STOWE ELECTRIC DEPARTMENT 2017 IRP

Transcription

1 STOWE ELECTRIC DEPARTMENT 2017 IRP SUBMITTED BY: Energy New England, LLC On behalf of Stowe Electric Department CONTRIBUTORS: Michelle Coscia, Peter Gomez, and Timothy Hebert of Energy New England; Ellen Burt, Stowe Electric Department Manager

2 TABLE OF CONTENTS A Executive Summary...1 A.1 Overview... 1 A.2 IRP Outline... 2 A.2.1 Resources Requirements... 2 A.2.2 Stowe s Renewable Supply Portfolio... 3 A.2.3 Resource Alternatives... 4 A.2.4 Comparative Tradeoff Analysis and Risk... 5 A.2.5 Stowe s Target Resource Portfolio... 7 B Introduction...9 B.1 Overview of Stowe Electric Department... 9 B.1.1 Overview of Town of Stowe... 9 B.1.2 Stowe Demographics B.1.3 Stowe Climate C Long Term Energy and Demand Forecasts and Scenarios C.1 Demand Forecasting (Submitted by ITRON Inc.) C.2 Background C.3 Forecast Summary C.4 Forecast Approach C.4.1 Customer Class Sales Forecast C Residential Average Use Model C Commercial Average Use Model C Adjustment for VEIC Efficiency Program Savings C Customer Forecast C.4.2 Baseline Energy and Demand Forecast C.4.3 Adjusted Energy and Demand Forecast C Solar Load Adjustment C Cold Climate Heat Pump Impact C.5 Forecast Data and Assumptions C.5.1 Sales, Customer, and Load Data C.5.2 Weather Data C Peak-Day Weather Variables C.5.3 Economic Data C.5.4 Appliances Saturation and Efficiency Trends C.5.5 VEIC Energy Efficiency Program Savings... 40

3 C.5.6 Solar Load Forecast C Solar Saturation Model C.5.7 Cold Climate Heat Pump Forecast D Portfolio Planning Approach and External Influences D.1 Regional Resource Portfolio and Marginal Supply D.2 Market Conditions D.2.1 Capacity Market D.2.2 Energy Market D.2.3 Natural Gas in New England D Reliance on Natural Gas for Electricity Generation in the Northeast D Market Fundamentals Influencing Spot and Forward Pricing of Natural Gas and Wholesale Electricity in New England D Natural Gas in New England - Summary D.2.4 Transmission Market D.3 Assessment of Environmental Impact D.3.1 Emerging Technologies D Distributed Generation (DG) D Electric Vehicle Penetration D Energy storage D Fuel Switching D.3.2 Environmental attributes D.3.3 Assessment of Carbon Impacts D Emission Calculation D Emission Trends E Data Models and Information E.1 RES Optimization Model 71 E.2 Portfolio Optimization Model- Lacima F Assessment of Resources F.1 Existing Energy Resources F.1.1 J.C. McNeil Generating Station F.1.2 New York Power Authority (NYPA) F.1.3 Vermont Electric Power Producers, Inc. (VEPPI) F.1.4 Sustainably Priced Energy Enterprise Development SPEED and Standard Offer F.1.5 Stony Brook Combined Cycle F.1.6 NEW -Hydro Quebec Contract... 77

4 F.1.7 Brown Bear II Hydro (Old Miller Hydro Contract) F.1.8 Saddleback Ridge Wind Project F.1.9 NextEra Seabrook offtake F.1.10 Nebraska Valley Solar Farm F.1.11 Snowmaking Procurement Energy Only Load Following G Renewable Energy Standard (RES) G.1.1 Tier I G.1.2 Tier II G.1.3 Tier III G.1.4 Renewable Energy Credit Arbitrage G.1.5 Snow Making Potential RES Cost G.2 RES modeling G.2.1 Model Assumptions G RES Tier Compliance rates use the CPI adder G Existing REC Market uses the CPI adder G Class I MA REC Market uses the MA compliance rate (using the CPI adder) and the REC market is a percentage of the compliance rate G.2.2 Model Outputs G.3 Existing Capacity Resources G.4 Capacity modeling G.4.1 Model Assumptions G.5 Assessment of Alternative Resources G.6 Smart Rates H Assessment of the Transmission and Distribution System H.1 T & D System Evaluation H.1.1 Substations: H Wilkens Substation H Houston Substation H Lodge Substation H.2 T & D Equipment Selection and Utilization H.3 Implementation of T & D Efficiency Improvements H.4 Maintenance of T & D System Efficiency H.5 Other T & D Improvements H.5.1 Bulk Transmission H.5.2 Sub-Transmission H.5.3 Distribution

5 H.5.4 Grid Modernization H.6 Vegetation Management Plan H.7 Studies and Planning H.8 Emergency Preparedness and Response H.9 Reliably H.10 Assessment of Outage Events and Trends in I Integrated Analysis and Plan of Action I.1 Evaluation of Portfolio Scenarios I.2 Preferred Plan I.2.1 Optimal Scenario I.2.2 Least Cost Scenario I.2.3 Greatest Cost Scenario I.2.4 Other optional Scenario I.3 Implementation or Action Plan I.4 Ongoing Maintenance and Evaluation A Appendix A A.1 Model Results A.1.1 Town Peak Forecast Model A.1.2 Residential Average Use Model A.1.3 Commercial Average Use Model A.1.4 Residential Customer Forecast Model A.1.5 Commercial Customer Model A.1.6 Saturation Model B Appendix B B.1 Model Description B.1.1 Residential Model B Constructing XHeat B Constructing XCool B Constructing XOther B.1.2 Commercial Model B XHeat B XCool B XOther B.1.3 Peak Model B Heating and Cooling Variables B Base Load Variable

6 B Peak Model C Appendix C D Appendix D E Appendix E F Appendix F G Appendix G H Appendix H I Appendix I J Appendix J

7 FIGURES Figure 1: Energy Supply Gap... 3 Figure 2: Renewable Portfolio... 4 Figure 3: 20 year Total Portfolio Cost Comparison for each Portfolio s RES NPV... 6 Figure 4: Risk/Cost Tradeoff Bubble Plot... 7 Figure 5: 20 Year Annual Energy and Total (Inclusive of RES Compliance Costs) Costs of IRP and Competing Alternate Resource Portfolios... 8 Figure 6: Stowe s most commonly used house-heating fuel Figure 7: Common Industries for Males and Females in Stowe vs. Vermont Figure 8: Stowe s Unemployment History Figure 9: Stowe s Average Temperatures Figure 10: Average Climate in Stowe Figure 11: 2016 System Hourly Demand (MW) Figure 12: 2016 System Hourly Demand (MW) Figure 13: Town Average Daily MWh VS. Average Daily Temperature Figure 14: Forecast Framework Figure 15: Town Baseline Hourly Load Forecast ( ) Figure 16: Town Baseline Hourly Load Forecast Figure 17: Town Baseline December Peak-Day Load Forecast Figure 18: Baseline System Hourly Load Forecast Figure 19: Residential SAE Model Overview Figure 20: Residential End-Use Intensities (kwh per Household) Figure 21: Aggregated End-Use Energy Intensities Figure 22: XHeat Figure 23: XCool Figure 24: XOther Figure 25: Predicted Residential Average Use Figure 26: EIA New England Commercial Energy Intensity Projections Figure 27: Commercial XHeat Figure 28: Commercial XCool Figure 29: Commercial XOther Figure 30: Commercial Average Use Forecast Figure 31: Residential EE Program Impact Adjustment Figure 32: Commercial EE Program Adjustment Figure 33: Customer Forecast Figure 34: Peak-Day Heating Requirements Figure 35: Peak-Day Cooling Requirements Figure 36: Peak-Day Base Load Requirements Figure 37: Baseline Town Demand Forecast Figure 38: Solar Hourly Load Forecast ( ) Figure 39: 2027 Solar Hourly Load Forecast Winter Peak-Day Figure 40: 2027 Solar Load Forecast Summer Peak-Day Figure 41: Heat Pump Program Hourly Load Impacts

8 Figure 42: Heat Pump Program Impacts Figure 43: 2027 Heat Pump Hourly Load Peak-Day Figure 44: Baseline and Adjustment Forecast Comparison - Winter Week, Figure 45: Baseline and Adjustment Forecast Comparison - Summer Week, Figure 46: Monthly HDD Figure 47: Monthly CDD Figure 48: Daily Normal HDD and CDD Figure 49: Peak-Day Normal HDD and CDD Figure 50: Payback Curve Figure 51: Solar Share Forecast Figure 52: Solar Share Forecast Figure 53: Projected Heat Pump Penetration (%) Figure 54: Heat Pump Usage (MWh) Figure 55: Capacity Supply Obligation by Fuel Type Figure 56: Rest of Pool Capacity Auction Clearing Prices Figure 57: Forward Capacity Price Simulation Range Figure 58: Vermont LMP Scatterplot Correlation to Northeast Natural Gas Prices Figure 59: ISO New England HUB PEAK FWD CURVE HISTORY Figure 60: Mass Hub ATC LMP, Monthly Simulated Range Jan 2018 to Dec Figure 61: Vermont Zone ATC, Monthly simulated Range Jan 2018 to December Figure 62: Vermont to Mass Hub Basis, Monthly Simulated Range, 5x Figure 63: Vermont to Mass Hub Basis, Monthly Simulated Range, 7x Figure 64: Vermont to Mass Hub Basis, Monthly Simulated Range, 2x Figure 65: New England Resource Mix Percent of Total System Capacity by Fuel Type Figure 66: Link between Regional Prices for Natural Gas and Wholesale Electricity Figure 67: EIA AEO2017 Natural Gas Consumption and Production History/Projections Figure 68: Natural Gas Forward Curve History Figure 69: Natural Gas, Monthly Simulated Range Jan 2018 to Dec Figure 70: Algonquin Citygates, Monthly Simulated Range Jan 2018 to Dec Figure 71: Algonquin to Henry Hub Basis, Monthly Simulated Range Jan 2018 to Dec Figure 72: RNS Forecasted Rates Figure 73: VT Load Forecast Figure 74: ISO-NE Total PV Installed Capacity Survey Results Figure 75: Stowe s Time Traveled to Work Figure 76: Annual Energy Outlook 2017 Conventional Cars vs. Alternative Fuel Cars Figure 77: Stowe Commonly Used Heating Fuel Figure 78 ISO-NE System Energy Generation Percentage by Fuel Source Figure 79 SED CO 2 Emissions and Carbon Free Portfolio Figure 80 SED CO 2 Emissions for RES Figure 81: Energy Resources in Figure 82: Stowe s yearly projected resource distribution Figure 83: Energy Provided by Standard Offer Projects... 77

9 Figure 84: Stowe s Potential RES Credit (Cost) Cash Flow Figure 85: Stowe s Tier I Forecast Figure 86: Stowe s Tier II Forecast Figure 87: Snowmaking Potential RES Cost Cash Flow Figure 88: Net Present Value of RES for Stowe Figure 89: Stowe s Capacity Forecast Figure Model Prices for Capacity Forecast Figure 91: Currently Served by SED Figure 92: SED s Nebraska Valley Solar Farm (1MWAC) Figure 93: SED Line Losses Figure 94: SED Tree Trimming, Last 5 Years: (Note: Underground facilities in black) Figure 95: Cost and Risk Tradeoff Bubble Plot Figure 96: Capacity Rates Figure 97: Optimal Scenario # Figure 98: Tier I with Scenario # Figure 99: Tier II with Scenario # Figure 100: Tier III with Scenario # Figure 101: VT LMP Historic Averages of Around the Clock Figure 102: Actual and Predicted Monthly Peak Demand (MW)

10 TABLES Table 1: Comparative Portfolio... 5 Table 2: System Energy and Demand Forecast Table 3: Baseline Customer Class Forecast Table 4: Moody's January 2017 Economic Projections Table 5: Residential End-Use Intensities (kwh/household) Table 6: Commercial End-Use Intensities (kwh/sqft) Table 7: Residential EE Program (MWh) Table 8: Non-Residential EE Program (MWh) Table 9: Solar Custom and Share Forecast Table 10: Solar Capacity and Generation Forecast Table 11: Heat Pump Energy (MWh) Table 12: ISO Auction Results FCA 7 through FCA Table 13: Stowe s RNS Forecast Table 14: Impact of Potential EV penetration in Stowe s work force Table 15: Capacity and Transmission Savings Table 16: Regional Annual CO 2 Emissions in lb/mwh Table 17: Stowe 2016 Resources Table 18: Stowe 2016 Current Resources Energy Cost Table 19: 2017 Avoided Cost Price CAPS for Standard Offer Table 20: Contract based on 218 MW Table 21: Contract based on 255 MW Table Model Inputs for RES Net Present Value Table 23: Capacity-Clearing Prices Table 24: SED TOU Rate Energy Charge Table 25: SED Capacitor Banks, Sizes, and Locations Table 26: SED Recloser settings Table 27: Scenario Simulation Summary Statistics by Ranking Table 28: Electric Space Heating Equipment Weights Table 29: Space Cooling Equipment Weights

11 A Executive Summary A.1 Overview Stowe Electric Department s (SED or Stowe) 2017 Optimal Integrated Resource Plan is filed pursuant to Vermont Statute 30 V.S.A. 202a. Stowe filed its previous IRP in The New England wholesale energy market continues to evolve, bringing various challenges and opportunities to Stowe and its customers. By operating within a changing environment, Stowe faces uncertainty and volatility that the energy market creates. The intent of the plan submitted herein is for Stowe to continue providing reliable, reasonably priced energy services while managing risk for the utility and its customers. To better navigate this ever-changing energy market, Stowe consults with Energy New England, LLC (ENE), who helps it participate in the ISO New England markets and gives guidance in the structuring of short and long-term power contracts. Stowe Electric Department, ITRON Inc., and Energy New England, LLC prepare this Integrated Resource Plan. Stowe uses the IRP as a key tool in developing its strategic plan. The strategic goal is to optimize Stowe s portfolio with a cost structure that stabilizes rates and improves the financial health, services, and environmental impact for the electric department. Stowe understands there will always be tradeoffs to consider when deciding on various issues concerning future projects and contracts. This planning process considers a number of key influencers to the energy market and several strategies that Stowe could utilize when continuing to build its long-term resource portfolio. Such concepts include: Incorporate future resources that balance low present value costs while reducing the environmental footprint of the portfolio. Stowe aims to construct a portfolio that is both fiscally and environmentally responsible for their customers. Currently, more than half Stowe s portfolio is carbon free or carbon neutral and with the new Renewable Energy Standard (RES), Stowe intends to seek out future resources that serve to fill RES needs while being economical. Consider long-term resources that provide protection against adverse market conditions. Stowe will seek flexible pricing that will work to mitigate current commitment to substantially out-ofmarket resources. Stowe will seek out and review Vermont-based resources to help it comply with RES. In addition, behind-the-meter generation projects that will reduce emissions in Stowe will be priority for analysis, as they will enable Stowe to fill RES standards that begin in P age

12 A.2 IRP Outline Section A. Table of Content gives titles and page numbers per section of the report Section B. Executive Summary provides an overview of the report Section C. Forecasts and Scenarios includes load forecasts and scenarios Section D. Assessment of Environmental Impact gives value to the significant environmental attributes of the resource portfolio. Section E. Data Models and Information Section F. Assessment of Resources reviews existing resources as well as supply options, models and integration of new resources in order to select the preferred portfolio. Section G. Renewable Energy Standard Analysis Section H. Assessment of the Transmission and Distribution System evaluates system improvement of efficiency and reliability for bulk transmission, grid modernization and vegetation management. Section I. Integrated Analysis and Plan of Action is an assessment of demand, supply, finances, transmission, and distribution to find the least-cost portfolio and preferred plan of action. A.2.1 Resources Requirements Stowe has seen a steady increase in sales numbers, with an 8% growth of retail sales from 2014 to Although Stowe has life of unit contracts in their portfolio, they do have a supply gap to address in the future. While this IRP analyzes various portfolio options, it also addresses both coverage and Renewable Energy Standard requirements. The benefits of certain resources in the RES program will have greater implications to SED s overall power costs. Therefore, assessment of resources is based on not only potential cost, but RES offset as well. 2 P age

13 Figure 1: Energy Supply Gap The Base Case load forecast (black line in Figure 1) has load maintaining steady. This includes adjustments for expected future energy efficiency improvements, impacts of solar, electric vehicles, and heat pump penetration. This forecast removes the variable mountain load, only because all mountain costs are billed back, and they never become a detriment to Stowe s ratepayers. Stowe intends to continue to explore ways to supply its portfolio with renewable best benefit solutions. A.2.2 Stowe s Renewable Supply Portfolio Currently, Stowe has over 50% renewable in their supply portfolio. This includes unit entitlements and Purchase Power Agreements (PPAs) that have qualified Renewable Energy Certificates (RECs) and/or State-approved RECs for RES. Figure 2 shows the base caseload applied and matches it to the forecasted output of SED s renewable resources. While Stowe's current portfolio is largely renewable based, the one exception is the Seabrook offtake contract. Although Seabrook s attributes do not count towards RES compliance, it is still a carbon-free energy source in SED s portfolio. When focusing on alternative resources, SED will continue to search out renewable generation and, at the same time, keep in mind overall power costs that may affect customer rates. With RES compliance part of the power costs, renewable generation has the ability to offset RES compliance costs. 3 P age

14 Figure 2: Renewable Portfolio A.2.3 Resource Alternatives Stowe will always try to seek resources for its portfolio that lower cost and are beneficial to State or ISO cost. With Renewable Energy Standard beginning in 2017, SED has begun to seek fair and equitable ways to promote energy efficiency. The IRP process selected combinations of potential resources for evaluation. Together, Stowe and ENE chose twelve scenarios using an optimization algorithm, which is explained in section I.2. ENE s simulation models can be found in section Data Models and Information tested each portfolio for performance within simulated in market environments. The evaluation review chose the ideal scenario using four major criteria: 1) Least Cost: Net Present Value (NPV) of the total portfolio; this includes energy cost of both current resources and potential scenario resources 4 P age

15 2) Renewable Energy Standard: Mean of each scenario based on current RES coverage and resources for each scenario. 3) Standard Deviation: Risk of each scenario relative variation of the expected NPV of Total Portfolio Cost and RES, as measured by the standard deviation and various tradeoff considerations 4) Spot Market Exposure: The relative spot market exposure to Stowe based on each scenario. Stowe maintains a portfolio hedge plan of at least 80% coverage. A.2.4 Comparative Tradeoff Analysis and Risk The Energy New England Portfolio Simulation Model used a couple of simulation-based models that estimate future values of the input variables. The simulation approach to portfolio modeling provides a powerful, unbiased, and dynamic tool to measure the future performance of Stowe s resource portfolio under different market conditions and identifies the factors to which the performance is most sensitive. The RES was a large weight within each scenario model. The RES section of SED s energy portfolio has the largest risk if left unhedged. The I.1 Evaluation of Portfolio Scenarios section describes the details of all twelve scenarios. Table 1 below shows a few scenarios the IRP process analyzed. Table 1: Comparative Portfolio NPV Total Cost Total RES Std Dev Spot Exposure Target Scenario Deviation Least Cost Scenario #1 ($67,904,715.22) ($4,012,367) $ 2,795,522 64% High Cost Scenario #12 ($75,294,817.81) $3,957,634 $ 2,094,589 73% Optimal Scenario Scenario #4 ($74,172,340.64) $2,087,307 $ 2,124,866 72% Here are the highlights of the most competitive resource combination along with Stowe s current resource portfolio: I. Scenario 1 (Base Case) maintains the current portfolio as status quo and does not procure new resources. This set a baseline for comparing alternatives. In the current market environment, this approach can be effective, but requires comparison to a multitude of potential future market states. This scenario is the least cost scenario. II. Scenario 11 is the current portfolio with an extension of Brown Bear Hydro, refurbished Moscow Mills Hydro, Ryegate extension, an additional 1 MW of solar, an additional 1MW of a virtual combined cycle plant and 3% of load coming from a large wind project. This provides the greatest coverage with the less risk due to a low standard deviation. The renewables all help comply with SED s RES. III. Scenario 6 is the current portfolio with an extension of the Hydro Quebec contract and a PPA with VEPPI s Dodge Falls project when the contract expires in 2021 of 2MW. This scenario 5 P age

16 IV. provides RES compliance, but limits the resources within the portfolio to just one large renewable. Scenario 4 is the current portfolio with 1.5 MW of a 5MW wind project in Vermont. This scenario is most optimal due to its ability to maintain RES costs because of the value of the RECs. This project would qualify for both Tier II and I. Although a PPA cost would most likely be high, at or around $0.12 kwh, the RES offset makes it much a more attractive option for SED. These select scenarios provide an analysis of both RES and energy coverage at various levels and price. Using the previously mentioned four major criteria during evaluations allows Stowe to fulfill its goals of compliance and risk coverage in order to help provide reliable, reasonably priced energy. However, one must be cognizant of the fact that with more renewables, although helpful towards RES, there is a reliability risk as well because of higher prices to SED s energy cost. The following figure shows the results of the simulations in a box plot 1 format, which provides a quick visual summary of the mean value, the minimum and maximum values, and the relative amount of relative variation around the expected cost of RES to Stowe for each scenario. Figure 3: 20 year Total Portfolio Cost Comparison for each Portfolio s RES NPV Another method for comparative tradeoff analysis is to rank the portfolios by their standard deviations and then plot them in risk/return 2 space. This plots the expected values along the x-axis and the risk 1 Box-and-Whisker diagram, the white area, or the box, represents the upper and lower quartiles (25 th and 75 th percentiles) of values, the black line is the 50 th percentile of the data, and the thin black lines, or the whiskers, represent the minimum and maximum values of the sample data. 6 P age

