INTEGRATED RESOURCE PLAN

Transcription

1 INTEGRATED RESOURCE PLAN Date Issued: September, 2015

2 ii 2015 Integrated Resource Plan

3 2015 Integrated Resource Plan Table of Contents TABLE OF CONTENTS... III LIST OF TABLES... VI LIST OF FIGURES... VII CROSS REFERENCE TABLE... IX LIST OF ACRONYMS... XI EXECUTIVE SUMMARY INTRODUCTION OVERVIEW IRP PROCESS INTRODUCTION TO PSO LOAD FORECAST AND FORECAST METHODOLOGY SUMMARY OF PSO LOAD FORECAST FORECAST ASSUMPTIONS Economic Assumptions Price Assumptions Specific Large Customer Assumptions Weather Assumptions Demand-Side Management (DSM) Assumptions OVERVIEW OF FORECAST METHODOLOGY DETAILED EXPLANATION OF LOAD FORECAST METHODOLOGY General Customer Forecast Models Short-term Forecasting Models Long-term Load Forecasting Models Final Monthly Internal Energy Forecast Forecast Methodology for Seasonal Peak Internal Demand LOAD FORECAST RESULTS AND ISSUES Load Forecast Peak Demand and Load Factor Energy Efficiency (EE) Impacts on the Load Forecast Blended Load Forecast Large Customer Changes Price Elasticity LOAD FORECAST TRENDS & ISSUES Changing Usage Patterns Energy Efficiency Embedded in the Load Forecast LOAD FORECAST SCENARIOS iii

4 2015 Integrated Resource Plan Low Load Sensitivity Case High Load Sensitivity Case EXISTING RESOURCES AND TRANSACTIONS CURRENT RESOURCES EXISTING PSO GENERATING RESOURCES CAPACITY NEEDS ASSESSMENT ENVIRONMENTAL ISSUES AND IMPLICATIONS Introduction Regional Haze Rule (RHR) Mercury and Air Toxics Standards Rule (MATS) Cross-State Air Pollution Rule (CSAPR) National Ambient Air Quality Standards (NAAQS) Coal Combustion Residuals (CCR) Rule Effluent Limitations Guidelines (ELG) Clean Water Act 316(b) Rule Carbon Dioxide (CO 2 ) and Greenhouse Gas (GHG) Regulations Impact of Environmental Rules on Planning Process CURRENT DEMAND SIDE PROGRAMS Background Impacts of Existing and Future Codes and Standards Current Demand Response (DR)/Energy Efficiency (EE) Programs Demand Reduction Energy Efficiency (EE) Conservation Voltage Reduction (CVR) Energy Conservation Distributed Generation AEP-SPP TRANSMISSION Transmission System Overview Current AEP-SPP Transmission System Issues Recent AEP-SPP Bulk Transmission Improvements Impacts of New Generation Summary of Transmission Overview MODELING PARAMETERS MODELING AND PLANNING PROCESS AN OVERVIEW METHODOLOGY FUNDAMENTAL MODELING INPUT PARAMETERS Commodity Pricing Scenarios Long-Term CO 2 Forecast DEMAND-SIDE MANAGEMENT (DSM) PROGRAM SCREENING & EVALUATION PROCESS Overview Achievable Potential Determining Future Demand Side Programs Evaluating Incremental Demand-Side Resources iv

5 2015 Integrated Resource Plan IDENTIFY AND SCREEN SUPPLY-SIDE RESOURCE OPTIONS Capacity Resource Options New Supply-Side Capacity Alternatives Baseload/Intermediate Alternatives Peaking Alternatives Renewable Alternatives Cogeneration & Combined Heat & Power (CHP) INTEGRATION OF SUPPLY-SIDE AND DEMAND-SIDE OPTIONS WITHIN PLEXOS MODELING Optimize Expanded DSM Programs Optimize Other Demand-Side Resources RESOURCE PORTFOLIO MODELING THE PLEXOS MODEL - AN OVERVIEW Key Input Parameters PLEXOS OPTIMIZATION Modeling Options and Constraints Optimized Portfolios SELECTED PLAN Demand-Side Management (DSM) Results CO 2 Emissions RISK ANALYSIS Stochastic Modeling Process and Results CONCLUSIONS AND FIVE-YEAR ACTION PLAN PLAN SUMMARY PSO Five-Year Action Plan CONCLUSION APPENDIX EXHIBIT A: DEMAND PORTFOLIO FILING SUMMARY EXHIBIT B: 2015 FUEL SUPPLY PORTFOLIO AND RISK MANAGEMENT PLAN EXHIBIT C: NEW GENERATION TECHNOLOGY OPTIONS EXHIBIT D: OPTIMIZATION MODEL COST OUTPUTS EXHIBIT E: OKLAHOMA CORPORATE COMMISSION TECHNICAL CONFERENCE TRANSCRIPT v

6 2015 Integrated Resource Plan List of Tables Table 1. PSO Actual and Forecast Internal Energy (GWh) Requirements by Sector with Growth Rates (%) Table 2. PSO Summer, Winter, and Annual Peak Demand (MW), Internal Energy Requirements (GWh) and Load Factor (%) Table 3. PSO EE in Load Forecast Energy (GWh) and Coincident Peak Demand (MW) Table 4. PSO Short-Term Load Forecast Blended Forecast vs. Long-Term Model Results Table 5. PSO Going-In Capacity, Demand and Reserves (CDR) Table 6. PSO Owned Generation Assets as of August, Table 7. Forecasted View of Relevant Energy Efficiency Code Improvements Table 8. Energy Efficiency (EE) Market Barriers Table 9. Residential Sector Energy Efficiency (EE) Measure Categories Table 10. Commercial Sector Energy Efficiency (EE) Measure Categories Table 11. Incremental Demand-Side Residential Energy Efficiency (EE) Bundle Summary Table 12. Incremental Demand-Side Commercial Energy Efficiency (EE) Bundle Summary Table 13. Conservation Voltage Reduction (CVR) Tranche Profiles Table 14. New Generation Technology Options with Key Assumptions Table 15. Cumulative SPP Capacity Additions (MW) for Five Commodity Pricing Scenarios Table 16. Cumulative SPP Capacity Additions (MW) for Load Sensitivity Scenarios Table 17. Cumulative SPP Capacity Additions (MW) and Average Annual Energy Position (GWh) for Selected Plan Table 18. Cumulative CVR Additions in Selected Plan (MW) Table 19. Risk Factors and Their Relationships Table 20. Cumulative Present Worth (CPW) of Selected Plan and Combustion Turbine Substitution Portfolio ($M) Table 21. Summary of PSO Selected Plan Resource Additions from a Capacity (MW) Viewpoint vi

7 2015 Integrated Resource Plan List of Figures Figure 1. PSO Service Territory (Shown in Red) Figure 2. PSO Internal Energy Requirements and Peak Demand Forecasting Method Figure 3. PSO Residential and Commercial Normalized Use per Customer (kwh) Figure 4. PSO Projected Cooling Efficiencies, Figure 5. PSO Select Appliance Efficiencies, Figure 6. PSO Going-In SPP Capacity Position (MW) and Obligation (MW) Figure 7. PSO Capacity Positions (MW) net of SPP Requirements Figure 8. Codes & Standards Impact on PSO Retail Load, post Figure 9. Forecasted Solar Installed Costs (nominal) for Oklahoma (Excluding Federal & State Incentives) Figure 10. PSO Cumulative Rooftop Solar Additions/Projections Figure 11. Long-term Power Price Forecast Process Flow Figure 12. SPP On-Peak Energy Prices (Nominal $/MWh) Figure 13. SPP Off-Peak Energy Prices (Nominal $/MWh) Figure 14. Panhandle East Pipeline Natural Gas Prices (Nominal $/mmbtu) Figure 15. Panhandle East Pipeline Natural Gas Prices (Real $/mmbtu) Figure 16. PRB 8,800 BTU/lb. Coal Prices (Nominal $/ton, FOB) Figure 17. CO 2 Prices (Nominal $/metric ton) Figure Projected Residential Energy Consumption (GWh) by End Use Figure Projected Commercial Energy Consumption (GWh) by End Use Figure 20. Forecasted Solar Installed Costs (nominal) for Oklahoma (Excl. Federal & State Incentives) Figure 21. LCOE (nominal $/MWh) for Wind Resource Tranches Included in PSO Model Figure 22. PSO EE and CVR Energy (GWh) for Existing and Incremental Programs Figure 23. PSO Cumulative Rooftop Solar Additions/Projections Figure 24. PSO Projected 'Mass-Based" CO2 Emissions under the Selected Plan Figure 25. Variable Input Prices for Gas (nominal $/mmbtu) Figure 26. Variable Input Prices for Power (nominal $/MWh) vii

8 2015 Integrated Resource Plan Figure 27. Variable Input Prices for CO 2 (nominal $/metric ton) Figure 28. PSO Portfolio Revenue Requirement at Risk (RRaR) ($000 s), Figure PSO Nameplate Capacity Mix Figure PSO Nameplate Capacity Mix Figure PSO Energy Mix Figure PSO Energy Mix Figure 33. PSO Annual SPP Capacity Position (MW) viii

9 2015 Integrated Resource Plan Cross Reference Table Requirement from OAC 165: (c) Location of PSO s Response (1) Schedule A: An electric demand and energy forecast IRP Section 2.5 (2) Schedule B: A forecast of capacity and energy IRP Sections 3.3, 6.1 contributions from existing and committed supply- and demand-side resources (3) Schedule C: A description of transmission capabilities IRP Section 3.6 and needs covering the forecast period (4) Schedule D: An assessment of need for additional IRP Section 3.3 resources (5) Schedule E: A description of the supply, demand-side IRP Sections 4.4, 4.5 and transmission options available to the utility to address the identified needs (6) Schedule F: A fuel procurement plan, purchased-power Appendix, Exhibit B procurement plan, and risk management plan (7) Schedule G: An action plan identifying the near-term IRP Section (i.e., across the first five [5] years) actions that the utility proposes to take to implement its proposed resource plan (8) Schedule H: Any proposed RFP(s), supporting Appendix, Exhibit B documentation, and bid evaluation procedures by which the utility intends to solicit and evaluate new resources (9) Schedule I: A technical appendix for the data, IRP pp. Exhibit C assumptions and descriptions of models needed to understand the derivation of the resource plan (10) Schedule J: A description and analysis of the adequacy IRP Section 3.6 of its existing transmission system to determine its capability to serve load over the next ten (10) years, including any planned proposed changes to existing transmission facilities ix

10 2015 Integrated Resource Plan (11) Schedule K: An assessment of the need for additional resources to meet reliability, cost and price, environmental or other criteria established by the Commission, the State of Oklahoma, the Southwest Power Pool, North American Electric Reliability Council, or the Federal Energy Regulatory Commission. This assessment should address both base line forecast condition and important uncertainties, including but not limited to load growth, fuel prices, and availability of planned supplies (12) Schedule L: An analysis of the utility s proposed resource plan and any alternative scenarios necessary to demonstrate how the preferred plan best meets the planning criteria. Technical appendices should be included to document the planning analysis and assumptions used in preparing this analysis (13) Schedule M: A description and analysis of the Utility s consideration of physical and financial hedging to determine the Utility s ability to mitigate price volatility for the term covered by the IRP IRP Section 3.3 IRP pp. Sections 5.3, 5.4, 6.0 Appendix, Exhibit B x

11 2015 Integrated Resource Plan List of Acronyms ACRONYM DEFINITION AACEI Association for the Advancement of Cost Engineering International AC Alternating Current A/C Air Conditioning ACI Activated Carbon Injection AD Aeroderivative AECC Arkansas Electric Cooperative Corporation AEP American Electric Power AP Achievable Potential ARIMA Autoregressive Integrated Moving Average ARRA American Recovery and Reinvestment Act ATSS Aggregated Transmission Service Study BART Best Available Retrofit Technology BNEF Bloomberg New Energy Finance BTU British Thermal Unit CAFE Corporate Average Fuel Economy CAIR Clean Air Interstate Rule CCR Coal Combustion Residuals CCS Carbon Capture and Sequestration CDR Capacity Demand and Reserves CERA Cambridge Energy Research Associates CHP Combined Heat and Power CO 2 Carbon Dioxide COS Cost of Service CPP Clean Power Plan CPW Cumulative Present Worth CSAPR Cross-State Air Pollution Rule CVR Conservation Voltage Reduction DC Direct Current DG Distributed Generation DOE Department of Energy DR Demand Reduction DSI Dry Sorbent Injection DSM Demand-side Management EE Energy Efficiency EGU Electric Generating Units EHV Extra High Voltage EIA Energy Information Administration EIEA2008 Energy Improvement and Extension Act of 2008 EISA Energy Independence and Security Act ELG Effluent Limitation Guidelines EPA Environmental Protection Agency EPAct Energy Policy Act EPRI Electric Power Research Institute ERCOT Electric Reliability Council of Texas FERC Federal Energy Regulatory Commission xi

12 2015 Integrated Resource Plan ACRONYM FF FGD FRB GDP GE GHG GI GRDA GWh HAP HCl HHV HRSG HVAC HVDC IGCC IRP ITC ITP kv kw kwh lb LCOE LHV MAR MATS mmbtu MOPC MW MWac MWh MWh-g NAAQS NERC NGCC NGCT NO 2 NO x NPDES NSPS NTC O&M OATT OCC DEFINITION Fabric Filters Flue Gas Desulfurization Federal Reserve Board Gross Domestic Product General Electric Greenhouse Gas Generation Interconnection Grand River dam Authority Gigawatt-hour High Achievable Potential Hydrochloric Acid Higher Heating Value Heat Recovery Steam Generator Heating, Ventilation, and Cooling High Voltage Direct Current Integrated Gasification Combined-Cycle Integrated Resource Plan Investment Tax Credit Integrated Transmission Planning Kilovolt Kilowatt Kilowatt-hour Pound Levelized Cost of Energy Lower Hating value Market Acceptance Ratio Mercury and Air Toxics Standard Million BTU Market and Operations Policy Committee Megawatt Alternating Current Megawatts Megawatt-Hour Megawatt-Hour, Gross National Ambient Air Quality Standards North American Electric Reliability Corporation Natural Gas Combines Cycle Natural Gas Combustion Turbine Nitrogen Dioxide Nitrogen Oxide National Pollutant Discharge Elimination System New Source Performance Standards Notification to Construct Operations and Maintenance Open Access Transmission Tariff Oklahoma Corporation Commission xii

13 2015 Integrated Resource Plan ACRONYM ODEQ OG&E OK PIF PIRA PM PPA PRB PSIG PSO PTC PUD PV RE RHR RRaR RSC RTO SAE SCR SD SEER SIP SO 2 SOFA SPP STEP SWEPCO DEFINITION Oklahoma Department of Environmental Quality Oklahoma Gas and Electric Energy Corporation Oklahoma Program Implementation Factor Petroleum Industry Research Associates Particulate Material Power Purchase Agreement Powder River Basin Pounds per Square Inch, Gage Public Service Company of Oklahoma Production Tax Credit Public Utility Division Photovoltaic Reciprocating Engine Regional Haze Rule Revenue Requirement at Risk Regional State Committee Regional Transmission Organization Statistically Adjust End-Use Selective Catalytic Reduction Standard Deviation Seasonal Energy Efficiency Ratio State Implementation Plan Sulfur Dioxide Separated Overfire Air Southwest Power Pool SPP Transmission Expansion Plan Southwestern Electric Power Company xiii

14 2015 Integrated Resource Plan Executive Summary The Integrated Resource Plan (IRP) is based upon the best available information at the time of preparation. However, changes that may impact this plan can, and do, occur without notice. Therefore this plan is not a commitment to a specific course of action, since the future is highly uncertain, particularly in light of the economic conditions, the movement towards increasing use of renewable generation and end-use efficiency, as well as current and future environmental regulations, including the U.S. Environmental Protection Agency s (EPA) Clean Power Plan (CPP). The implementation action items as described herein are subject to change as new information becomes available or as circumstances warrant. An IRP explains how a utility company plans to meet the projected capacity and energy requirements of its customers. Per Oklahoma Administrative Code 165, Chapter 35, Subchapter 37, Public Service Company of Oklahoma (PSO or Company) is required to provide an IRP every three years. PSO s 2015 IRP has been developed using the Company s current assumptions for: Customer load requirements peak demand and energy; Commodity prices coal, natural gas, on-peak and off-peak power prices, capacity and emission prices; Supply-side alternative costs including fossil fuel and renewable generation resources; and Demand-side program costs and impacts. In addition, PSO must consider the impact of the ongoing promulgation of environmental rules, including greenhouse gas emissions, which could result in the Company taking additional supply- and demand-side compliance measures. Along with the uncertainty created by increasing environmental requirements, the electric utility industry is beginning a transition driven by emerging technologies including renewable energy, both large-scale and distributed, within the planning horizon. In aggregate, these uncertainties will likely influence the Company s decision whether or not to construct new long-lived central plant generation. 1

15 2015 Integrated Resource Plan Summary of PSO Resource Plan PSO s total internal energy requirements are forecasted to increase at a rate of 0.9% over the IRP planning period (through 2024). PSO s corresponding summer peak internal demand is forecasted to increase at a rate of 0.7%, with annual peak demand expected to continue to occur in the summer season through Figure ES - 1, below, shows PSO s going-in (i.e. before resource additions) Southwest Power Pool (SPP) capacity position over the planning period. Though PSO has a small capacity shortfall in 2021, by 2022 PSO is anticipated to experience a significant capacity shortfall based upon the expiration of Power Purchase Agreements (PPAs). Figure ES - 1. PSO Going-In SPP Capacity Position (MW) and Obligation (MW) To determine the appropriate level and mix of incremental supply- and demand-side resources required to address the indicated going-in capacity deficiencies, PSO utilized the Plexos linear program optimization model to develop least cost resource portfolios under a variety of commodity pricing and load scenarios. Although the IRP planning period is limited to 10 years (through 2024), Plexos modeling was performed through the year 2045 with infinite end-effects, so as to properly consider various long-run costs for the resource alternatives. 2

16 2015 Integrated Resource Plan PSO used the modeling results to develop a Selected Plan or Plan. To arrive at the Selected Plan, using Plexos, PSO developed optimal portfolios based on five long-term commodity price forecasts and two load sensitivities. The Selected Plan balances cost and other factors such as risk and environmental regulatory considerations, to cost effectively meet PSO s demand and energy obligations. In summary, the Selected Plan: PSO s Selected Plan Adds 50MW (nameplate) of utility-scale solar resources per year in , for a total of 200MW (nameplate) of utility-scale solar by the end of the planning period. Adds 100MW (nameplate) wind energy in Implements customer and grid energy efficiency programs, including Conservation Voltage Reduction, reducing energy requirements by 517GWh and capacity requirements by 99MW by Fills long-term needs through the addition of 390MW of natural gas combined-cycle generation by repowering Northeastern Unit 4, or equivalent in 2022, and an additional 870MW of natural gas combined-cycle resources in the years Fills short-term need of 266MW in 2021 through a short-term PPA. Anticipates retirement of Southwestern Units 1 (68MW) and 2 (79MW) in 2021 and 2023, respectively. Anticipates retirement of Weleetka Units 4, 5, and 6 (150MW total) in Note: The modeling for this IRP was conducted prior to the issuance of the EPA CPP final rule. PSO capacity changes over the 10-year planning period associated with the Selected Plan are shown in Figure ES - 2 and Figure ES - 3. Their relative impacts to PSO s annual energy position are shown in Figure ES - 4 and Figure ES

