costs, i.e. the heat rate.
Question - Is carbon cost the major driver of risk reduction resulting from the acquisition of the hydro plants?
Answer - Simulating carbon costs would increase uncertainty and hence the risk premium. We add the carbon cost in 2021 which makes hydropower cost effective compared to the market or combined cycle alternative portfolios. However, we use a deterministic carbon cost and do not directly assign uncertainty to it. Thus, the risk premium from carbon emitting electricity generation does not grow. This assumption reduces the benefit of the hydro acquisition.
Question - Have you run other dates than 2021 for imposing the carbon costs?
Answer - No. We could do so. We could for example use a 2019 date.
Comment - The political viability of a $50/Mwh carbon tax is uncertain.
Comment - We do not need a carbon tax for carbon to affect regional power prices.
Comment - NWE must make a reasoned and prudent business decision based on what we know now. Utilities are not building coal plants because of regulatory risk associated with doing so. Other pollutants such as sulfur and mercury are subject to emission standards rather than emission taxes. The question is not whether we will have a carbon tax but whether emitting carbon will raise generation costs.
Question - In addition to the deterministic carbon cost scenario which you are now using, when the advanced approval application for the hydro plant acquisition, could you also include a sensitivity case for carbon costs?
Answer - PowerSimm could do so.
Question - Does the Basin Creek plant run only when it is in the money?
Answer –Generally yes. Updated answer: Basin Creek is also called on for reserves and when it is called on it may or may not be in the money.
Question - Why does the net revenue from the Yellowstone Energy Limited Partnership (YELP) and the Colstrip Energy Limited Partnership (CELP) qualifying facility projects drop off beginning in 2017?
Answer - The net revenue drops with the decline in the implied heat rate.
Question - Do you model Colstrip Units 3 and 4 separately taking into account the reciprocal ownership agreement?
Answer - We scale the full unit amounts based on the ownership shares. When the units make money, they run.
Question - You said that generation is displaced when it is not in the money. Is the net position based on theoretical capability?
Answer - The plants are run based on a market rather than a system reliability perspective. Subject to operational constraints, units are run in the model when they make money, i.e., when market prices exceed generation variable cost of production.
Question - Why is exposure to the market a bad thing?
Answer - With and without the hydro plants, NWE resources do not meet loads, i.e. NWE is short. Exposure to market prices can result in rate shocks.
Question - Why do you use the net present value of risk to judge the value of market exposure?
Answer - One cannot beat the forward price curve. If you could lock in $30 power for 30 years, then you could beat the forward price curve. Such contracts are generally not available because producers would generally prefer to receive higher returns consistent with the market.
Comment - NWE also faces counter party risk when contracting for power. If the counter party defaults then you don’t get the MWs.
Response - Sophisticated utilities factor in credit risk.
Comment - You list the capacity factor of a combined cycle plant in the high 50% range. In the last resource plan, the capacity factor was in the high 80% range.
Response - The baseload capacity factor realized in PowerSimm increases with carbon costs. We also constrain the combined cycle plant operation as follows: two hour cold start, 4 hour run time, 1 hour shut down time, and four hour down time.
Question - Does the load forecast include economic development and energy efficiency?
Answer by Todd Guldseth - As we have previously discussed with the ETAC, the load forecast does not include business cycles. It is based on historic growth.
Question - The analysis you are presenting today focuses on the hydro plant acquisition. What will be in the resource plan?
Answer - We are focused on completing the work that supports the hydro plant asset purchase. We will also address in the plan resource adequacy, compliance with the renewables portfolio standard, as well as the selection of the preferred resource portfolio using PowerSimm.
Comment - The proposed asset purchase allows this plan to focus on an actual resource, the PPL Montana hydro plants, with specific cost and operational characteristics.
Comment - The resource plan question brings to mind something we discussed at the last ETAC meeting, what is planning all about. The plan is built using generic resources. Deciding whether to acquire a specific resource is analyzed a different way. Acquiring the hydro plants raises issues such as the plant ages and climate change which may change flows. NWE will have to show why this specific acquisition is a good idea. Other factors may influence business and corporate decision makers such as the logic in reclaiming part of Montana’s heritage. The planning process does not replace business decisions.
Response by Dave Fine - We are examining whether the hydro plants fulfill NWE’s resource needs and fit into our resource portfolio. The hydro plants have known cost and performance.
Question - What are the independent variables distributions of which are drawn from in individual PowerSimm stochastic runs?
Answer - Each run draws from distributions of weather (temperature), water year, and gas prices.
Next Steps
NWE will complete its advanced approval filing and resource plan. A date for the next ETAC meeting to review the resource plan was not set.
Disclaimer
Committee members provide advice to NWE as individual professionals; the advice they provide does not bind the agencies or organizations that the members serve.
November 13, 2013 NWE ETAC Meeting SummaryPage 1