17 on the y-axis. For this analysis a bubble chart was used, where each bubble is a point on the chart and represents a portfolio s relative position based on its respective expected value, X, and standard deviation, Y. This allows for a comparison and evaluation of portfolios based on their location on the chart namely, which quadrant they fall within from the output of the modeling. For example, if comparing portfolios on risk vs. least cost, the lower left quadrant should contain the portfolios with both lower costs and risk, and the upper right quadrant should hold the higher cost and higher risk portfolios. The additional benefit of using a bubble chart is that the relative size of each bubble also represents that relative variation of each portfolio. Not only does the quadrant show a portfolio s merit, but displays the size of a portfolio s bubble according to its relative risk. Figure 4 shows the bubble plot comparison for least cost and risk. Figure 4: Risk/Cost Tradeoff Bubble Plot A.2.5 Stowe s Target Resource Portfolio Based on the comparative analysis, the optimal portfolio is Portfolio 4 (Scenario #4) for SED s Integrated Resource Plan. The caveat is that specific resource volumes will be determined relative to Stowe s load requirements throughout the term of this plan. These volumes will need adjusting to effectively balance the cost and environmental performance while avoiding the purchase of too many resources at certain times of the year. Material changes to Stowe s load, whether efficiency driven or not, will have an impact on the volume and nature of new resources pursued. 2 risk/return space is term used in Portfolio Theory when finding the Min-Variance portfolio, where return is term used when portfolio consists of equity assets; in the IRP context we use the implied improvement (savings/benefit) in Total Cost metrics by pursuing an alternative resource portfolio as a proxy for return. 7 P age

18 Stowe s optimal Integrated Resource Plan: Portfolio 4 = Stowe s current existing resources, 1.5MW of Vermont based Wind PPA (Unit is considered to be a Tier II qualified new Distributed Generation project), and without a regularly planned schedule of forward market purchases, market purchases will be dictated by the dynamics of SED load and rate stability considerations. The results point to the enhanced economic and environmental performance that is achievable by allocating resources to one of the alternative portfolios. An expected performance, such as lower average cost and lower greenhouse gas emissions, has a more reliable estimation when choosing resource combinations that exhibit relatively lower values of variations in the sample data. The most competitive portfolios strike a balance with resources that improve the environmental performance towards Vermont s Renewable Energy Standard and take advantage of the current market environment, which provide lower costs over time and across various market environments. Figure 5 shows how the selected IRP portfolio (Scenario #4) expects to enhance Stowe s annual cost structure over the next 20 years. Figure 5: 20 Year Annual Energy and Total (Inclusive of RES Compliance Costs) Costs of IRP and Competing Alternate Resource Portfolios The plan incorporates the following time line and action points: 1. Continue to explore ways to promote energy efficiency and conservation for Tier III compliance purposes. 2. Monitor load growth or contraction on an ongoing basis. 8 P age

19 3. Continue market purchases as needed in a low commodity price environment over the next several years. This is especially relevant for the Stowe Mountain Snow Making contract. 4. Continue to investigate adding in-state renewable resources. 5. Continue to review renewable resource alternatives, including wind, biomass, and hydro, to both diversify and comply with RES within SED's portfolio. Technology improvements, the relative cost of market power, i.e. higher fossil fuel, and renewable energy credit prices will make these resources more attractive and affect their reviews. 6. Continue to procure short-term market contracts as needed to mitigate Stowe s exposure to shortterm price volatility and to enhance rate stability. 9 P age B Introduction B.1 Overview of Stowe Electric Department In March of 1763, the Village of Stowe was founded and the first settlement took place in As the Village of Stowe grew, it added most of the Town of Mansfield in 1840, and the rest of Mansfield in 1855 along with the Town of Sterling. The first electric department was established in 1911 as the Village of Stowe Electric Light and Power System. In 1996, the Village of Stowe and the Town of Stowe merged and with that, the Town of Stowe Electric Department ( SED ) became an enterprise division of the Town. Currently, SED's consumer base consists entirely of residents and businesses within the Town of Stowe. Over 4,000 Residential and Commercial customers rely on SED to provide reliable energy at affordable prices. During the second half of 2008, SED contracted Energy New England LLC (ENE) to manage its wholesale power supply entitlements. In recent years, SED had a VELCO transmission expansion and upgrade, called the Lamoille Country Project. This upgrade consisted of 10 miles of new 115kV lines installed between Duxbury and Stowe. SED also benefited from the construction of a new 115/34.5kV substation. The entire upgrade resulted in a more efficient electrical usage by creating greater reliability to the Town of Stowe. SED consistently looks to the future and is investigating in carbon reducing alternatives such as installing electric charging stations, building and owning a solar farm, and installing smart meters so customers can make informed decisions on energy usage. SED is committed to exploring all avenues, which will give the most reliable energy and service at the most affordable cost to its consumers. SED's Board of Electric Commissioners are engaged and get involved within the community. Its ratepayers are always first in mind. The utility supports environmentally viable and economical power from local sources, and evaluates all sustainable contracts for of purchased power from renewal sources that fall within its budget. B.1.1 Overview of Town of Stowe Located in Lamoille County, the Town of Stowe is mainly a winter recreational area whose activities center on Mount Mansfield. Stowe s winter sport availability is a substantial revenue generator for the town, with a significant amount of its revenues derived from Stowe Mountain Resort.

20 The Town of Stowe also capitalizes on the landscape s exceptional beauty and scenery, enabling Stowe to develop an extensive year-round tourist economy. The annual transition from summer to fall with its beautiful foliage spectrum has become a popular tourist attraction. Since winter is a strong tourist season for Stowe, it is important to understand the main fuel source that Stowe s residences are using. Stowe s housing and condominium heating source representation is found in Figure 6 below. These facts will become important when Stowe looks for ways to implement energy efficiency within the service territory for Tier III compliance. 10 P age

21 Figure 6: Stowe s most commonly used house-heating fuel 3 With 2008 s capital improvement investment in the Lamoille county reliability project by VELCO, which upgraded 10 miles of new 115kW lines and added a new 115/34.5 kv substation, SED elevated its dependability to its customers. The upgrades allowed SED to plan and build future projects, such as electric charging stations, and especially benefited the Mt. Mansfield snow making usage. B.1.2 Stowe Demographics As of , the population in Stowe, VT was 4,314, with the median residential age of 44.9 years. Within the occupied residential housing market, 72% owner occupied while 28% are renter occupied. 5 The median household income in 2015 was $70,022 (vs. $56,990 for Vermont), while the median house or condominium value was $431,760 (vs. $223,700 for Vermont). Stowe s main industry for jobs, for both male and female in 2015 was within the accommodation and food services industry as shown below in Figure 7, this is due to the heavy importance of tourism for the town. The second common industry is technical and educational services P age

22 Figure 7: Common Industries for Males and Females in Stowe vs. Vermont 6 Stowe s 2015 unemployment rate was 3.5% (vs. 3.9% for Vermont). Stowe s unemployment history is found in Figure 8 below P age

23 Figure 8: Stowe s Unemployment History B.1.3 Stowe Climate Stowe s climate is also important to take into consideration when planning future generation and/or location of generation. Stowe s average climate, found below in Figure 9: Stowe s Average Temperatures provides insight into which months are the highest heating and cooling driven months. Figure 9: Stowe s Average Temperatures 7 The data compiled by the city-data.com website, which uses over 4,000 weather stations, shown below in the graphs of Figure 10: Average Climate in Stowe provides additional information. By analyzing wind speed and cloud coverage, Stowe is able to make educated assumptions of resource optimization within Stowe. Although renewable generation has benefits to Stowe, it is important to choose the resource that will benefit Stowe the most by providing the greatest output P age

24 Figure 10: Average Climate in Stowe 8 C Long Term Energy and Demand Forecasts and Scenarios C.1 Demand Forecasting (Submitted by ITRON Inc.) The Town of Stowe Electric Department (Stowe) contracted Itron, Inc. (Itron) through Energy New England (ENE) to develop a twenty-year energy and demand forecast to support the IRP planning process. This document provides an overview of the sales and energy trends, forecast results, forecast assumptions, and methodology P age

25 C.2 Background Stowe serves approximately 3,300 residential customers and 780 commercial customers; this includes the Stowe Mountain Resort (Mountain). Snow making is a large share of the load served; snowmaking and auxiliary loads accounted for 15% of Stowe s 2016 electric sales. Stowe has a relatively large commercial customer class with the commercial sector accounting for approximately 55% of system sales. The residential sector accounts for 30% of sales and mountain snowmaking the remaining 15% of system sales. Stowe electric sales were hit hard by the Great Recession. In 2009, electric sales fell 8.0% and the number of commercial customers dropped 2.0%. However, since that time, Stowe has seen a strong resurgence in customer and electric sales growth. Total sales have increased from 57,879 MWh in 2009 to 68,213 MWh in 2016 a 17.9% gain in sales. The system added 66 new commercial accounts a 7.0% gain over 2009, and added 88 residential accounts a 5.8% gain over A large contributor to this growth has been the expansion at the Stowe Mountain Resort; this expansion includes a new ski lodge, hotel, learning center, a condominium development, and performance art center. The economic recovery has also helped spur growth as visitation has increased significantly over the last five years contributing to hotel and restaurant sales growth and increase occupancy rate in the second-home market; roughly half the homes in Stowe are second-homes. Sales growth is expected to slow over the longer term as Resort expansion work slows along with the regional economic growth. Further, the state energy efficiency efforts through VEIC and further adoption of behind the meter (BTM) solar are expected to translate into declining long-term customer average usage. Stowe is a winter peaking utility with significant load variation in the winter months; this variation is largely driven by snowmaking. Figure 11 shows 2016 system hourly demand. Figure 11: 2016 System Hourly Demand (MW) For forecasting purposes, snowmaking load (Mountain Load) is separated from residential and commercial sales (Town Load). Figure 12 shows the 2016 Town hourly load demand. 15 P age

26 Figure 12: 2016 System Hourly Demand (MW) The 2016 Town load profile is typical of past years. Stowe generally peaks during the Christmas holiday period. Figure 13 shows town daily load (MWh) against daily average temperature. As depicted, even though it is colder in January and February, loads are higher in December. Figure 13: Town Average Daily MWh VS. Average Daily Temperature C.3 Forecast Summary Both customer and average use growth has been strongly impacted by improving state and regional economic growth and the Resort expansion activity. In the Baseline Forecast system, energy and demand are expected to flatten out. Customer growth should slow with slower household and economic growth. Customer average use is expected to decline and begin tracking state usage trends as a result of improving end-use efficiency, strong growth in statewide efficiency activity and increase in solar market penetration. 16 P age

27 Table 2 shows projected annual energy and peak demand. Baseline Forecast reflects current state economic and household projections, New England end-use saturation and efficiency projections (EIA 2017 residential and 2016 commercial forecasts), end-use intensity adjustments for state appliance saturation surveys, and VEIC current efficiency program savings projections. The Adjusted Forecast incorporates expected impact of additional BTM solar load growth and state efforts promoting coldclimate heat pump adoption. Table 2: System Energy and Demand Forecast Baseline Adjusted Year MWh Peak (MW) MWh Peak (MW) , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , % -0.4% 0.0% 0.0% % 0.0% 0.0% 0.4% In the Baseline Forecast, average use declines faster than customer growth, strong end-use efficiency gains from end-use efficiency standards and VEIC efficiency program activity drive average use lower. Winter peak demand is especially impacted by the strong decline in lighting intensity due to the expected impact of new lighting standards and VEIC efficiency programs targeting lighting and heating. The Adjusted Forecast is lower than the baseline forecast through 2025 as a result of strong solar load growth. Adjusted energy is greater than baseline after this point as cold-climate heat pump sales outweigh solar load loss. Winter peak demand is stronger in the Adjusted Forecast as a result of VEIC s cold-climate heat pump program. 17 P age

28 C.4 Forecast Approach Energy and demand forecasts are derived using a bottom-up framework as depicted in Figure 14. Figure 14: Forecast Framework The process entails first developing residential and commercial sales forecasts and using heating, cooling, and base-use energy forecasts derived from the class sales forecast to drive Town-level demand requirements. Energy requirements are calculated by applying a loss factor to the class sales forecast. The baseline hourly load forecast can then be calculated by combining peak, energy, and system hourly load forecast. Figure 15 through Figure 17 show resulting Town hourly load forecast. Figure 15: Town Baseline Hourly Load Forecast ( ) 18 P age

29 Figure 16: Town Baseline Hourly Load Forecast 2027 Figure 17: Town Baseline December Peak-Day Load Forecast 2027 The Baseline System load forecast is derived by adding the Town and Mountain hourly load forecasts. Figure 18 shows the 2027 Baseline system hourly load forecast. Figure 18: Baseline System Hourly Load Forecast 2027 System Town Mountain C.4.1 Customer Class Sales Forecast The forecast process begins with developing long-term residential and commercial sales forecasts. Customer heating, cooling, and base-use energy requirements are then used to calculate energy requirements and drive system peaks through a monthly peak-demand regression model. Over the longterm, structural change as well as changes in economic and weather, conditions drive customer usage. 19 P age

30 Improvements in end-use efficiency resulting from new appliance and business equipment efficiency standards and state energy efficiency programs have had a significant impact on customer usage across the state. The impact of end-use efficiency improvements are captured through monthly customer average use models estimated using a Statistically Adjusted End-Use (SAE) model framework. The SAE model is estimated using a linear regression specification that relates customer average use to estimates of heating (XHeat), cooling (XCool), and base-use (XOther) energy requirements. The end-use variables are constructed by combining structural elements such as end-use saturation, average end-use stock efficiency, and index for housing thermal shell improvements with economic drivers, weather conditions, and price. Figure 19 shows the residential SAE model specification. Figure 19: Residential SAE Model Overview AC Saturation Central Room AC AC Efficiency Thermal Efficiency Home Size Income Household Size Price Heating Saturation Resistance Heat Pump Heating Efficiency Thermal Efficiency Home Size Income Household Size Price Saturation Levels Water Heat Appliances Lighting Plug Loads Appliance Efficiency Income Household Size Price Cooling Degree Days Heating Degree Days Billing Days XCool XHeat XOther AvgUse = a + b XCool + b XHeat + b XOther + e m c m h m o Estimate monthly model with historical billed sales data m m The model detail is provided in Appendix A. C Residential Average Use Model The primary forecast driver is the end-use intensities, which are incorporated into the model variables. End-use intensities are measured in kwh per household in the residential sector and kwh per square foot in the commercial sector; energy intensity estimates incorporate both end-use saturation trends (share of homes that own a given appliance or business equipment) and annual kwh usage for that enduse. For example, roughly 75% of the homes in Vermont use an electric dryer. On average, an electric dryer in 2016 used 725 kwh per year. The energy intensity can then be calculated as: Dryer EI = share of homes that have an electric dryer X annual usage = 76% 725kkh = 551kkh 20 P age

31 The dryer energy intensity will change over time as the saturation increases and the annual kwh declines with improvements in appliance efficiency. If end-use efficiency is improving faster than end-use ownership then the energy intensity will decline over time. This is the case for most residential and commercial end-uses. Figure 20 shows the primary residential end-use intensities. Figure 20: Residential End-Use Intensities (kwh per Household) Over the last ten years, there has been significant decline in electric heating, lighting, refrigeration intensities, counter these end-uses, TV, water heating, and miscellaneous intensities have been increasing. Figure 21 shows aggregated end-use intensity trends when mapped to cooling, heating, and other use. 21 P age

32 Figure 21: Aggregated End-Use Energy Intensities Heating Cooling Base Total % 0.3% -0.2% -0.4% % -0.1% -1.0% -1.0% % 0.1% -0.7% -0.7% Across all end-uses, energy intensity has averaged 0.4% annual decline, largely as a result of new appliance standards. Cooling intensity has actually increased as strong growth in room air conditioning has outweighed air conditioning unit efficiency gains. Over the next ten-years, intensities are expected to fall at an even greater rate, as miscellaneous energy growth slows significantly and new lighting standards drive base-use intensity lower. In the residential sector, end-use intensities are combined with real personal income number of persons per household, and price. The rationale is that end-use utilization is partly driven by increase in real income, number of occupants, and real electricity prices. While small, the expected impact is that households will increase electricity use with increase in income, usage will decline as household size fall, and as prices increases, household electricity use will decline. Price is excluded from this model, as we assume real electricity prices are constant. Household income, household size, and price are integrated in the constructed model variables - XHeat, XCool, and XOther - by imposing elasticities on income, household, and price. The primary economic driver, real household income, has an elasticity of 0.20; a 1.0% change in real income translates into a 0.2% change in XHeat, XCool, and XOther. XHeat also incorporates actual and normal monthly HDD, XCool actual and normal monthly CDD, and XOther the number of days in the months. Figure 22 through Figure 24 show the constructed model variables. 22 P age

33 Figure 22: XHeat Figure 23: XCool Figure 24: XOther 23 P age

34 The average use forecast is derived using a linear regression model that relates average use to XOther, XHeat, and XCool. Figure 25 shows predicted residential average use. Figure 25: Predicted Residential Average Use C Commercial Average Use Model Similar end-use variables XOther, XHeat, and XCool are constructed for the commercial sector; economic drivers are combined with base, heating, and cooling end-use intensity projections. Energy intensity in the commercial sector is measured on a kwh per sqft basis. Historical and forecasted intensities are derived from the EIA s Annual Energy Outlook (AEO) for New England. Figure 26 compares total building intensity projections used in the 2014 and 2017 long-term Vermont energy and demand forecast. Figure 26: EIA New England Commercial Energy Intensity Projections Total14 Total17 kwh/sqft Years Total14 Total % -1.1% % -1.6% % -0.4% % -0.2% P age

35 Updated intensity estimates show the decline in commercial energy intensity was stronger than initially thought with commercial intensity averaging 1.1% decline. Most of this decline can be attributed to decline in lighting and business equipment intensities. On the business equipment side, PCs have been replaced with laptops and computing requirements that have transitioned from local servers to data centers that are significantly more efficient. EIA expects the decline in commercial energy intensity to slow significantly through the forecast period. Model variables are used in estimating a commercial average use model. Construction of the model variables are provided in Appendix A and models statistics in Appendix B. The commercial economic variables incorporate actual and forecasted state GDP and employment. Figure 27 through Figure 29 show the commercial end-use model variables. Figure 27: Commercial XHeat Figure 28: Commercial XCool 25 P age

36 Figure 29: Commercial XOther The drop in February XOther reflects fewer number of days in February. The commercial average use forecast is derived from a linear regression model that relates monthly commercial average use to XOther, XHeat, and XCool. Figure 30 shows the resulting commercial average use forecast. Figure 30: Commercial Average Use Forecast C Adjustment for VEIC Efficiency Program Savings Both residential and commercial average use forecasts also reflect the expected energy savings from Vermont Efficiency Investment Corporation (VEIC) energy efficiency program activity. VEIC provides energy savings estimates by customer class and end-use. We assume that 80% of future EE program 26 P age

37 savings is already embedded in the forecast; historical savings used in estimating the forecast models includes the impact of prior-year EE programs and also impact the EIA intensity forecasts as EIA calibrates annual intensity estimates to steadily declining customer usage. Figure 31 and Figure 32 show the intensity impact of the EE adjustments. Figure 31: Residential EE Program Impact Adjustment Baseline Adjusted Baseline Adjusted % -0.4% % -1.4% Figure 32: Commercial EE Program Adjustment Baseline Adjusted Baseline Adjusted % -1.7% % -0.8% The tables show the compounded annual growth rates. EE impacts reduce residential average annual growth by 0.4% (-1.0% to -1.4%) and doubles the commercial average use decline from -0.4% to -0.8%. C Customer Forecast The sales forecast is derived by combining the average use forecast with customer forecast. Coming out of the recession, Stowe has experienced strong residential and commercial customer growth. Customer growth has been nearly twice as high as that of the state. Customers are forecasted using a monthly regression model that relates number of customers to a mix of household, employment and state GDP 27 P age

38 projections. Customer growth slows over the forecast period with slowing economic growth. Figure 33 summarizes residential and commercial customer forecasts. Figure 33: Customer Forecast Customers - Average Annual Growth Residential Commercial Total % 0.85% 0.80% % 0.64% 0.56% Residential Commercial Table 3 summarizes baseline customer class sales and customer forecast. The baseline forecast is adjusted for future EE program impacts. Table 3: Baseline Customer Class Forecast 28 P age Residential Commercial Year Sales (MWh) Customers AvgUse (kwh) Sales (MWh) Customers AvgUse (kwh) ,719 3,237 6,401 38, , ,575 3,252 6,941 39, , ,636 3,256 6,952 41, , ,650 3,297 6,871 42, , ,313 3,342 6,976 44, , ,080 3,367 6,854 42, , ,014 3,380 6,809 42, , ,913 3,405 6,729 42, , ,324 3,421 6,526 42, , ,885 3,429 6,383 42, , ,687 3,444 6,296 42, , ,622 3,473 6,226 42, , ,632 3,499 6,182 42, , ,415 3,519 6,085 42, , ,270 3,537 6,014 42, , ,189 3,555 5,960 42, , ,211 3,576 5,931 42, , ,128 3,597 5,874 43, , ,008 3,618 5,806 43, , ,920 3,640 5,747 43, , ,913 3,661 5,712 43, , ,804 3,683 5,649 43, , ,779 3,705 5,608 43, , ,768 3,728 5,571 44, , ,821 3,751 5,551 44, , ,762 3,774 5,502 44, , % 0.6% 1.7% 2.9% 0.9% 2.0% % 0.5% -1.4% 0.1% 0.6% -0.6% % 0.6% -0.8% 0.4% 0.7% -0.3%

39 C.4.2 Baseline Energy and Demand Forecast The Town Baseline energy forecast is calculated by applying historical monthly average loss factors to the monthly sales forecast. The total system energy forecast is the sum of the Town energy and Mountain energy forecasts. The Mountain energy use is primarily energy used for snowmaking; energy forecast is based on average sales over the last five year. Adjusted for line losses, Mountain energy is a little over 11,000 MWh per year. System peak requirements can be expected to change as underlying heating, cooling, and non-weather sensitive (base-use) energy requirements change. To capture the impact of changing end-use sales growth on peak, the Baseline peak demand is estimated with a monthly regression model that relates monthly peak-demand to peak-day HDD and CDD, and system heating, cooling, and base-use load requirements. System-level heating, cooling, and base-use energy requirements are derived from the residential and commercial sales forecasts. The peak model variables are defined as the interaction of peak-day CDD and HDD with cooling and heating energy requirements and estimated baseload requirements. Figure 34 shows estimated peak-day heating requirements. Figure 34: Peak-Day Heating Requirements Similar peak-day load estimates are generated for cooling and non-weather sensitive use (base-use). Constructed variables are shown in Figure 35 and Figure P age