17 2015 Integrated Resource Plan Figure ES PSO Nameplate Capacity Mix Figure ES PSO Nameplate Capacity Mix 4

18 2015 Integrated Resource Plan Figure ES PSO Energy Mix Figure ES PSO Energy Mix Figure ES - 2 through Figure ES - 5 indicate that this Selected Plan would reduce PSO s reliance on solid fuel-based generation, and increase reliance on demand-side, natural gas, and renewable resources. Specifically, over the 10-year planning horizon the Company s nameplate capacity mix attributable to solid fuel-fired assets declines from 18.7% to 8.5%, and natural gas assets would increase from 57.8% to 62.8%. Solar assets make up 3.1% of the capacity mix and 5

19 2015 Integrated Resource Plan wind assets increase to 18.8%. Demand-side resources are added to the mix at 2.9% of total nameplate capacity resources. PSO s energy output attributable to solid fuel generation decreases from 54.7% to 16.4% over the planning period, while energy from natural gas resources increases from 13.2% to 46.9%. The Selected Plan introduces solar resources, attributing to 3.0% of total energy. Reliance on thermal PPA energy is decreased from 16.7% to 2.6% based on the planning assumption that PPA s will be replaced with newly constructed natural gas combined-cycle generation. However, the final PPA percentages may change once a Request for Proposal process is conducted to determine if there are more cost effective market opportunities that exist to meet the capacity need in 2022 and beyond. Figure ES - 6 shows PSO s annual SPP capacity position over the planning period, per the Selected Plan. Thermal PPA s replace the capacity and a portion of the energy from the retiring Northeastern Unit 4 in Some of these PPA s expire in 2021 and new natural gas combined-cycle generation is added, along with the repower of Northeastern Unit 4 into a natural gas combined-cycle. 6

20 2015 Integrated Resource Plan Figure ES - 6. PSO Annual SPP Capacity Position (MW), per the Selected Plan Table ES- 1 provides a summary of the Selected Plan, which resulted from analysis of optimization modeling under the load and commodity pricing scenarios: 7

21 2015 Integrated Resource Plan Table ES- 1. Summary of PSO Selected Plan Resource Additions from a Capacity (MW) Viewpoint IRP Perio d (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15)=(1)to(14), ex(7) (16) (17) (18) (19) (20) (Cumulative) (Cumulative) Firm Capacity Resource ADDITIONS RETIREMENTS (Cumul.) SPP Planning Coal & Gas- Year (A) Steam (Frame) CTs Recip Engines New-Build PPA (L) Energy Efficiency (EE) CVR DR Wind (D) Solar (E) NET Above Wind (D) Solar (E) NE4 Repower NGCC 'Embedded' Federal EE Regulations (B) Existing DSM Programs (C) New Existing DSM Utility- Programs (C) Scale Distrib. 'RESOURCE CHANGE (J) SPP Rqrmnt (K) % Minimum Margin Utility-Scale Distrib. Yr. MW MW MW MW MW MW MW MW MW MW MW MW MW MW MW MW MW MW MW % (467) (F) % (467) % (467) % (467) % (467) % (467) % (535) (G) (274) % (685) (H) (274) % (764) (I) (274) % (A) SPP Planning Year is effective 6/1/XXXX. (B) (C) (D) (E) 'TOTAL' Energy Efficiency ( ) Represents estimated energy efficiency levels already 'embedded' into PSO's long-term load & peak demand forecast based on emergence of PRIOR-ESTABLISHED Federal efficiency standareds (EPAct 2005; 2007 EISA, 2009 ARRA). Represents estimated contribution from current/known PSO DSM-EE and Demand Response (Interruptible, AMI DLC/ELM) program activity also reflected in the Company's long-term load and demand forecast. Wind resource capacity credits, as a percentage of namplate rating, vary for each project and may be modified in the future when operational data becomes available. Due to the intermittency of solar resources, only 35% of solar resource 'nameplate' MW rating are included for capacity resource determination purposes. RETIREMENTS: (F) (G) (H) (I) Northeastern Unit 4 retirement effective approximately April 16, Southwestern Unit 1 retirement effective December 31, Weleetka Units 4,5, and 6 retirement assumed December 31, Southwestern Unit 2 retirement assumed December 31, Public Service of Oklahoma 2015 Integrated Resource Plan Cumulative Resource Changes Selected Plan Portfolio Resulting (Cumulative) PSO Reserves 'NAMEPLATE' ADDITIONS 'TOTAL' Solar ( ) (J) Excludes cumulative annual changes in PSO SPP 'Load Responsibility' (coincident peak demand) and 3rd-party resources which also impacts relative capacity resource position. (K) SPP minimum 13.6% as a function of peak demand. (L) PPA's include Thermal Generation ' 8

22 2015 Integrated Resource Plan PSO Five-Year Action Plan Steps to be taken by PSO in the near future as part of its Five-Year Action Plan include: 1. Continue the planning and regulatory actions necessary to implement economic energy efficiency programs in Oklahoma. 2. Conduct a Request For Proposals (RFP) to explore opportunities to add wind generation in the near future to take advantage of the Federal Production Tax Credit. 3. Conduct an RFP to explore adding cost effective utility-scale solar resources. 4. Initiate the RFP process to evaluate PSO s options for replacing the existing PPAs when they expire in 2021 and Evaluate the greenhouse gas rules. Work with the Oklahoma Executive Branch, Oklahoma Department of Environmental Quality, and the Office of the Attorney General on Oklahoma s response to the EPA s greenhouse gas rule. 6. Be ready to adjust this Action Plan and future IRPs to reflect changing circumstances Conclusion PSO s Selected Plan provides the Company with an increasingly diversified portfolio of supply- and demand-side resources which provides flexibility to adapt to future changes to the power market, technology, and environmental regulations. The addition of efficient natural gasfired generation along with increased renewables and demand-side management mitigates fuel price and environmental compliance risk. Inasmuch as there are many assumptions, each with its own degree of uncertainty, which had to be made in the course of resource portfolio evaluations, material changes in these assumptions could result in modifications. The action plan presented in this IRP is sufficiently flexible to accommodate possible changes in key parameters, including load growth, environmental compliance assumptions, fuel costs, and construction cost estimates, which may impact this IRP. By minimizing PSO s costs in the optimization process, the Company s model produced optimized portfolios with the lowest, reasonable impact on customers rates. 9

23 2015 Integrated Resource Plan 1.0 Introduction 1.1 Overview This report presents the Integrated Resource Plan (IRP or Plan) for Public Service Company of Oklahoma (PSO or Company) including descriptions of assumptions, study parameters, and methodologies. The results incorporate the integration of traditional and renewable supply-side resources and demand-side management (DSM) programs. The goal of the IRP process is to demonstrate the amount, timing and type of resources required to ensure a reliable supply of power and energy to customers at the least reasonable cost. In addition to developing a long-term strategy for achieving reliability/reserve margin requirements as set forth by the Southwest Power Pool (SPP), capacity resource planning is critical to PSO due to its impact on: Determining Capital Expenditure Requirements which represents one of the basic elements of the Company s long-term business plan. Regulatory Filings this planning process is a critical component of recovery filings that will reflect input based on a prudent planning process. Integration with other Strategic Business Initiatives generation/capacity resource planning is naturally integrated with the Company s current and anticipated environmental compliance, transmission planning, and other planning initiatives. 1.2 IRP Process This IRP covers the processes and assumptions required to develop the Plan for PSO. The IRP process consists of the following steps: Description of the Company, the resource planning process in general, and the implications of current issues as they relate to resource planning. Provide projected growth in demand and energy which serves as the underpinning of the plan. Identify and discuss renewable energy and demand-side options. 10

24 2015 Integrated Resource Plan Identify current supply resources, including projected changes to those resources (e.g., up-rates, de-rates, or retirements), and transmission system integration issues. Identify and describe supply-side resource options. Describe the analysis and assumptions that will be used to develop the plan such as Regional Transmission Organization (RTO) reserve margin criteria, and fundamental modeling parameters. Perform resource modeling and use the results to develop portfolios, including the selection of the ultimate plan. Present the findings and recommendations, Action Plan and, finally, Plan implications on PSO. 1.3 Introduction to PSO PSO s customers consist of both retail and wholesale customers located in the state of Oklahoma (Figure 1). The Company serves a population of approximately 1.9 million in a 29,000 square-mile area. Currently, PSO has approximately 541,000 retail customers in the State of Oklahoma. The peak load requirement of PSO s total retail and wholesale customers is seasonal in nature, with distinctive peaks occurring in the summer and winter seasons. PSO s historical all-time highest recorded peak demand was 4,410MW, which occurred in August 2012; and the highest recorded winter peak was 3,015MW, which occurred in February The most recent actual PSO summer and winter peak demands are 4,030MW and 2,973MW, occurring on July 7, 2014, and January 8, 2015, respectively. Figure 1. PSO Service Territory (Shown in Red) 11

25 2015 Integrated Resource Plan 2.0 Load Forecast and Forecast Methodology 2.1 Summary of PSO Load Forecast The PSO load forecast was developed by American Electric Power s (AEP) Economic Forecasting organization and completed in June The load forecast is the culmination of a series of underlying forecasts that build on each other. The economic forecast provided by Moody s Analytics is used to develop the customer count forecast which is then used to develop the sales forecast which is ultimately used to develop the peak load and internal energy requirements forecast. Over the next 10 year period ( ), PSO s service territory is expected to see population and non-farm employment growth of 0.6% 2 and 0.7% per year, respectively. Not surprisingly, PSO is projected to see customer count growth of 0.6% per year as well. Over the same forecast period, PSO s retail sales are projected to grow at 0.9% per year with stronger growth expected from the industrial class (+2.2% per year) while the residential class experiences a modest decline -0.1% per year over the forecast horizon. The projected growth in PSO s internal energy over the next 10 years is expected to be stronger (0.9% per year) than the expected growth in customer counts (0.6% per year). This variation is the result of strong industrial growth in the early years of the forecast. Finally, PSO s peak demand is expected to grow at an average rate of 0.7% per year through Forecast Assumptions Economic Assumptions The load forecasts for PSO incorporate a forecast of U.S. and regional economic growth provided by Moody s Analytics. The load forecasts utilized Moody s Analytics economic forecast issued in January Moody s Analytics projects moderate growth in the U.S. 1 The load forecasts (as well as the historical loads) presented in this report reflect the traditional concept of internal load, i.e., the load that is directly connected to the utility s transmission and distribution system and that is provided with bundled generation and transmission service by the utility. Such load serves as the starting point for the load forecasts used for generation planning. Internal load is a subset of connected load, which also includes directly connected load for which the utility serves only as a transmission provider. Connected load serves as the starting point for the load forecasts used for transmission planning. 2 All yearly rates are average compound annual growth rates unless specified otherwise. 12

26 2015 Integrated Resource Plan economy during the forecast period, characterized by a 2.2% annual rise in real Gross Domestic Product (GDP), and moderate inflation as well, with the implicit GDP price deflator expected to rise by 2.1% per year. Industrial output, as measured by the Federal Reserve Board's (FRB) index of industrial production, is expected to grow at 1.6% per year during the same period. Moody s projected employment growth of 0.7% per year during the forecast period and real regional income per-capita annual growth of 1.8% for the PSO service area Price Assumptions The Company utilizes an internally developed service area electricity price forecast. This forecast incorporates information from the Company s financial plan for the near term and the U.S. Energy Information Administration (EIA) outlook for the West South Central Census Region for the longer term. These price forecasts are incorporated into the Company s energy sales models, where appropriate Specific Large Customer Assumptions PSO s customer service engineers maintain frequent contact with industrial and commercial customers about their needs and future plans. Expected load additions or reductions are collected and incorporated into the Company s load projections where appropriate based on these discussions Weather Assumptions The Company also includes weather as an explanatory variable in its energy sales and peak demand models. These models reflect historical weather for the model estimation period and normal weather for the forecast period Demand-Side Management (DSM) Assumptions Inherent in the historical data used to specify the load forecast models are the impacts of past customer energy conservation and load management behaviors. Energy usage is being impacted by a combination of federal and/or state efficiency mandates in addition to company sponsored energy efficiency and DSM programs. The statistical adjusted end-use models 13

27 2015 Integrated Resource Plan incorporate changing saturations and efficiencies of the various end-use appliances which results in a certain amount of energy efficiency to be embedded into the load forecast. In addition to the embedded energy efficiency, the Company also accounts for Commission approved DSM program impacts in the load forecasting process. New or incremental DSM resources over-and-above those levels are analyzed and projected separately as part of the IRP development process. 2.3 Overview of Forecast Methodology PSO's load forecasts are derived from mostly econometric, statistically adjusted end-use models in addition to analysis of time-series data. This is helpful when analyzing future scenarios and developing confidence bands in addition to objective model verification by using standard statistical criteria. PSO utilizes two sets of econometric models: 1) a set of monthly short-term models which extends for approximately 24 months and 2) a set of monthly long-term models which extends for approximately 30 years. The forecast methodology leverages the relative analytical strengths of both the short- and long-term methods to produce a reasonable and reliable forecast that is used for various planning purposes. For the first full year of the forecast, the values are generally governed by the short-term models. The short-term models are regression models with time series errors which analyze the latest sales and weather data to better capture the monthly variation and patterns in energy sales for short-term applications like capital budgeting and resource allocation. While these models generally produce more accurate forecasts in the short run, without logical ties to economic factors they are less capable of capturing structural trends in electricity consumption that are more important for longer term resource planning applications. The long term models are econometric, and Statistically Adjusted End-use (SAE) models which are specifically equipped to account for structural changes in the economy as well as changes in customer consumption due to increased energy efficiency. The long-term forecast models incorporate regional economic forecast data for income, employment, households, output, and population. 14

28 2015 Integrated Resource Plan The short-term and long-term forecasts are then blended to ensure a smooth transition from the short-term to the long-term forecast horizon for each major revenue class. There are some instances when the short-term and long-term forecasts diverge, especially when the longterm models are incorporating a structural shift in the underlying economy that is expected to occur within the first 24 months of the forecast horizon. In these instances, professional judgment is used to ensure that the final forecast that will be used in the peak models is reasonable. The class level sales are then summed and adjusted for losses to produce monthly net internal energy sales for the system. The demand forecast model utilizes a series of algorithms and load shapes to allocate the monthly net internal energy to hourly demand. The inputs into forecasting hourly demand are internal energy, weather, 24-hour load profiles and calendar information. A flow chart depicting the sequence of models used in projecting PSO s electric load requirements as well as the major inputs and assumptions that are used in the development of the load forecast is shown in Figure 2 below. Figure 2. PSO Internal Energy Requirements and Peak Demand Forecasting Method 15

29 2015 Integrated Resource Plan 2.4 Detailed Explanation of Load Forecast Methodology General This section provides a more detailed description of the short-term and long-term models employed in producing the forecasts of PSO s energy consumption, by customer class. Conceptually, the difference between short and long term energy consumption relates to changes in the stock of electricity-using equipment and economic influences, rather than the passage of time. In the short term, electric energy consumption is considered to be a function of an essentially fixed stock of equipment. For residential and commercial customers, the most significant factor influencing the short-term is weather. For industrial customers, economic forces that determine inventory levels and factory orders also influence short-term utilization rates. The short-term models recognize these relationships and use weather and recent load growth trends as the primary variables in forecasting monthly energy sales. Over time, demographic and economic factors such as population, employment, income, and technology influence the nature of the stock of electricity-using equipment, both in size and composition. Long-term forecasting models recognize the importance of these variables and include all or most of them in the formulation of long-term energy forecasts. Relative energy prices also have an impact on electricity consumption. One important difference between the short-term and long-term forecasting models is their treatment of energy prices, which are only included in long-term forecasts. This approach makes sense because although consumers may incur increased electricity costs from energy price fluctuations, there is little they can do to impact them in the short-term. They already own a refrigerator, furnace or industrial equipment that may not be the most energy-efficient model available. In the long term, however, these constraints are lessened as durable equipment is replaced and as price expectations come to fully reflect price changes Customer Forecast Models The Company also utilizes both short-term and long-term models to develop the final customer count forecast. The short-term customer forecast models are time series models with intervention, when needed, using Autoregressive Integrated Moving Average (ARIMA) methods of estimation. These models typically extend for 24 months into the forecast horizon. 16

30 2015 Integrated Resource Plan The long-term residential customer forecasting models are also monthly but extend for 30 years. The explanatory economic and demographic variables include mortgage interest rates, real personal income, population and households are used in various combinations for each jurisdiction. In addition to the economic explanatory variables, the long-term customer models employ a lagged dependent variable to capture the adjustment of customer growth to changes in the economy. There are also binary variables to capture monthly variations in customers, unusual data points and special occurrences. The short-term and long-term customer forecasts are blended as was described earlier to arrive at the final customer forecast that will be used as a primary input into both short-term and long-term usage forecast models Short-term Forecasting Models The goal of PSO's short-term forecasting models is to produce an accurate load forecast for the first full year into the future. To that end, the short-term forecasting models generally employ a combination of monthly and seasonal binaries, time trends, and monthly heating cooling degree-days in their formulation. The heating and cooling degree-days are measured at weather stations in the Company's service area. The forecasts rely on ARIMA models. The estimation period for the short-term models was January 2005 through January Residential and Commercial Energy Sales Residential and commercial energy sales are developed using ARIMA models to forecast usage per customer and number of customers. The usage models relate usage to lagged usage, lagged error terms, heating and cooling degree-days and binary variables. The customer models relate customers to lagged customers, lagged error terms and binary variables. The energy sales forecasts are a product of the usage and blended customer forecasts Industrial Energy Sales Short-term industrial energy sales are forecast separately for the largest industrial customers of PSO and for the remainder of industrial energy. These short-term industrial energy sales models relate energy sales to lagged energy sales, lagged error terms and binary variables for each of the Company s jurisdictions. The industrial models are estimated using ARIMA 17

31 2015 Integrated Resource Plan models. The short-term industrial energy sales forecast is a sum of the forecasts for the largest industrial customers and the forecasts for the remainder of the manufacturing customers. Customer service engineers also provide input into the forecast for specific large customers All Other Energy Sales The All Other Energy Sales category for PSO includes public authorities, public street and highway lighting (or other retail sales for those two items combined) and sales to its wholesale customer. PSO s wholesale requirements customer is the Town of South Coffeyville. Both the other retail and municipal models are estimated using ARIMA models. PSO's short-term forecasting model for other retail energy sales includes binaries, heating and cooling degree-days, and lagged energy sales. The sales-for-resale model includes binaries, heating and cooling degree-days, and lagged energy sales. Off-system sales and/or sales of opportunity are not relevant to the net energy requirements forecast as they are not requirements load or part of the IRP process Long-term Load Forecasting Models The goal of the long-term forecasting models is to produce a reasonable load outlook for up to 30 years in the future. Given that goal, the long-term forecasting models employ a full range of structural economic and demographic variables, electricity and natural gas prices, weather as measured by annual heating and cooling degree-days, and binary variables to produce load forecasts conditioned on the outlook for the U.S. economy, for the PSO service-area economy, and for relative energy prices. Most of the explanatory variables enter the long-term forecasting models in a straightforward, untransformed manner. In the case of energy prices, however, it is assumed, consistent with economic theory, that the consumption of electricity responds to changes in the price of electricity or substitute fuels with a lag, rather than instantaneously. This lag occurs for reasons having to do with the technical feasibility of quickly changing the level of electricity use even after its relative price has changed, or with the widely accepted belief that consumers make their consumption decisions on the basis of expected prices, which may be perceived as functions of both past and current prices. 18