40 Figure 35: Peak-Day Cooling Requirements Figure 36: Peak-Day Base Load Requirements The peak-day end-use load requirements estimates are used in developing a monthly peak-day regression model. Figure 37 shows the Baseline forecast results. Stowe remains a winter peaking utility; summer peak demand is increasing, while winter peak demand is decreasing. Declines in winter peaks reflect a decline in lighting and heating intensities, which are the result of new efficiency standards and VEIC s EE program activity. 30 P age

41 Figure 37: Baseline Town Demand Forecast Winter Summer Average Annual Growth Rate Summer Winter % -0.40% % 0.01% C.4.3 Adjusted Energy and Demand Forecast The Baseline Forecast is adjusted for both BTM solar and cold-climate heat pump impacts; cold climate heat pumps are being promoted as part of VEIC s efforts to reduce the cost and impact of energy use and a means with which electric distribution utilities can meet state Tier 3 CO 2 goals. The forecast is adjusted for new solar adoptions and heat pump saturation beginning in the first forecast year. The impact of prior-year adoptions are embedded in historical load data and incorporated in the Baseline hourly load forecast. Given the scale of expected solar load growth and program-related heat-pump adoptions, the system hourly load may potentially change over time; this would impact both the timing and level of peak demand. The Adjusted hourly load forecast is derived by combining the Baseline hourly load forecast with solar and heat pump hourly load forecasts. C Solar Load Adjustment Figure 38 through Figure 40 show the incremental BTM solar hourly load forecast. The BTM solar capacity forecast is based on a regression model that relates saturation (i.e., share of the homes that have a solar system) to simple payback. The capacity forecast is translated to monthly generation and hourly load forecasts based on a typical solar load profile for Stowe. 31 P age

42 Figure 38: Solar Hourly Load Forecast ( ) Figure 39: 2027 Solar Hourly Load Forecast Winter Peak-Day Figure 40: 2027 Solar Load Forecast Summer Peak-Day Given Stowe is expected to peak in December at night, solar has no impact on winter peak demand. C Cold Climate Heat Pump Impact VEIC has recently launched a program to promote the adoption of cold climate heat pumps; the program is intended to help meet the Vermont goals to reduce CO 2 emissions from heating end-users. Program-driven heat pump loads are added to the Baseline Forecast. The heat pump program is 32 P age

43 expected to have a significant positive impact on winter peak demand. Figure 41 through Figure 43 show program induced heat-pump hourly load impacts. Figure 41: Heat Pump Program Hourly Load Impacts Figure 42: Heat Pump Program Impacts 2027 Figure 43: 2027 Heat Pump Hourly Load Peak-Day The hourly solar forecast is subtracted from the Baseline hourly load forecast and the heat pump hourly load forecast is added to the Baseline hourly load forecast. Figure 44 and Figure 45 compare the 2027 Baseline and Adjusted system hourly load forecasts. 33 P age

44 Figure 44: Baseline and Adjustment Forecast Comparison - Winter Week, 2027 Winter Adjusted Baseline Figure 45: Baseline and Adjustment Forecast Comparison - Summer Week, 2027 Summer Baseline Adjusted The winter adjusted hourly load forecast is higher than the Baseline Forecast as increases in heat-pump heating loads outweigh BTM solar impacts. BTM solar impacts can be seen in summer load profile Adjusted summer hourly load profile is lower than the Baseline profile over the day-time hours. The impact of solar load, however, is somewhat mitigated by heat-pump cooling load increases. C.5 Forecast Data and Assumptions C.5.1 Sales, Customer, and Load Data Monthly residential and commercial average use models are estimated from historical billed sales and customer counts. These models are estimated using nine years of historical monthly billed sales and customer data from 2008 to The peak demand model is based on peak data derived from Stowe s town hourly load data, which covers the period January 1, 2011 to December 31, Town hourly load data is also used in estimating the Baseline Town hourly load profile. The Mountain hourly load forecast is based on historical hourly load data for serving the Resort snowmaking equipment. 34 P age

45 C.5.2 Weather Data Monthly HDD and CDD are calculated from historical daily maximum and minimum temperature data for the Burlington Airport. Normal monthly HDD and CDD are calculated for the most recent 20-year period ( ). Normal degree-day estimates are used to generate the Baseline hourly load profile. Figure 46 and Figure 47 show historical and normal monthly HDD and CDD, and Figure 48 shows daily normal HDD and CDD. Figure 46: Monthly HDD 35 P age

46 Figure 47: Monthly CDD Figure 48: Daily Normal HDD and CDD C Peak-Day Weather Variables Normal peak-day CDD and HDD are based on temperature data from the Burlington Airport and are calculated by evaluating peak-month HDD and CDD over a twenty year period (1997 to 2016). The process entails using a rank and average approach as described below: 36 P age

47 1. Calculate daily HDD and CDD over the ten year period. 2. Find the highest HDD and CDD that occur in each month. This results in twelve monthly HDD and twelve monthly CDD for each year. 3. Rank the monthly HDD and CDD in each year from the highest value to the lowest value. 4. Average across the annual rankings average the highest HDD values in each year, the second highest in each year, the third highest, the lowest HDD values in each year. This results in twelve HDD values and twelve CDD values. 5. Assign the HDD and CDD values to specific months based on past weather patterns. The highest HDD is assigned to January and the highest CDD value is assigned to August. Figure 49 shows the calculated peak-day normal HDD (base 55 degrees) and CDD (base 65 degrees). Figure 49: Peak-Day Normal HDD and CDD PkHDD PkCDD C.5.3 Economic Data State economic forecasts drive the energy and demand forecasts. While Stowe is a small part of the state in terms of economic activity and energy consumption, sales and customer growth are strongly correlated with state economic activity. The energy and demand forecast is based on Moody s Economy.com January 2017 economic forecast for Vermont. Table 4 summarizes the primary economic drivers. 37 P age

48 Table 4: Moody's January 2017 Economic Projections Population Households Real Personal Employment GSP (Mil $) Year (Thou) (Thou) Income (Mil $) (Thou) ,214 25, ,950 25, ,395 25, ,150 25, ,249 26, ,038 26, ,451 26, ,620 26, ,047 27, ,766 27, ,173 27, ,416 28, ,733 28, ,925 28, ,998 28, ,167 28, ,479 29, ,951 29, ,342 30, ,621 30, ,889 31, ,206 31, ,559 31, ,891 32, ,215 32, ,545 33, ,883 33, ,231 34, ,585 34, ,936 34, ,283 35, ,632 35, % 0.5% 1.5% 0.7% 0.3% % 0.5% 0.9% 1.2% 0.6% % 0.4% 1.0% 1.2% 0.6% C.5.4 Appliances Saturation and Efficiency Trends Residential end-use saturation, average end-use efficiency, and intensity projections are based on the EIA 2016 AEO. Commercial end-use intensities are pulled from the 2016 AEO commercial end-use forecast database for the New England Census Division. Residential end-use intensities are adjusted to reflect Vermont-specific end-use saturations and estimated annual appliance kwh use. Residential and commercial end-use intensities are summarized in Table 5 and Table P age

49 Table 5: Residential End-Use Intensities (kwh/household) Year Heating Cooling Base Total , ,679 8, , ,633 8, , ,595 8, , ,565 8, , ,611 8, , ,650 8, , ,715 8, , ,686 8, , ,552 8, , ,547 8, , ,557 8, , ,538 8, , ,492 8, , ,422 7, , ,217 7, , ,101 7, , ,018 7, , ,944 7, , ,883 7, , ,809 7, , ,746 7, , ,697 7, % 0.3% -0.2% -0.4% % -0.1% -1.4% -1.3% Table 6: Commercial End-Use Intensities (kwh/sqft) Year Heating Cooling Base Total % -0.3% -1.7% -1.7% % -0.9% -0.7% -0.8% 39 P age

50 C.5.5 VEIC Energy Efficiency Program Savings End-use intensity projections are adjusted for VEIC s expected energy efficiency (EE) program savings. VEIC provides historical and forecasted EE savings by end-use. The forecast reflects VEIC s March 2017 expected program savings. As discussed earlier, a significant amount of future EE savings is already embedded in the forecast. Table 7and Table 8 show total EE program savings by class and primary enduse. Table 7: Residential EE Program (MWh) Year Heating Cooling Other Total , ,530 44, , ,158 46, , ,133 38, , ,211 39, , ,831 37, , ,416 36, , ,094 37, , ,046 34, , ,119 35, ,216 1,016 24,152 36, ,854 1,049 27,603 40, ,670 1,083 28,023 41, ,427 1,117 28,221 42, ,603 1,116 29,517 44, ,684 1,194 29,690 44, ,806 1,194 33,124 48, ,785 1,253 33,558 48, ,939 1,253 34,512 49, ,947 1,254 35,778 50, ,076 1,256 36,694 52, % 9.7% -4.8% -2.3% % 2.1% 4.3% 3.6% 40 P age

51 Table 8: Non-Residential EE Program (MWh) Year Heating Cooling Other Total ,839 2,923 60,563 69, ,492 3,272 66,403 76, ,880 3,542 65,818 76, ,549 3,672 70,811 82, ,530 3,766 73,284 84, ,509 3,919 70,023 81, ,471 4,527 73,423 85, ,469 5,044 74,408 86, ,451 5,262 74,090 86, ,449 5,662 76,063 89, ,445 6,725 75,004 89, ,436 7,371 75,291 90, ,431 7,631 75,111 90, ,514 7,642 76,378 91, ,534 7,774 76,862 92, ,699 8,129 76,630 92, ,699 8,129 75,492 91, ,666 7,995 75,132 90, ,665 7,996 75,616 91, ,516 8,569 75,088 91, % 7.6% 2.6% 2.8% % 4.2% -0.1% 0.2% The forecast is adjusted for 20% of future EE program savings; estimated state-level models indicate that the 80% of projected savings are embedded in the unadjusted forecasts. C.5.6 Solar Load Forecast The energy and peak forecast incorporates the impact of expected photovoltaic adoption. Although relatively small in magnitude compared to the rest of Vermont, Stowe has experienced a steady increase in BTM solar load growth. This growth is expected to increase over time as solar system costs continues to decline. C Solar Saturation Model The primary factor driving solar system adoption is the favorable economics from the customers perspective. Simple payback is used to reflect customer economics. The simple payback reflects the length of time needed for a customer to recover the cost of installing a solar system - the shorter the payback, the higher the system adoption rate. The payback calculation is a function of the total installed cost, annual savings from reduced energy bills, and incentive payment for generated power. Payback is calculated for a typical 5 kw residential solar system. The resulting payback curve can be seen in Figure P age

52 Figure 50: Payback Curve Current system payback is roughly 8 years. By 2027, payback is expected to fall to 6 years. The most significant factor driving the payback trend downwards is system costs (expressed on an installed dollar per watt basis). System costs have been declining rapidly over the last five years. In 2010, the average residential solar system cost $6.21 per watt; by 2015, costs have dropped to $3.55 per watt. For the forecast, we assume that system costs continue to decline 10% annually through 2021, at which point costs continue to decline at 1% a year. The market saturation model relates the share of customers that have adopted solar systems to simple payback, payback squared, and payback cubed. A cubic model specification is chosen to impose an S- shaped adoption curve. Figure 51 shows the resulting customer share forecast. 42 P age

53 Figure 51: Solar Share Forecast As of December 2016, there were 59 solar customer accounts representing a 1.4% solar saturation rate. The solar accounts continue to grow. Given declining payback, we estimate that the solar saturation rates doubles within three years to 2.8% and by 2027 is close to 5%. The number of solar systems is derived by multiplying the saturation forecast to with the customer forecast. Table 9 shows the solar customer forecast and corresponding share. Table 9: Solar Custom and Share Forecast Year Customers Share % % % % % % % % % % % % % % % % % % % % % 43 P age

54 The installed solar capacity forecast is the product of the solar customer forecast and an average system size. The energy forecast is adjusted for incremental new solar capacity beginning in January Based on installed systems, the average size is 8.3 kw. The capacity forecast is depicted Figure 52. Figure 52: Solar Share Forecast The capacity forecast is translated into a monthly generation forecast by applying monthly solar load factors to the capacity forecast. The monthly load factors are derived from a typical PV load profile for Stowe, VT. The PV shape is from the National Renewable Energy Laboratory (NREL) and represents a typical meteorological year (TMY). Table 10 summarizes solar capacity and generation forecasts. Table 10: Solar Capacity and Generation Forecast 44 P age Year Capacity (MW) Generation (MWh) , , , , , , , , , , , , , , , , , , , ,612

55 Solar has no impact on the annual peaks due to the fact that Stowe is a winter peaking utility. C.5.7 Cold Climate Heat Pump Forecast VEIC has recently launched a program to promote the adoption of cold climate heat pumps; help meet the Vermont goals to reduce CO 2 emissions from heating end-uses. Heat pump saturation projections are based on VEIC s long-term heat pump sales forecast. We assume that an equivalent share of homes in Stowe will participate in the VEIC program. Program-driven heat pump loads are added to the Baseline Forecast. Since the baseline forecast already includes some heat pumps (based on EIA s 2017 New England projections), only heat pump usage over and above the built-in amount is added to the baseline forecast. Figure 53 summarizes heat pump penetration projections. Figure 53: Projected Heat Pump Penetration (%) Additional saturation is combined with unit energy consumption to come up with energy requirements for heating and cooling heat pump usage. The resulting energy requirements are summarized in Table P age

56 Table 11: Heat Pump Energy (MWh) Year Heating Cooling Total , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , % 362.1% 33.3% % 4.5% 6.2% Given lower starting unit energy consumption, heat pump cooling usage is significantly lower than heating. In addition, we assume that some cooling equipment is displaced by heat pumps, resulting in lower overall impact in the summer. On the heating side, heat pumps are expected to displace propane and wood heating, without significantly affecting existing electric heating. Figure 54 compares heating with cooling heat pump energy. 46 P age

57 Figure 54: Heat Pump Usage (MWh) Not surprisingly, the heat pump program is expected to have a significant positive impact on winter peak demand, but very little effect on the summer peak. D Portfolio Planning Approach and External Influences D.1 Regional Resource Portfolio and Marginal Supply The New England ISO meets a majority of both its base load and its peak load with natural gas fueled units. As seen below in Figure 55, natural gas is about 43% of the resource fuel type used to cover the New England demand. New England has rapidly surged its reliance on natural gas. ISO NE now gets about 50% of its power from gas, versus 10-15% a decade ago. From alone, gas increased its share of New England's power generation from 43% to 49%. Nearly 30 gas plants have been built in the region since This fact is only set to increase with the shuttering of nuclear and coal plants, such as Vermont Yankee in Vernon VT, Salem Harbor in Salem MA and soon the Brayton Point in Somerset MA. The 65-acre site of Salem Harbor will be the location of Footprint Power s 674 MW quick-start combined-cycle gas turbine project. These gas plants are able to provide efficient, low-emission power to New England while the renewable generating fleet is still relatively small P age

58 Figure 55: Capacity Supply Obligation by Fuel Type 11 35,000 Summer Generating Resource CSO by Fuel Type and Demand Resource CSO (MW) Capacity Supply Obligation (MW) 30,000 25,000 20,000 15,000 10,000 5, Demand Resources Wind Other Renewables Hydro Pumped Storage Coal Nuclear Oil Natural Gas D.2 Market Conditions D.2.1 Capacity Market The Forward Capacity Market (FCM) began on June 1, The FCM s goal is to acquire a sufficient amount of resources to meet the future demand. The FCM auctions take place three years in advance of actual settlement. The FCM auction designs clearing prices that will attract new generation and demand response assets as well as support the existing resources. The evolution with FCM has been within the zonal classifications. In the beginning, there was Rest of Pool and Maine. Beginning on June 1, 2016 there were Rest of Pool, Maine, Connecticut, and NEMA/Boston capacity zones. Currently, in the latest auction #11, there were Rest of Pool, Northern New England, and Southeast New England. Stowe has been in Rest of Pool until auction 11, where they now are under Northern New England. Historically, there has been price separation from zone to zone. The zones that were import constrained (NEMA) had CELT Report by ISO-NE 48 P age

59 larger clearing prices. Seen in Figure 56 are the clearing prices for the Rest of Pool Location that will affect Stowe s capacity charges. Figure 56: Rest of Pool Capacity Auction Clearing Prices Although the clearing prices increased in auction 8, Stowe did not see the price spike earlier as the NEMA location had. The zone location will also affect resource compensation, meaning where the unit resides will determine the compensation, which will not be a one for one on the load charge rates. This brings up the importance on self-supplying resources that are qualified to do so. In FCM 8, Stowe has self-supplied NYPA, Stony Brook, NextEra s Seabrook, McNeil, and VEPPI. This will guarantee a 1 to 1 offset of Stowe s load charges. Stowe s capacity portfolio is below in Figure 89. Stowe will assess the capacity market when researching different portfolio scenarios. Placement of generation and settlement of generation will come into play. Resources that directly offset peak usage for Stowe will be most attractive, because it will lower Stowe s obligation and give them the largest benefit. When forecasting the future capacity rates of the cost relations to portfolio scenarios for Stowe s IRP, the process included the analyzation of historical clearing prices and what factors drove those prices. Table 12 below shows how much capacity was needed and how much the clearing prices were affected by new Demand resources and New Generation. In the auctions where new resources were needed the most, the clearing prices were greater. Currently the system has sufficient resource to meet electric demand in and therefore it caused the lowest price settlement than the past three auctions. 49 P age

60 Table 12: ISO Auction Results FCA 7 through FCA Energy New England utilized a Monte Carlo simulation technique to estimate future capacity clearing prices for Rest of Pool capacity zone. Simulation results are found in Figure 57. More information regarding the forecast can be found in G.4 Capacity modeling. Appendix F contains the simulation output using historical year weighting. Figure 57: Forward Capacity Price Simulation Range P age

61 D.2.2 Energy Market The ISO determines the cost of the spot markets power prices. These prices drive the relative performance of Stowe s dispatchable resources within its portfolio, which ISO NE dispatches in either the day-ahead or the real-time spot markets for each operating day. If Stowe s dispatchable resources are bid at prices below LMP at their locations, they will be dispatched, and thus offset spot market energy purchases for Stowe. If they do not dispatch, then Stowe will purchase greater amounts of energy that day from the spot market. The benefit of the dispatchable resources is that if they are not economically dispatched, Stowe will buy those remaining MW s at a lower cost through LMP purchases. Within Stowe s scenario modeling, the Vermont load zone Locational Marginal Prices (LMP), where Stowe must purchase its load charges, are projected based on assumptions. These assumptions include natural gas and oil prices, as well as implied heat rates for the future. Calculations utilize regional delivered natural gas prices and implied heat rates due to the high frequency of natural gas fired resources setting marginal energy prices in New England. The link between energy prices in New England, specifically the Vermont Zone, is captured in Figure 58, which shows a.989 correlation between Vermont Zone 5x16 monthly average LMPs with monthly average northeast delivered natural gas prices. Figure 58: Vermont LMP Scatterplot Correlation to Northeast Natural Gas Prices VT 5X16 LMP / Avg NE Gas $160 $140 $120 $100 $80 $60 $40 $20 Scatterpot of Monthly Average Northeast Natural Gas Price ($/Mmbtu) vs. Monthly Average Vermont 5x16 LMP ($/MWh), $0 $0.00 $2.00 $4.00 $6.00 $8.00 $10.00 $12.00 $14.00 $16.00 Avg NE Gas / VT 5X16 LMP Correlation The aforementioned assumptions construct Energy New England s forward curve of power prices in New England. In the portfolio optimization model, this forward curve is set to a mean (expected outcome); then, by modeling the historical periodic movement of LMP at the Mass Hub and the Vermont nodal basis, the model produces 1000 s of simulations of LMP at the Vermont Load Zone. The simulations become a range of probabilistic outcomes (bucketed into percentiles) of simulated LMPs around the forward curve (the mean) to determine the probabilistic costs for open market purchases. Stowe s chosen portfolio scenario and future resource decisions will influence the nature of its interaction with 51 P age

62 the spot market. Stowe is able to reduce its spot market activities by procuring renewable resources and short and longer-term market purchases. Figure 59: ISO New England HUB PEAK FWD CURVE HISTORY Figure 60: Mass Hub ATC LMP, Monthly Simulated Range Jan 2018 to Dec P age

63 Figure 61: Vermont Zone ATC, Monthly simulated Range Jan 2018 to December 2037 Figure 62: Vermont to Mass Hub Basis, Monthly Simulated Range, 5x16 $2.00 $1.50 Vermont to MassHub Basis Differentials, 5x16 5% - 95% Mean $1.00 $0.50 $/MWh $0.00 -$0.50 -$1.00 -$1.50 -$2.00 -$2.50 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 53 P age

64 Figure 63: Vermont to Mass Hub Basis, Monthly Simulated Range, 7x8 $1.50 $1.00 Vermont to MassHub Basis Differentials, 7x8 5% - 95% Mean $0.50 $0.00 $/MWh -$0.50 -$1.00 -$1.50 -$2.00 -$2.50 -$3.00 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 64: Vermont to Mass Hub Basis, Monthly Simulated Range, 2x16 $3.00 $2.00 Vermont to MassHub Basis Differentials, 2x16 5% - 95% Mean $1.00 $/MWh $0.00 -$1.00 -$2.00 -$3.00 -$4.00 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 54 P age

65 D.2.3 Natural Gas in New England D Reliance on Natural Gas for Electricity Generation in the Northeast Over the last decade, the reliance on natural gas for electricity generation has grown significantly in the Northeast; nearly doubling its share of the region s total electricity generation. As of 2016, over 40% of regional electricity generation was reported to be fueled primarily by natural gas, and another 10% of generation list natural gas as its secondary fuel source. Nationally, 2016 became the first year on record in which natural gas surpassed coal as the primary fuel used for power generation in the United States (~34% of nations electricity). The predominant reason for natural gas surpassing coal as the fuel of choice for a majority of electricity generation regionally has been due to the development of increased access to low-cost natural gas (resulting from improvements in drilling technologies such as horizontal drilling and hydraulic fracturing) from the Marcellus Shale and other regional shale plays within the Appalachian Basin. Furthermore, environmental policies such as the Regional Greenhouse Gas Initiative as well as state-driven renewable portfolio standards have also contributed to the dwindling reliance on coal throughout the region. Figure 65: New England Resource Mix Percent of Total System Capacity by Fuel Type 13 D Market Fundamentals Influencing Spot and Forward Pricing of Natural Gas and Wholesale Electricity in New England With natural gas positioning itself as the popular fuel source for electricity generation in the Northeast, it has subsequently become the marginal fuel source for wholesale electricity pricing. When low-cost natural gas delivered from the Algonquin City Gate is readily available and not in exceptionally high demand, this relationship between wholesale electricity prices and relatively low-cost natural gas is favorable to wholesale electricity consumers. However, natural gas remains one of the most volatile P age

66 commodities in which its price can change frequently and materially. The market fundamentals of supply and demand, which are mostly driven by seasonal weather cycles and production/storage data, largely influence the spot and forward market pricing of natural gas. Further augmenting the volatility of natural gas prices in the Northeast are seasons that induce significant heating/cooling demand, during which the availability of natural gas is not a certainty. The preeminent issue in the Northeast, which most notably reared its head in the winter of 13/ 14 (due to the Polar Vortex), is that of natural gas pipeline capacity constraints and their ability to plague the region s wholesale energy markets. When pipeline constraints and/or periods of exceptionally high demand hit the region, the basis price (the cost of moving a commodity from point A to B - in New England s case, moving natural gas from Henry Hub to the Algonquin City-Gates) increases, thus causing wholesale electricity prices to increase as well. Historically, the Northeast has experienced its most notable pipeline capacity constraints in the winter. However, the last several winters in New England have brought relatively mild weather, and in turn, the price spikes in the Algonquin City-Gates basis have been lower than in previous years, as shown in Figure 66. Figure 66: Link between Regional Prices for Natural Gas and Wholesale Electricity P age

67 D Natural Gas in New England - Summary The Northeast saw an additional pipeline capacity built in 2016 and anticipates more expansion in 2017 and beyond. The question is whether the capacity buildout can keep pace with demand. Increasing demand has come in several forms, for example, heating demand in the Northeast continues to be more reliant on natural gas as Local Distribution Companies (LDCs) continue to place customers on the preferred fuel. Demand increase is also a result of new natural-gas-fired generators replacing retiring non-gas-fired generation. Natural gas prices have come down over the last several years, as seen in the flattening of the forward curve shown in Figure 68. These price decreases are the result of enhancements in exploration and production technologies, increased supply and resources (i.e. Marcellus Shale play), and warmer-thannormal temperatures experienced over the past several winters. According to the AEO2017, shown in Figure 67 below, production is expected to grow and as exploration and production technologies become increasingly more efficient, thus driving prices down, so too will growth in consumption and net exports (including liquid natural gas (LNG)). Figure 67: EIA AEO2017 Natural Gas Consumption and Production History/Projections 15 It is hard to predict the levels at which natural gas spot and forward market pricing will reside since pricing will remain sensitive to advancements in E&P technology, the availability of resources, and seasonal weather cycles. However, ENE took into serious consideration the aforementioned market forces and scenarios when creating natural gas simulations P age

68 Figure 68: Natural Gas Forward Curve History Figure 69: Natural Gas, Monthly Simulated Range Jan 2018 to Dec P age

69 Figure 70: Algonquin Citygates, Monthly Simulated Range Jan 2018 to Dec 2037 Figure 71: Algonquin to Henry Hub Basis, Monthly Simulated Range Jan 2018 to Dec 2037 D.2.4 Transmission Market The third largest piece of Stowe s New England Independent System Operator (ISO) costs is the Open Access Transmission Tariff (OATT). Within the transmission category are various ancillary charges, the largest of those being the Regional Network Service (RNS). RNS is the service over the Pool Transmission Facilities, which the ISO provides to transmission customers to serve their loads. 16 These are monthly charges based on Stowe s regional network load value at VELCO s peak. Every summer, the ISO publishes the presentation from the Reliability Committee/Transmission Committee of the Rates Working Group for the RNS PRT Forecast. These going forward rates include current transmission projects. Figure 72 shows the latest published forecast on August 9 and 10, 2016 ISO presentation. The P age

70 rates are steadily increasing, and therefore, Stowe s resource and efficiency become a larger importance. If Stowe can reduce consumption and do so at the critical coincident peak of VELCO, it could potentially save on its transmission charges to the ISO. Using the most recent forecasted rates and Stowe s three-year monthly peaks, ENE created a forecast of Stowe s transmission impact, shown in Table 13. With projected RNS costs totaling over 1 million a year, Stowe s desired portfolio will have a mix of load reduction resources and energy efficiency load savings. Figure 72: RNS Forecasted Rates Table 13: Stowe s RNS Forecast D.3 Assessment of Environmental Impact The New England Independent System Operator (ISO) is responsible for the reliable and economical operation of New England s electric power system. It also administers the region s wholesale electricity markets and manages the comprehensive planning of the regional power system. 17 Stowe can use the information with the ISO s Regional System Plan for its own planning purposes Regional System Plan ( 60 P age

71 D.3.1 Emerging Technologies VELCO creates a long-range transmission plan, which within the analysis is a discussion of how emerging technologies can affect the future load of the state. VELCO balances the impacts for efficiency and standard offer against electric vehicles and fuel switching. VELCO s 2015 Plan states it was updated with forecast adjusted for the effects of energy efficiency, demand response, the Standard Offer and net metering programs, and future load increases due to heat pumps and electric vehicles. 18 In Figure 73, VELCO assesses the MW impacts each technology can do to the state s load. Analyzing the trends, it can be reasonably assumed Stowe s load will increase or decrease at the same rate, if within Stowe, any technology enhancements include these same components. Figure 73: VT Load Forecast 19 D Distributed Generation (DG) The ISO New England website describes DG as Generation provided by relatively small installations directly connected to distribution facilities or retail customer facilities. A small (24 kilowatt) solar photovoltaic (PV) system installed by a retail customer is an example of distributed generation. 20 The ISO reached out for PV data within the Vermont utilities to help determine the DG affect and Burlington, GMP, Stowe, VEC, VPPSA, and WEC provided data as of December 31, Vermont s data totaled MW, and Stowe s contribution was.31 MW. In Figure 74 below are the survey results from all the New England States PV data along with the Vermont data Vermont Long-Range Transmission Plan 6/25/ P age

72 Figure 74: ISO-NE Total PV Installed Capacity Survey Results As of December 31, 2016, Stowe has 57 installed net-metered solar projects on residential accounts. It expects another 19 projects to be installed within the next two years. The total installed kw is , with an additional kw in the queue. Stowe s internal PV net-metered customers and the Standard Offer resources, which are DG amongst the VT utilities, both reduce Stowe s load. With the Standard Offer Program as of December 31, 2016, there has been MW s of PV projects accepted as well as MW s of Biomass, Farm and Landfill Methane, and Hydroelectric generation reducing the Vermont Utility load by each municipal s pro rata share each hour. Stowe s share percentage beginning in 11/1/2015 was %. Going forward, DG within both Vermont and within Stowe will help count towards Stowe s RES compliance obligation. D Electric Vehicle Penetration A majority of Stowe residents (74%) travel to work by mode of a car, with 8% of the population carpooling to work 22. Time traveled for the majority of residents is greater than 25 minutes to work (34%), which could lead one to believe that, in theory and without constraints, Stowe s residents could use the current plugin electric vehicle (EV) or plugin hybrid electric vehicle (PHEV) technology in order to reduce gas usage due to longer commutes to work P age

73 Figure 75: Stowe s Time Traveled to Work 40 to 44 minutes 5% 35 to 39 minutes 4% Stowe Time Travel To Work 60 to 89 minutes 2% 45 to 59 minutes 8% 30 to 34 minutes 8% 90 or more minutes 3% Less than 5 minutes 8% 5 to 9 minutes 20% 25 to 29 minutes 4% 10 to 14 minutes 15% 20 to 24 minutes 5% 15 to 19 minutes 18% Kelley Blue Book lists the many different electric car options 23, such as a 2017 Ford Focus, a Ford Fusion Energi, a Chevrolet Volt, a Nissan LEAF, and a Toyota Prius Prime. These all offer a gas-fee option from the standard gasoline-fueled vehicles. The 2017 Volt now gets 53 miles of battery power and travel up to 420 miles. With a car like this, one would have to expect a full recharge to take 4.5 hours with a 240- volt charger or 13 hours with a standard 120-volt plug. This vehicle charges at a rate of three kwh per mile; the Volt uses about 13.5 kwh per charge. Assumptions for this IRP include 1) the average speed of the Stowe driver is 35 MPH, 2) there are an average of 250 work travel days a year, and 3) the use of an average discharge rate of three miles per kwh. For a 2012 Nissan Leaf, its average rated efficiency of 99 MPGe translates to 34 kilowatt-hours per 100 miles. Just multiply that by your electric cost. 24 Table 14 below shows the impact of potential EV penetration. With 100% penetration, Stowe s average annual load may increase by 5,718 MWhs; whereas a low case of 25% penetration might add 1,429 MWhs. Currently, Stowe has seen yearly an increase of 1.45 to MWH s from 2014 through In year, 2017 Stowe expects to surpass the historical usage. Stowe monitors credit card zip codes, and therefore can attribute the EV increase to tourist usage P age

74 Table 14: Impact of Potential EV penetration in Stowe s work force Time Traveled to Work # % Miles using kwh round kwh used for EV Usage EV Usage EV Usage AVG 35 MPH trip the year 100% 50% 25% Less than 5 minutes 180 8% ,443 87,443 43,722 21,861 5 to 9 minutes % , , ,566 93, to 14 minutes % , , , , to 19 minutes % , , , , to 24 minutes 113 5% , , ,195 79, to 29 minutes 95 4% , , ,837 66, to 34 minutes 181 8% , , , , to 39 minutes 86 4% , , ,936 81, to 44 minutes 101 5% , , , , to 59 minutes 181 8% ,037,562 1,037, , , to 89 minutes 48 2% , , , , or more minutes 58 3% , , , , ,717,813 5,717,813 2,858,906 1,429,453 kwh/yr MW/hr The US Energy Information Administration (EIA) estimates car usage, of both conventional and alternative fuels, in a forecast that extends through the year When necessary, Stowe will add additional charging stations beyond the 10 various locations already located within municipality in an effort to promote and accommodate electric vehicles. EV will remain as high interest for Stowe because EV stations and usage will count towards Stowe s compliance of the Tier III Renewable Energy Standard. Figure 76: Annual Energy Outlook 2017 Conventional Cars vs. Alternative Fuel Cars D Energy storage Storage technology for electrical energy is growing in popularity. This technology offers users the ability to meet demand whenever needed and, more importantly, enables user to call upon it during peak P age

75 energy events. SED could use this energy to reduce their load during these events and help reduce peak load. Energy storage could not only save SED on load cost, but it could also reduce their transmission and capacity charges within the ISO. Table 15 below shows how a system using a one MW storage capability at the critical peak times can result in large yearly savings. See section D.2, Market Conditions, for the forecasted rates used to calculate a one MW reduction. ENE also forecasted the capacity reduction using an estimated 40% reserve adder. With these assumptions, SED would not only reduce its peak by the 1 MW, it would ultimately reduce it by the storage amount plus the ISO reserve adder, making storage a more appealing tool for cost savings. Table 15: Capacity and Transmission Savings Project Assumptions MW Commerical Operation Date Load Zone Est Reserve Margin RNS Ratio (10/12 months etc) 1 1/1/2017 VT 40% 83% The greatest benefit of energy storage is its ability to heighten the capacity factor of renewable generation, such as solar. These devices can also help make renewable energy, whose power output cannot be controlled by grid operators, smooth and dispatchable. 26 When solar production is low and a peak event is on the horizon, energy storage can supplement the solar output, and thereby, enable load reduction during the critical time P age Total ISO Year Capacity Savings , , , , , , , , , , , , , ,104 ISO RNS Savings Total Savings $ $ 81,727 $ 81,727 $ $ 95,113 $ 188,713 $ $ 100,936 $ 236,687 $ $ 104,781 $ 205,902 $ $ 107,794 $ 213,767 $ $ 110,894 $ 228,998 $ $ 114,082 $ 232,186 $ $ 117,362 $ 235,466 $ $ 120,737 $ 238,841 $ $ 124,208 $ 242,312 $ $ 127,780 $ 245,884 $ $ 129,285 $ 247,389 $ $ 129,285 $ 247,389 $ $ 129,285 $ 247,389 $ $ 129,285 $ 247,389 Grand Total $ 1,617,484 $ 1,722,556 $ 3,340,040

76 D Fuel Switching Unfortunately, owners of coal-fired power plants cannot easily switch fuels. A coal boiler is designed to burn coal, not natural gas. Even if a coal plant was modified to accept natural gas, the resultant fuel efficiency would be horrible and production costs would remain elevated. 27 Costs would be large for buildings that use oil as a heating source, if they wanted to fuel switch as well. They cannot just switch to natural gas if there are no pipelines to connect to the homes or businesses, etc. Referencing citydata.com s reporting as of March 29, 2017, complied data shows the max fuel source for Stowe is oil, kerosene at 58%. IF consumers switch to electric heating and cooling options due to economics this has the potential to increase Stowe s load. Although fuel switching among power generators is becoming more and more noticeable, due to either economics or Federal policies, individual home and business fuel switching is less common. Even if power generators desire to fuel switch, sometimes it is just not available. Unfortunately, owners of coal-fired power plants cannot easily switch fuels. A coal boiler is designed to burn coal, not natural gas. Even if a coal plant were modified to accept natural gas, the resultant fuel efficiency would be the resultant fuel efficiency would be horrible and production costs would remain elevated. 28 The same could be said for a home that is using oil as a heating source, they cannot just switch to natural gas if there are no pipelines to connect to their homes or businesses. Referencing to the city-data.com as of March 29, 2017 they have complied the max fuel source for Stowe is oil, kerosene at 58%. Depending on what consumers can switch to, either it economically sound it could increase Stowe s load if more opt for electric heating and cooling options. Figure 77: Stowe Commonly Used Heating Fuel D.3.2 Environmental attributes Environmental attributes are defined as characteristics of a program or project (such as particulate emissions, thermal discharge, waste discharge) that determine the type and extent of its short-term and P age

77 long-term impacts on its environment. 29 Projects qualify their attributes in different state classifications, based on year, fuel type, and emissions to name a few. These attributes are then marketable on a current platform called NEPOOL Generation Information System. Projects with qualifying attributes trade them to participants within the New England ISO, who apply them towards their renewable portfolio to meet compliance rules Beginning in 2017, Vermont has incorporated a renewable energy standard program, or RES, that requires utilities to meet various obligations of renewable attributes. The State goal is to obtain 90% of its energy from renewable sources by Additional RES information is found in the Renewable Energy Standard (RES) section G below. D.3.3 Assessment of Carbon Impacts Energy New England began the carbon assessment by reviewing the historical carbon intensity of SED s power mix from 2010 through 2016 and comparing it to the forecasts for the given years. ENE quantified SED s yearly non-emitting MWH totals by combining its NYPA allocations and REC retention and compared this total against their total yearly retail sales data, which includes snowmaking loads. ENE collected ISO-NE s final emission reports to incorporate the carbon impact of the regional system for each year. 31 Even though there are other components of GHG such as CH 4 and N 2 O, ENE chose to focus on CO 2 because in the U.S., CO 2 emissions represent more than 99 percent of the total CO 2 -equivalent GHG emissions from all commercial, industrial, and electricity generation combustion sourcesco 2 emission rates. 32 D Emission Calculation ENE chose to calculate SED s emission rates using ISO-NE s yearly ISO New England Electric Generator Air Emissions Report. Although the report is published on a lag, the methodology used to create the emission rate best suits SED s portfolio emission estimates. The ISO uses a total system emission rate calculation method that is based on the emissions by all the ISO New England generators during a calendar years worth of production. They use actual run time for on and off peak generation at the emission rate for each month. The emission rate uses 76% of the reported CO 2 from actual US EPA s Clean Air Market Division (CAMD) database, as well as the Regional Greenhouse Gas Initiative (RGGI). They also use EPA s egrid annual emission rates as a means of accounting for units for which this information is not available. All units that are dispatched are included in the emission rate calculation. The calculation is: Annual System Emission Rate (lb/mwh) = Total Annual Emissions (lb) all generators Total Annual Energy (MWh) all generators Using ISO data is important because not all generation is operational at the same or all of the time. The ISO tracks the air emissions from the NE system Grid while taking into consideration: P age

78 Forced and scheduled maintenance outages Fuel and emission allowance costs Imports and exports to and from NE region System energy consumption Water availability, etc. Incorporating these factors set ISO emissions methods apart from those of other data sources such as egrid. EPA s egrid states Emissions and emission rates in egrid represent emissions and rates at the point(s) of generation... they do not take into account any power purchases, imports, or exports of electricity into a specific state or any other grouping of plants, and they do not account for any transmission and distribution losses between the points of generation and the points of consumption. Also, egrid does not account for any pre-combustion emissions associated with the extraction, processing, and transportation of fuels and other materials used at the plants or any emissions associated with the construction of the plants. 33 D Emission Trends Figure 78 shows the fuel mix in the ISO New England control area in 2006 compared to We use 2015, as it is the most recent period for which the ISO regional emissions report is available. Coal has decreased the most over the period, dropping from 15% to 4%. Oil generation was cut in half from 4% to 2%. These changes resulted from a combination of tightening emission requirements, relatively higher operating and maintenance expenses of solid fuel and older thermal generating facilities compared to natural gas, and market forces, such as low natural gas prices in the past several years. The latter is due to the merchant generator boom that occurred in the late 1990 s and early 2000 s. This resulted in the building out of thousands of MW of high efficiency natural gas fired generating capacity. This moved natural gas to become the dominant marginal fuel in New England, where it now sets the marginal wholesale electricity price 60% of the time or more. This means that all generating technologies fortunes are affected by the price and availability of natural gas P age

79 Figure 78 ISO-NE System Energy Generation Percentage by Fuel Source Table 16 shows New England s average yearly CO 2 emission rates. Following the build out of merchant, gas fired generating capacity in the late 1990 s and early 2000 s, these rates continue to trend downward slightly as the underlying resource mix changes with less reliance on coal and oil generation. These rates were used to determine SED s supply emission profile for its open position and bilateral commodity energy contracts since these purchases are not tagged to a particular generator. Table 16: Regional Annual CO 2 Emissions in lb/mwh Annual System (NE) Co2 Emission lb/mwh SED s current carbon reduction power supply portfolio includes New York Power Authority, and all retained RECs such as Hydro Quebec, and Seabrook. Figure 79 shows that SED s total portfolio represents about 30,000 tons of CO 2 in 2010 and drops to about 15,000 tons of CO 2 in P age

80 Figure 79 SED CO 2 Emissions and Carbon Free Portfolio With the implementation of Renewable Energy Standards in 2017 SED will be increasing their nonemitting portfolio by retaining and retiring RECs. ENE used an average of the ISO-NE emission rates from 2013 through 2015, and held it consent throughout the IRP timeline. ENE also assumed SED would be 100% compliant with Tier I, II, and III. Figure 80 shows that these assumptions maintains SED s carbon footprint around 10,000 tons of CO 2. Achieving the RES targets reduces SED s carbon emissions by 30% from 2016 levels. This decrease directly follows the State goals set in August 2015 at the New England Governors and Eastern Canadian Premiers to set targets of decreasing carbon in the region by 35% to 45% from 1990 levels by In 2025, SEDs CO 2 emissions reduction totals 57%. This exceeds the target established by the Vermont Comprehensive Energy Plan of meeting 25% of energy needs using renewable sources by P age

81 Figure 80 SED CO 2 Emissions for RES E Data Models and Information E.1 RES Optimization Model In performing the RES portfolio integration and identifying an optimal REC position, Energy New England performed Monte Carlo simulations using commercial statistical software package to run optimization algorithms that identify the percentile of each outcome to SED s portfolio. The Energy New England Portfolio Simulation Model is a stochastic simulation based model that utilizes the Monte Carlo simulation technique to estimate future values of the input variables. This method allows a view into the probability distribution of outputs. The reason for the quantitative modeling is to determine the sensitivity of Stowe s portfolio cost to the change in market conditions and to identify an optimal combination of resources that will provide Stowe with the highest probability of having a competitive and low-cost resource portfolio. The model allows the use of inputs that will represent extreme cases as well as mild cases per resource. ENE reviewed and analyzed these extreme cases in the stress testing results. ENE used this model for both the Capacity Market and the RES modeling sections within the IRP. E.2 Portfolio Optimization Model- Lacima Lacima is a specialist provider of risk management, valuation and optimization software and services for multi commodity trading organizations. Lacima enhances analytical capabilities around risk analysis by providing a platform from which to run 1000 s of scenarios, as compared to traditional empirical modeling, which limits the number of variables that can be efficiently modified. From Lacima s simulations, probability ranges are drawn from the output data. 71 P age