32 2015 Integrated Resource Plan There are several techniques, including the use of lagged price or a moving average of price that can be used to introduce the concept of lagged response to price change into an econometric model. Each of these techniques incorporates price information from previous periods to estimate demand in the current period. The general estimation period for the long-term load forecasting models was The long-term energy sales forecast is developed by blending the short-term forecast with the long-term forecast. The energy sales forecast is developed by making a billed/unbilled adjustment to derive billed and accrued values, which are consistent with monthly generation Supporting Models In order to produce forecasts of certain independent variables used in the internal energy requirements forecasting models, several supporting models are used, including a natural gas price model for the State of Oklahoma. These models are discussed below Consumed Natural Gas Pricing Model The forecast price of natural gas used in the Company's energy models comes from a model of state natural gas prices for three primary consuming sectors: residential, commercial, and industrial. In the state natural gas price models sectoral prices are related to the West South Central Census region, with the forecast being obtained from U.S. Department of Energy (DOE) /EIA s 2015 Annual Energy Outlook. The natural gas price model is based upon historical data Residential Energy Sales The residential usage model is estimated using an SAE model, which was developed by Itron, a consulting firm with expertise in energy modeling. This model assumes that usage will fall into one of three categories: heating, cooling and other. The SAE model constructs variables to be used in an econometric equation where residential usage is a function of Xheat, Xcool and Xother variables. The Xheat variable is derived by multiplying a heating index variable by a heating use variable. The heating index incorporates information about heating equipment saturation; heating equipment efficiency standards and trends; and thermal integrity and size of homes. The heating 19

33 2015 Integrated Resource Plan use variable is derived from information related to billing days, heating degree-days, household size, personal income, gas prices and electricity prices. The Xcool variable is derived by multiplying a cooling index variable by a cooling use variable. The cooling index incorporates information about cooling equipment saturation; cooling equipment efficiency standards and trends; and thermal integrity and size of homes. The cooling use variable is derived from information related to billing days, heating degree-days, household size, personal income, gas prices and electricity prices. The Xother variable estimates the non-weather sensitive sales and is similar to the Xheat and Xcool variables. This variable incorporates information on appliance and equipment saturation levels; average number of days in the billing cycle each month; average household size; real personal income; gas prices and electricity prices. The appliance saturations are based on historical trends from PSO s residential customer survey. The saturation forecasts are based on DOE/EIA forecasts and analysis by Itron. The efficiency trends are based on DOE forecasts and Itron analysis. The thermal integrity and size of homes are for the West South Central Census Region and are based on DOE and Itron data. The number of billing days is from internal data. Economic and demographic forecasts are from Moody s Analytics and the electricity price forecast is developed internally. The SAE residential model is estimated using a linear regression model. This monthly model was for the period January 2005 through January It is important to note, as will be discussed later in this document, that this modeling has incorporated the reductive effects of the Energy Policy Act of 2005 (EPAct), the Energy Independence and Security Act of 2007 (EISA), American Recovery and Reinvestment Act of 2009 (ARRA) and Energy Improvement and Extension Act of 2008 (EIEA2008) on residential and commercial energy usage. The long-term residential energy sales forecast is derived by multiplying the blended customer forecast by the usage forecast from the SAE model Commercial Energy Sales Long-term commercial energy sales are forecast using a SAE model. This models is similar to the residential SAE models, where commercial usage is a function of Xheat, Xcool and Xother variables. 20

34 2015 Integrated Resource Plan As with the residential model, Xheat is determined by multiplying a heating index by a heat use variable. The variables incorporate information on heating degree-days, heating equipment saturation, heating equipment operating efficiencies, square footage, average number of days in a billing cycle, commercial output and electricity price. The Xcool variable uses measures similar to the Xheat variable, except it uses information on cooling degree-days and cooling equipment, rather than those items related to heating load. The Xother variable measures the non-weather sensitive commercial load. It uses nonweather sensitive equipment saturations and efficiencies, as well as billing days, commercial output and electricity price information. The saturation, square footage and efficiencies are from the Itron base of DOE data and forecasts. The saturations and related items are from EIA s 2014 Annual Energy Outlook. Billing days and electricity prices are developed internally. The commercial output measure is real commercial gross regional product from Moody s Analytics. The equipment stock and square footage information are for the West South Central Census Region. The SAE is a linear regression for the period which is January 1996 through January As with the residential SAE model, the effects of EPAct, EISA, ARRA and EIEA2008 are captured in this model Industrial Energy Sales The Company uses the following economic and pricing explanatory variables: service area gross regional product manufacturing, FRB industrial production index for manufacturing, and service area industrial electricity prices. In addition binary variables for months or special occurrences are incorporated into the model. Based on information from customer service engineers there may be load added or subtracted from the model results to reflect plant openings, closures or load adjustments. The last actual data point for the industrial energy sales model is January All Other Energy Sales The forecast of other retail energy sales relates energy sales to service area population and binary variables. 21

35 2015 Integrated Resource Plan Wholesale energy sales are modeled relating energy sales to service area population and heating and cooling degree-days. Binary variables are necessary to account for discrete changes in energy sales that result from events such as the addition of new customers Final Monthly Internal Energy Forecast Blending Short and Long-Term Forecasts Forecast values for 2015 and 2016 are taken from the short-term process. Forecast values for 2017 are obtained by blending the results from the short-term and long-term models. The blending process combines the results of the short-term and long-term models by assigning weights to each result and systematically changing the weights so that by July of 2017 the entire forecast is from the long-term models. The goal of the blending process is to leverage the relative strengths of the short-term and long-term models to produce the most reliable forecast possible. However, at times the short-term models may not capture structural changes in the economy as well as the long-term models, which may result in the long-term forecast being used for the entire forecast horizon Large Customer Changes The Company s customer service engineers are in continual contact with the Company s large commercial and industrial customers about their needs for electric service. These customers relay information about load additions and reductions. This information will be compared with the load forecast to determine if the industrial or commercial models are adequately reflecting these changes. If the changes are different from the model results, then add factors may be used to reflect those large changes that are different from those from the forecast models output Losses and Unaccounted-For Energy Energy is lost in the transmission and distribution of the product. This loss of energy from the source of production to consumption at the premise is measured as the average ratio of all Federal Energy Regulatory Commission (FERC) revenue class energy sales measured at the premise meter to the net internal energy requirements metered at the source. In modeling, 22

36 2015 Integrated Resource Plan company loss study results are applied to the final blended sales forecast by revenue class and summed to arrive at the final internal energy requirements forecast Forecast Methodology for Seasonal Peak Internal Demand The demand forecast model is a series of algorithms for allocating the monthly internal energy sales forecast to hourly demands. The inputs into forecasting hourly demand are blended revenue class sales, energy loss multipliers, weather, 24-hour load profiles and calendar information. The weather profiles are developed from representative weather stations in the service area. Twelve monthly profiles of average daily temperature that best represent the cooling and heating degree-days of the specific geography are taken from the last 30 years of historical values. The consistency of these profiles ensures the appropriate diversity of the company loads. The 24-hour load profiles are developed from historical hourly company or jurisdictional load and end-use or revenue class hourly load profiles. The load profiles were developed from segregating, indexing and averaging hourly profiles by season, day types (weekend, midweek and Monday/Friday) and average daily temperature ranges. In the end, the profiles are benchmarked to the aggregate energy and seasonal peaks through the adjustments to the hourly load duration curves of the annual 8,760 hourly values. These 8,760 hourly values per year are the forecast load of PSO and the individual companies of AEP that can be aggregated by hour to represent load across the spectrum from end-use or revenue classes to total AEP-East, AEP-West, or total AEP system. Net internal energy requirements are the sum of these hourly values to a total company energy need basis. Company peak demand is the maximum of the hourly values from a stated period (month, season or year). 2.5 Load Forecast Results and Issues Load Forecast Table 1 presents PSO's annual internal energy requirements, disaggregated by major category (residential, commercial, industrial and other energy, which is comprised of other retail sales, wholesale sales and losses) on an actual basis for the years and on a forecast basis for the years Table 1 includes annual growth rates for both the historical and 23

37 2015 Integrated Resource Plan forecast periods. The data for 2015 are three months actual and nine months forecast. As mentioned earlier, PSO s internal energy requirements are expected to increase at a rate of 0.9% per year through Most of the projected growth is coming from the industrial class (+2.2% per year), while the commercial class is expected to see modest growth (+0.5% per year) and the residential class sales experience a slight decline (-0.1% per year) through PSO s Industrial sales started to experience significant growth in 2014 (+3.0%) and are expected to continue to see strong growth through Most of the growth in industrial sales is related to oil & gas related activity that has already occurred or is expected to come online by Table 1. PSO Actual and Forecast Internal Energy (GWh) Requirements by Sector with Growth Rates (%) Year Residential Growth Rate Commercial Growth Rate Industrial Growth Rate Other* Energy Requirements Growth Rate Total Internal Energy Requirements ACTUAL , , , , , , , , , , , , , , , FORECAST , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , *Other energy requirements include other retail sales, wholesale sales and losses. Note: 2015 data are three months actual and nine months forecast Growth Rate Peak Demand and Load Factor Table 2 provides PSO s seasonal peak demands, annual peak demand, internal energy requirements and annual load factor on an actual basis for the years and on a forecast basis for the year The table also shows annual growth rates for both the historical and forecast periods. The data for 2015 are three months actual and nine months forecast. 24

38 2015 Integrated Resource Plan Table 2. PSO Summer, Winter, and Annual Peak Demand (MW), Internal Energy Requirements (GWh) and Load Factor (%) Year Summer Peak Demand Preceding Winter Peak Demand Annual Peak Demand Internal Energy Requirements Load Factor ACTUAL ,410 2,530 4,410 19, ,987 2,632 3,987 18, ,030 3,015 4,030 19, FORECAST ,222 2,974 4,222 19, ,304 2,909 4,304 19, ,349 2,958 4,349 20, ,362 2,966 4,362 20, ,380 2,984 4,380 20, ,385 2,985 4,385 20, ,414 3,016 4,414 20, ,440 3,037 4,440 20, ,463 3,057 4,463 20, ,486 3,064 4,486 21, Note: 2015 data are three months actual and nine months forecast Energy Efficiency (EE) Impacts on the Load Forecast Table 3 provides the impacts of planned EE programs on energy and demand incorporated in PSO s load forecast. These values are determined by taking the most recent approved EE program details and correlating them with PSO s load forecast, in order to generate the true impact of the programs on the Company s energy use and peak demand. Table 3. PSO EE in Load Forecast Energy (GWh) and Coincident Peak Demand (MW) Year Energy (GWh) Summer* Demand (MW) *Demand coincident with Company's seasonal peak demand. 25 Winter* Demand (MW)

39 2015 Integrated Resource Plan Blended Load Forecast As noted above, at times the short-term models may not capture structural changes in the economy as well as the long-term models, which may result in the long-term forecast being used for the entire forecast horizon. Table 4 provides an indication of which retail models are blended and which strictly use the long-term model results. In addition, the wholesale forecast utilizes the long-term forecast model results. Table 4. PSO Short-Term Load Forecast Blended Forecast vs. Long-Term Model Results Class Residential Commercial Industrial Other Retail Retail Model Long-Term Blend Long-Term Blend Large Customer Changes The Company s customer service engineers are in continual contact with the Company s large commercial and industrial customers about their needs for electric service. These customers will relay information about load additions and reductions. This information will be compared with the load forecast to determine if the industrial or commercial models are adequately reflecting these changes. If the changes are different from the model results, then add factors may be used to reflect those large changes that are different from those from the forecast models output Price Elasticity The long term load forecast models include electricity price as one of many explanatory variables. The coefficient of the electricity price variable is an estimate of the price elasticity, which is simply a measure of how responsive customers are to changes in price. The formula for price elasticity is simply the percentage change in the quantity demanded divided by the percentage change in price. If the change in demand is greater than the change in price, the elasticity estimate would be greater than 1 and it would be described as elastic demand. If the change in demand is less than the change in price, the elasticity estimate would be less than 1 and it would be classified as inelastic demand. The demand for electricity is very inelastic. For 26

40 2015 Integrated Resource Plan the Residential class, the long term elasticity estimate is approximately 0.1. For the Commercial class, the modeled price elasticity is 0.15 and the elasticity estimate for the Industrial class is For comparison, the estimated long term elasticity for gasoline is 0.6 while the elasticity for restaurant meals is (Note: technically each of these elasticity estimates are negative values based on the inverse relationship between price and quantity demanded. The convention by economists when describing the elasticity is to report the absolute value of these elasticity estimates.) 2.6 Load Forecast Trends & Issues Changing Usage Patterns Over the past decade, there has been a significant change in the trend for electricity usage from prior decades. Figure 3 presents PSO s historical and forecasted residential and commercial usage per customer between 1991 and During the first decade shown ( ), residential usage per customer grew at an average rate of 1.6% per year, while the commercial usage grew by 0.1% per year. Over the next decade ( ), growth in residential usage slowed to 0.6% per year while the commercial class usage actually declined by 1.1% per year. In the last decade shown ( ) residential usage is projected to decline at a rate of 0.1% per year while the commercial usage is relatively flat. 3 O Sullivan, Arthur, Steven M. Sheffrin, & Stephen J. Perez Survey of Economics: Principles, Applications, and Tools. Prentice Hall 2012 Table 4.2 Price Elasticities of Demand for Selected Products pg

41 2015 Integrated Resource Plan Figure 3. PSO Residential and Commercial Normalized Use per Customer (kwh) Energy Efficiency Embedded in the Load Forecast The SAE models are designed to account for changes in the saturations and efficiencies of the various end-use appliances. Every 3-4 years, the Company conducts a Residential Appliance Saturation Survey to monitor the saturation and age of the various appliances in the residential home. This information is then matched up with the saturation and efficiency projections from the EIA which includes the projected impacts from the various enacted federal policy mentioned earlier. The result of this is a base-load forecast that already includes some significant reductions in usage as a result of projected EE. For example, Figure 4 below shows the assumed cooling efficiencies embedded in the statistically adjusted end-use models for cooling loads. It shows that the average SEER (Seasonal Energy Efficiency Ratio) for central air conditioning is projected to increase from 13.1 in 2010 to over 14.8 by The chart shows a similar trend in projected cooling efficiencies for heat pump cooling as well as room air conditioning units as well. 28

42 2015 Integrated Resource Plan Figure 4. PSO Projected Cooling Efficiencies, Cooling loads are certainly not the only appliances assumed to see significant increases in appliance efficiency. Figure 5 below shows the projected energy usage for lighting as well as clothes dryers and in both instances, you see a dramatic decline in the average energy usage. 29

43 2015 Integrated Resource Plan Figure 5. PSO Select Appliance Efficiencies, Load Forecast Scenarios The base case load forecast is the most likely load growth assumption that the Company uses for planning. There are a number of known and unknown potentials that could drive load growth to be different from the base case. While potential scenarios could be quantified at varying levels of assumptions and preciseness, the Company has chosen to frame the possible outcomes around the base case. Forecast sensitivity scenarios have been established which are tied to respective high and low economic growth cases. The high and low economic growth scenarios are consistent with scenarios laid out in the EIA s 2015 Annual Energy Outlook. While other factors may affect load growth, this analysis only considered high and low economic growth. The economy is seen as a crucial factor affecting future load growth Low Load Sensitivity Case The Low Load forecast reflects the impact of low economic growth for the region and is consistent with the low economic growth presented by EIA. 30

44 2015 Integrated Resource Plan The Low Load forecast projects firm peak load growth to average -0.1% per year on a compound basis. Total energy growth is also projected to average about 0.1% per year. The load factor is unchanged from the Base Case at about 53%. The Low Load forecast for energy is 5.8% below the Base Case forecast in High Load Sensitivity Case The High Load forecast represents a scenario of more sustained regional growth. As with the Low Load Forecast the high economic growth scenario is consistent with EIA high growth in its economic scenario. The High Load forecast projects firm peak load growth to average 1.0% per year. Energy growth is also projected to average 1.2 % per year with a load factor of 53%. The High Load forecast for energy is 5.2% above the Base Case forecast in

45 2015 Integrated Resource Plan 3.0 Existing Resources and Transactions 3.1 Current Resources An initial step in the IRP process is the demonstration of the capacity resource requirements. This needs assessment must consider projections of: Existing capacity resources current levels and anticipated changes Anticipated changes in capability Changes resulting from decisions surrounding unit disposition evaluations Regional and sub-regional capacity and transmission constraints/limitations Current DSM Load and Peak Demand SPP capacity reserve margin and reliability criteria 3.2 Existing PSO Generating Resources The underlying minimum reserve margin criterion to be utilized in the determination of PSO s capacity needs is based on the current SPP minimum capacity margin of 12 percent. 4 As a function of peak demand this converts to an equivalent reserve margin of 13.6 percent. 5 The SPP minimum capacity margin, and its adequacy at the level of 12 percent, is currently being reviewed by an SPP-led Capacity Margin Task Force, of which AEP is a participant. Table 5 provides a summary of the Company s Capacity, Demand and Reserves (CDR) report for the 10-year period through the year 2024 assuming no new capacity additions. This is considered to be the Company s gong-in position. In addition to identifying current projected peak demand requirements of its internal customers, it also identifies the MW capability of resources that are projected to be required to meet the minimum SPP reserve margin criterion. For instance, at the beginning of the first forecasted SPP planning year, , it indicates PSO is expected to rely on 4,245MW of owned generating capability (seasonal ratings) to achieve this threshold. 4 Per Section of the Southwest Power Pool Criteria (Latest Revision: April 28, 2015) / (1 0.12) = For capacity planning/reporting purposes, SPP operates on a June (Year X) -through- May (Year X+1) planning year basis. 32

46 2015 Integrated Resource Plan Table 5. PSO Going-In Capacity, Demand and Reserves (CDR) CAPABILITY (MW) 14 ACT Plant Capabilities 4,272 4,245 3,778 3,778 3,778 3,778 3,778 3,778 3,710 3,560 3,481 Adjustments to Plant Capabilities Net Plant Capability 4,272 4,245 3,769 3,746 3,746 3,746 3,746 3,746 3,678 3,528 3,449 Sales Without Reserves Purchases Without Reserves ,122 1,162 1,182 1,208 1, Total Capability 4,873 4,847 4,891 4,908 4,928 4,954 4,954 4,624 4,037 3,887 3,808 DEMAND (MW) 14 ACT Peak Demand 4,030 4,222 4,304 4,349 4,362 4,389 4,413 4,453 4,482 4,507 4,533 Active DSM Firm Demand 3,996 4,131 4,225 4,257 4,260 4,290 4,319 4,365 4,399 4,429 4,454 Other Demand Adjustments Native Load Responsibility 3,996 4,107 4,201 4,234 4,237 4,268 4,297 4,342 4,376 4,406 4,431 Sales With Reserves Purchases With Reserves Load Responsibility 3,957 4,068 4,162 4,195 4,198 4,229 4,258 4,303 4,337 4,367 4,392 PSO RESERVES 14 ACT Reserve Capacity % Reserve Margin % Capacity Margin Reserve Above 12% Capacity Margin Table 6 below identifies the generating resources identified in the CDR. Future plans surrounding these assets must take into account each unit s useful service life. Unit retirements are incorporated in PSO s plans based upon each unit s in-service date along with the anticipated service life. Retirement dates are continually reviewed and adjusted with respect to a unit s ability to maintain safe and reliable operation, as well as external factors such as environmental regulations. PSO currently utilizes several additional capacity entitlements to meet the minimum SPP reserve margin requirement. Beginning in 2012, PSO began to receive approximately 520MW of generating capacity under a 10-year Power Purchase Agreement (PPA) with Exelon Generating Company LLC, from the Green Country Generating Station located in Jenks, OK. 33