82 Lacima uses Mean-Reversion Jump Diffusion to model the periodic movement of any historical time series. Using a random number generator, the regression formulas simulate 1000 s scenarios over a specified period, all with a forward projection acting as the target. The analytical tool utilized in Lacima was Gross Margin at Risk (GMaR). A GMaR analysis is the evaluation of the payoffs of all contracts in an energy portfolio, over a specific period of time, using simulations of all inputs in a model for that period, starting from forward expectations given by values in a forward curve or singular base case forecast (i.e. load). Then, the process aggregates the contract results (payoffs, cost, revenue, volumes) by time buckets and summarizes them across simulations into percentiles. For example, if 1,000 simulations are performed, but only the 5th percentile, the Mean, and the 95th percentile are to be reported, the report will contain three values for each output (one for each percentile), instead of 1,000 outputs. GMaR allows one to find the distribution of each energy NPV cost for each scenario. ENE used this model for SED s Evaluation of Portfolio Scenarios section. F Assessment of Resources F.1 Existing Energy Resources Stowe s portfolio consists of several existing resources, including long-term contracts and entitlements, which provide supplier, fuel source, and term diversity. See Table 17 for a brief description of each resource. Each resource includes capacity information, annual production, fuel, location, and termination date. Table 17: Stowe 2016 Resources 2016 Total KWh's by Resource Resource FCM Description MWH kwh's % of Total Resources Fuel Location Termination Niagara 0.49 Block 3,379 3,379, % Hydro Roseton 9/1/2025 St. Lawrence Block 70 70, % Hydro Roseton 4/30/2032 VEPPI PURPA 3,901 3,901, % Wood/Hydro VT Nodes Exp. Varies VEPPI-SPEED Load Reducer - 0.0% Behind meter Life of Unit HQ Contract ISO Bilateral 16,992 16,992, % Hydro HQ Highgate /31/2038 McNeil Wood Unit 9,175 9,175, % Wood Essex Life of Unit Stony 1A/1B/1C Dispatchable 1,764 1,764, % Natural Gas/Oil Stonybrk 115 Life of Unit Miller / Brown Bear Hydro 0.21 Run of River 1,941 1,941, % Hydro TopSham.Milr 5/31/2021 Saddleback Wind Wind 2,135 2,135, % Wind LUDDN_LN Exp Market Contracts-ENE ISO Bilateral % Seabrook Offtake ISO Bilateral 17,228 17,228, % Seabrook 555 Exp Market Contracts-ENE snow ISO Bilateral 10,781 10,780, % Mass hub 4/30/2014 ISO Energy Net Interchange 13,604 13,604, % Totals 80,973 80,972, % Beech Hill Solar Project Load Reducer , % Hydro Stowe Life of Unit 72 P age

83 Table 18: Stowe 2016 Current Resources Energy Cost 2016 Energy Cost by Resource Resource $/MWH Niagara $ 4.91 St. Lawrence $ 5.39 HQ PPA Contract $ Seabrook Offtake $ Beech Hill Solar Project $ - VEPPI $ McNeil $ Stony 1A/1B/1C $ Miller Hydro Purchase $ Saddleback Wind $ Bilateral Purchase- Mtn Jan-Apr $ Bilateral Purchase- Mtn Oct-Dec $ Figure 81, below, represents Stowe s 2016 resources by fuel type format. This pie chart shows that 13% of Stowe coverage was from market purchases. Figure 81: Energy Resources in 2016 In Stowe s resource forecast, found in Figure 82, ENE uses specific resource knowledge in order to estimate generation. In the long-term bilateral purchases, such as Brown Bear Hydro, the estimate is an average of historical run times. ENE conservatively estimated the HQ bilateral at the lower MW value of 73 P age

84 218. ENE expects McNeil to run between a 47-67% annualized capacity factor due to the NOx upgrade. The Saddleback Ridge Wind, VEPPI, and NYPA forecasts are each calculated using an average of historical generation, with VEPPI adjusted for expiring units. Seabrook offtake is a steady bilateral that makes up about 20% of Stowe s portfolio. ENE used a generic solar forecast when estimating the solar projection for Stowe. Once the Solar project has been in operation for a year, ENE will review actual data against the forecasted output and make necessary changes to the forecast then. This resource forecast results in very modest exposure to the spot market for Stowe, see Figure 82. Furthermore, that exposure is limited to the pricing of Stowe s Stony Brook entitlements, so long as the units are available for dispatch in high LMP times, Stowe will have a great coverage percent when it needs it most. Figure 82: Stowe s yearly projected resource distribution F.1.1 J.C. McNeil Generating Station The McNeil wood-fired generation station is located in Burlington, Vermont and has a maximum capability of 53 MW. Stowe s unit entitlement for energy, capacity, and ancillary products stems from a power purchase agreement with the Vermont Public Power Supply Authority for the life of the unit. Wood is the primary fuel source, with natural gas as an alternate. Plant startup utilizes either natural gas 74 P age

85 or fuel oil. With the NOx improvement, McNeil renewable credits are qualified in Connecticut Class I category. This has increased McNeil s run time as well as lower the overall cost of the unit. With the McNeil s bonds paid off in June 2015, fixed costs for the plant have decreased. The variable cost structure is due to ISO NE dispatching the unit regularly when the price of wood is competitive with natural gas. F.1.2 New York Power Authority (NYPA) The New York Power Authority provides preference hydroelectric power to New York s neighboring states. Two contracts provide this power to Vermont. The first is a one MW entitlement to the Saint Lawrence project in Massena, New York. The second is for a 14.3 MW entitlement in the Niagara project located in Niagara Falls, NY. The Saint Lawrence contract was renegotiated after its most recent end date of April 30, 2032 and the Niagara contract through September 1, 2025.The energy, capacity, and transmission payments required to deliver this entitlement to Vermont are at prices that are very competitive to the New England power markets. F.1.3 Vermont Electric Power Producers, Inc. (VEPPI) Stowe receives power from a group of independent power producer projects (IPPs) under Order of the Vermont PUC. VEPPI included a number of small hydroelectric facilities and one biomass. There were 19 VEPPI units, as of December 31, have expired, leaving 9 remaining. VEPPI assigns the energy generated by these facilities using a load ratio basis that compares Stowe s electric sales to other utilities in Vermont on an annual basis. The VEPPI contracts have varying maturities, with the last VEPPI contract scheduled to end in Stowe s current pro rata share of the VEPPI production is %, which started November 1, 2016 and will run through October 31, The prior percent, which ran from November 1, 2015 through October 31, 2016, was %. The VEPPI contracts prices have relatively high-energy rates and modest fixed costs. Note, the wood-fired Ryegate unit that was once within the VEPPI production expired on October 31, The utilities negotiated a 10 years contract for power. The contract now will terminate on November 1, F.1.4 Sustainably Priced Energy Enterprise Development SPEED and Standard Offer SPEED Standard Offer is a program established under Vermont Public Utility Commission Rule The program s goal is to achieve renewable energy and long-term, stably priced contacts. Vermont utilities will purchase power from the SPEED projects, which are behind the meter. Each utility will have their percent share (Stowe s share for November 1, 2015 through October 31, 2016 was % and decreased to % for November 1, 2016 through October 31, 2017) of load reduced by the output of the generation. Stowe receives a modest capacity credit and renewable energy credits for these resources. The cost paid to the SPEED projects are set based on the generation type. Section of Rule defines Speed Projects (those that qualify to serve a Vermont utility s SPEED requirement) as: 75 P age (SPEED projects are new electric generating projects that produce renewable energy. A new project means a project brought on-line after December 31, A SPEED project must use a

86 technology that relies on a resource that is being consumed at a harvest rate at or below its natural regeneration rate. Obvious examples of SPEED projects are utility scale wind farms, hydroelectric projects less than 200 MW, wood-to-energy projects, landfill gas-to-energy projects, etc. Combined Heat and Power (CHP) projects are SPEED projects if they meet certain efficiency standards or if they are fueled with a renewable resource. Projects that use a mix of fossil fuels and renewable fuels, such as a diesel generator that is partially fueled with bio-diesel, may qualify as SPEED in proportion to the amount of renewable fuel (in this case bio-diesel) that is used. The incremental energy produced by an expansion or modification of a pre-existing renewable energy project will be considered as a SPEED project. In May of 2009, as the SPEED Program progressed and implemented modifications, it changed into the Standard Offer program. This change began a feed-in-tariff to encourage the development of SPEED resources by making contracts long term and at fixed prices to qualified renewable energy projects. By May of 2012, the Vermont Energy Act of 2012 expanded the program to MW over a 10-year span with a new pricing mechanism for qualified projects. The 2017 RFP for the Standard Offer Program within the Public Utility Commission Docket No contained the avoided cost price caps. These prices are found below in Table 19. Each CAP is subject to a location and a fuel type. Figure 83 shows the current fuel source breakdown of the Standard Offer Projects. The complete list of projects is in Appendix C. Table 19: 2017 Avoided Cost Price CAPS for Standard Offer 76 P age

87 Figure 83: Energy Provided by Standard Offer Projects F.1.5 Stony Brook Combined Cycle Stowe is entitled to just under 6 MW of the Stony Brook combined cycle facility. This is a natural gas and #2 oil fired generation facility located in Ludlow, Massachusetts. Its total capacity is 350 MW in the winter. During the winter months, the unit is challenged in sourcing natural gas, so it generally will run on fuel oil during that time. That typically limits unit generation to non-winter months, concentrated around the summer New England peak load season. The build out of newer, high-efficiency combined cycle facilities in the past 10 years has served to limit Stony Brook s run time. Built as an intermediate unit in 1981, it now generally provides peaking duty. The unit heat rate is in the 8,500 BTU/KWH range, and the fact that the unit runs relatively little during the year is a testament to the impact that merchant generation has had in New England. While power prices have been falling due to natural gas storage increases, it has reduced the run time for peaking units, because locational marginal prices have been far below bid price. Stated in MMWEC s 2009 audited financials, the Stony Brook Intermediate Series A Bonds were paid in full as of July 1, This has helped reduce Stowe s fixed cost obligation for its entitlement. ENE did not include Stonybrook as a cost or coverage among SED s scenarios because of the low amount of output from the unit. In addition, the times Stonybrook is used to hedge peak hours where it can run in the money, can be a benefit for SED. F.1.6 NEW -Hydro Quebec Contract This contract began on November 1, 2012, for energy and renewable credits. The contract calls for 218 MW, with Stowe s portions vary during different periods as shown below in Table 20. The contract pricing will be flexible and competitive to the market price because it will follow the defined Energy 77 P age

88 Market index and the cost of power on the forward market. The pricing is based on market prices and inflation. The contract structure carries limits on year-to-year price fluctuations. Given the greater degree of market price volatility exhibited since the original Hydro Quebec contract was agreed, this pricing approach should be beneficial to Stowe as the contract will be limited to how out of market it might become for both Hydro Quebec and Stowe. This is an important contract quality in the current market environment and it reduces potential rate pressure to Stowe. In addition to the price flexibility, this will continue to provide very low carbon energy to Stowe, helping it maintain a market price based green energy procurement strategy. Table 20: Contract based on 218 MW Schedule Highgate has been working on increasing the transfer capability. Once the ISO-NE has approved the scheduled to finalize upgrade, the MW s will increase to 255 MW. With this adjustment, the contract will then shift to the second option of bilateral amounts. Table 21 below shows what will be the new portion for Stowe. Table 21: Contract based on 255 MW Start Date Final Delivery Date Stowe Entitlement (MW) Period 1 11/1/ /31/ Period 2 11/1/ /31/ Period 3 11/1/ /31/ Period 4 11/1/ /31/ Period 5 11/1/ /31/ Period 6 11/1/ /31/ Schedule Start Date Final Delivery Date Stowe Entitlement (MW) Period 1 11/1/ /31/ Period 2 11/1/ /31/ Period 3 11/1/ /31/ Period 4 11/1/ /31/ Period 5 11/1/ /31/ Period 6 11/1/ /31/ F.1.7 Brown Bear II Hydro (Old Miller Hydro Contract) Stowe signed a purchase power agreement for 2.613% of the Worumbo (Miller Hydro) Project. The contract states that Stowe will receive its percent of the Miller Hydro output per month. The contract price is for energy delivered to the Maine Zone and for capacity settled at the Maine location. The PPA terminated on May 1, P age

89 Brown Bear Hydro purchased the Miller Hydro Project and renegotiated a PPA, which began on June 1, Stowe has the same 2.613% share of the unit, but the PPA only includes energy going forward. This will terminate on May 31, Brown Bear Hydro is a run of river unit that, over the past 3 to 5 years, averaged an annual production of 90,000 MWH per year. This resource should equate to roughly 3% of Stowe s energy. F.1.8 Saddleback Ridge Wind Project Stowe purchased 2.172% of the Saddleback Wind Project, a 33 MW project with a 20-year PPA. This is roughly 3% of Stowe s load. The project allows Stowe to buy energy, capacity, and RECs. Saddleback Wind went full Commercial on September F.1.9 NextEra Seabrook offtake Beginning January 1, 2015 and going through December 31, 2034, Stowe will receive.16% (or a max of 2 MW) of around the clock from the NextEra Seabrook Resource. This contract provides Stowe with the same PPA percentage of capacity as well. F.1.10 Nebraska Valley Solar Farm Stowe built a 1 MW AC ground mounted solar electric generation project. Estimated output is approximately 1,568 MWh per year. This is about 1-2% of Stowe s annual energy requirement. The greatest benefit to Stowe from this project is the ability to use the renewable energy credits towards Tier 2 of the Renewable Energy Standard. Considered as distributed generation, or behind Stowe s meter, additional benefits include energy, capacity, and transmission. The project began operation in September P age

90 F.1.11 Snowmaking Procurement Energy Only Load Following In the case of Mount Mansfield, since the snowmaking load requirements are intermittent due to the nature of snowmaking demands, ENE s analysis showed a load-following energy product provides the best solution. It reduces Stowe s price risk for what can be a significant load during the winter months and at the same time, provides a vehicle to mitigate potential true-up payments that may be made from Stowe to the Mountain or vice versa. A load-follow energy product also has a greater potential to allow a lower overall cost as the energy markets continue to fall. 80 P age G Renewable Energy Standard (RES) In July 2015, using the 2011 Vermont Comprehensive Energy Plan, the State of Vermont established Act 56 (H. 40) in order to detail the State s goals and place direction on how utilities will reach these goals. The RES requires utilities to buy or retain renewable energy credits and energy transformation projects, and it set yearly percentage goals of retail sales to be covered by them. In lieu of renewable credits or transformation projects, a utility can meet its obligation by paying an alternative compliance payment at rates set by the State. The compliance rates adjust annually for inflation using CPI. There are three tiers to the RES program: Tier I : Meet a 75% by 2032 total renewable energy requirement (55% in 2017) o Any class of tradeable renewable attributes that are delivered in New England qualify o Approved Unit generations that will qualify towards compliance are McNeil, Hydro Quebec bilateral, and NYPA. Tier II: Meet 10% of sales with distributed generation in 2032 (1% in 2017) o New Vermont based unit that is 5 MWs or less or renewable generation Tier III: Municipal utilities must meet 10 2/3 % of sales with "energy transformation projects" in 2032 (2% in 2019) o Excess Tier II-qualifying distributed generation or project that reduces fossil fuel consumed by their customers and emission of greenhouse gases qualifies for compliance

91 Beginning in 2017, Vermont Statue Title 30, Chapter 89 (30 V.S.A ) will begin the RES for the Vermont distribution utilities. There are three tiers that SED will comply with either in renewable energy credits or compliance payments. Analyzing SED s current portfolio, ENE estimated the cost impact to Stowe s retail sales forecast, as shown below in Figure 84. Compliance of RES heavily influenced the selection of portfolio scenarios for the IRP. The analysis is on SED s load, not including the mountain s snowmaking load. The snowmaking load will be addressed as a pass through, whereas all obligations to RES will be billed back to the Mountain. Figure 84: Stowe s Potential RES Credit (Cost) Cash Flow G.1.1 Tier I Currently Stowe s resource portfolio contains about 55% renewable generation. This percentage comes from qualified generation that is either State approved, such as McNeil generation, HQ and the New York Power Authority contract for RES, or as generation, that has tradeable renewable energy credits. Figure 85 below shows SED s Tier I forecast. As the percentage requirement increases, the compliance gap increases. Using this forecast of current contracts, one can assess new projects. When looking forward to future purchases, Stowe can analyze the cost of retaining a project s renewable energy credits against possible future Compliance payment rates. 81 P age

92 Figure 85: Stowe s Tier I Forecast G.1.2 Tier II Currently, Stowe s distributed generation resource portfolio is about 2% renewable generation, mostly made up by Stowe s Nebraska Valley Solar project, which is 1 MW of distributed generation behind SED s transmission system. The renewable energy credits will qualify to Tier II compliance. The Commission shall allow a provider that has met the required amount of renewable energy in a given year, commencing with 2017, to retain tradeable renewable energy credits created or purchased in excess of that amount for application to the provider s required amount of renewable energy in one of the following three years. 36 With this three year banking policy, SED is able to maintain Tier II compliance until As the compliance percentage increases, SED s gap also escalates as shown in Figure 86. Analyzing this shortage is important when determining new distributed generation. Stowe will need to balance what the potential compliance payment charges may be against building or purchasing from a Tier II qualified project V.S.A. 8004(c) 82 P age

93 Figure 86: Stowe s Tier II Forecast G.1.3 Tier III Tier III is for energy transformation projects. This category is set to encourage projects that will help reduce fossil fuel usage and reduce greenhouse gas emissions. Currently, Stowe has ten Electric Vehicle charging stations, seven of which will qualify for Tier III compliance. The Public Utility Commission approved a conversion methodology developed by the Department of Public Service that utilities will use to equate fossil fuel reduction into MWHs of electric energy. The conversion uses the most recent year s approximate heat rate for electricity net generation from the total fossil fuels category as reported by the U.S. Energy Information Administration in its Monthly Energy Review. 37 Stowe can collaborate with Efficiency Vermont in sharing the savings with EV programs that are within SED s territory. Examples of these projects could include building weatherization; air source or geothermal heat pumps and high-efficiency heating systems; industrial process fuel efficiency improvements; increased use of biofuels; biomass heating systems; electric vehicles or related Infrastructure; and infrastructure for storage of renewable energy on the electric grid. 38 Stowe will be addressing energy efficiency programs as well to help decrease fossil fuel usage and comply with this RES requirement. We will begin with the base case as being open for purposes of filling it completely with the optimal scenario. G.1.4 Renewable Energy Credit Arbitrage The rules regarding Tier I qualification is that a provider, such as SED, may use renewable energy with environmental attributes attached or any class of tradeable renewable energy credits generated by any renewable energy plant whose energy is capable of delivery in New England. (Act 56 of 2015). Because 37 Docket No P age

94 of this rule, Stowe has the ability to create REC arbitrage. The meaning of arbitrage is the simultaneous purchase and sale of the same securities, commodities, or foreign exchange in different markets to profit from unequal prices. 39 Stowe can assess the market, and if its renewable energy credits are more valuable to sell in its qualified markets than buying other class RECs, SED will sell the RECs it owns and buy back another class or state REC that is available at lower prices. This ability can help SED buy down RES compliance payments in other Tiers where it may have a shortfall. G.1.5 Snow Making Potential RES Cost Because ENE did not model the snowmaking load into SED s energy or RES portfolio, ENE has modeled their impact as a separate entity. All snowmaking charges will be a pass through in their rate structure. Figure 87: Snowmaking Potential RES Cost Cash Flow G.2 RES modeling The Energy New England Portfolio Simulation Model, which is a stochastic simulation based model that utilizes the Monte Carlo simulation technique to estimate future values of the input variables, was used to asses SED s RES positions. The process then used the ranges of estimated values to identify the key drivers of the REC portfolio performance. The stochastic simulation approach to portfolio modeling provides a powerful, unbiased, and dynamic tool to measure the future performance of Stowe s REC portfolio under different conditions and identifies the factors to which the performance is most sensitive. A major benefit of using a simulation method is the ability to apply thousands of different scenario conditions across all of the model inputs, which ultimately produces a distribution of possible outcomes P age

95 G.2.1 Model Assumptions Table Model Inputs for RES Net Present Value G G G RES Tier Compliance rates use the CPI adder Existing REC Market uses the CPI adder Class I MA REC Market uses the MA compliance rate (using the CPI adder) and the REC market is a percentage of the compliance rate G.2.2 Model Outputs Appendix D contains the modeling report for the RES base case Net Present Value. Appendix E contains the modeling report for the RES snowmaking load Net Present Value. Figure 88: Net Present Value of RES for Stowe 85 P age

96 G.3 Existing Capacity Resources Stowe currently has around 57% of their Forward Capacity Market Obligation covered with capacity resources, as seen below in Figure 89. The open capacity position is charged at the net regional clearing price for the month. The rates are known through May 2020, as seen in Table 23, for Rest of Pool location, where Stowe will be charged for Capacity. The Auction took place in February The most recent Capacity auction took place on February 6, 2017 for FCM 11, which will begin on June 1, 2020 through May 31, The latest FCM 11 window for self-supply designation closed on October 31, The new auction for FCM 11 had three locations- Southeast New England SENE (encompassed NEMA, SEMA, and RI), Northern New England NENE (encompasses ME, VT and NH), and the Rest of Pool Zone (CT, and WMASS). In this auction, all locations cleared at the same price of $5.30/kW-month, which is the lowest price since FCA 7 back in This auction had no price separation in any zone except for New Brunswick. Southeast New England was modeled as constrained and Northern was modeled as export constrained. The auction outcome was a surplus of 1,760 MW above the forecasted installed capacity requirement. This was a large increase of surplus over FCA 10 s amount of 350 MW. Also in this auction, there were no new large generators, however, 640 MW of new demand response (DR) resource reductions cleared. DR consists of both dispatchable load reducers and energy efficiency reductions. Lastly, there was only a small amount of new wind (6 MW) and solar (5 MW) in this auction. The FCM 11 auction will also have a greater Pay for Performance incentive, which has a respectively stiffer penalty. The rate paid and rewarded is the same $2,000 MWH. Figure 89: Stowe s Capacity Forecast 86 P age

97 Table 23: Capacity-Clearing Prices Load Obligation Charge FCA Date RoP FCA $ FCA $ FCA $ FCA $ FCA $ FCA $ G.4 Capacity modeling The Energy New England Portfolio Simulation Model, which is a stochastic simulation based model that utilizes the Monte Carlo simulation technique to estimate future values of the input variables, was used to asses SED s Capacity positions. The process then uses the ranges of estimated values to identify the key drivers of the Capacity portfolio performance. The stochastic simulation approach to portfolio modeling provides a powerful, unbiased, and dynamic tool to measure the future performance of Stowe s Capacity portfolio under different conditions and identifies the factors to which the performance is most sensitive. A major benefit of using a simulation method is the ability to apply thousands of different scenario conditions across all of the model inputs, which ultimately produces a distribution of possible outcomes G.4.1 Model Assumptions The IRP s capacity forecast is shown in the Capacity Market section. Below are the $/kw-mo. forecasted charges that ENE s simulation exported for each IRP year. The historical data (June 2010 through May 2021) used includes clearing prices and payment rate percentages of the historical clearing price to the payment rates. ENE used a risk simulation table that weighted five scenarios based on the percentage of the past three year FCM clearing prices. Using FCA 9 through FCA 11 was the most ideal because they are the results from the most recent capacity parameters. 87 P age

98 Figure Model Prices for Capacity Forecast Category: Stochastic Spot FCM Price, $kw-mo Stochastic Spot FCM Price, $kw-mo / 5/31/2022 FCM RiskTriang(N11,N10,N9) Stochastic Spot FCM Price, $kw-mo / 5/31/2023 FCM RiskTriang(O11,O10,O9) Stochastic Spot FCM Price, $kw-mo / 5/31/2024 FCM RiskTriang(P11,P10,P9) Stochastic Spot FCM Price, $kw-mo / 5/31/2025 FCM RiskTriang(Q11,Q10,Q9) Stochastic Spot FCM Price, $kw-mo / 5/31/2026 FCM RiskTriang(R11,R10,R9) Stochastic Spot FCM Price, $kw-mo / 5/31/2027 FCM RiskTriang(S11,S10,S9) Stochastic Spot FCM Price, $kw-mo / 5/31/2028 FCM RiskTriang(T11,T10,T9) Stochastic Spot FCM Price, $kw-mo / 5/31/2029 FCM RiskTriang(U11,U10,U9) Stochastic Spot FCM Price, $kw-mo / 5/31/2030 FCM RiskTriang(V11,V10,V9) Stochastic Spot FCM Price, $kw-mo / 5/31/2031 FCM RiskTriang(W11,W10,W9) Stochastic Spot FCM Price, $kw-mo / 5/31/2032 FCM RiskTriang(X11,X10,X9) Stochastic Spot FCM Price, $kw-mo / 5/31/2033 FCM RiskTriang(Y11,Y10,Y9) Stochastic Spot FCM Price, $kw-mo / 5/31/2034 FCM RiskTriang(Z11,Z10,Z9) Stochastic Spot FCM Price, $kw-mo / 5/31/2035 FCM RiskTriang(AA11,AA10,AA9) Stochastic Spot FCM Price, $kw-mo / 5/31/2036 FCM RiskTriang(AB11,AB10,AB9) Stochastic Spot FCM Price, $kw-mo / 5/31/2037 FCM RiskTriang(AC11,AC10,AC9) G.5 Assessment of Alternative Resources When assessing different portfolio strategies, Stowe s focus is RES compliance. Therefore, the scenarios that were heavily focused on were to include either one or a combination of wind, solar, and hydro. We analyzed small Tier II compliant renewables against Tier I compliant to see which suited SED s portfolio the best. The IRP removed the interruptible snowmaking load from the scenarios due to the unique fact that the snowmaking tariff is a cost pass-through. We address the impact of RES to the mountain load but do not include it when making portfolio decisions for the net Stowe load. The tariff is designed to be a load-following obligation where all costs will flow through Stowe Mountain Resort. Stowe Electric will be handling the Mountain s compliance obligations for their portion of both the RES and load coverage pursuant to SED s renewable energy obligations. The goal for the resources was to get SED to the max RES compliance. Tier I was not the important driver to fill, both tier II and III, due to their large alternative compliance payments, were SED s priority. G.6 Smart Rates After its last cost of service study, completed in 2015, Stowe introduced a residential time of use rate with a critical peak pricing component. This rate was set to entice customers to become more energy 88 P age

99 efficient at costly times of the day while simultaneously communicating the dynamics of the wholesale electricity marketplace from which Stowe secures its power. By reducing usage, these customers would see reductions in their electric bills. This option became possible after SED implemented its fleet of AMI smart meters. In addition, by collecting 15-minute meter data, SED has the ability to view load patterns. The TOU is set seasonally from summer (June-September) in hour s noon to 8pm and winter (October- May) in hours 4pm to 8pm. Critical peak periods can be called on a day-ahead basis for the peak hours for up to 15 days during a given summer season. Participating customer would be contracted through a combination of , text, phone call to be made aware of the event. For the rates mentioned, see Table 24 below. These peak hours have the potential to help SED directly in saving on coincident peak load with the ISO-NE and VELCO. Table 24: SED TOU Rate Energy Charge H Assessment of the Transmission and Distribution System H.1 T & D System Evaluation Stowe Electric Department ( SED ) is a municipally owned electric utility providing service to 4,156 customers in the Town of Stowe, Vermont. The service territory spans 63 square miles. Some areas within the Town of Stowe are served by Vermont Electric Coop or Morrisville Water & Light. The primary make-up of the customer base is residential and small commercial with some larger vacation resorts as well as Stowe Mountain Resort (Mount Mansfield) making up the balance. 89 P age

100 Figure 91: Currently Served by SED SED s system consists of 8.1 miles of 34.5kV transmission line, 120 miles of overhead distribution and twenty-five (25) miles underground distribution lines. SED serves an average of twenty-eight (28) customers per mile of distribution line. SED owns three (3) substations and receives our primary service through a VELCO 115kV interconnection but can also receive service through a backup interconnection with GMP s 34.5kV subtransmission line if needed. H.1.1 Substations: SED decommissioned the last of its 4kV substations, Dewey Hill substation, in 2012 and now has three primary 12.47kV distribution substations that are fed from the 34.5kV transmission system and are able to tie and back-up each other supporting 75-80% of our customers. Location L.S. Voltage H.S. Voltage Transformer Sizes Wilkens 12, kV 2 x 5 MVA Houston 12, kV 2 x 7.5 MVA Lodge 12, kV 1 x 7.5 MVA TOTAL: 32.5 MVA 90 P age

101 H Wilkens Substation This substation was built in 1996 and consists of two 12.47kV distribution feeders (Circuit 1 and Circuit 2). Each circuit is regulated by the three 167kVA voltage regulators and each protected by a separate circuit recloser. The station transformer sizes are 2 x 5 MVA, which are fed underground from the VELCO/Stowe 34.5kV ring bus through a circuit switcher. The substation was designed low profile and all equipment is constructed in metal ground mounted equipment and is not located in the flood plain. It is in good condition and has good working clearances. 91 P age

102 H Houston Substation This substation was built in and consists of two 12.47kV distribution feeders (Circuit 5 & Circuit 6). Circuit 5 consists of three 333kVA voltage regulators and Circuit 6 has three 250kVA voltage regulators. Both are protected by circuit reclosers. The Circuit 5 voltage regulators were upgraded in May 2017 and the Circuit 6 voltage regulators are scheduled to be upgraded to 333kVA in the fall of Both station transformers for each circuit were upgraded in 2015 from 5MVA to 7.5MVA units pursuant to PUC Docket The substation is of wooden pole and cross arm construction, is in good condition, and has good working clearances. The pole structures for the distribution lines leaving the substation were re-built in April Both circuits originally shared common pole structures but are now separated and on individual poles. A new three-gang switch was also incorporated so that each circuit can be easily back fed through this switch and the buses isolated. A redundant station service transformer and transfer switch were installed so secondary equipment can remain energized during bus outages. This substation is not located in the flood plain (NOTE: See T & D System Evaluation, Statement 9). 92 P age

103 H Lodge Substation This substation has two 12.47kV distribution feeders (Circuit 7 and Circuit 8) which share three 333kVA voltage regulators and one 7.5MVA station transformer. Each feeder is protected by a circuit recloser. Lodge substation also contains a 34.5kV bus where the transmission line continues and feeds Stowe Mountain Resort. This 34.5kV circuit includes three 500kVA voltage regulators, a grounding transformer bank, and is protected by a circuit recloser. Two 3600kVAR capacitor banks are in place for the 34.5kV transmission line in the substation as well. The substation is wood pole and cross arm construction. The 34.5kV bus was re-built in 2003, it is in good condition, and has desired working clearances. The 12.47kV bus clearance will be studied in 2018 for a rebuild. This substation is not located in the flood plain. The Vermont Department of Public Service updated the Vermont Comprehensive Energy Plan ( CEP ) in The 2016 CEP included guidance for IRPs. Relevant to this section of Stowe s IRP, the CEP included specific questions that utilities are to use to evaluate their transmission and distribution systems. Stowe s assessment per those questions follows below. 1) The utility s power factor goal(s), the basis for the goals(s), the current power factor of the system, how the utility measures power factor, and any plans for power factor correction. SED currently does not have the equipment to accurately measure and monitor power factor within our system. A system study was performed in 1997 by PLM and the system power factor was estimated to be 98%. PLM recommended the installation of additional capacitor banks on the system in order to improve our power factor goal to 99.8%, which we later completed. 93 P age

104 Table 25: SED Capacitor Banks, Sizes, and Locations Cap Bank # Location Pole # Switched Circuit Size Voltage C1 Lodge Sub Sub Y 34.5kV Line 3600 kvar 34.5KV C2 Lodge Sub Sub Y 34.5kV Line 3600 kvar 34.5KV Moscow Rd 2-87 Y kvar 12470V 5-42-C1 Mountain Rd 5-42 Y kvar 12470V 1N-ED-11-C1 Weeks Hill Rd 1N-ED-11 Y kvar 4160v 5-D2-C1 Cottage Club Rd 5-D2 N kvar 12470V 2) Distribution circuit configuration, phase balancing, voltage upgrades where appropriate, and opportunities for backup. Each of SED s six 12.47kV feeders have been reconfigured to back up other feeders with bus ties at the substations or tie points on the lines. Much of the main feeder lines have been re-conductored in the past 10 years during 4kV conversions to the system and new transformers were also installed at which point phase balancing was done during those upgrades. Loads are recorded on a monthly basis at the substation reclosers and reviewed for phase balancing. 3) Sub transmission and distribution system protection practices and philosophies. Protection for the 34.5kV transmission line is provided at the breakers on the VELCO/Stowe 34.5kV substation ring bus and are maintained and monitored by VELCO. SED has recloser protection on all of our distribution circuits. Recloser settings are as follows: Table 26: SED Recloser settings Circuit Control Form Min Trip TCC1 TCC2 Operations to Lockout TCC Sequence 1 4C Phase Ground ALT Phase 120 ALT Ground 60 2 F6 Phase Ground C Phase Ground ALT Phase 110 ALT Ground C Phase Ground ALT Phase 110 ALT Ground A Phase 480 A K Ground F6 Phase Ground WCAX- URD F6 Phase Ground P age

105 SED uses fusing on all main lines, side taps, and transformers to minimize the number of customers affected by system faults. Arresters are used to protect all aerial transformers, capacitors, and primary underground equipment. All of the above equipment is saved in the VELCO GeoDE GIS system. Stowe s 2011 IRP states that the last protection coordination study was conducted in SED has plans to conduct an engineering study the results of which will be incorporated into the next IRP. This study will include an evaluation of protection settings on SED s distribution system 4) The utilities planned or existing smart grid initiatives such as advanced metering infrastructure, SCADA, or distribution automation. Please See Section Grid Modernization 5) Re-conductor lines with lower loss conductors. SED s main feeder lines have been re-conductored during the 4kV conversions to the system over the past 10 years. Standard conductor sizes are 336 AAC for three phase main lines, 1/0 AAAC or ACSR for all branched side taps. SED uses 1/0 URD jacketed primary cable with full neutral placed in conduit for all underground-branched side taps. 6) Replacement of conventional transformers with higher efficiency transformers. It has been established SED practice to purchase rebuilt transformers from T & R Electric Supply Co. out of South Dakota at a fraction of the cost of new transformers. The cost of new units is at least double the cost of re-builts, carries a shorter warranty period, and is not readily available. This information coupled with the fact that SED is an at-cost provider and is not allowed a rate of return like investorowned utilities supports our judgement to continue purchasing rebuilt transformers. Though SED plans to continue purchasing re-builts, for the immediate future, we also plan to conduct an updated costbenefit analysis and will incorporate the results into the next IRP. 7) The utility s distribution voltage settings (on a 120V base) and whether the utility employs, or plans to employ, conservation voltage regulation or volt/var optimization. All circuits are bus regulated with a set point of 122V-124V, +/-1.0V-1.5V volts at the substation and our AMI meters monitor customer voltage and provide alarms when voltage does not meet SED requirements. Capacitor banks have been installed on our system to provide volt/var support where needed. As stated above in response to Item 3, SED plans to conduct an engineering study to evaluate any potential advantages from implementing conservation voltage regulation. The study will assess whether such benefits can be gained through deployment of CVR in a wide range of scenarios, such as system-wide implementation or feeder-specific applications. The results of the study will be incorporated into the next IRP 8) Implementation of a distribution transformer load management (DTLM) or similar program. 95 P age

106 SED does not have a DTLM program at this time. Currently SED uses traditional transformer sizing methods and uses Load Data Loggers to monitor customer loading where necessary. 9) A list of the location of all substations that fall within the 100 and 500 year flood plains, and a plan for protection or relocation of these facilities. None of SED s three substations are located in the flood plain. During an upgrade of the Houston Substation station transformers in 2014 and 2015 per Docket 8466, it was determined by the VT DEC Watershed Management Division after a survey of the facility that the Houston Substation elevation was above the 100 and 500 year flood elevations. No additional flood proofing measures were required by the VT DEC Watershed Management Division at that time, however a recommendation was made for SED to work with the Watershed Management Division to take protective steps if SED decides to rebuild or relocate this substation in the future. See Docket No SED will work closely with the Watershed Management Division should the utility decide to rebuild or relocate this substation in the future. 10) A discussion of whether the utility has Damage Prevention Program (DPP), or plans to develop and implement a DPP, if none exists. SED is in the beginning phases of developing its own DPP. It is currently analyzing DPPs from other Vermont utilities to evaluate how to integrate its existing Dig Safe procedures into a formal DPP. SED, as a member utility in the Dig Safe program, requires customers and contractors to contact Dig Safe for all underground construction activity. All SED facilities are located and marked by SED personnel. SED utilizes its own underground locating equipment and our line maintainers are fully trained on its use. Additionally, this equipment has GPS capability and is used to capture and store GPS coordinates of the underground system during the locating of cables. The coordinates are then uploaded into a GIS mapping system for future reference. 11) The location criteria and extent of the use of animal guards. SED policy is to install animal guards on all new construction and line rebuilds. Animal guards are also installed on existing services whenever maintenance is done on these services. SED evaluates outages on a regular basis to determine if animal guards in those areas would be beneficial. 12) The location criteria and extent of use of fault indicators, or the plans to install fault indicators, or a discussion as to why fault indicators are not applicable to the specific system. SED requires all primary underground developments with more than three pad mount transformers, particularly long underground or loop feed systems, to install fault indicators at each transformer or elbow cabinet. Fault indicators have been installed on the overhead transmission line in strategic locations, such as road crossings and before underground risers. No fault indicators are currently installed on overhead distribution lines. SED s overhead distribution lines are relatively small and well protected by reclosers and fusing and faults can usually be easily located. SED evaluates outages on a regular basis to determine if fault indicators in those areas would be beneficial. 96 P age

107 13) A Pole inspection program, the plans to implement a pole inspection program, or a discussion as to why a pole inspection program is not appropriate to the specific utility. SED has an informal pole inspection program. Many of the distribution poles have been replaced during voltage conversion and re-conductoring projects in the last 10 years leaving just a few outlying areas where maintenance is being conducted. SED line maintainers patrol the lines and conduct surveys on a weekly basis to determine which poles may need to be replaced and/or may need work. This information is added to GeoDE GIS system provided by VELCO. SED is also using this tool to keep track of pole replacements. As line maintainers consistently work to supplement the data currently contained in the system, SED feels that the database effectively serves the utility s needs to keep track of its poles and in time will help to identify those areas of SED s system which may command specific attention. 14) The impact of distributed generation on system stability. SED s total installed net metering capacity as of December 31, 2016 was 490.4kW with an additional kW currently permitted or with a filed application. Additionally SED installed and commissioned a 1MWac solar array at our Nebraska Valley facility in August 2016, which is interconnected on SED s Circuit 2. A system impact study was conducted for this and also for a proposed 496kW unit on the same circuit. It was determined that any additional distributed generation on this particular circuit would result in reverse flow at our Wilkens substation. This will require replacement of the substation recloser to handle the reverse flow and other existing equipment will need to be analyzed as well. Figure 92: SED s Nebraska Valley Solar Farm (1MWAC) 97 P age

108 H.2 T & D Equipment Selection and Utilization SED solicits quotations from three sources before making purchases for all major equipment. Purchase decisions are made on price and reliability. SED also evaluates the functionality and suitability of equipment before a decision is made to purchase. SED will continue to purchase rebuilt transformers from T & R Electric Supply Co. out of South Dakota at a fraction of the cost of new transformers. SED will also conduct a new cost-benefit analysis to ensure that this practice remains consistent with least-cost principles. SED maintains a substantial inventory of distribution transformer sizes, both pole and pad mounted, on hand for new installations and replacements. An inventory of critical units such as step downs and voltage regulators is also available for emergency replacements. Inventory is reviewed periodically to keep counts at suitable levels. Currently SED uses traditional transformer sizing methods based on the size of the home. We also request anticipated load information with applications for new service and seek assistance from outside engineers when the anticipated load is larger than a typical service. SED will also use Load Data Loggers to monitor customer loading where necessary. As SED develops Outage Management and GIS Systems and integrates them with our AMI and Customer Service systems, actual transformer load data will be monitored and used for the proper sizing of transformers. H.3 Implementation of T & D Efficiency Improvements SED continues to see improvement in efficiency on our distribution system. Line losses have decreased since 2008: Figure 93: SED Line Losses SED s main feeder lines have been re-conductored over the past 10 years during 4kV conversions to the system. Standard conductor sizes are 336 AAC for three phase main lines, 1/0 AAAC or ACSR for all 98 P age

109 branched side taps and 1/0 URD jacketed primary cable with full neutral and in conduit for all underground branched side taps. Capacitor banks have also been installed specified areas to maintain voltages. H.4 Maintenance of T & D System Efficiency SED continues to convert the few remaining sections of our distribution system that are still operating at 4kV to 12.47kV Poles, equipment and wires are evaluated before the start of a project to determine if full, partial, or no replacement is required. Typically SED will replace conductor types and sizes that do not conform to our current standards, with a particular focus on aging conductors that are reaching the end of their useful life, such as Copperweld. Below is the estimated remaining length of 4 kv in Stowe s system which we anticipate will be replaced in the next 5 years through regular maintenance i. Weeks Hill Rd (1.1 miles) ii. Percy Hill Rd (0.6 miles) iii. West Hill Rd (2.1 miles) iv. Stowe Hollow Rd (1.0 miles) v. Upper Hollow Rd (1.6 miles) Substation inspections are completed on a monthly basis and equipment problems are documented and addressed as they occur. Oil samples are drawn from substation transformers and reclosers on an annual basis and analyzed. A system wide infrared study is conducted on an annual basis as well. Results are analyzed and questionable equipment is repaired or replaced where needed. SED has scheduled to replace or install fifteen new three gang switches in strategic locations over the next three years. Located at circuit tie points and heavy concentrated load areas, they will be used for sectionalizing and isolating lines during outages and maintenance operations. The new switches will have the capability of having motorized operators installed in the future for remote monitoring. H.5 Other T & D Improvements H.5.1 Bulk Transmission The new VELCO 115kV line and new VELCO/Stowe substation was completed in December 2009 and energized in January The new line provides a stronger feed into SED s system and greatly improves reliability to the Stowe Mountain Resort. Before the 115kV line was installed, SED frequently had to have the Mountain limit snowmaking to stabilize the system but has not had to do so since. No further upgrades are being considered by Stowe Electric at this time. H.5.2 Sub-Transmission SED s 34.5kV transmission line is fed from the VELCO 34.5kV ring bus in the new Stowe/VELCO substation. Two existing 34.5kV feeds remain on the 34.5kV ring bus as back up to the 115kV feed. SED is working with a line engineer to design and rebuild a 2,500-foot section of the line. The work would include pole replacements and re-sagging the existing lines. 99 P age

110 H.5.3 Distribution SED is continuing to make upgrades on the two major circuits fed from our Houston Substation. In 2015/16 the two 5MVA stations transformers were replaced and upgrade to two 7.5MVA units. Six 333KVA voltage regulators were purchased in August 2016 to replace the existing 250KVA units. SED line-maintenance staff has rebuilt the pole structures that deliver power from the substation. The two circuits originally shared single pole structures, in March 2017, line work was completed, and Circuits 5 and 6 were separated on individual poles. SED also added switching flexibility with the installation of a new switch between both feeders to further enhance the load serving capabilities at this substation by creating a new tie point. SED is taking a proactive approach for handling direct burial primary cable failures. SED digitally mapped approximately 85% of our underground system in 2016 and has identified the age of the cables in those areas. SED has purchased new equipment in anticipation of potential failures of old underground primary. We have purchased new underground locating equipment and have trained our line maintainers on fault locating to help decrease our restoration times in such an event. We have also purchased a CAT305 mini excavator and trailer. This will not only help reduce our response time as we will no longer need to rent such equipment, but it also means that SED has the in-house ability to replace larger sections of this aging underground when needed. We will continue with our practice of installing new cabling in conduit for added protection and ease of replacement in the future. SED installed close to 9,000 feet of new underground primary cable and PVC conduit in four locations of our system in 2016 thus eliminating outages from those areas caused by failed primary cabling. Much of SED s distribution lines are located along the roadside and much of those that are cross-country cannot be relocated because of the remoteness. Currently SED has plans to relocate two cross-country sections to roadside existing poles and right of ways. SED works with the telephone and cable TV utilities on utilizing and maintaining their existing infrastructure. SED is a member utility with NJUNS and utilizes the online portal to coordinate pole transfers with telephone and cable utilities. SED is currently reviewing our in-house procedures to better document pole replacements in the field and to better coordinate transfer work in NJUNS. SED is also in negotiations with Fairpoint to purchase the poles they own in Stowe, approximately a third of the total. This will give SED better controls on coordinating pole transfers and removing existing double sets. H.5.4 Grid Modernization In October 2016, SED purchased the old Moscow Mills property, which will become the future home for SED s new headquarters and operations center. This old industrial site was home to a machine shop, a saw mill, several outbuildings and a residence and site storage for construction materials. During the first half of 2017, SED has been developing site and building plans. Demolition of two buildings will begin in September 2017 and construction of the new buildings will begin in November SED feels that our efforts to clean up and re-develop this dilapidated industrial site will significantly improve the area. SED will build a new state of the art complex using energy efficient technologies. The buildings will be designed to enhance the neighborhood. SED will work closely with Efficiency Vermont on lighting, 100 P age