47 2015 Integrated Resource Plan Table 6. PSO Owned Generation Assets as of August, 2015 Additionally, PSO currently has a total of 690MW (nameplate rating) of wind capacity from five wind facilities in which the Company is receiving energy, capacity, and renewable energy credit attributes under separate renewable energy PPAs. Starting in 2016 PSO will begin delivery for approximately 600MW (nameplate) of new wind generation. For capacity resource planning purposes, however, an important distinction is that SPP criteria also dictates that intermittent resources such as wind may only recognize a small portion of such nameplate capacity rating. Specifically, the latest SPP Criteria (April 28, 2015 Revision), Section , describes the approach for establishing such seasonal ratings for wind (and solar) resources. Using those guidelines, capacity credit of 83MW is used capacity planning purposes in Capacity Needs Assessment COAL UNITS NAMEPLATE (MW) LOCATION IN-SERVICE YEAR Oklaunion Unit 1 102* Vernon, TX 1986 Northeastern Unit Oologah, OK 1979 Northeastern Unit Oologah, OK 1980 GAS STEAM UNITS NAMEPLATE (MW) LOCATION IN-SERVICE YEAR Northeasten Unit Oologah, OK 1970 Riverside Unit Jenks, OK 1974 Riverside Unit Jenks, OK 1976 Southwestern Unit 1 78 Anadarko, OK 1952 Southwestern Unit 2 79 Anadarko, OK 1954 Southwestern Unit Anadarko, OK 1967 Tulsa Unit Tulsa, OK 1956 Tulsa Unit Tulsa, OK 1958 GAS COMBINED-CYCLE UNITS NAMEPLATE (MW) LOCATION IN-SERVICE YEAR Comanche Unit Lawton, OK 1973 Northeastern Unit Oologah, OK 1961** GAS COMBUSTION TURBINE UNITS NAMEPLATE (MW) LOCATION IN-SERVICE YEAR Weleetka Unit 4 55 Weleetka, OK 1975 Weleetka Unit 5 54 Weleetka, OK 1976 Weleetka Unit 6 54 Weleetka, OK 1976 Riverside Unit 3 76 Jenks, OK 2008 Riverside Unit 4 76 Jenks, OK 2008 Southwestern Unit 4 85 Anadarko, OK 2008 Southwestern Unit 5 85 Anadarko, OK 2008 ** Represents PSO's PSO s 15.62% ownership stake in Oklaunion ** Repowered from Natural Gas Steam to Natural Gas Combined Cycle in 2001 Based on the assessment of the AEP-SPP current resources and peak demand projections (Section 2.5.2); a capacity needs assessment can be established that will determine the amount and timing of capacity resources for this IRP. 34

48 2015 Integrated Resource Plan Figure 6 summarizes the going-in capacity position through the 10-year 2015 IRP window. Figure 7 compares the demand (line) and total capacity (bar) trends over the period, illustrating PSO s net capacity position with respect to the company s load obligation, and with respect to SPP s 12% capacity margin requirement. 35

49 2015 Integrated Resource Plan Figure 6. PSO Going-In SPP Capacity Position (MW) and Obligation (MW) Figure 7. PSO Capacity Positions (MW) net of SPP Requirements 36

50 2015 Integrated Resource Plan 3.4 Environmental Issues and Implications Introduction There are two main current EPA rules requiring PSO to install control equipment to meet emission limits in specific time frames: (1) the Regional Haze Rule (RHR) and (2) the Mercury and Air Toxics Standard (MATS) rule. These rules were finalized on December 28, 2011, and February 16, 2012, respectively. These rules require PSO to make substantial reductions in emissions of Sulfur Dioxide (SO 2 ), Nitrogen Oxide (NO x ), mercury, acid gases, and particulate matter (PM). PSO has chosen technologies that enable it to comprehensively comply with the requirements of the RHR and MATS, particularly at the coal-fired Northeastern Units 3 and 4, while preserving its flexibility to address future environmental requirements cost-effectively. To meet this desire, PSO executed an environmental compliance strategy to resolve PSO s emission control requirements for Northeastern Units 3 and 4 and to ensure sufficient resources to meet customer demands Regional Haze Rule (RHR) On July 6, 2005, the EPA published the final Regional Haze Regulations and Guidelines for Best Available Retrofit Technology Determinations (BART). The Federal Clean Air Act and the Regional Haze Rule require certain states, including Oklahoma, to make reasonable progress toward the prevention of any future, and the remedying of any existing, impairment of visibility in mandatory Class I Federal areas. Moreover, the Regional Haze Rule requires the State of Oklahoma to develop programs to address regional haze in each mandatory Class I Federal area located within the State and in each mandatory Class I Federal area located outside the State which may be affected by emissions from within the State. Air pollutants emitted by BART eligible sources in Oklahoma, which may reasonably be anticipated to cause or contribute to visibility impairment, in any mandatory Class I Federal area are NO x, SO 2, PM-10, and PM EPA also provided guidance on what level of control is reasonable for certain BARTeligible sources, including Electric Generating Units (EGUs), and published presumptive BART emission rates for SO 2 and NOx based on the types of cost-effective controls available. In November 2012, PSO reached an agreement with the Federal EPA, the State of Oklahoma and other parties that would provide for submission of a revised Regional Haze State 37

51 2015 Integrated Resource Plan Implementation Plan (SIP) requiring the retirement of one coal-fired unit of PSO s Northeastern Station no later than April 2016, and the installation of Dry Sorbent Injection (DSI), Activated Carbon Injection (ACI), and Fabric Filter (FF) emission control systems on the second coal-fired Northeastern unit by April 2016, with retirement of the second unit no later than EPA proposed to approve this revision to the Oklahoma Regional Haze SIP on August 21, 2013, with final approval published in the Federal Register on March 7, The retirement of Northeastern Unit 4 and the installation of emission retrofits at Northeastern Unit 3 have been modeled as a base assumption. In addition, PSO has installed a combination of low NO X burners and Separated Over- Fire Air (SOFA) combustion modifications on Northeastern Units 2, 3, and 4, and Southwestern Unit 3. The Comanche Units 1G1 and 1G2 burner retrofit is scheduled to begin construction in August 2015 with an in service date of June Mercury and Air Toxics Standards Rule (MATS) The final MATS Rule became effective on April 16, 2012, and required compliance by April 16, This rule regulates emissions of hazardous air pollutants from coal and oil-fired electric generating units. Hazardous air pollutants regulated by this rule are: 1) mercury; 2) certain non-mercury metals such as arsenic, lead, cadmium and selenium; 3) certain acid gases, including Hydrochloric Acid (HCl); and 4) certain organic hazardous air pollutants. The MATS Rule establishes stringent emission rate limits for mercury, filterable Particulate Matter as a surrogate for all non-mercury toxic metals, and HCl as a surrogate for all acid gases. Alternative emission limits were also established for the individual non-mercury metals and for SO 2 (alternate to HCl) for generating units that have operating Flue Gas Desulfurization (FGD) systems. The rule regulates organic hazardous air pollutants through work practice standards. Because the time required to complete construction of the controls on Northeastern Unit 3 and secure replacement capacity for the retiring Unit 4, exceeded the initial compliance deadline, PSO requested and received approval of a one-year extension from Oklahoma Department of Environmental Quality (ODEQ). Oklaunion installed ACI and a calcium bromide fuel additive system for MATS purposes. 38

52 2015 Integrated Resource Plan On November 25, 2014, the U.S. Supreme Court granted petitions to hear state and industry challenges against the EPA s MATS Rule to decide whether EPA unreasonably refused to consider costs in determining that it is appropriate to regulate hazardous air pollutants emitted by coal- and oil-fired electric generating units. The Supreme Court determined on June 29, 2015, that EPA must consider costs when deciding whether it is appropriate and necessary to regulate emissions under MATS. The decision did not vacate the MATS rule, but remanded the rule to the D.C. Circuit Court for further proceedings. MATS requirements remain effective unless otherwise ordered by the lower court Cross-State Air Pollution Rule (CSAPR) EPA developed the CSAPR to reduce the interstate transport of SO 2 and NOx from 28 states in the eastern half of the country to address associated concerns related to National Ambient Air Quality Standards (NAAQS) for ozone and particulate matter. CSAPR was finalized in 2011 as a replacement for the Clean Air Interstate Rule (CAIR). Along with other requirements, the final CSAPR established state-specific annual emission budgets for SO 2 and annual and seasonal budgets for NO x. Based on this budget, each emitting unit within an affected state was allocated a specified number of NO x and SO 2 allowances for the applicable compliance period, whether annual or ozone season. Allowance trading within and between states is allowed on a regional basis. Phase I of the CSAPR was originally intended to go into effect in January, The program was delayed as a result of complicated and lengthy litigation. Although the D.C. Circuit issued a decision in 2014 vacating and remanding the rule to EPA, the U.S. Supreme Court found the flaws identified by the D.C. Circuit did not justify vacating the rule. On remand, the D.C. Circuit held that the 2014 budgets for SO2 in four states, and the seasonal NOx budgets in 11 states were much more stringent than necessary to eliminate any significant contribution to any downwind non-attainment area. Oklahoma was added to the CSAPR program in a supplemental rule that is still on appeal before the D.C. Circuit. Phase 1 of the program took effect on January 1, 2015, and CSAPR remains in effect while EPA evaluates what changes to make to the rule. Unless modified, the CSAPR Phase 2 emission budgets will be applicable beginning in The rule currently is in effect in Oklahoma, and PSO and other electric 39

53 2015 Integrated Resource Plan generating companies have additional NO x emission reduction and compliance obligations as a result of the implementation of the CSAPR. The NOx controls installed on PSO s units to comply with the RHR also assist in compliance with the CSAPR National Ambient Air Quality Standards (NAAQS) The Clean Air Act requires EPA to establish and periodically review the NAAQS designed to protect public health and welfare. Several NAAQS have been recently revised or are under review, which could lead to more stringent SO 2 and NO X limits. This includes NAAQS for SO 2 (revised in 2010), Nitrogen Dioxide (NO 2 ) (revised in 2010), fine PM (revised in 2012), and ozone (proposed revision in 2014). The scope and timing of potential requirements is uncertain. Since both SO 2 and NO x are precursors for PM 2.5, and NO x is an ozone precursor, the controls being installed at Northeastern Unit 3 and the retirement of Northeastern Unit 4 will assist Oklahoma in achieving or maintaining these NAAQS. The precise impacts will not be known until the designations processes are completed and ODEQ completes its work to determine what additional reductions, if any, are required Coal Combustion Residuals (CCR) Rule EPA signed the final CCR Rule on December 19, This rule regulates CCR as a nonhazardous waste under Subtitle D of the Resource Conservation and Recovery Act and will become effective in October Preliminary review of this extensive rule indicates it is applicable to new and existing CCR landfills and CCR surface impoundments. It contains requirements for liner design criteria for new landfills, surface impoundment structural integrity requirements, CCR unit operating criteria, groundwater monitoring and corrective actions, closure and post-closure care, and recordkeeping, notification and internet posting obligations. EPA has not included a mandatory liner retrofit requirement for existing, unlined CCR surface impoundments, however operations must cease if groundwater monitoring data indicate there has been a release from the impoundment that exceeds applicable groundwater protections standards. PSO anticipates the need to conduct a dry bottom ash conversion for Northeastern Unit 3 in the 40

54 2015 Integrated Resource Plan 2017 to 2019 time frame, as well as close multiple ponds at the Oklaunion Plant prior to 2019 to comply with CCR Effluent Limitations Guidelines (ELG) The EPA proposed an update to the Effluent Limitation Guidelines and Standards (ELG Rule) for the steam electric power generating category in the Federal Register on June 7, The proposed ELG Rule would require more stringent controls on certain discharges from certain electric utility steam generating units or EGUs and will set technology-based limits for waste water discharges from power plants with a main focus on process water and wastewater from FGD, fly ash sluice water, bottom ash sluice water and landfill/pond leachate. The ELG Rule is currently a proposed rule. Per a Consent Decree entered into with various environmental groups resulting from litigation, EPA has agreed to sign a decision taking final action on the ELG Rule by September 30, Clean Water Act 316(b) Rule A final rule under Section 316(b) of the Clean Water Act was issued by EPA on August 15, 2014, with an effective date of October 14, 2014, and affects all existing power plants withdrawing more than two million gallons of cooling water per day. The rule offers seven technology options to comply with a standard that addresses impingement of aquatic organisms on cooling water intake screens and requires site-specific studies to determine appropriate compliance measures to address entrainment of organisms in cooling water systems for those facilities withdrawing more than 125 million gallons per day. The overall goal of the rule is to decrease impacts on fish and other aquatic organisms from operation of cooling water systems. Additional requirements may be imposed as a result of consultation with other federal agencies to protect threatened and endangered species and their habitats. Facilities with existing closed cycle recirculating cooling systems may not be required to make any technology changes. This determination would be made by the applicable state environmental agency during the plants next National Pollutant Discharge Elimination System (NPDES) permit renewal cycle. If additional capital investment is required, the magnitude is expected to be relatively small 41

55 2015 Integrated Resource Plan compared to the investment that could be needed if the plants were not equipped with cooling towers. PSO s generating plants may be required to make investments to upgrade cooling water intake structures as a result of this rule, and any requirement for this relatively modest cost will be determined through each plant s NPDES permitting cycle. At this time, the 316(b) Rule is not expected to require major capital investment, such as the addition of cooling towers, at any PSO plants Carbon Dioxide (CO 2 ) and Greenhouse Gas (GHG) Regulations On August 3, 2015, EPA finalized two rulemakings to regulate CO 2 emissions from fossil fuel-based electric generating units. EPA finalized New Source Performance Standards (NSPS) under Section 111(b) of the Clean Air Act that apply to new fossil units, as well as separate standards for modified or reconstructed existing fossil steam units. Separately, EPA finalized a rule referred to as the Clean Power Plan (CPP), which establishes CO 2 emission guidelines for existing fossil generation sources under Section 111(d) of the Clean Air Act. EPA also issued for public comment a proposed Federal Implementation Plan to implement the Clean Power Plan if states fail to submit or do not develop an approvable state plan for compliance. EPA finalized NSPS for new sources at 1,400 pounds CO 2 per megawatt-hour gross (lb/mwh-g) for new coal units based on the agency s assumed use of carbon capture and storage technology. Reconstructed coal units have a limit of 1,800 or 2,000 lb/mwh based on the size of the unit. The NSPS for modified coal units is site-specific based on historical operations. For new and reconstructed natural gas combined cycle (NGCC) units, the NSPS was finalized at 1,000 lb/mwh-g based on the use of efficient combustion turbine designs. No limit was proposed for modified NGCC or simple cycle units. The Final Clean Power Plan establishes separate, uniform national CO 2 emission performance rates for fossil steam units (coal-, oil-, and gas based units) and for stationary combustion turbines (which EPA defines as natural gas combined cycle units). The rates were established based on EPA s application of three building blocks as the best system of emission reduction (BSER) for existing fossil generating units. Block 1 assumes efficiency improvements at existing coal units. Building Block 2 assumes the increased use of NGCC units that would 42

56 2015 Integrated Resource Plan displace coal-based generation. And block 3 entails the expansion of renewable energy sources that would displace generation from both coal and NGCC units. Excluded from the BSER process are consideration of nuclear energy, simple cycle gas turbines, and the previously proposed building block 4 related to energy efficiency measures, From the national emission performance rates, EPA also developed equivalent statespecific emission rate goals and equivalent state-specific mass-based goals as alternatives. The final 2030 state emission rate goal for Oklahoma increased to 1,068 lb/mwh-net from the originally proposed 895 lb/mwh-net rate. With respect to mass based goals, the final 2030 goal increased to approximately 40,488,000 short tons from the proposed approximate 34,052,000 short tons. EPA included interim rates in the final rule, but extended the initial compliance period start from 2020 to Additionally, for Oklahoma the initial mass-based goal at the start of the program increased by about 30%. The proposed rule would have become effective in 2020 and set a mass-based goal of ~37,867,000 short tons, while the final rule becomes effective in 2022 with a mass-based goal of ~49,387,000 short tons. States that decide to develop a State Plan to implement the Clean Power Plan have the option of developing either an emissions standards approach that would apply directly to the affected units, or a state measures approach that would incorporate other elements into the compliance strategy. An initial draft State Plan must be submitted to EPA by September 6, Extensions for submitting a final State Plan are available through September 6, If States do not submit an approvable plan to EPA, EPA will adopt a Federal Implementation Plan, based on model rules that will be open for public comment when published in the Federal Register. The model rules are expected to be finalized in the summer of PSO is currently in the process of reviewing these rulemakings and must undertake significant new analyses to understand the impacts of the final CPP. PSO, AEP, and other stakeholders will be working in the coming months and years to better understand the requirements of the Final CPP and to work with state agencies on the state s response to the final CPP. 43

57 2015 Integrated Resource Plan Impact of Environmental Rules on Planning Process The estimated, potential impact of the previously described rules, with the exception of the CPP, may be factored into the analysis of potential resource plans by adding the incremental cost to comply with the rules, and retiring units where it is not economical to comply. The Final CPP was issued on August 3, The Company will work in the coming months and years to develop reasonable plans toward compliance, but at this time it is not possible to determine with any certainty what the final impact will be on PSO s generating fleet. 3.5 Current Demand Side Programs Background Current DSM refers to, for the purposes of this IRP, utility programs, including tariffs, which encourage reduced energy consumption, either at times of peak consumption or throughout the day or year. Programs or tariffs that reduce consumption at the peak are Demand Reduction (DR) programs, while around-the-clock measures are typically categorized as EE programs. The distinction between DR and EE is important, as the solutions for accomplishing each objective are typically different, but not necessarily mutually exclusive. Included in the load forecast discussed in Section of this report are the demand and energy impacts associated with PSO s embedded DSM programs that have been previously approved by the Oklahoma Corporation Commission (OCC) and the portfolio of measures currently proposed in docket Public Utility Division (PUD) As will be discussed later, the potential for additional demand-side resources, over and above the levels embedded in the load forecast, are modeled on the same economic basis as supply-side resources Impacts of Existing and Future Codes and Standards The EISA requires, among other things, a phase-in of heightened lighting efficiency standards, appliance standards, and building codes. The increased standards are having a pronounced effect on energy consumption. Many of the standards already in place impact lighting. For instance, beginning in 2013 and 2014 common residential incandescent lighting options have begun their phase out as have common commercial lighting fixtures. Given that lighting options have comprised a large portion of utility-sponsored energy efficiency programs 44

58 2015 Integrated Resource Plan over the past decade, this pre-established transition is already incorporated into the SAE longterm load forecast modeling previously described in Section and may greatly affect the market potential of utility energy efficiency programs in the near and intermediate term. Table 7 illustrates the current schedule for the implementation of new EISA codes and standards. Table 7. Forecasted View of Relevant Energy Efficiency Code Improvements The impact of emerging codes and standards on PSO s load forecast can be seen in Figure 8. Over the planning period codes and standards are forecasted to reduce retail load by over 5%. 45