111 heating and cooling for the buildings. A rooftop photovoltaic system will be installed to provide our energy needs as well as a Level 1 EV charging station for use by our customers and employees. Also located on the property, an old grist mill built in 1899 that recently housed a run of the river hydroelectric generator. The grandfathered hydroelectric facility was flooded during hurricane Irene and removed. SED s goal is to remodel the mill and install a new turbine and generator to produce electricity again and also to clean up the shore lines on both sides for public use. SED completed the installation of AMI meters and AMI and MDM systems in The AMI meters communicate over a mesh RF network back to collectors placed in strategic locations throughout our system. The MDM and customer billing systems were replaced in the first quarter of A new IP based phone system was installed in June SED currently uses VELCO s GeoDE geographical information system. SED line crews conduct surveys on sections of SED s transmission and distribution system, collect equipment information, pole and conductor data and GPS coordinates, and upload it into the GIS. SED plans to deploy mobile devices for the line maintainers to refer to when in the field so it can be used with a new outage management system. Fiber optic cable has been installed from our Wilkens substation along our 34.5kV transmission line with terminations at Houston and Lodge substations, continues to the top of the mountain, and terminates in the WCAX building. SED plans to install fiber from Wilkins to our new operations and office complex. SED will also start replacing the reclosers in 2018 at each substation with units utilizing digital relaying that will provide feeder status, voltages, load data, and power factor back to the new office complex. Currently SED manually updates VTOutages.org. SED plans to add an outage management system (OMS) capable of displaying real time system status, voltage, and load information. We recognize the value of implementing a system which will be able to provide such data but we do not have a set timeline for implementation. It also presents a significant opportunity to improve customer service as it will also include a web based information portal for use by our customers as well as provide automatic updates to VTOutages.org. H.6 Vegetation Management Plan SED continues to see positive results from our vegetation management program. Management efforts since Stowe s last IRP along with a modest increase in funding during that period have allowed us to develop our tree trimming program nearly down to a five-year cycle for both transmission and distribution and improving our annual results. This exceeds the stated goal from the last IRP which was to achieve a 5-year tree-trimming cycle for subtransmission and a 7-year cycle for our distribution system. Lands within the SED right-of-ways either are owned by private individuals or are by the State of Vermont. A perpetual easement is the most common type of utility right-of-way document and most easements at SED are 50 feet on aerial distribution and 100 feet on aerial transmission. There are also a 101 P age

112 large number of distribution lines which are located near roadways. These varying conditions, as well as the considerable efforts of the last few years to achieve a consistent vegetation management cycle, means that some areas of SED s network have had higher tree-trimming costs, as is reflected in the table below. This also initiates that the utility s anticipated future tree trimming budgets may be able to cover more miles of SED s distribution system as more fully-grown areas, and therefore those areas with a higher cost per mile, have already been addressed. A variety of vegetation is present along SED right-of-ways ranging from open agricultural land, lowgrowing shrubs and brush, as well as full grown trees. The most common forest types in wooded areas along SED right-of-ways are northern hardwoods, spruce-fir, eastern hemlock, yellow birch and white pine. SED has mapped our entire system by year to help coordinate pre-season line surveys with tree trimming assignments for the year: Figure 94: SED Tree Trimming, Last 5 Years: (Note: Underground facilities in black) Tree trimming activities are conducted by qualified line clearing contractors who are bound by contract to adhere to the American National Standard Institutes (ANSI) Standard A300. SED staff conducts routine maintenance inspections and contract administration to ensure that maintenance activities are conducted in accordance with established standards. The contract work is augmented by SED line maintainers cutting danger trees and some trim work during the slower winter months. Line crews continually monitor our overhead lines for danger trees. Danger trees may also be identified by 102 P age

113 contracted tree crews or brought to our attention by customers and landowners. In the event that there are no contracted tree crews currently working in SED s territory that can be redirected to evaluate and deal with a danger tree, line maintainers regularly cut danger trees as they are identified. Note: Actual Expense is for all poles in the system. SED contributes 60% of the entire cost. SED does not apply herbicides to any right-of-ways, but does use them through a licensed applicator within the fences of each of our three substations. H.7 Studies and Planning Following are SED distribution system future upgrades: Job Location Estimated Priority Replace VRs at Houston Sub Houston Sub $60, < 1 year Replace Cap Bank Controller at Lodge Sub Lodge Sub $5, < 1 year Install new tie conductor at Stoney Brook Stoney Brook $20, years Install new tie point for Circuits 6 & 7 Mountain Rd $100, years New reclosers and controllers at substations All Subs $200, years New transformer (Circuit 1) at Wilkens Sub Wilkens Sub $150, years Notchbrook URD replacement - Phase 3 Notchbrook $3, < 1 year Rebuild Lodge Substation - 12KV Structure Lodge Sub $50, years Re-configure (Circuit 6B) 1Ø (Maintenance) Houston Sub $0.00 < 1 year Complete 1Ø Aerial Primary Brook Rd (Circuit 6B) Brook Rd $15, < 1 year Luce Hill move 3Ø cross country section out to road Luce Hill Rd. $15, < 1 year River Rd. - replace poles and crossarms (Maintenance) River Rd. $ years Tamarac Rd. - replace poles and crossarms (Maintenance) Tamarac Rd. $ years Barrows Rd. - replace poles and crossarms (Maintenance) Barrows Rd. $ years Barrows Rd H.S. Riser Pole - replace and re-configure High School $ years Convert Sanborn Rd. to 3Ø up to Robinson Springs Sanborn Rd $15, years Move main line out to Mountain Rd. (across from Rusty Nail) Mountain Rd $15, years Weeks Hill 3Ø to Percy Farm Conversion - Phase 1 Weeks Hill $20, years Weeks Hill 3Ø to Percy Farm Rebuild - Phase 2 Weeks Hill $30, years Birch Hill - Winterbird Rd. URD replacement Birch Hill Rd. $10, years H.8 Emergency Preparedness and Response Customers have 24/7 access to SED for all emergencies by calling our main phone number. After hours, calls are handled by SED s answering service, which has direct phone contact with on-call linemen, the Director of Operations and General Manager for a response. The on-call lineman will call in additional SED personnel if needed depending on severity of the situation. Customers with significant loading also have direct 24/7 cell contact with the General Manager and the Director of Operations. 103 P age

114 In the event SED crews require additional outside help, SED is a member of the Northeast Public Power Association s mutual aid program and has access to other local Municipal utility crews. Further help is available from Green Mountain Power and Vermont Electric Coop. For planned outages, SED uses several forms of communications to inform customers in advance: phone calls, s, and doorknockers. Information may also be posted quickly on Front Porch Forum, SED s website, Twitter, and Facebook page, as well as in the local newspaper and radio stations when time permits. Currently SED manually updates VTOutages.org. SED plans to add an outage management system (OMS) capable of displaying real time system status, voltage, and load information. We recognize the value of implementing a system which will be able to provide such data but we do not have a set timeline for implementation. It also presents a significant opportunity to improve customer service as it will include a web based information portal for use by our customers as well as provide automatic updates to VTOutages.org. SED participates in the Fall Vermont Joint Utilities/State Agencies Emergency Prep Program and the Lamoille County Emergency Response Table Top Exercise. SED also participates in the VELCO statewide emergency preparation conference calls when scheduled. H.9 Reliably SED serves over 92% of residents and 100% of businesses located within Stowe, Vermont and currently serves 4,156 customer meters, net of voltage and current meters, station service meters, and any meters at a retail customer s premises beyond the customer s first meter. SED s SQRP reliability goals are: SAFI 0.9 and CAIDI 3.3 SED s system did not experience any major storms as defined in the Stowe Service Quality and Reliability Plan. SED s system performance ratings for 2016 are as follows: All outages included: SAIFI 0.2 and CAIDI 2.3 These numbers represent the best annual system performance ratings that SED has achieved in recent memory and are a product of the concerted multi-year efforts of line crews and office staff to improve the reliability of Stowe s network to provide acceptable electrical service to our customers. These indices represent an improvement of almost 10% over 2015 when SED experienced 71 outage events. Stowe s 2016 indices are a continuation of a general upward trend in performance. SAIFI and CAIDI ratings were 0.7 and 0.8 in 2015, 1.1 and 1.3 in 2014, and 1.1 and 2.9 in 2013, excluding major storms. 104 P age

115 H.10 Assessment of Outage Events and Trends in 2016 Overall, SED s system experienced 65 outage events on our system and 1,762 customer hours out in This shows a steady improvement over previous years. Our system experienced 71 outages and 2,338 customer hours out in 2015, and 118 outages and 5,782 customer hours out in SED attributes a considerable portion of this improvement to the consistent focus on vegetation management, as trees have consistently been a major contributing factor on the outage indexes. 105 P age

116 Trees caused 15 outages and 565 customer hours out in This amounted to a decrease in both indices: a 50% decrease in the number of outages from 2015 and a roughly, 13% decrease in customer hours out. Two of the tree outages were significant contributing events to SED s overall CAIDI increase. These outages, which resulted in 313 and 172 customer hours out, occurred when large trees fell from outside our right-of-way late in the evening and in the early morning hours. One caused considerable damage to conductors and required substantial repair work. Company initiated outages was tied as the first highest category in 2016 with 15 outages as well and 426 customer hours out, slightly less than 2015 at 16 outages but a notable improvement in customer hours out from 856. Five of the company initiated outages in 2016 resulted in 384 customer hours and therefore made up the bulk of the total customer hours out in that category. These outages were necessary as part of the development of the utility s new 1 MW ac solar facility. Approximately four miles of circuit had to be rebuilt and reconductored in order to ensure that the system could carry generation from the project before it was energized last August. Equipment failures were the third highest cause of outages on SED s system in 2016 at 12 total evets and 357 customer hours out. This was a decrease in both numbers from SED s system experienced 16 events and 856 customer hours out that year. Three of the outages this year were caused by failed direct buried primary cable that was installed in the 1970s and 1980s and lead to 330 customer hours out. One accident was a significant contribution to the increase in the 2016 CAIDI number with 274 customer hours out. A dump truck pulled down all three phases, broke one pole, and spun another. All of this leads to a considerable amount of time to repair. 106 P age

117 At Stowe Electric, we never stop working to improve our customers experiences, our community, and our employees. Every day we are focused on operating and maintaining Stowe s electrical system to keep the lights on and provide reliable service. I Integrated Analysis and Plan of Action I.1 Evaluation of Portfolio Scenarios ENE s portfolio simulation models evaluated twelve (12) scenarios that consisted of varying amounts of resources and fuel type. Scenario #1 is the base case, which is the do nothing current portfolio. ENE analyzed each scenario from both the energy perspective and the RES contribution to compliance perspective. Below are all the scenarios, categorized by number for clarification. Portfolio Scenarios: Scenario # 1 = Current Portfolio with no additional resource procurement Scenario #2 = Current Portfolio, with 100 kw Moscow Mills Hydroelectric Unit and an additional 1MW Solar Project Scenario #3 = Current Portfolio, with.44 MWs biomass beginning in year 2022 (Ryegate Extension) Scenario #4 = Current Portfolio, with 1.5MW of Vermont based Wind PPA (Unit is considered to be a Tier II qualified new Distributed Generation project (<5MWs)) Scenario #5 = Current Portfolio, (Moscow, Solar, Ryegate, Large (>5MWs) Wind VT based Wind PPA (3% of SED Load)) 107 P age

118 Scenario #6 = Current Portfolio, with Renegotiated HQ output (Additional 25% of load) and Dodge Hydro PPA (2 MWs) Scenario #7 = Current Portfolio, with Moscow, Solar, Ryegate, Large (>5MWs) Wind VT based Wind PPA (3% of SED Load), and Miller Hydro (Brown Bear) Extension (3% of Load) Scenario #8 = Current Portfolio, with Moscow, Solar, Ryegate, Large (>5MWs) Wind VT based Wind PPA (3% of SED Load), Miller, and HQ Scenario #9 = Current Portfolio, with 5% load follow forward purchase and Ryegate Extension Scenario #10 = Current Portfolio, with 5% load follow forward purchase, Moscow, and Solar Scenario #11 = Current Portfolio, with Virtual Combined Cycle MW off-take of 1 MW, Moscow, Solar, Ryegate, Large (>5MWs) Wind VT based Wind PPA (3% of SED Load), and Miller Hydro Scenario #12 = Current Portfolio, with 1.5MW of Vermont based Wind PPA (Unit is considered to be a Tier II qualified new Distributed Generation project (<5MWs)), Moscow, and Solar The NPV of each scenario Cost and the risk tradeoff is below in Figure 95. With the stochastic models of Lacima ENE was able to rank each portfolio by the NPV of each scenario using energy cost and RES value. Using the Monte Carlo simulation allowed ENE the use of multiple variables, such as compliance payment rates, LMP, and load. ENE then performed iterations of these inputs and developed a probability of returns. Next, ENE analyzed these returns to determine the optimal scenario for Stowe that would not largely increase costs and maintain coverage of around 80%. Figure 95: Cost and Risk Tradeoff Bubble Plot The four primary factors that used for comparative analysis are: (Also found in A.2.3 Resource Alternatives) 108 P age

119 1) Least Cost: Net present value of the total portfolio; this includes energy cost of both current resources and potential scenario resources 2) Renewable Energy Standard: Mean of each scenario based on current RES coverage and resources for each scenario. 3) Standard Deviation: Risk of each scenario relative variation of the expected NPV of Total Portfolio Cost, and RES, as measured by the standard deviation and various tradeoff considerations 4) Spot Market Exposure: The relative spot market exposure to Stowe based on each scenario. Stowe maintains a portfolio hedge plan of at least 80% coverage. Table 27: Scenario Simulation Summary Statistics by Ranking NPV Total Cost Rank Total RES Rank Std Dev Rank Spot Exposure Target Rank Deviation Scenario #1 ($67,904,715.22) 1 ($4,012,367) 12 $ 2,795, % 12 Scenario #2 ($70,586,528.62) 4 ($2,101,875) 9 $ 2,573, % 11 Scenario #3 ($69,988,488.39) 2 ($3,068,746) 10 $ 2,680, % 10 Scenario #4 ($74,172,340.64) 8 $2,087,307 2 $ 2,124, % 8 Scenario #5 ($74,606,655.28) 9 ($253,673) 6 $ 2,415, % 5 Scenario #6 ($73,149,973.85) 6 ($711,959) 7 $ 2,393, % 3 Scenario #7 ($74,896,269.33) 10 ($201,228) 5 $ 2,368, % 4 Scenario #8 ($75,294,817.81) 11 ($120,484) 3 $ 2,253, % 2 Scenario #9 ($70,470,305.22) 3 ($3,071,060) 11 $ 2,598, % 7 Scenario #10 ($71,068,345.44) 5 ($2,098,355) 8 $ 2,500, % 9 Scenario #11 ($74,023,474.85) 7 ($196,824) 4 $ 2,244, % 1 Scenario #12 ($77,542,208.99) 12 $3,957,634 1 $ 2,094, % 6 The analytical process was to determine the most optimal scenario for Stowe that both, maintained energy costs with reasonable renewable alternatives, and helped curb the large cost impact of RES to SED. The ranking per category is based solely on the most optimal of that category. ENE chose to consider more than category rank to determine the best solution for Stowe. To determine the scenarios that would financial benefit Stowe ENE analyzed how each scenario ranked in each category, the mean cost of each portfolio, and the risk to Stowe for each scenario. ENE s integration models were used to run 1,000 iterations of each potential portfolio for both energy and RES impact. ENE determined how the cost, stability, and environmental impact to Stowe would be for each scenario. The goal was to manage overall portfolio cost by minimizing the RES compliance payments. Although ENE chose to focus on complying with RES as the first goal, the next evaluation was the comparative between risk and cost of the portfolio. I.2 Preferred Plan I.2.1 Optimal Scenario The IRP process found the optimal scenario to be scenario #4. This scenario was the current SED portfolio with an additional 1.5 MW of a small 5 MW Vermont based wind project. The projected cost used in ENE s Lacima modeling for a PPA on a small wind project with RECs was a $.120 kwh flat rate. The stochastic model data is below in Figure 97. The Output for the RES impact is found in Appendix G. The rate used in the scenario can be justified using the PUC s Standard Offer Win 1.5 MW Cash Flow 109 P age

120 Model 40, whereas a developer could build a 1.5 MW wind project for $0.112 kwh. If the Rest of Pool clearing price for FCA8-FCA21 is added, one can calculate an average capacity rate of $.070 KWH. These rates can be found below in Figure 96. These numbers, together, get one to the estimated PPA rate of $.120 kwh in the model for a conservative approach. The State model assumptions are found in Appendix H. Although this option does not include a lot of diversity in resources, it does bring more diversity to SED s current portfolio. It adds 7% additional hedging opportunity at a steady price, which can help SED maintain a rate that does not fluctuate. The largest benefit of a resource like is the RES value. This project would be qualified as Tier I and Tier II. Even if it offered SED excess of Tier II, moving it to Tier III, and removing the Tier I; remains a huge benefit to SED s RES compliance because Tier II and III have the largest compliance price associated with them. This scenario fills SED s RES requirement the most because it fills the most expensive tiers, Tier II and Tier III. With the REC arbitrage, SED will be able to fill the minimal shortfall in the last year with the extra benefit from selling high and buying low at the beginning of the program. Figure 98 through Figure 100 are the RES resulting coverage from scenario #4. Figure 96: Capacity Rates Figure 97: Optimal Scenario #4 NPV Total Cost Rank Total RES Rank Std Dev Rank Spot Exposure Target Rank Deviation Scenario #4 ($74,172,340.64) 8 $2,087,307 2 $ 2,124, % P age

121 Figure 98: Tier I with Scenario #4 Figure 99: Tier II with Scenario #4 111 P age

122 Figure 100: Tier III with Scenario #4 I.2.2 Least Cost Scenario The least cost scenario is #1. This is actually Stowe s current portfolio, with doing no additional hedging or building of renewable projects. The reason for this outcome is largely due to the low forward price curves, seen in Figure 61. The Lacima model is mapping the open position to forward prices that are reasonably low compared to historical actual prices. Vermont average LMP for around the clock is below in Figure 101. The current NPV of scenario #1 is low enough to carry the extreme high cost of the NPV of RES to Stowe if it does nothing, and fill all the shortfall RES compliance at compliance payments. Figure 101: VT LMP Historic Averages of Around the Clock 112 P age

123 I.2.3 Greatest Cost Scenario The greatest cost scenario is #12. This scenario includes the current portfolio with a PPA for 1.5 MW of a small 5 MW VT wind project, the refurbished Moscow Mills hydro project, and an additional 1 MW solar project in SED. Although this scenario offers more diversity of resources, the resources are very costly to SED s portfolio. The wind, based at $.120 kwh, is largely offset by the value of the project towards SED s RES compliance. The solar and hydro are new builds that have hefty capital costs associated with them. ENE modeled the Moscow Mills at a price of $.130 kwh. This mirrors the PUC cash flow of hydro projects. 41 The Cash Flow model can be found in Appendix I. For the solar project, ENE modeled it at a price of $.150 kwh. This also is in line with the PUC Solar price cap model 42, which can be found in Appendix J. The cost of new builds, such as these resources, do not offset the RES cost, and therefore, does not make these resources appealing to SED if it wants to maintain low cost rates for its customers. I.2.4 Other optional Scenario Other scenarios that could benefit SED are scenario #6 and #8. Scenario #6 is SED s current portfolio with a hopeful option for a PPA with Dodge Falls Hydro, once its contract with VEPPI ends in 2020, of 2 MW, as well as an extension of the new HQ bilateral of the current 25% of load. ENE s Dodge estimated cost of a PPA was set at $.08 kwh. The estimate was based on the current Brown Bear contract of around $.050 kwh with a $.03 kwh REC cost for SED, because these PPAs are more valuable with RECs, in order for RES compliance. The HQ extension was set at a continued escalating rate of the existing HQ contract. ENE estimated both these contracts to contribute to Tier I of RES, but with Tier II at a higher cost of non-compliance, SED is really looking to fill more of its shortfall in this tier. Scenario #8 also contains an extension of HQ, but brings in an extension of Brown Bear, Ryegate, as well as new solar, and Moscow hydro. Lastly, it contains a PPA with a large 10 MW wind project of 3% of load for SED. ENE set the prices of HQ, Ryegate, and Miller Hydro in a continuation the escalation of the existing contracts. Whereas Moscow and Solar are new builds at rates of $.130 kwh and $.150 kwh, the new wind project was set at a rate of $.100 kwh. This was slightly less than a PPA with a small wind farm of $.120 kwh due to economies of scale of a larger project. #8 would be the second most optimal plan for SED, and thirdly would be scenario #6. The top three-scenarios, #4, #6, and #8, were selected by ENE based on their diversity to SED s current portfolio and the value to SED against a potential large charge of RES compliance. These resources are a good value based on their benefits to both, towards RES, and to portfolio stability with locked-in longterm contract rates. SED searches for options that will help stabilize costs just as much as trying to find low cost resources. With the compliance of RES now a large driver of decision making, more expensive renewable resources are more valuable, and SED can now justify adding these to its portfolio instead of locking in market contracts. I.3 Implementation or Action Plan Based on trade-offs of each scenario, Scenario #4 has the greatest amount of RES benefit and limited energy cost escalation for SED s Integrated Resource Plan. The components of the optimal and other P age

124 ideal scenarios are balanced to maintain SED at 80% coverage from 2018 through This target is SED s risk tolerance, because the municipal knows if market prices increase high enough it has Stonybrook as a Peaker unit that will help mitigate price spikes. This scenario allows SED to add other potential types of resources to a portfolio that is already RES compliant and is economically reasonable. Vermont based resources will be the most sought after. SED will evaluate each potential resource on cost and benefit to both energy and RES. After 2031, depending on both the energy and renewable markets, SED will have another option window to look into for additional products in order to comply with any new regulations. Reviewing Vermont based resources will be the key to Stowe s RES compliance and reducing their environmental impact. This option reduces environmental carbon footprint for Vermont and SED s customers. It would provide a long-term energy price point that SED can lock into its rates so it can monitor rate increases more efficiently if needed. Lastly, it will provide SED a RES compliance that will reduce its exposure to any compliance payments, which could increase costs to the ratepayers. I.4 Ongoing Maintenance and Evaluation Stowe will update this IRP on a scheduled basis per regulatory requirement and make any necessary adjustments. The implementation of the plan will include an annual review of factors that could initiate an adjustment, such as major shifts in the New England supply stack, new generation and carbon capture technology, fundamental changes to the natural gas market, and regulatory changes, including ISO New England market design. In the next IRP, Stowe will use the recommendations in the Vermont CEP and guidance from the Department of Public Service when addressing and setting a path to helping Vermont meet its goals. V.S.A states the RES program is to promote renewable energy goals of Balancing the benefits, lifetime costs, and rates of the State's overall energy portfolio to ensure that to the greatest extent possible the economic benefits of renewable energy in the State flow to the Vermont economy in general, and to the rate-paying citizens of the State in particular P age

125 A Appendix A A.1 Model Results A.1.1 Town Peak Forecast Model Estimation Period: January 2011 to December 2016 A.1.2 Residential Average Use Model Estimation Period: January 2008 to December P age

126 A.1.3 Commercial Average Use Model Estimation Period: January 2008 to December P age

127 A.1.4 Residential Customer Forecast Model Estimation Period: January 2008 to December P age

128 ResCust_Var = Household Index^.4*Non-Man Employment Index^.6 A.1.5 Commercial Customer Model Estimation Period: January 2008 to December 2016 ComCust_Var = Household Index^.7*Output Index^ P age

129 A.1.6 Saturation Model Estimated May 2011 through December P age

130 B Appendix B B.1 Model Description The traditional approach to forecasting monthly sales for a customer class is to develop an econometric model that relates monthly sales to weather, seasonal variables, and economic conditions. From a forecasting perspective, econometric models are well suited to identify historical trends and to project these trends into the future. In contrast, the strength of the end-use modeling approach is the ability to identify the end-use factors that are drive energy use. By incorporating end-use structure into an econometric model, the statistically adjusted end-use (SAE) modeling framework exploits the strengths of both approaches. There are several advantages to this approach. The equipment efficiency and saturation trends, dwelling square footage, and thermal shell integrity changes embodied in the long-run end-use forecasts are introduced explicitly into the short-term monthly sales forecast. This provides a strong bridge between the two forecasts. By explicitly introducing trends in equipment saturations, equipment efficiency, dwelling square footage, and thermal integrity levels, it is easier to explain changes in usage levels and changes in weather-sensitivity over time. Data for short-term models are often not sufficiently robust to support estimation of a full set of price, economic, and demographic effects. By bundling these factors with equipment-oriented drivers, a rich set of elasticities can be incorporated into the final model. B.1.1 Residential Model The statistically adjusted end-use modeling framework begins by defining energy use (Usey,m) in year (y) and month (m) as the sum of energy used by heating equipment (Heaty,m), cooling equipment (Cooly,m), and other equipment (Othery,m). Formally, = Heat Cool Other (1) USE y,m y,m + y,m + y,m Although monthly sales are measured for individual customers, the end-use components are not. Substituting estimates for the end-use elements gives the following econometric equation. USE m = a + b XHeat + b XCool + b XOther + e (2) 1 m 2 m 3 m m XHeatm, XCoolm, and XOtherm are explanatory variables constructed from end-use information, dwelling data, weather data, and market data. As will be shown below, the equations used to construct these X-variables are simplified end-use models, and the X-variables are the estimated usage levels for each of the major end uses based on these models. The estimated model can then be thought of as a statistically adjusted end-use model, where the estimated slopes are the adjustment factors. B Constructing XHeat As represented in the SAE spreadsheets, energy use by space heating systems depends on the following types of variables. 120 P age

131 Heating degree days Heating equipment saturation levels Heating equipment operating efficiencies Thermal integrity and footage of homes Average household size, household income, and energy prices The heating variable is represented as the product of an annual equipment index and a monthly usage multiplier. That is, XHeat y, m HeatIndexy, m HeatUsey, m = (3) Where: XHeaty,m is estimated heating energy use in year (y) and month (m) HeatIndexy,m is the monthly index of heating equipment HeatUsey,m is the monthly usage multiplier The heating equipment index is defined as a weighted average across equipment types of equipment saturation levels normalized by operating efficiency levels. Given a set of fixed weights, the index will change over time with changes in equipment saturations (Sat), operating efficiencies (Eff), building structural index (StructuralIndex), and energy prices. Formally, the equipment index is defined as: HeatIndex y = StructuralIndex y Weight Type Type Sat Sat Type y Type 09 Eff Eff Type y Type 09 (4) The StructuralIndex is constructed by combining the EIA s building shell efficiency index trends with surface area estimates, and then it is indexed to the 2009 value: StructuralIndex y BuildingShellEfficiencyIndexy SurfaceArea y = (5) BuildingShellEfficiencyIndex SurfaceAre 09 a 09 The StructuralIndex is defined on the StructuralVars tab of the SAE spreadsheets. Surface area is derived to account for roof and wall area of a standard dwelling based on the regional average square footage data obtained from EIA. The relationship between the square footage and surface area is constructed assuming an aspect ratio of 0.75 and an average of 25% two-story and 75% single-story. Given these assumptions, the approximate linear relationship for surface area is: SurfaceAre a y = Footage y (6) 121 P age

132 For electric heating equipment, the SAE spreadsheets contain two equipment types: electric resistance furnaces/room units and electric space heating heat pumps. Examples of weights for these two equipment types for Vermont are given in Table 28. Table 28: Electric Space Heating Equipment Weights Equipment Type Weight (kwh) Electric Resistance Furnace/Room units 545 Electric Space Heating Heat Pump 4 Data for the equipment saturation and efficiency trends are presented on the Shares and Efficiencies tabs of the SAE spreadsheets. The efficiency for electric space heating heat pumps are given in terms of Heating Seasonal Performance Factor [BTU/Wh], and the efficiencies for electric furnaces and room units are estimated as 100%, which is equivalent to 3.41 BTU/Wh. Heating system usage levels are impacted on a monthly basis by several factors, including weather, household size, income levels, prices, and billing days. The estimates for space heating equipment usage levels are computed as follows: HDDy, m HHSize y Incomey Elec Pr ice y, m HeatUse, = (7) y m 09 09,7 09,7 Pr HDD HHSize Income Elec ice09,7 Where: HDD is the number of heating degree days in year (y) and month (m). HHSize is average household size in a year (y) Income is average real income per household in year (y) ElecPrice is the average real price of electricity in month (m) and year (y) By construction, the HeatUsey, m variable has an annual sum that is close to 1.0 in the base year (2009). The first term, which involves heating degree days, serve to allocate annual values to months of the year. The remaining terms average to 1.0 in the base year. In other years, the values will reflect changes in the economic drivers, as transformed through the end-use elasticity parameters. The price impacts captured by the Usage equation represent short-term price response. B Constructing XCool The explanatory variable for cooling loads is constructed in a similar manner. The amount of energy used by cooling systems depends on the following types of variables. Cooling degree days Cooling equipment saturation levels Cooling equipment operating efficiencies Thermal integrity and footage of homes Average household size, household income, and energy prices 122 P age

133 The cooling variable is represented as the product of an equipment-based index and monthly usage multiplier. That is, XCool y, m CoolIndexy CoolUsey, m = (8) Where XCooly,m is estimated cooling energy use in year (y) and month (m) CoolIndexy is an index of cooling equipment CoolUsey,m is the monthly usage multiplier As with heating, the cooling equipment index is defined as a weighted average across equipment types of equipment saturation levels normalized by operating efficiency levels. Formally, the cooling equipment index is defined as: CoolIndex y = StructuralIndex y Weight Type Type Sat Sat Type y Type 09 Eff Eff Type y Type 09 (9) For cooling equipment, the SAE spreadsheets contain three equipment types: central air conditioning, space cooling heat pump, and room air conditioning. Examples of weights for these three equipment types for Vermont are given in Table 29. Table 29: Space Cooling Equipment Weights Equipment Type Weight (kwh) Central Air Conditioning 67 Space Cooling Heat Pump 1 Room Air Conditioning 184 The equipment saturation and efficiency trends data are presented on the Shares and Efficiencies tabs of the SAE spreadsheets. The efficiency for space cooling heat pumps and central air conditioning (A/C) units are given in terms of Seasonal Energy Efficiency Ratio [BTU/Wh], and room A/C units efficiencies are given in terms of Energy Efficiency Ratio [BTU/Wh]. Cooling system usage levels are impacted on a monthly basis by several factors, including weather, household size, income levels, and prices. The estimates of cooling equipment usage levels are computed as follows: CoolUse y, m CDD = CDD y, m 09 HHSize HHSize y 09,7 0.2 Income Income y 09,7 0.2 Elec Pr ice Elec Pr ice y, m 09, (10) 123 P age

134 Where: CDD is the number of cooling degree days in year (y) and month (m). HHSize is average household size in a year (y) Income is average real income per household in year (y) ElecPrice is the average real price of electricity in month (m) and year (y) By construction, the CoolUse variable has an annual sum that is close to 1.0 in the base year (2009). The first term, which involves cooling degree days, serve to allocate annual values to months of the year. The remaining terms average to 1.0 in the base year. In other years, the values will change to reflect changes in the economic driver changes. B Constructing XOther Monthly estimates of non-weather sensitive sales can be derived in a similar fashion to space heating and cooling. Base-use energy requirements are driven by: Appliance and equipment saturation levels Appliance efficiency levels Average number of days in the billing cycle for each month Average household size, real income, and real prices The explanatory variable for other uses is defined as follows: XOther y, m OtherEqpIndex y, m OtherUsey, m = (11) The first term on the right hand side of this expression (OtherEqpIndexy) embodies information about appliance saturation and efficiency levels and monthly usage multipliers. The second term (OtherUse) captures the impact of changes in prices, income, household size, and number of billing-days on appliance utilization. End-use indices are constructed in the SAE models. A separate end-use index is constructed for each end-use equipment type using the following function form. Sat Type y Type Type y OtherEqpIn dex y m Weight, = Sat Type 05 1 UEC 1 UEC Type 09 MoMult Type m (12) 124 P age

135 Where: Weight is the weight for each appliance type Sat represents the fraction of households, who own an appliance type MoMultm is a monthly multiplier for the appliance type in month (m) Eff is the average operating efficiency the appliance UEC is the unit energy consumption for appliances This index combines information about trends in saturation levels and efficiency levels for the main appliance categories with monthly multipliers for lighting, water heating, and refrigeration. The appliance saturation and efficiency trends data are presented on the Shares and Efficiencies tabs of the SAE spreadsheets. Further monthly variation is introduced by multiplying by usage factors that cut across all end uses, constructed as follows: OtherUse y, m BDays = 30.5 y, m HHSize HHSize y 09, Income Income y 09,7 0.2 Elec Price Elec Price y, m 09,7 0.1 B.1.2 Commercial Model The commercial statistically adjusted end-use model framework begins by defining energy use (Usey,m) in year (y) and month (m) as the sum of energy used by heating equipment (Heaty,m), cooling equipment (Cooly,m) and other equipment (Othery,m). Formally, USE y,m Heat y,m + Cool y,m + = Other (1) y,m Although monthly sales are measured for individual customers, the end-use components are not. Substituting estimates for the end-use elements gives the following econometric equation. USE m = a + b XHeat + b XCool + b XOther + e (2) 1 m 2 m 3 m m XHeatm, XCoolm, and XOtherm are explanatory variables constructed from end-use information, weather data, and market data. B XHeat As represented in the Commercial SAE spreadsheets, energy use by space heating systems depends on the following types of variables. Heating degree days, Heating equipment saturation levels, Heating equipment operating efficiencies, Commercial output, employment, population, and energy price. 125 P age

136 The heating variable is represented as the product of an annual equipment index and a monthly usage multiplier. That is, XHeat = HeatIndex HeatUse (3) y,m y y,m Where: XHeaty,m is estimated heating energy use in year (y) and month (m), HeatIndexy is the annual index of heating equipment, and HeatUsey,m is the monthly usage multiplier. The heating equipment index is composed of electric space heating equipment saturation levels normalized by operating efficiency levels. The index changes over time with changes in heating equipment saturations (HeatShare) and operating efficiencies (Eff). Heating Index is calculated as: HeatSharey Eff y HeatIndex y = HeatEI04 (4) HeatShare04 Eff is used as a base year for normalizing the index. The ratio on the right is equal to 1.0 in Increases in saturation drive the index higher, increases in efficiency drives the index lower. Base-year heating intensity is defined as: kwh HeatEI04 = (5) Sqft Heating By definition, base-year heating sales are the product of the space heating intensity and the total commercial floor space. In practice, we let the model determine the overall contribution of heating by estimating a coefficient for the heating energy intensity estimate. Variations from this value in other years will be proportional to saturation and efficiency variations around their base values. Heating system usage levels are impacted on a monthly basis by several factors, including weather, economic activity, prices and billing days. HeatUse is defined as: 0.2 HDDy, m EconVary, m Pricey, m HeatUse, y m = 04 04,7 Pr HDD EconVar ice04,7 (6) Where: HDD is the number of heating degree days in month (m) and year (y) EconVar is Output in month (m), and year (y). Price is the average real price of electricity in month (m) and year (y). By construction, the HeatUsey,m variable has an annual sum that is close to one in the base year (2004). The first term, which involves heating degree days, serves to allocate annual values to months of the 126 P age

137 year. The remaining terms average to one in the base year. In other years, the values will reflect changes in commercial output and prices, as transformed through the end-use elasticity parameters. B XCool The explanatory variable for cooling loads is constructed in a similar manner. The amount of energy used by cooling systems depends on the following types of variables. Cooling degree days, Cooling equipment saturation levels, Cooling equipment operating efficiencies, Commercial output, employment, population, and energy price. The cooling variable is represented as the product of an equipment-based index and monthly usage multiplier. That is, XCool = CoolIndex CoolUse y,m y y,m (7) Where: XCooly,m is estimated cooling energy use in year (y) and month (m), CoolIndexy is an index of cooling equipment, and CoolUsey,m is the monthly usage multiplier. As with heating, the cooling equipment index depends on equipment saturation levels (CoolShare) normalized by operating efficiency levels (Eff). Formally, the cooling equipment index is defined as: CoolSharey Eff y CoolIndex y = CoolEI 04 (8) CoolShare04 Eff04 Data values in 2004 are used as a base year for normalizing the index, and the ratio on the right is equal to 1.0 in In other years, it will be greater than one if equipment saturation levels are above their 2004 level. This will be counteracted by higher efficiency levels, which will drive the index downward. Estimate of base year cooling intensity is defined as: kwh CoolEI 04 = (9) Sqft Cooling By definition, base-year cooling sales are the product of the space cooling intensity and the total commercial floor space. Here, we let the model determine the overall contribution of cooling by estimating a coefficient for the cooling energy intensity estimate. Variations from this value in other years will be proportional to saturation and efficiency variations around their base values. 127 P age

138 Cooling system usage levels are impacted on a monthly basis by several factors, including weather, economic activity levels, and prices. The estimates of cooling equipment usage levels are computed as follows: 0.2 CDDy, m EconVary, m Pricey, m CoolUse, y m = 04 04,7 Pr CDD EconVar ice04,7 (10) Where: CDD is the number of heating degree days in month (m) and year (y). 0.2 EconVar is Output in month (m), and year (y). Price is the average real price of electricity in month (m) and year (y). By construction, the CoolUse variable has an annual sum that is close to one in the base year (2004). The first term, which involves cooling degree days, serves to allocate annual values to months of the year. The remaining terms average to one in the base year. In other years, the values will change to reflect changes in commercial output and prices. B XOther Monthly estimates of non-weather sensitive sales can be derived in a similar fashion to space heating and cooling. Based on end-use concepts, other sales are driven by: Equipment saturation levels, Equipment efficiency levels, Average number of days in the billing cycle for each month, and Real commercial output and real prices. The explanatory variable for other uses is defined as follows: XOther = OtherIndex OtherUse (11) y,m y,m y,m The first term on the right hand side of this expression embodies information about equipment saturation levels and efficiency levels. The equipment index for other uses is defined as follows: Type Share y Type Eff Type y OtherIndex = y, m EI 04 Type (12) Type Share04 Type Eff 04 Where: EI is the intensity for each equipment type, Share represents the fraction of floor stock with an equipment type, and Eff is the average operating efficiency. 128 P age

139 This index combines information about trends in saturation levels and efficiency levels for the main base use equipment categories. The intensities are defined as follows. Type kwh BaseEI04 = (13) Sqft Type Further monthly variation is introduced by multiplying by usage factors that cut across all end uses, constructed as follows: BDays y, m Output y, m Pr icey. m OtherUse y, m = ,7 Pr (14) Output ice04,7 Where: BDays is the number of billing days in month (m) and year (y). Output is GDP in month (m), and year (y). Price is the average real price of electricity in month (m) and year (y). By construction, the OtherUse variable has an annual sum that is close to one in the base year (2004). The first term, billing days, serves to allocate annual values to months of the year. The remaining terms average to one in the base year. In other years, the values will change to reflect changes in commercial output and prices. B.1.3 Peak Model Over the long-term, peak demand is driven by underlying growth in cooling, heating, and base use demand and peak-producing weather conditions. The coefficients from the SAE class sales forecast model, the constructed end-use variables, and peak-day weather conditions are used to develop a set of monthly peak-demand model variables. B Heating and Cooling Variables As system cooling and heating requirements change, the impact of temperature on peak demand changes as well. Cooling and heating loads in turn are driven by economic activity, customer growth, and improving end-use efficiency that are captured in the residential and commercial model variables. From the estimated class sales models, cooling load is calculated as: CoolLoadm = (ResCoolm+ ComCoolm)/Daysm/24 Normalized residential and commercial cooling (ResCool and ComCool) are calculated using normal monthly CDD. Cooling requirements are divided by the number of days per month (Days) and hours per day (24) to express the variable on an average hourly MW basis. The peak model cooling variable (PeakCoolVar) is calculated by interacting the CoolLoad with the peakday CDD (base 65 degrees): PeakCoolVarm = CoolLoadm * PkDayCDD65m Heating load is calculated in a similar manner: 129 P age

140 HeatLoadm = (ResHeatm+ ComHeatm)/Daysm/24 The peak model heating variable is then calculated by interacting HeatLoad with the peak day HDD (base 55 degrees): PeakHeatVarm = HeatLoadm * PkDayHDD55m B Base Load Variable The demand impact of non-weather sensitive load is captured by the base load variable (BaseLoad). Like the heating and cooling load variable, BaseLoad is calculated from the customer class SAE models. BaseLoadm = (ResOtherm+ComOtherm)/Daysm/24 The base load variable is divided by the number of days per month (Days) and hours per day (24) to express the variable as an average hourly MW load. B Peak Model The monthly peak model is estimated using a linear regression model. The model is estimated from January 2011 to December The model also includes monthly binaries for Apr, May, November, and December interacted with BaseLoad; this is designed to capture monthly variation in base load on monthly peak demand. The model variables are statistically significant and explain historical monthly peak loads relatively well with an Adjusted R-Squared of Figure 102 shows actual and predicted peak demand. Figure 102: Actual and Predicted Monthly Peak Demand (MW) The higher forecast starting point reflects the normal peak-day HDD. The normal HDD is based on twenty-years of minimum peak-day temperatures; the last six-year winter peaks have been warmer than normal. 130 P age

141 C Appendix C STANDARD OFFER PROJECTS OPERATING 131 P age

142 D Appendix D RES Analysis Base Case 132 P age

143 133 P age E Appendix E

144 F Appendix F Capacity Simulation for Alternative Simulation based on historical year Output Results Name Cell Sim# Graph Min Mean Max 5% 95% Range: FCM Prices FCM Prices / 2022 I FCM Prices / 2022 I FCM Prices / 2022 I FCM Prices / 2022 I FCM Prices / 2022 I FCM Prices / 2023 J FCM Prices / 2023 J FCM Prices / 2023 J FCM Prices / 2023 J FCM Prices / 2023 J FCM Prices / 2024 K FCM Prices / 2024 K FCM Prices / 2024 K FCM Prices / 2024 K FCM Prices / 2024 K FCM Prices / 2025 L FCM Prices / 2025 L FCM Prices / 2025 L FCM Prices / 2025 L FCM Prices / 2025 L P age

145 FCM Prices / 2026 M FCM Prices / 2026 M FCM Prices / 2026 M FCM Prices / 2026 M FCM Prices / 2027 N FCM Prices / 2027 N FCM Prices / 2027 N FCM Prices / 2027 N FCM Prices / 2027 N FCM Prices / 2028 O FCM Prices / 2028 O FCM Prices / 2028 O FCM Prices / 2028 O FCM Prices / 2028 O FCM Prices / 2029 P FCM Prices / 2029 P FCM Prices / 2029 P FCM Prices / 2029 P FCM Prices / 2029 P P age

146 FCM Prices / 2030 Q FCM Prices / 2030 Q FCM Prices / 2030 Q FCM Prices / 2030 Q FCM Prices / 2030 Q FCM Prices / 2031 R FCM Prices / 2031 R FCM Prices / 2031 R FCM Prices / 2031 R FCM Prices / 2031 R FCM Prices / 2032 S FCM Prices / 2032 S FCM Prices / 2032 S FCM Prices / 2032 S FCM Prices / 2032 S P age

147 FCM Prices / 2033 T FCM Prices / 2033 T FCM Prices / 2033 T FCM Prices / 2033 T FCM Prices / 2033 T FCM Prices / 2034 U FCM Prices / 2034 U FCM Prices / 2034 U FCM Prices / 2034 U FCM Prices / 2034 U FCM Prices / 2035 V FCM Prices / 2035 V FCM Prices / 2035 V FCM Prices / 2035 V FCM Prices / 2035 V FCM Prices / 2036 W FCM Prices / 2036 W FCM Prices / 2036 W FCM Prices / 2036 W FCM Prices / 2036 W FCM Prices / 2037 X FCM Prices / 2037 X FCM Prices / 2037 X FCM Prices / 2037 X FCM Prices / 2037 X P age

148 138 P age G Appendix G