59 2015 Integrated Resource Plan Figure 8. Codes & Standards Impact on PSO Retail Load, post Current Demand Response (DR)/Energy Efficiency (EE) Programs PSO currently has DR and EE programs in place. For the year 2015, the Company anticipates 69MW of peak demand reduction (total company basis); consisting of 8MW and 61MW of passive EE and active DR peak demand reductions, respectively Demand Reduction Peak demand, measured in MW, can be thought of as the amount of power used at the time of the Company s maximum power usage. PSO s maximum (system peak) demand is likely to occur on the hottest summer weekday of the year, in the afternoon. This happens as a result of the near-simultaneous use of electric cooling by the majority of customers, as well as the normal use of other appliances, commercial equipment, and (industrial) machinery. At other times during the day, and throughout the year, the use of power is less. 7 Passive demand reductions are achieved via around-the-clock energy efficiency program activity as well as voluntary price response programs; while Active DR is centered on focused summer peak reduction initiatives, including interruptible contracts and electric load management/direct load control programs. 46

60 2015 Integrated Resource Plan As peak demand grows with the economy and population, new capacity must ultimately be built. To defer construction of new power plants, the amount of power consumed at the peak must be reduced. This can be addressed several ways via both active and passive measures: Interruptible loads (Active DR). This refers to a contractual agreement between the utility and a large consumer of power, typically an industrial customer. In return for reduced rates, an industrial customer allows the utility to interrupt or reduce power consumption during peak periods, freeing up that capacity for use by other consumers. Direct load control (Active DR). Very much like an (industrial) interruptible load, but accomplished with many more, smaller, individual loads. Commercial and residential customers, in exchange for monthly credits or payments, allow the energy manager to deactivate or cycle discrete appliances, typically air conditioners, hot water heaters, lighting banks, or pool pumps during periods of peak demand. These power interruptions can be accomplished through radio signals that activate switches or through a digital smart meter that allows activation of thermostats and other control devices. Time-differentiated rates (Active DR). This offers customers different rates for power at different times during the year and even the day. During periods of peak demand, power would be relatively more expensive, encouraging conservation. Rates can be split into as few as two rates (peak and off-peak) to as often as 15-minute increments in what is known as real-time pricing. Accomplishing real-time pricing requires digital (smart) metering. EE measures (Passive DR). If the appliances that are in use during peak periods use less energy to accomplish the same task, peak energy requirements will likewise be less. Line loss mitigation (Passive DR). A line loss results during the transmission and distribution of power from the generating plant to the end user. To the extent which load is decreased, these losses can be reduced, and less energy is required from the generator. What may not be apparent is that, with the exception of EE and line loss mitigation measures, DR programs do not significantly reduce the amount of energy consumed by customers. Less energy may be consumed at the time of peak demand, but that energy will be consumed at some other time during the day. For example, if rates encourage customers to avoid running their clothes dryer at 4:00p.m. they will run it at some other time of the day. This is often referred to as load shifting. 47

61 2015 Integrated Resource Plan Energy Efficiency (EE) EE measures save money for customers by reducing energy consumption. The trade-off is the up-front investment in a building/appliance/equipment modification, upgrade, or new technology. If the consumer feels that the new technology is a viable substitute and will pay them back in the form of reduced bills over an acceptable period, they will adopt it. EE measures most commonly include efficient lighting, weatherization, efficient pumps and motors, efficient Heating, Ventilation and Cooling (HVAC) infrastructure, and efficient appliances. Often, multiple measures are bundled into a single program that might be offered to either residential or commercial/industrial customers. EE measures will reduce the amount of energy consumed but may have limited effectiveness at the time of peak demand. Many EE measures are viewed as a readily deployable, relatively low cost, and clean energy resource that provides many benefits. However, as summarized in Table 8, market barriers to EE exist for the potential participant. Table 8. Energy Efficiency (EE) Market Barriers High First Costs High Information or Search Costs Consumer Education Performance Uncertainties Energy-efficient equipment and services are often considered high-end products and can be more costly than standard products, even if they save consumers money in the long run. It can take valuable time to research and locate energy efficient products or services. Consumers may not be aware of energy efficiency options or may not consider lifetime energy savings when comparing products. Evaluating the claims and verifying the value of benefits to be paid in the future can be difficult. Transaction Costs Additional effort may be needed to contract for energy efficiency services or products. Access to Financing Split Incentives Product/Service unavailability Externalities Lending industry has difficulty in factoring in future economic savings as available capital when evaluating credit-worthiness. The person investing in the energy efficiency measure may be different from those benefiting from the investment (e.g., rental property) Energy-efficient products may not be available or stocked at the same levels as standard products. The environmental and other societal costs of operating less efficient products are not accounted for in product pricing or in future savings Source: Eto, Goldman, and Nadel (1998): Eto, Prahl, and Schlegel (1996); and Golove and Eto (1996) 48

62 2015 Integrated Resource Plan To overcome many of the participant barriers noted above, a portfolio of EE programs may often include several of the following elements: Consumer education Technical training Energy audits Rebates and discounts for efficient appliances, equipment and buildings Industrial process improvements Utility efficiency improvements directly saving energy for customers The level of incentives (rebates or discounts) offered to participants is a major determinant in the pace of EE measure adoption. Additionally, the speed with which programs can be rolled out also varies with the jurisdictional differences in stakeholder and regulatory review processes. The lead time can easily exceed a year for getting programs implemented or modified. This IRP begins adding new demand-side resources in 2019 that are incremental to approved or mandated programs Conservation Voltage Reduction (CVR) An emerging energy efficiency technology known as CVR is a form of voltage control that allows the grid to operate more efficiently at lower voltages. In a CVR installation sensors and intelligent controllers monitor load flow characteristics and control capacitor and voltage regulating equipment to optimize power factor and voltage levels. Power factor is the ratio of real power to apparent power, and is a characteristic of electric power flow which is controlled to optimize power flow on an electric network. Power factor optimization also improves energy efficiency by reducing losses on the system. CVR is a technology that allows the utility to systematically reduce voltages in its distribution network, resulting in a proportional reduction of load on the network. Voltage optimization can allow a reduction of system voltage that still maintains minimum levels needed by customers, thereby allowing customers to use less energy without any changes in behavior or appliance efficiencies. 49

63 2015 Integrated Resource Plan In 2011 and 2012, PSO deployed CVR technology on 11 circuits in the city of Owasso as part of a pilot demonstration that also included other grid management technologies. Subsequently, CVR technology was expanded to two additional circuits in PSO conducted an evaluation of 2013 CVR performance, and additional evaluations, including impacts of the technology on customers, were performed by an independent and nationally recognized third party, Pacific Northwest National Labs. The results of the study showed energy savings between approximately 2% and 7% and demand savings between 2% and 5%. The load forecast includes one year of CVR installations, taking effect in CVR has been modeled as a separate energy efficiency resource. The results of this modeling are discussed in Section Energy Conservation Often used interchangeably with efficiency, conservation results from foregoing the benefit of electricity either to save money or simply to reduce the impact of generating electricity. Higher rates for electricity typically result in lower consumption. Inclining block rates, or rates that increase with usage, are rates that encourage conservation Distributed Generation Distributed Generation (DG) typically refers to small scale customer-sited generation downstream of the customer meter. Common examples are combined heat and power (CHP), residential and small commercial solar applications, and even wind. Currently, these sources represent a small component of demand-side resources, even with available Federal Investment Tax Credits (ITCs). The economics of DG, particularly solar, continue to improve. Figure 9 charts the fairly rapid decline of expected installed solar costs, based on a combination of AEP market intelligence and the Bloomberg New Energy Finance s (BNEF) Installed Cost of Solar forecast. These are costs shown without accounting for the 30% ITC (reduced to a 10% credit in 2016 for utility and commercial projects and 0% for residential). 50

64 2015 Integrated Resource Plan Figure 9. Forecasted Solar Installed Costs (nominal) for Oklahoma (Excluding Federal & State Incentives) Figure 10 demonstrates the historical installed rooftop solar capacity, or DG, for PSO by jurisdiction along with projected rooftop solar capacity additions. Figure 10. PSO Cumulative Rooftop Solar Additions/Projections 51

65 2015 Integrated Resource Plan The assumed growth rate for rooftop solar within PSO is 5% per year. The assumed growth rate is an estimate and is based on both the declining cost for rooftop solar as well as the historical additions in PSO. Potential future changes to federal and state tax incentives create uncertainty in the future costs of rooftop solar. However, a number of customers may decide to install such systems due to personal motivations, regardless of costs. 3.6 AEP-SPP Transmission Transmission System Overview The portion of the AEP Transmission System operating in SPP (AEP-SPP zone, or AEP- SPP) consists of approximately 1,300 miles of 345 kv, approximately 3,600 miles of 138 kv, and approximately 2,300 miles of 69 kv. The AEP-SPP zone is also integrated with and directly connected to ten other companies at 89 interconnection points, of which 71 are at or above 138 kv and to Electric Reliability Council of Texas (ERCOT) via two high voltage direct current (HVDC) ties. These interconnections provide an electric pathway to provide access to offsystem resources, as well as a delivery mechanism to neighboring systems Current AEP-SPP Transmission System Issues The limited capacity of interconnections between SPP and neighboring systems, as well as the electrical topology of the SPP footprint transmission system, influences the ability to deliver non-affiliate generation, both within and external to the SPP footprint, to AEP-SPP loads and from sources within AEP-SPP balancing authority to serve AEP-SPP loads. Moreover, a lack of seams agreements between SPP and its neighbors has significantly slowed down the process of developing new interconnections. Despite the robust nature of the AEP-SPP transmission system as originally designed, its current use is in a different manner than originally designed, in order to meet SPP requirements, which can stress the system. In addition, factors such as outages, extreme weather, and power transfers also stress the system. This has resulted in a transmission system in the AEP-SPP zone that is constrained when generation is dispatched in a manner substantially different from the original design of utilizing local generation to serve local load. SPP uses models developed from data provided by all load serving entities to study the reliability needs of the SPP footprint. 52

66 2015 Integrated Resource Plan The SPP Transmission Planning Process SPP works with its members to determine and construct the transmission infrastructure needed in the near and long-term planning horizon to maintain electric reliability, meet public policy mandates and provide economic benefits. SPP does not own or build transmission assets. The SPP Open Access Transmission Tariff (OATT) contains the rules that govern transmission construction by SPP members. SPP s transmission planning services include development of regional transmission expansion plans, oversight of transmission upgrade construction in accordance with approved plans, and development and implementation of cost allocation methodologies to ensure appropriate recovery by the constructing and managing utilities. SPP s construction oversight includes monitoring project status and costs through quarterly reporting by the constructing utilities and ensuring proper adherence to cost estimates and construction completion need dates. The SPP Transmission Expansion Plan (STEP) is a compilation of SPP-directed projects based on studies performed by SPP to determine upgrades needed to maintain reliability and provide economic benefit into the future. SPP's transmission planning processes seek to identify system limitations and needs, develop cost effective transmission solutions, and ensure timely completion of needed system expansion within reasonable cost expectations. Rather than looking at the needs of just one utility, SPP assesses needs from a larger, regional perspective and determines necessary new transmission infrastructure that would provide the most net benefit to the region. The STEP is developed through an open stakeholder process with AEP participation. SPP studies the transmission system, checking for base case and contingency overload and voltage violations in all of the SPP base case load flow models, plus models which include power transfers. SPP s transmission planning processes evaluate transmission system needs and determine necessary transmission solutions according to one or more methods prescribed in the SPP OATT. The projection of SPP-directed transmission projects in the SPP region as determined by these planning processes is reported annually in the STEP. Transmission projects contained in the STEP are the result of one or more of the following processes or sources: 1) Transmission service study process, 53

67 2015 Integrated Resource Plan 2) Generator interconnection study process, 3) Requests for Sponsored Upgrades, 4) Integrated Transmission Planning (ITP) process. These topics are critical to meeting mandates of either the SPP strategic plan or the nine planning principles in FERC Order 890. As an RTO under the domain of the Federal Energy Regulatory Commission (FERC), SPP must meet FERC requirements and the SPP OATT, or Tariff. SPP acts independently of any single market participant or class of participants. It has sufficient scope and configuration to maintain electric reliability, effectively perform its functions, and support efficient and non-discriminatory power markets. Regarding short-term reliability, SPP has the capability and exclusive authority to receive, confirm, and implement all interchange schedules. It also has operational authority for all transmission facilities under its control. The 10-year RTO regional reliability assessment continues to be a primary focus. STEP projects are categorized by the following designations: Balanced Portfolio Projects identified through the Balanced Portfolio Process; Generation Interconnect Projects associated with a FERC-filed Interconnection Agreement; Interregional- Projects developed with neighboring Transmission Providers; ITP - Projects needed to meet regional reliability, economic, or policy needs in the ITP study process; ITP Non-OATT Projects to maintain reliability for SPP members not participating under the SPP OATT Transmission service Projects associated with a FERC-filed Service Agreement; Zonal Reliability - Projects identified to meet more stringent local Transmission Owner criteria; and Zonal-Sponsored Projects sponsored by facility owner with no Project Sponsor Agreement Transmission service: SPP s transmission service study process is known as the Aggregate Transmission Service Study (ATSS) process, and is described in Attachment Z1 of 54

68 2015 Integrated Resource Plan the SPP Tariff. Under this process, a transmission customer desiring to use SPP s transmission system to deliver energy from a specific generation resource or group of resources to a consumer or group of consumers must first reserve and ultimately pay for the requested transmission capacity. To reserve this capacity, the Transmission Customer must submit a transmission service request to SPP containing specific information defining the intended usage of SPP s transmission system. SPP will evaluate the available capacity of the transmission grid, including capacity from approved planned upgrades, to accommodate the request considering all existing committed uses of the transmission system. If available capacity is sufficient, the transmission service request is accepted. If available capacity is insufficient to grant the transmission service request, SPP will identify the transmission system upgrades necessary to accommodate the request. If upgrades are required to accommodate the transmission service request and if the Transmission Customer agrees to pay the applicable transmission service rates and any upgrade costs directly assigned to it, SPP will direct the upgrades to be constructed. AEP participates in this process both as a transmission customer and transmission owner. Generation Interconnection (GI): The GI study process is described in Attachment V of the SPP Tariff. Under this process, a GI customer desiring to interconnect a new generation resource to the SPP transmission grid must first submit a GI request to SPP. The request must include basic information about the generator characteristics and an expected in-service date. The GI study process determines transmission upgrades necessary to interconnect the new generation resource to the transmission grid but does not determine what transmission facilities are necessary to deliver energy from the new resource to a specific consumer or set of consumers. Capacity from approved planned transmission upgrades is assumed to be available in the GI study process. If incremental or other upgrades are required to accommodate the GI request and the GI customer agrees to pay its share of upgrade costs assigned to that GI customer, SPP will direct the incremental or other upgrades to be constructed. AEP participates in this process if the AEP SPP transmission system requires upgrades on its system for any generating facility to connect to the transmission grid, even if an AEP affiliate is the requesting GI customer. Sponsored Upgrades: A Sponsored Upgrade is one requested by a specific entity (Sponsor) that does not result from any of the other SPP planning processes. If a Sponsored 55

69 2015 Integrated Resource Plan Upgrade is requested, SPP will evaluate its impact on transmission system reliability and identify any necessary mitigation of those impacts. The Sponsor must agree to assume the cost responsibility of the Sponsored Upgrade and any necessary mitigation. The Sponsored Upgrade is considered approved for construction once the SPP Board endorses the requested upgrade and upon execution of a contract that financially commits the Sponsor to the upgrade. To date, AEP has not requested a Sponsored Upgrade to be evaluated by SPP. Integrated Transmission Planning (ITP): ITP is a three-year study process that assesses long- and near-term infrastructure needs of the SPP Transmission System. The intent of ITP is to bring about continued development of a redundant, flexible, and reasonably balanced transmission network in a cost-effective manner that will provide efficient, reliable access to the region s diverse generating resources. The ITP process, as described in Attachment O of the SPP Tariff, promotes transmission investment that will meet reliability, economic, and public policy needs. The first phase of the ITP study process is the ITP 20-Year Assessment (ITP20). The ITP20 is a long-term assessment focusing on an EHV backbone for the transmission system and is conducted over the first half of the three-year ITP cycle. Because of the long-term nature of the ITP20, NTCs are generally not issued at its conclusion. The second phase of the ITP study process is the ITP 10-Year Assessment (ITP10). The ITP10 is conducted over the second half of the three-year ITP cycle and is a value-based approach that assesses policy, economic and reliability needs for a 10-year horizon. The third phase of the ITP study process is the ITP Near- Term (ITPNT) and is an annual assessment of short-term reliability needs and solutions. The ITPNT is performed every year, overlapping the ITP20 and ITP10 study efforts. The SPP stakeholders are an integral part of the development of assumptions and the selection of projects in the ITP process. Under the SPP Bylaws, with very few exceptions, all SPP meetings are required to be open meetings. These meetings, including discussions about the development of the ITP, are routinely attended by representatives of a broad range of stakeholders, including state regulatory staff members and SPP customers and members, including vertically integrated utilities, transmission owners, generation owners, marketers, cooperatives, and municipal utilities. As allowed by the SPP Bylaws, any of the parties involved in this case could also attend these meetings. In addition, the ITP projects are reported to the Markets and Operations Policy Committee (MOPC), Regional State Committee (RSC) and SPP 56

70 2015 Integrated Resource Plan Board as part of the annual STEP report. The MOPC is required to recommend any ITP projects for approval by the SPP Board. The SPP Board is required by Attachment O of the SPP Tariff to approve the ITP projects before SPP issues Notifications to Construct (NTCs) for those projects. AEPSC, on behalf of PSO and its affiliate Southwestern Electric Power Company, has representatives either attending or serving in a voting role on many SPP committees, working groups and task forces. The Transmission Working Group and the Economic Studies Working Group provide input into the ITP Process pursuant to the SPP ITP Manual. The work of these two groups is key to the development of the inputs, analysis, and review of the results of the ITP process. AEPSC also provides recommendations on transmission solutions to reliability and other problems identified during the course of the ITP process. AEPSC also has members on the MOPC that review the ITP projects for a recommendation to the SPP Board for approval. The process of assuring that the cost for transmission projects is reasonable begins with being sure that the most cost-effective project is selected for construction. Due to the iterative and open nature of the SPP transmission expansion planning processes, projects are reviewed multiple times by the stakeholders, customers, and interested parties. By the time a project reaches the SPP Board for approval, the project has been repeatedly reviewed by many stakeholders, including customers and members comprised of vertically integrated utilities, transmission owners, generation owners, marketers, cooperatives and municipal utilities, as well as state regulatory agency staff. It is this rigorous and open review that assures SPP, its members and its customers that the most cost-effective projects are being selected for construction to meet the needs of the SPP region. SPP identifies those projects within the STEP that require construction to begin within the next four years and issues a NTC to the Designated Transmission Owner (DTO) so that the necessary facilities are constructed. The NTC is the formal notification from the SPP instructing the DTO to build the project within the required time frame. The 2015 STEP identified 568 transmission network upgrades with a total cost of approximately $5.7 billion. At the heart of SPP s STEP process is its ITP process, which represented approximately 56% of the total cost in the 2015 STEP. The ITP process was designed to maintain reliability 57

71 2015 Integrated Resource Plan and provide economic benefits to the SPP region in both the near and long-term. The ITP process was conducted in three phases. The first phase recommended a long-term transmission plan for a 20-year horizon, incorporating a proposed extra-high voltage supply system. The second phase of the ITP process resulted in a recommended portfolio of transmission projects for comprehensive regional solutions, local reliability upgrades, and the expected reliability and economic needs of a 10-year horizon. Finally, the third phase of the ITP process studied the reliability of the SPP transmission system in the near-term, identifying upgrades based on a 6- year planning horizon PSO-SWEPCO Interchange Capability Operational experience and internal assessments of company transmission capabilities indicate that, when considering a single contingency outage event, the present firm capability transfer limit from PSO to Southwestern Electric Power Company (SWEPCO) and from SWEPCO to PSO is about 200MW. As much as 900MW from PSO to SWEPCO and 700MW from SWEPCO to PSO may be available for economical energy transfers when no transmission facilities are out of service. However, the intra-company available transmission capability between the two companies is available to all transmission users under the provisions established by FERC Order 888 and subsequent orders. Thus, there is some question as to whether, in the future, as SPP grants further transmission rights, any transfer capability will in fact be available without further upgrades to the transmission system. As previously indicated, each company s capacity additions are planned so that each meets its own reserve requirement over the long-term. Any capacity transfers (i.e. reserve sharing ) should be considered for short time frames only. Specifically, the practice has been that, as the last step of the planning process, the respective PSO and SWEPCO expansion plans are adjusted to take advantage of any surplus of one company that might match a potential deficit of the other, and thereby delay some of the identified new capacity. Because of the sizes, demand growth rates, and peak coincidence of the two companies, it rarely appears that either company would ever have more than 200MW of surplus capacity in any year that could be transferred to the other company. 58

72 2015 Integrated Resource Plan AEP-SPP Import Capability Currently the capability of the transmission system to accommodate large incremental firm imports to the AEP-SPP area is limited. Generally, the transfers are limited by the facilities of neighboring systems rather than by transmission lines or equipment owned by AEP. Increasing the import capabilities with AEP-SPP s neighboring companies could require a large capital investment for new transmission facilities by the neighboring systems or through sponsored upgrades by SPP transmission owners. An analysis of the cost of the upgrades cannot be performed until the capacity resources are determined. For identified resources, the cost of any transmission upgrades necessary on AEP s transmission system can be estimated by AEP once SPP has identified the upgrade. AEP s West Transmission Planning group can identify constraints on third-party systems through ad hoc power flow modeling studies, but West Transmission Planning does not have information to provide estimates of the costs to alleviate those third-party constraints SPP Studies that may Provide Import Capability Within the STEP, some projects that may lead to improved transfer capability between AEP-SPP and neighboring companies and regions include: A Chisholm-Gracemont 345 kv line across western Oklahoma from a new Chisholm kv station near existing wind generation facilities west of Elk City to Gracemont station near Anadarko A new Layfield kv station in northwestern Louisiana (previously referred to as Messick) A Sooner-Cleveland 345 kv line in northern Oklahoma, west of the Tulsa area (completed) A Seminole-Muskogee 345 kv line in eastern Oklahoma (completed) A Sunnyside-Hugo-Valliant 345 kv line across southeastern Oklahoma (completed) A Tuco-Woodward 345 kv line from the Texas Panhandle to northwestern Oklahoma (completed) 59

73 2015 Integrated Resource Plan Besides the annual STEP process, SPP also performs other special studies or area studies on an as needed basis. One SPP study that resulted in approved projects that may lead to improved transfer capability between AEP-SPP and neighboring companies and regions is the Priority Projects study. This study included several 345 kv lines which provide pathways from areas of existing and potential wind generation development in the west to the existing 345 kv system further to the east. Among the projects approved as a result of this study are: Double circuit Spearville-Thistle 345 kv line in southwestern Kansas (completed) Double circuit Thistle-Wichita 345 kv line in southern Kansas (completed) Double circuit Woodward-Thistle 345 kv line from northwestern Oklahoma to southwestern Kansas (completed) Double circuit Hitchland-Woodward 345 kv line in northwestern Oklahoma (completed) A Valliant-Northwest Texarkana 345 kv line from southeastern Oklahoma to northeastern Texas Recent AEP-SPP Bulk Transmission Improvements Over the past several years, there have been several major transmission enhancements initiated to reinforce the AEP-SPP transmission system. These enhancements include: Northwest Arkansas Northwest Arkansas is one of the fastest growing areas on the AEP-SPP Transmission System. The approximate 1,300MW of load in this area, about 49% of which is AECC load, is supplied primarily by the SWEPCO and Arkansas Electric Cooperative Corporation (AECC) jointly-owned Flint Creek generating plant, the SWEPCO Mattison generating plant, the Grand River Dam Authority (GRDA)-Flint Creek 345 kv line, and the Clarksville-Chamber Springs 345 kv line. Wal-Mart s international headquarters and its supplying businesses offices and Tyson s headquarters are all located in this area. A new 345 kv line has been completed from Flint Creek to the new Shipe Road 345/161 kv substation along with a 161 kv line connecting Shipe Road substation to East Centerton substation 60

74 2015 Integrated Resource Plan Port of Shreveport (Port), Louisiana A 138 kv loop was completed in 2012 around the Port to increase system reliability and to serve the increasing area load. This loop extends approximately 33 miles from Wallace Lake Station to the Port to Bean Station to Caplis Station to McDade Station to Haughton Station to Red Point Station. In order to serve a new industrial customer, Benteler Steel/Tube, two 138 kv lines of approximately three to four miles each, have been built from the Port to the Benteler Steel/Tube plant. McAlester, Oklahoma area The Canadian River kv substation has been completed northwest of McAlester, along Oklahoma Gas and Electric Energy Corporation s (OG&E) Pittsburg-Muskogee 345 kv line. A 69 kv line was converted to 138 kv line for approximately 17 miles from the Canadian River substation to the McAlester City substation. This relieves 138 kv line loadings in the area and provides voltage support. Cornville/Rush Springs, Oklahoma area The Cornville-Lindsay Water Flood radial line, approximately 33 miles, has been rebuilt and converted to 138 kv operation. A 138 kv connection, approximately 10 miles, is being built from this line to an existing radial that serves Rush Springs Natural Gas from the existing Cornville-Duncan 138 kv line. This will complete a 138 kv loop, improving reliability of the transmission system in this area. These major enhancements are in addition to several completed or initiated upgrades to 138 kv and 69 kv transmission lines to reinforce the AEP-SPP transmission system Impacts of New Generation Integration of additional generation capacity within the AEP-SPP zone will likely require significant transmission upgrades. At most locations, any additional generation resources will aggravate existing transmission constraints. Specifically: Western Oklahoma/Texas Panhandle Until recently, there were very few EHV transmission lines in this area, though that is changing due to the 345 kv projects discussed above. The area is one of the highest wind density areas within the SPP footprint. The potential wind farm capacity for this area has been estimated to exceed 61

75 2015 Integrated Resource Plan 4,000MW. Several wind farms have already been built, and several more are in the development stages. Wind generation additions in the SPP footprint in this region will likely require significant transmission enhancements, including EHV line and station construction, to address thermal, voltage, and stability constraints. PSO/SWEPCO Interface - There is one 345 kv EHV line linking PSO s service area with the majority of SWEPCO s generation resources in its service area. An SPP approved project, mentioned above, to build a 345 kv line approximately 76 miles from Valliant substation to Northwest Texarkana substation will improve transfer capability by forming a second 345 kv path between PSO and SWEPCO s transmission system in northeastern Texas. Significant generation additions to the AEP-SPP transmission facilities (or connection to neighbor s facilities) may require significant transmission enhancements, possibly including EHV line and station construction, to address thermal, voltage, and stability constraints. Tulsa Metro Area the Tulsa metro area load is supplied primarily by the PSO Northeastern, Riverside, and Tulsa Power Station generating plants. Additionally, Oklahoma Gas & Electric Company has large generation plants located to the southeast and southwest of Tulsa, and there are large merchant plants just east and south of Tulsa. The Grand River Dam Authority has a large plant located to the east of Tulsa. Generation additions in the Tulsa area would likely require significant enhancements in the EHV and sub-transmission system to address thermal, voltage and stability constraints. SPP Eastern Interface there are only five east-west EHV lines into the SPP region, which stretches from the Gulf of Mexico (east of Houston) north to Des Moines, Iowa. This limitation constrains the amount of imports and exports along the eastern interface of SPP with neighboring regions. It also constrains the amount of transfers from the capacity rich western SPP region to the market hubs east and north of the SPP region. Significant generation additions near or along the SPP eastern interface would likely require significant transmission enhancements, including EHV line and station construction, to address thermal and stability constraints should such generation additions adversely impact existing transactions along the interface. 62

76 2015 Integrated Resource Plan Integration of generation resources at any location within the AEP-SPP zone will require significant analysis by SPP to identify potential thermal, short circuit, and stability constraints resulting from the addition of generation. Depending on the specific location, EHV line and station construction, in addition to connection facilities, could be necessary. Other station enhancements, including transformer additions and breaker replacements may be necessary. Some of the required transmission upgrades could be reduced or increased in scope if existing generating capacity is retired concurrent with the addition of new capacity. For example, if SWEPCO s Flint Creek Generating Plant were retired, rather than upgraded as SWEPCO proposed in Case No U, SWEPCO s transmission system would require significant upgrades to support the delivery of power from remote generating plants, provide transfer capability, and supply reactive power for voltage support Summary of Transmission Overview In the SPP region, the process of truly integrating Generation and Transmission planning is still developing. AEP continues to stand ready to engage in that process. AEP also continues supporting the SPP STEP and ITP transmission expansion processes, which include some projects which may improve import capability. Such capability improvements are more likely to be within SPP, but less so between SPP and neighboring regions, partly due to lack of seams agreements which slows the development of new interconnections as discussed above. PSO and SWEPCO have been open to such imports as evidenced by the issuing of recent Requests For Proposals (RFPs) for non-site specific generation types. These RFPs allow bidding entities to offer generation coupled with transmission solutions, which would be subject to SPP approvals. 63

77 2015 Integrated Resource Plan 4.0 Modeling Parameters 4.1 Modeling and Planning Process An Overview The objective of a resource planning effort is to recommend a system resource expansion plan that balances least-cost objectives with planning flexibility, asset mix considerations, adaptability to risk, conformance with applicable North American Electric Reliability Corporation (NERC) and RTO criteria. In addition, given the unique impact of fossil-fired generation on the environment, the planning effort must ultimately be in concert with anticipated long-term requirements as established by the EPA-driven environmental compliance planning process. The information presented with this IRP includes descriptions of assumptions, study parameters, methodologies, and results, including the integration of traditional supply-side resources, renewable energy resources and DSM programs. In general, assumptions and plans are continually reviewed and modified as new information becomes available. Such continuous analysis is required by multiple disciplines across PSO and AEP to ensure that market structures and governances, technical parameters, regulatory constructs, capacity supply, energy adequacy and operational reliability, and environmental mandate requirements are constantly reassessed to ensure optimal capacity resource planning. Further impacting this process are a growing number of federal initiatives that address many issues relating to industry restructuring, customer choice, and reliability planning. Currently, fulfilling a regulatory obligation to serve native load customers represents one of the cornerstones of the PSO IRP process. Therefore, as a result, the objective function of the modeling applications utilized in this process is the development of a least-cost plan, with cost being more accurately described as revenue requirement under a traditional ratemaking construct. That does not mean, however, that the most appropriate plan is the one with the absolute least cost over the planning horizon evaluated. Other factors were considered in the 64

78 2015 Integrated Resource Plan determination of the Plan. To challenge the robustness of the IRP, sensitivity analyses were performed to address these factors. This overall process reflects consideration of options for maintaining and enhancing rate stability; economic development; and service reliability. 4.2 Methodology The IRP process aims to address the gap between resource needs and current resources. Given the various assets and resources that can satisfy this expected gap, a tool is needed to sort through the myriad of potential combinations and return an optimum solution. Plexos is the primary modeling application used by PSO for identifying and ranking portfolios that address the gap between needs and current available resources. 8 Given the cost and performance parameters around sets of potentially available proxy resources both supply and demand side and a scenario of economic conditions that include long-term fuel prices, capacity costs, energy costs, emissionbased pricing proxies including CO 2, as well as projections of energy usage and peak demand, Plexos will return the optimal suite of proxy resources (portfolio) that meet the resource need. Portfolios created under similar pricing scenarios may be ranked on the basis of cost, or the cumulative present worth (CPW), of the resulting stream of revenue requirements. The least cost option is considered the optimum portfolio for that unique input parameter scenario. 4.3 Fundamental Modeling Input Parameters The AEP Fundamental Analysis group derives long-term power (energy) price forecasts from a proprietary model known as AURORA xmp. Having similarities to Plexos, AURORA xmp is a long-term fundamental production cost-based energy and capacity price forecasting tool developed by EPIS, Inc., that is driven by comprehensive, user-defined commodity input parameters. For example, nearer-term unit-specific fuel delivery and emission allowance price forecasts, based upon actual transactions, which are established by AEP Fundamental Analysis and AEP Fuel, Emissions and Logistics, are input into AURORA xmp. Estimates of longer-term natural gas and coal pricing are provided by AEP Fundamental Analysis in conjunction with 8 Plexos is a production cost-based resource optimization model, which was developed and supported by Energy Exemplar, LLC. The Plexos model is currently licensed for use in 37 countries throughout the world. 65

79 2015 Integrated Resource Plan input received from consultants, industry groups, trade press, governmental agencies and others. Similarly, capital costs and performance parameters for various new-build generating options, by duty-type are vetted through AEP Engineering Services and incorporated into the tool. Other information specific to the thousands of generating units being modeled is researched from Velocity Suite, an on-line information database maintained by Ventyx, an ABB Company. This includes data such as unit capacity, heat rates, retirement dates and emission controls status. Finally, the model maintains and determines region-specific resource adequacy based on regional load estimates provided by AEP Economic Forecasting, as well as current regional reserve margin criterion. AEP uses AURORA xmp to model long-term (market) energy and capacity prices for the entire U.S. eastern interconnect as well as ERCOT. The projection of a CO 2 pricing proxy is based on assumptions developed in conjunction with the AEP Strategic Policy Analysis organization. Figure 11 shows the Fundamentals process flow for solution of the long-term commodity forecast. The input assumptions are initially used to generate the output report. The output is used as feedback to change the base input assumptions. This iterative process is repeated until the output is congruent with the input assumptions (e.g., level of natural gas consumption is suitable for the established price and all emission constraints are met). 66

80 2015 Integrated Resource Plan Input Output Fuel Forecast Longterm Capacity Expansion Load Forecast Annual Dispatch Generate Report Emission Totals Fuel Burn Totals Market Prices Emissions Forecast Capital Cost Forecast Emission Retrofits Feedback Recycle Figure 11. Long-term Power Price Forecast Process Flow Commodity Pricing Scenarios Five commodity pricing scenarios were developed by AEP Fundamental Analysis for PSO to enable Plexos to construct resource plans under various long-term pricing conditions. In this report, the five distinct long-term commodity pricing scenarios that were developed for Plexos are: a Base scenario view, a plausible Lower Band view, a plausible Higher Band view; a High Carbon view; and a No Carbon view. The scenarios are described below with the results shown in Figure 12 through Figure Base Scenario This scenario recognizes the following major assumptions: MATS Rule effective beginning in 2015; Initially lower natural gas price due to the emergence of shale gas plays; and CO 2 emission pricing proxy begins in 2022 and was assumed to be at $15 per metric ton, growing with inflation. 67

81 2015 Integrated Resource Plan Each of the pricing forecasts includes a CO 2 impact as a result of the implementation of any prospective carbon reduction regulation. The Base, High Band and Low Band cases all reflect the fundamental view that a CO 2 proxy could be modeled as a $15/tonne dispatch cost penalty beginning in 2022 because it results in significant reduction of national CO 2 emissions when combined with newly promulgated EPA regulations and standards such as MATS, morestringent Corporate Average Fuel Economy (CAFE) standards and others. Given that any plan to reduce GHG emissions must be accompanied by a thorough assessment of the impact on the electric grid, allow adequate time for implementation, respect the authority of states and other federal agencies, and preserve a balanced, diverse mix of fuels for electricity generation, 2022 is the earliest reasonable projection as to when such regulation could become effective. The specific effects of the MATS Rule are modeled in the development of the long-term commodity forecast by retiring the smaller, older solid-fuel (i.e., coal and lignite) units which would not be economic to retrofit with emission control equipment. The retirement time frame modeled is 2015 through Those remaining solid-fuel generating units will have some combination of controls necessary to comply with the EPA s rules. Incremental regional capacity and reserve requirements will largely be addressed with new natural gas plants. One effect of the expected retirements on the emission control retrofit scenario is an over-compliance of the previous CSAPR emission limits. This will drive the emission allowance prices for SO 2 and NO X to zero by 2018 or Low Band Scenario This scenario is best viewed as a plausible lower natural gas/solid-fuel/energy price profile compared to the Base view. In the near term, Low Band natural gas prices largely track the Base but, in the longer term, natural gas prices represent an even more significant infusion of shale gas. From a statistical perspective, this long-term pricing scenario is approximately one (negative) standard deviation (-1.0 SD) from the Base scenario and illustrates the effects of coalto-gas substitution at plausibly lower gas prices. Like the Base scenario, proxied CO 2 mitigation/pricing is assumed to start in 2022 at a $15 per metric ton (real dollars). 68

82 2015 Integrated Resource Plan High Band Scenario Alternatively, this High Band scenario offers a plausible, higher natural gas/solidfuel/energy price profile compared to the Base scenario. High Band natural gas prices reflect certain impediments to shale gas developments including stalled technological advances (drilling and completion techniques) and as yet unseen environmental costs. The pace of environmental regulation implementation is in line with the Base scenario and Low Band. Analogous to the Low Band scenario, this High Band view, from a statistical perspective, is approximately, one (positive) standard deviation (+1.0 SD) from the Base. Also, like the Base and Low Band scenarios, CO 2 pricing is assumed to begin in 2022 at the same $15 per metric ton pricing proxy High Carbon Scenario Built upon the assumption of a $25 per metric ton (66% higher than the Base scenario) CO 2 mitigation pricing proxy beginning in 2022, the High Carbon scenario includes correlative price adjustments to natural gas and solid-fuel due to changes in consumption that such heightened CO 2 pricing levels would create. This results in some additional retirements of coal-fired generating units around the implementation period. Natural gas and, to a lesser degree, renewable generation is built as replacement capacity No Carbon Scenario This scenario does not consider the prospects of a carbon tax. While also including the necessary correlative fuel price adjustments, it serves as a baseline to understand the impact on unit dispatch and, with that, the attendant impact on energy prices associated with the Base and High Carbon scenarios. The following set of figures illustrates the range of such long-term pricing projections, on a nominal dollar basis, by major commodity through the year

83 2015 Integrated Resource Plan Figure 12. SPP On-Peak Energy Prices (Nominal $/MWh) Figure 13. SPP Off-Peak Energy Prices (Nominal $/MWh) 70

84 2015 Integrated Resource Plan Figure 14. Panhandle East Pipeline Natural Gas Prices (Nominal $/mmbtu) Figure 15. Panhandle East Pipeline Natural Gas Prices (Real $/mmbtu) 71

85 2015 Integrated Resource Plan Figure 16. PRB 8,800 BTU/lb. Coal Prices (Nominal $/ton, FOB) Figure 17. CO 2 Prices (Nominal $/metric ton) 72

86 2015 Integrated Resource Plan Long-Term CO 2 Forecast Each of the pricing forecasts includes a CO 2 impact as a result of the implementation of any prospective carbon-reduction rules or legislation. The Base, Low Band and High Band longterm pricing scenarios each reflect the fundamental view that a CO 2 penalty could be proxied with a $15 per metric ton dispatch cost burden applicable to fossil-fired generating units dispatch beginning in By contrast, the High Carbon scenario includes a higher, $25 per metric ton dispatch cost burden which also commences in Recognizing the relative higher carbon emission from a solid-fuel unit versus a natural gas-fired unit, the relative Plexos - modeled impact on variable/dispatch costs was reflected. Although the EPA publicized the Final CPP on August 3, 2015, it is too early to justify substantive changes to CO 2 impact modelling. Any plan to reduce GHG emissions must be accompanied by a thorough assessment of the impact on the electric grid, allow adequate time for implementation, respect the authority of states and other federal agencies, and preserve a balanced, diverse mix of fuels for electricity generation. 4.4 Demand-Side Management (DSM) Program Screening & Evaluation Process Overview The process for evaluating DSM impacts for PSO is divided into two components: existing programs and future activity. Existing programs are those that are known or are reasonably welldefined and follow a pre-existing process for screening and determining ultimate regulatory approval. The impacts of such existing DSM programs are propagated throughout the long-term PSO load forecast and were discussed in Section 3.5. Future program impacts which are, naturally, less-defined, are developed with a dynamic modeling process using more generic cost and performance parameter data. For PSO, the potential future DSM activity was developed based on the Electric Power Research Institute s (EPRI) 2014 U.S. Energy Efficiency Potential Through 2035 report. This comprehensive report served as the basic underpinning for the establishment of potential EE bundles, developed for residential and commercial customers that were then introduced as a resource option in the Plexos optimization model. Industrial programs were not developed or modeled based on the thought that industrial customers, by and large, will self-invest in energy 73

87 2015 Integrated Resource Plan efficiency measures based upon unique economic merit irrespective of the existence of utilitysponsored program activity Achievable Potential The amount of available EE is typically described in three sets: technical potential, economic potential, and achievable potential. The previously-cited EPRI report breaks down the achievable potential into a higher utility cost High Achievable Potential (HAP) and an Achievable Potential (AP). Briefly, the technical potential encompasses all known efficiency improvements that are possible, regardless of cost, and thus, whether it is cost-effective (i.e., all EE measures would be adopted if technically feasible). The logical subset of this pool is the economic potential. Most commonly, the total resource cost test is used to define economic potential. This compares the avoided cost savings achieved over the life of a measure/program with its cost to implement it, regardless of who paid for it and regardless of the age and remaining economic life of any system/equipment that would be replaced (i.e., all EE measures would be adopted if economic ). The third set of efficiency assets is that which is achievable. As highlighted above, the HAP is the economic potential discounted for market barriers such as customer preferences and supply chain maturity; while AP is additionally discounted for programmatic barriers such as program budgets and execution proficiency. Of the total Technical potential, typically only a fraction is ultimately achievable and only then over time due to the existence of market barriers. The question of how much effort and money is to be deployed towards removing or lowering the barriers is a decision made by state governing bodies (legislatures, regulators or both). The achievable potential range is typically a fraction of the economic potential range. This achievable amount must be further split between what can or should be accomplished with utility-sponsored programs and what should fall under codes and standards. Both amounts are represented in this IRP as reductions to what would otherwise be the load forecast. 74

88 2015 Integrated Resource Plan Determining Future Demand Side Programs Incremental Energy Efficiency (EE) To determine the economic EE activity to be modeled that would be in addition to assumed EE programs already reflected in the load forecast, a determination was made as to the potential level and cost of such incremental EE activity as well as the ability to expand current programs. Figure 18 and Figure 19 show the going-in make-up of projected consumption in 2018 for PSO s residential and commercial sectors, respectively. It was assumed that the incremental programs modeled would be effective in 2019, due to the time needed to develop specific program cost and measures and receive any necessary regulatory approval to implement such programs. Figure Projected Residential Energy Consumption (GWh) by End Use 75

89 2015 Integrated Resource Plan Total = 5,318 Figure Projected Commercial Energy Consumption (GWh) by End Use The current programs target certain end-uses in both sectors. Future incremental EE activity can further target those areas or address other end-uses. To determine which end-uses are targeted, and in what amounts, PSO looked at the previously-cited 2014 EPRI Report. This report provides comprehensive information on a multitude of current and anticipated end-use measures including their cost, energy savings, Market Acceptance Ratios (MARs) and Program Implementation Factors (PIFs). PSO utilized this data to develop bundles of future EE activity for the demographics and weather-related impacts of its service territory. A bundle is a combination of energy efficiency measures intended to provide energy savings to customers at a reasonable cost. Table 9 and Table 10, from the EPRI Report, list the individual measure categories considered for both the residential and commercial sectors, respectively. 76

90 2015 Integrated Resource Plan Table 9. Residential Sector Energy Efficiency (EE) Measure Categories Central Air Conditioning Programmable Thermostat Storm Doors Dehumidifier Air-Source Heat Pumps Water Heating External Shades Dishwashers Ground-Source Heat Pumps Faucet Aerators Ceiling Insulation Clothes Washers Room Air Conditioning Pipe Insulation Foundation Insulation Clothes Dryers Air Conditioning Maintenance Heat Pump Maintenance Low-Flow Showerheads Foundation Insulation Refrigerators Dishwashers (Domestic Hot Water) Wall Insulation Freezers Attic Fan Furnace Fans Windows Cooking Furnace Fans Lighting Linear Fluorescent Reflective Roof Televisions Ceiling Fan Lighting Screw-in Reflective Roof Personal Computers Whole-House Fan Enhanced Customer Bill Presentment Duct Repair Smart Plug Strips, Reduce Standby Wattage Duct Insulation Infiltration Control 77

91 2015 Integrated Resource Plan Table 10. Commercial Sector Energy Efficiency (EE) Measure Categories Heat Pumps Central Air Conditioning Chiller Cool Roof Variable Speed Drive on Pump Economizer Energy Management System Roof Insulation Duct Insulation Water Heater Water Temperature Reset Computers Fans, Energy-Efficient Motors Fans, Variable Speed Control Programmable Thermostat Variable Air Volume System Lighting Screw-in High-Efficiency Compressor Anti-Sweat Heater Controls Floating Head Pressure Controls Servers Duct Testing and Sealing Installation of Glass Doors Displays HVAC Retrocommissioning High-Efficiency Vending Machine Copiers Printers Efficient Windows Icemakers Other Electronics Lighting Linear Fluorescent Reach-in Coolers and Freezers What can be derived from the tables is that the 2014 EPRI report has taken a comprehensive approach to identifying available EE measures. From this information, PSO has developed proxy EE bundles for both residential and commercial customer classes to be modeled within Plexos. These bundles are based on measure characteristics identified within the EPRI report and PSO customer usage, and are shown in Section Conservation Voltage Reduction (CVR) As discussed in Section 3.5.6, CVR equipment is an additional resource that reduces enduse consumption. This resource is assumed to be available in amounts that can be reasonably installed and tested in a given year. Estimates of CVR opportunities were developed and then grouped into viably-sized tranches to be modeled within Plexos Advanced Metering Infrastructure (AMI) SWEPCO and AEP recognize the potential value of smart meters and the possible options it may provide customers with respect to managing energy consumption. SWEPCO also realizes, as with most new advancements in technology, the initial cost of the technology is considerable and there are unforeseen complications with the implementation of the technology. In an attempt to minimize any negative impact to customers, SWEPCO is currently analyzing the 78

92 2015 Integrated Resource Plan benefits of this new technology through the implementation programs currently being deployed by several sister companies at AEP Demand Response (DR) The current level of DR is maintained throughout the Plan and was discussed in Section Looking into the future, other options, including expanded residential DR, may be considered Distributed Generation (DG) DG resources were evaluated assuming a residential rooftop solar photovoltaic (PV) resource, as this is the primary distributed resource. Solar has favorable characteristics in that it produces the majority of its energy at near-peak usage times. As previously described in Section 3.5.8, DG resources (i.e., rooftop Solar) are included in the model at an assumed growth rate based on the current level of federal incentives, future estimated costs of rooftop solar and historical rooftop solar additions Evaluating Incremental Demand-Side Resources The Plexos model allows the user to input incremental EE, DG and CVR as resources, thereby considering such alternatives in the model on equal-footing with more traditional supplyside generation resource options Incremental Energy Efficiency (EE) Modeled Table 11 and Table 12 list the cost and energy profiles of measure bundles for both residential and commercial EE, respectively, that were constructed for modeling purposes. 79

93 2015 Integrated Resource Plan Table 11. Incremental Demand-Side Residential Energy Efficiency (EE) Bundle Summary Bundle Installed Cost ($/kwh) Yearly Potential Energy Savings (MWh) Yearly Potential Energy Savings (MWh) Yearly Potential Energy Savings (MWh) Yearly Potential Energy Savings (MWh) Bundle Life Thermal Shell - AP ,193 1,893 2,597 3, Thermal Shell - HAP ,848 20,548 24,093 13, Water Heating - AP , Water Heating - HAP ,104 3,185 2,995 1, Appliances - AP ,075 1,788 1,789 1, Appliances - HAP ,937 9,756 8,130 2, Cooling - AP ,610 7,199 9,041 3, Cooling - HAP ,398 19,636 19,259 9, Lighting - AP ,721 1, Lighting - HAP ,996 19,534 14,088 2, Table 12. Incremental Demand-Side Commercial Energy Efficiency (EE) Bundle Summary Bundle Installed Cost ($/kwh) Yearly Potential Energy Savings (MWh) Yearly Potential Energy Savings (MWh) Yearly Potential Energy Savings (MWh) Yearly Potential Energy Savings (MWh) Bundle Life Heating - AP , Heating - HAP , Cooling - AP ,538 2,207 2, Cooling - HAP ,472 2, Office Equipment - AP , Office Equipment - HAP ,750 1, Indoor Lighting - AP ,887 1,519 1,731 1, Indoor Lighting - HAP ,139 8,214 6,755 1, As can be seen from the tables, each program has both AP and HAP characteristics. The development of these characteristics is based on the 2014 EPRI EE Potential report that has been previously referenced. This report further identifies MARs and PIFs to apply to primary measure savings, as well as Application Factors for secondary measures. Secondary measures are not consumers of energy, but do influence the system that is consuming energy. The Residential Thermal Shell, Residential Water Heating and Commercial Cooling bundles in both AP and HAP include secondary measures. The MAR and PIF are utilized to develop the incremental AP program characteristics and the MAR only is used to develop the incremental HAP program characteristics. All of the bundles were offered into the Plexos model Conservation Voltage Reduction (CVR) Modeled Potential future PSO CVR circuits considered for modeling varied in relative cost and energy-reduction effectiveness. The circuits were grouped into 9 tranches based on the relative potential demand reduction of each tranche of circuits. The Plexos model was able to pick the 80

94 2015 Integrated Resource Plan most cost-effective tranches first and add subsequent tranches as merited. Typically, a CVR tranche includes approximately circuits. Table 13 illustrates all of the tranches offered into the model and the respective cost and performance of each. The costs shown are in 2015 dollars. Table 13. Conservation Voltage Reduction (CVR) Tranche Profiles Tranche Number of Capital Annual Demand Energy Reduction Circuits Investment O&M Reduction (kw) (MWh) 1 48 $ 15,990,000 $ 479,700 9,611 42, $ 16,750,000 $ 502,500 10,695 46, $ 14,810,000 $ 444,300 9,197 45, $ 15,740,000 $ 472,200 9,572 41, $ 15,760,000 $ 472,800 8,726 37, $ 17,400,000 $ 522,000 12,647 58, $ 14,150,000 $ 424,500 9,927 46, $ 14,830,000 $ 444,900 7,580 30, $ 14,020,000 $ 420,600 8,890 42, Demand Response (DR) Modeled Additional levels of DR were not modeled as an incremental resource within this Plan. However, DR associated with known and anticipated interruptible and real-time pricing initiatives have already been incorporated into PSO s future going-in capacity position, as shown in Section Distributed Generation Modeled DG resources were modeled as described Section Optimizing Incremental Demand-side Resources The Plexos software views demand-side resources as non-dispatchable generators that produce energy similar to non-dispatchable supply-side generators such as wind or solar. Thus, the value of each resource is impacted by the hours of the day and time of the year that it generates energy. 4.5 Identify and Screen Supply-side Resource Options Capacity Resource Options New construction supply-side alternatives were modeled to represent peaking and baseload/intermediate capacity resource options. To reduce the number of modeling 81

95 2015 Integrated Resource Plan permutations in Plexos, the available technology options were limited to certain representative unit types. However, it is important to note that alternative technologies with comparable cost and performance characteristics may ultimately be substituted should technological or marketbased profile changes warrant. The options assumed to be available for modeling analyses for PSO are presented in Table 14. When applicable, PSO may take advantage of economical market capacity and energy opportunities. Prospectively, these opportunities could take the place of currently planned resources and will be evaluated on a case-by-case basis New Supply-Side Capacity Alternatives As identified in Table 14, natural gas base/intermediate and peaking generating technologies were considered in this IRP as well as utility-scale solar and wind. Further details on these technologies are available in Exhibit C of the appendix. To reduce the problem size within Plexos, the number of alternatives explicitly modeled was reduced through an economic screening process which analyzed various supply options and developed a quantitative comparison for each duty-cycle type of capacity (i.e., baseload, intermediate, and peaking) on a forty-year, levelized basis. The options were screened by comparing levelized annual busbar costs over a range of capacity factors. In this evaluation, each type of technology is represented by a line showing the relationship between its total levelized annual cost per kw and an assumed annual capacity factor. The value at a capacity factor of zero represents the fixed costs, including carrying charges and fixed Operation and Maintenance (O&M) costs, which would be incurred even if the unit produced no energy. The slope of the line reflects variable costs, including fuel, emissions, and variable O&M, which increase in proportion to the energy produced. The best of class technology, for each duty cycle, determined by this screening process was explicitly modeled the Plexos. These generation technologies were intended to represent reasonable proxies for each capacity type (baseload, intermediate, peaking). Subsequent substitution of specific technologies could occur in any later plan, based on emerging economic or non-economic factors not yet identified. 82

96 2015 Integrated Resource Plan AEP continually tracks and monitors changes in the estimated cost and performance parameters for a wide array of generation technologies. Utilizing access to industry collaborative organizations such as EPRI and the Edison Electric Institute, AEP s association with architect and engineering firms and original equipment manufacturers as well as its own experience and market intelligence, AEP provides current estimates to the planning process. Table 14 offers a summary of the most recent technology performance parameter data developed. Table 14. New Generation Technology Options with Key Assumptions Type Base Load Nuclear Base Load (90% CO2 Capture New Unit) Pulv. Coal (Ultra-Supercritical) (PRB) IGCC "F" Class (PRB) Base / Intermediate Combined Cycle (1 - "F" Class) Combined Cycle (2 - "F" Class) Combined Cycle (2X1 "G" Class, w/duct firing & evap coolers) Combined Cycle Repowered (1X1 "G" Class, w/ evap coolers) Peaking Combustion Turbine (2 - "E" Class) Combustion Turbine (2 - "F" Class, w/evap coolers) Aero-Derivative (1 - Small Machine) Aero-Derivative (2 - Small Machines) Recip Engine Farm (3 Engines) Capability (MW) (a) Notes: (a) Capability at Standard ISO Conditions at 1,000 feet above sea level. SO2 (lb/mmbtu) Emission Rates Nox (lb/mmbtu) CO2 (lb/mmbtu) Capacity Factor (%) Overall Availability (%) Baseload/Intermediate Alternatives Coal and nuclear baseload options were evaluated by PSO but were not included in the ultimate Plexos resource optimization modeling analyses. For coal generation resources, existing and proposed EPA regulations, described in Section 3.4, currently make the construction of new coal plants economically impractical due to the implicit requirement of Carbon Capture and Sequestration (CCS) technology. New nuclear construction is currently financially impractical since it would potentially require an investment of, minimally, $6,000/kW. Intermediate generating sources are typically expected to serve a load-following and cycling duty and effectively shield base-load units from that obligation. Historically, many generators have relied on older, smaller, less-efficient/higher dispatch cost, subcritical coal-fired or gassteam units to serve such load-following roles. Over the last several years, these units staffs have made strides to improve ramp rates, regulation capability, and reduce downturn (minimum 83

97 2015 Integrated Resource Plan load capabilities). As the fleet continues to age and subcritical units are retired, other generation dispatch alternatives and new generation will need to be considered to cost effectively meet this duty cycle s operating characteristics Natural Gas Combined-Cycle (NGCC) An NGCC plant combines a steam cycle and a combustion gas turbine cycle to produce power. Waste heat (~1,100 F) from one or more combustion turbines passes through a Heat Recovery Steam Generator (HRSG) producing steam. The steam drives a steam turbine generator which produces about one-third of the NGCC plant power, depending upon the gas-tosteam turbine design platform, while the combustion turbines produce the other two-thirds. The main features of the NGCC plant are high reliability, reasonable capital costs, operating efficiency - at 45-60% Low Heating Value(LHV) - low emission levels, small footprint and shorter construction periods than coal-based plants. In the past 8 to 10 years, NGCC plants were often selected to meet new intermediate and base-load needs. NGCC plants may be designed with the capability of being islanded which would allow them, in concert with an associated diesel generator, to perform system restoration (black start) services. Although cycling duty is typically not a concern, an issue faced by NGCC when load-following is the erosion of efficiency due to an inability to maintain optimum air-to-fuel pressure and turbine exhaust and steam temperatures. Methods to address these include: Installation of advanced automated controls. Supplemental firing while at full load with a reduction in firing when load decreases. When supplemental firing reaches zero, fuel to the gas turbine is cutback. This approach would reduce efficiency at full load, but would likewise greatly reduce efficiency degradation in lower-load ranges. Use of multiple gas turbines coupled with a waste heat boiler that will give the widest load range with minimum efficiency penalty Northeastern 4 Repowering The Northeastern Unit 4 coal unit is being retired in This provides PSO an opportunity to consider the possibility to repower the unit. Repower consists of adding one or two natural gas fired combustion turbine and HRSG modules to the unit. Each natural gas fired combustion turbine would provide power to the grid as well as exhaust heat to produce useful 84

98 2015 Integrated Resource Plan steam within the HRSG. This steam would then be routed to the existing Unit 4 steam turbine and generator to produce additional power to the grid. Additional major systems/equipment utilized for the repower include the: condenser, cooling tower, land, substation interconnection and operations building/infrastructure. The performance characteristics of the 1x1 repowered configuration are shown in Table 15 and in Exhibit C. The 2x1 repowered configuration was not modeled, but based on its similar cost and performance profile, it would be expected to perform similarly to the 1x1 configuration. While the repower study which the established cost and performance parameters used for modeling was complete and comprehensive, it should be noted the study did not include a detailed engineering study. The cost estimate used for modeling complies with the Association for the Advancement of Cost Engineering International (AACEI) Class 4 estimate with an accuracy range of -30% to +50%. Therefore, its efficacy as compared to other options is uncertain and requires significant engineering to improve its accuracy Peaking Alternatives Peaking generating sources provide needed capacity and energy during high demand periods and/or periods in which significant shifts in the load (or supply) curve dictate the need for quick-response capability. The peaks occur for only a few hours each year so the capacity dedicated to serving this reliability function can be expected to provide relatively little energy over an annual load cycle. As a result, fuel efficiency and other variable costs applicable to these resources are of lesser concern. Ultimately, peaking resources requirements are manifested in the system load duration curve. Peaking resources in SPP can be advantageous when utilizes as Quick-Start Resources. A Quick-Start Resource is a generating unit which can be started, synchronized, and supply energy to the electric grid within 10 minutes from the time they are called upon to produce energy. When units perform as expected they are eligible for additional compensation. In certain situations, peaking capacity such as combustion turbines can provide backup and some have the ability to provide emergency (Black Start) capability to the grid. 85

99 2015 Integrated Resource Plan Simple-Cycle Natural Gas Combustion Turbines (NGCT) In industrial or frame-type combustion turbine systems, air compressed by an axial compressor is mixed with fuel and burned in a combustion chamber. The resulting hot gas then expands and cools while passing through a turbine. The rotating rear turbine not only runs the axial compressor in the front section but also provides rotating shaft power to drive an electric generator. The exhaust from a combustion turbine can range in temperature between 800 and 1,150 degrees Fahrenheit and contains substantial thermal energy. An NGCT system is one in which the exhaust from the gas turbine is vented to the atmosphere and its energy lost, i.e., not recovered as in a combined-cycle design. While not as efficient at 30-35% LHV, they are inexpensive to purchase, compact, and simple to operate Aeroderivatives (AD) AD units are aircraft jet engines used in ground installations for power generation. They are smaller in size, lighter weight, and can start and stop quicker than their larger industrial or frame counterparts. For example, the General Electric (GE) model 7EA frame machine requires 20 minutes to ramp up to full load while the smaller model LM6000 aeroderivative only needs 10 minutes from start to full load. However, the cost per kw of an aeroderivative is on the order of 20% higher than a frame machine. The AD performance operating characteristics of rapid startup and shutdown make the aeroderivatives well suited to peaking generation needs. AD units can operate at full load for a small percentage of the time allowing for multiple daily startups to meet peak demands, compared to frame machines which are more commonly expected to start up once per day and operate at continuous full load for 10 to 16 hours per day. The cycling capabilities provide AD units the ability to backup variable renewables such as solar and wind. This operating characteristic is expected to become more valuable over time as: a) the penetration of variable renewables increase; b) base-load generation processes become more complex limiting their ability to load-follow and; c) intermediate coal-fueled generating units are retired from commercial service. AD units weigh less than their industrial counterparts allowing for skid or modular installations. Efficiency is also a consideration in choosing an aeroderivative over an industrial turbine. AD units in the less than 100MW range are more efficient and have lower heat rates in 86

100 2015 Integrated Resource Plan simple cycle operation than industrial units of equivalent size. Exhaust gas temperatures are lower in the AD units. Some of the better known AD vendors and their models include GE's LM series, Pratt & Whitney's FT8 packages, and the Rolls Royce Trent and Avon series of machines Reciprocating Engines (RE) The use of RE units or internal combustion engines has increased over the last twenty years. According to EPRI, in 1993 about 5% of the total RE units sold were natural gas-fired spark ignition engines and post-2000 sales of natural gas-fired generators have remained above 10% of total units sold worldwide. Improvements in emission control systems and thermal efficiency have led to the increased utilization of natural gas-fired RE generators incorporated into multi-unit power generation stations for main grid applications. The RE generators have high efficiency, flat heat rate curves and rapid response makes this technology very well suited for peaking and intermediate load service and as back up to intermittent generating resources. Additionally, the fuel supply pressure required is in the range of 40 to 70 psig. This lower gas pressure allows this technology more siting flexibility. A further advantage of RE generators is that power output is less affected by increasing elevation and ambient temperature as compared to gas turbine technology. Also, a RE plant generally would consist of multiple units, which will be more efficient at part load operation than a single gas turbine unit of equivalent size because of the ability to shut down units and operate the remaining units at higher load. Common RE unit sizes have generally ranged from 8MW to 18MW per machine with heat rates in the range 8,100 to 8,600 Btu/kWh, using HHV. Regarding operating cost, RE generators have a somewhat greater variable O&M than a comparable gas turbine; however, over the long term, maintenance costs of RE are generally lower because the operating hours between major maintenance can be twice as long as gas turbines of similar size. The main North American suppliers for utility-scale natural gas-fired RE most recently have been Caterpillar and Wartsila Turbomachinery International, Jan/Feb. 2009; Gas Turbine World; EPRI Technical Assessment Guide. 87

101 2015 Integrated Resource Plan Renewable Alternatives Renewable generation alternatives use energy sources that are either naturally occurring (wind, solar, hydro or geothermal), or are sourced from a by-product or waste-product of another process (biomass or landfill gas). In the recent past, development of these resources has been driven either by a state renewable portfolio requirement or as in the case of Oklahoma highquality low-cost sources such as wind. Advancements in both solar PV and wind turbine manufacturing have reduced both installed and ongoing costs making them economically justified in much broader areas of the country Utility-Scale Solar Solar power comes in two forms to produce electricity: concentrating and photovoltaics. Concentrating solar which heats a working fluid to temperatures sufficient to power a turbine produces electricity on a large scale and is similar to traditional centralized supply assets in that way. Photovoltaics produce electricity on a smaller scale (typically 2kW to 20MW per installation) and can be distributed throughout the grid. The cost of installed solar projects has declined considerably in the past decade and is expected to continue to decline, as shown in Figure 20. This has been mostly a result of reduced panel prices that have resulted from manufacturing efficiencies spurred by accelerating penetration of solar energy in Europe, Japan, and California. With the trend firmly established, forecasts generally foresee declining nominal prices in the next decade as well. 10 Technical Assessment Guide Power Generation and Storage Technology Options, 2012; Electric Power Research Institute. 88

102 2015 Integrated Resource Plan Figure 20. Forecasted Solar Installed Costs (nominal) for Oklahoma (Excl. Federal & State Incentives) Not only are utility-scale solar plants getting less expensive, the costs to install solar panels in distributed locations, often on a rooftop, are lessening as associated hardware, such as inverters, racks, and wiring bundles become standardized. If the projected cost declines materialize, both distributed and utility-scale solar projects will be economically justifiable in the future. Utility solar plants require less lead time to build than fossil plants. There is not a defined limit to how much utility solar can be built in a given time. However, in practice, solar facilities are not added in an unlimited fashion. Solar resources were considered available resources with some limits on the rate with which they could be chosen. In the IRP modeling, the assumption was made that utility-scale solar resources were available up to 50MWac 11 of nameplate capacity starting in To provide 11 Manufacturers usually quote system performance in DC watts, however electric service from the utility is supplied in AC watts. An inverter converts the DC electrical current into AC electrical current. Depending on the inverter efficiency, the AC wattage may be anywhere from 80 to 95 percent of the DC wattage. 89

103 2015 Integrated Resource Plan some context, a typical commercial installation is 50 kw and effectively covers the surface of a typical big box retailer s roof. A 50MW utility-scale solar farm is assumed to consume nearly 350 acres, or 1,000 big box retailer roofs. A limit on solar capacity additions is needed because as solar costs continue to decrease relative to the market price of energy, there will come a point where the optimization model will theoretically pick an unlimited amount of solar resources. This 50MWac annual threshold recognizes that there is a practical limit as to the number of sites that can be identified, permitted and constructed by PSO in a given year. Certainly as PSO gains experience with solar installations, this limit would likely be modified (for example, it may be lower earlier and greater later). Solar resources useful capacity is less than its nameplate rating. This IRP assumes solar resources will have capacity valued at 35% of nameplate rating Wind Utility-scale wind energy is generated by turbines ranging from 1.0 to 2.5MW, with a 1.5MW turbine being the most common size used in commercial applications today. Typically, multiple wind turbines are grouped in rows or grids to develop a wind turbine power project which requires only a single connection to the transmission system. Location of wind turbines at the proper site is particularly critical as not only does the wind resource vary by geography, but its proximity to a transmission system with available capacity will factor into the cost. A variable source of power in most non-coastal locales, with capacity factors ranging from 30 percent (in the eastern portion of the U.S.) to over 50 percent (largely in more westerly portions of the U.S., including the Plains states), wind energy s life-cycle cost ($/MWh), excluding subsidies, is currently higher than the marginal (avoided) cost of energy, in spite of its negligible operating costs. Another consideration with wind power is that its most critical factors (i.e., wind speed and sustainability) are typically highest in very remote locations, which forces the electricity to be transmitted long distances to load centers necessitating the build out of EHV transmission to optimally integrate large additions of wind into the grid. For modeling purposes, wind was considered under various blocks or tranches for each year. There are two tranches of wind with different pricing. The first tranche of wind resources, 90

104 2015 Integrated Resource Plan Tranche A was modeled as a 100MW block with a levelized cost of energy (LCOE) with the Production Tax Credit (PTC) of $24/MWh in 2015 and a 55% capacity factor load shape. In 2017, after the expiration of the PTC, the LCOE of Tranche A increases to $47/MWh in nominal dollars with prices increasing 0.5%/year through Tranche A resources were assigned a capacity value of 20% of nameplate rating. The second tranche of wind resources, Tranche B, was modeled as a 100MW block with a LCOE with the PTC of $28/MWh in 2015$ and a 50% capacity factor load shape. In 2017, after the assumed expiration of the PTC, the LCOE of Tranche B increases to $51/MWh in nominal dollars with prices increasing 2%/year through Tranche B resources were assigned a capacity value of 10% of nameplate rating. Wind prices were developed based on the U.S. DOE s Wind Vision Report. 12 The expected magnitude of wind resources available per year was limited to 400MW (nameplate) with a limit of 500MW nameplate, over the planning period, incremental to that which is currently planned. This cap is based the DOE s Wind Vision Report chart on page 12 that suggest from numerous transmission studies that transmission grids should be able to support 20% to 30% of intermittent resources in the 2020 to 2030 timeframe. The cap for PSO allows the model to select up to 30% of generation capacity resources as wind-powered by Figure 21 illustrates the two tranches of wind resources modeled and the relative LCOE per MWh utilized for each tranche. 12 WindVision: A New Era for Wind Power in the United States (2015). Retrieved from 91

105 2015 Integrated Resource Plan Cost increase due to expiration of Production Tax Credit Figure 21. LCOE (nominal $/MWh) for Wind Resource Tranches Included in PSO Model Hydro The available sources of, particularly, larger hydroelectric potential have largely been exploited and those that remain must compete with the other uses, including recreation and navigation. The potentially lengthy time associated with environmental studies, Federal Army Corps of Engineer permitting, high up-front construction costs, and environmental issues (fish and wildlife) make hydro unattractive at this time. As such, no incremental hydroelectric resources were considered in this IRP Biomass Biomass is a term that typically includes organic waste products (sawdust or other wood waste), organic crops (corn, switchgrass, poplar trees, willow trees, etc.), or biogas produced from organic materials, as well as select other materials. Biomass costs will vary significantly depending upon the feedstock. Biomass is typically used in power generation through the utilization of the biomass fuel in a steam generator (boiler) that subsequently drives a steam turbine generator; similar to the same process of many traditional coal fired generation units. Some biomass generation facilities use biomass as the primary fuel, however, there are some 92

106 2015 Integrated Resource Plan existing coal-fired generating stations that will use biomass as a blend with the coal. Given these factors, plus the typical high cost and required feedstock supply and attendant long-term pricing issues, no incremental biomass resources were considered in this IRP Cogeneration & Combined Heat & Power (CHP) Cogeneration is a process where electricity is generated and the waste heat by-product is used for heating or other process, raising the net thermal efficiency of the plant. To take advantage of the increased efficiency associated with CHP, the host must have a ready need for the heat that is otherwise potentially wasted in the generation of electricity. The Company does not reflect new projects in its IRP until they are assessed for reasonableness. This IRP does not include any incremental CHP projects. 4.6 Integration of Supply-Side and Demand-Side Options within Plexos Modeling Optimize Expanded DSM Programs As described in Section 4.4, EE and CVR options that would be incremental to the current programs were modeled as resources within Plexos. In this regard, they are demand-side power plants that produce energy according to their end use load shape. EE options have an initial (program) cost with no subsequent annual operating costs. CVR options have both an initial program cost and an ongoing annual operating cost Optimize Other Demand-Side Resources Customer-sited DG, specifically rooftop solar, was not modeled. Instead, reductions in energy use and peak demand were built into the load forecast based on the adoption rates discussed in Section DG installation costs to PSO were zero, with all costs paid by the customer. 93

107 2015 Integrated Resource Plan 5.0 Resource Portfolio Modeling 5.1 The Plexos Model - An Overview Plexos Liner Program long-term optimization model, also known as LT Plan, served as the basis from which PSO s capacity requirement evaluations were examined and recommendations were made. The LT Plan model finds the optimal portfolio of future capacity and energy resources, including DSM additions, which minimizes the CPW of a planning entity s generation-related variable and fixed costs over a long-term planning horizon. By minimizing CPW the model will provide optimized portfolios with the lowest and most stable customer rates, while adhering to the Company s constraints. Plexos accomplishes this by using an objective function which seeks to minimize the aggregate of the following capital and production-related (energy) costs of the portfolio of resources: fixed costs of capacity additions, i.e., carrying charges on incremental capacity additions (based on an PSO-specific, weighted average cost of capital), and fixed O&M; fixed costs of any capacity purchases; program costs of (incremental) DSM alternatives; variable costs associated with PSO generating units. This includes fuel, start-up, consumables, market replacement cost of emission allowances, and/or carbon tax, and variable O&M costs; a netting of the production revenue earned in the SPP power market from PSO s generation resource sales and the cost of energy based on unique load shapes from SPP purchases necessary to meet PSO s load obligation. Plexos executes the objective function described above while abiding by the following possible constraints: Minimum and maximum reserve margins; Resource additions (i.e., maximum units built); Age and lifetime of power generation facilities; Retrofit dependencies (Selective Catalytic Reduction (SCR) and FGD combinations); Operation constraints such as ramp rates, minimum up/down times, capacity, heat rates, etc.; 94

108 2015 Integrated Resource Plan Fuel burn minimum and maximums; Emission limits on effluents such as SO 2 and NO X ; and Energy contract parameters such as energy and capacity. The model inputs that compose the objective function and constraints are considered in the development of an integrated plan that best fits the utility system being analyzed. Plexos does not develop a full regulatory Cost-of-Service (COS) profile. Rather, it typically considers only the relative load and generation COS that changes from plan-to-plan, and not fixed embedded costs associated with existing generating capacity and demand-side programs that would remain constant under any scenario. Likewise, transmission costs are included only to the extent that they are associated with new generating capacity, or are linked to specific supply alternatives. In other words, generic (nondescript or non-site-specific) capacity resource modeling would typically not incorporate significant capital expenditures for transmission interconnection costs Key Input Parameters Major underpinnings of this process are long-term forecasts of PSO s energy requirements and peak demand, as well as the price of various generation-related commodities, including energy, coal, natural gas and, potentially, CO 2 (i.e. carbon). The load and commodity forecasts were created internally within AEP. The PSO load forecast was created by the AEP Economic Forecasting organization, while the long-term commodity pricing forecast was created by the AEP Fundamental Analysis group. These groups have years of experience forecasting PSO and AEP system-wide demand and energy requirements and fundamental pricing for both internal operational and regulatory purposes. Moreover, the Fundamental Analysis group routinely performs peer review by comparing and contrasting its commodity pricing projections versus consensus pricing on the part of outside forecasting entities such as HIS Cambridge Energy Research Associates (CERA), Petroleum Industry Research Associates (PIRA) and the U.S. EIA. Additional critical input parameters include the installed cost of replacement capacity alternative options, as well as the attendant operating costs associated with those options. This data came from the AEP Engineering Services organization. 95

109 2015 Integrated Resource Plan 5.2 Plexos Optimization Modeling Options and Constraints The major system limitations that were modeled by use of constraints are elaborated on below. The Plexos Linear Program, LT Plan, optimization algorithm operates modeled constraints in tandem with the objective function in order to yield the least-cost resource plan. There are many variants of available supply-side and demand-side resource options and types. As a practical limitation, not all known resource types are made available as modeling options. A screening of available supply-side technologies was performed with the optimum assets made subsequently available as options. Such screens for supply alternatives were performed for duty cycle families (baseload, intermediate, and peaking). The selected technology alternatives from this screening process do not necessarily represent the optimum technology choice for that duty-cycle family. Rather, they reflect proxies for modeling purposes. Other factors which will determine the ultimate technology type (e.g., choices for peaking technologies) are taken into consideration. Based on the established comparative economic screenings, the following specific supply alternatives were modeled in Plexos for each designated duty cycle: Peaking capacity o NGCT, consisting of two E class turbines at 180MW total o AD, consisting of two units at 91MW total. Intermediate capacity o NGCC installation (2x1 G class turbine with duct firing and inlet air cooling) sized at 390MW (50% share), with a summer capacity rating with duct-firing and inlet air cooling of 435. o NGCC repowering of Northeastern Unit 4 (1x1 G class turbine with duct firing) sized at 339MW, with a summer capacity with duct-firing of 390MW. This 1x1 Combined-Cycle facility is model with provisions for adding an additional Combustion-Turbine and HRSG to enable 2x1 operations. Wind resources were made available in two tiers based on expected output capacity factor. The first tier had an expected capacity factor of 55% and the second tier had an expected capacity factor of 50%. Each tier was made up of 100MW blocks of nameplate capacity and a total of up to 400MW could be added 96

110 2015 Integrated Resource Plan annually (2-100MW tranches of each tier could be added each year). In 2016, a total of 400MW of Tier 1 and Tier 2 wind could be added at a cost that assumed PTCs were available. After 2016, the wind could be added at a cost that assumes PTCs are not available. A total of 500MW of nameplate wind capacity, incremental to that which is currently planned, could be installed over the planning period. Tier 1 wind had a SPP capacity credit of 20MW and Tier 2 capacity value was 10MW. Utility-scale solar resources were made available up to 50MW annually of incremental nameplate capacity applied to an externally-derived declining installed cost curve. Solar resources were assumed to have a SPP capacity credit equal to 35% of nameplate. DG, in the form of distributed solar resources in 5kW sizes, was embedded in amounts equal to approximately 5% of annual increases from current levels. EE resources incremental to those already incorporated into the Company s long-term load and peak demand forecast in up to 18 unique bundles of residential and commercial measures considering cost and performance parameters for both HAP and AP categories. EE resources were made available in CVR was available in 9 tranches of varying installed costs and number of circuits/sizes ranging from a low of 9MW, up to 15MW of demand savings potential Optimized Portfolios One of the key decisions to be made by PSO during the planning period is how to fill the significant need for resources in the early 2020s. Portfolios with various options addressing PSO s capacity and energy resource needs over time were optimized using the Base load and demand forecast, but under five different long-term commodity pricing scenarios: 1. Base pricing 2. High Band pricing 3. Low Band pricing 4. High Carbon pricing 5. No Carbon pricing Two sensitivity portfolio evaluations were conducted under Base commodity pricing, using two different long-term load (and peak demand) forecasts: 6. High Load sensitivity 7. Low Load sensitivity 97

111 2015 Integrated Resource Plan The results of all seven scenarios were reviewed in detail. Each portfolio provides insight into the implications and risks associated with actual market conditions changing from those forecasted in the base plan Optimization Modeling Results under the Base-Load Forecast Scenarios 1 through 5 were all optimized under the base-load forecast. The incremental capacity additions for each portfolio are shown below in Table 15. Note that all portfolios include a diversity of resource options such as natural gas fired generation, energy efficiency, and renewable resources. The capacity values listed for renewable resources represent the SPPcredited capacity values. 98

112 2015 Integrated Resource Plan Table 15. Cumulative SPP Capacity Additions (MW) for Five Commodity Pricing Scenarios Commodity Pricing Scenarios Base Base Load 825 1,260 1,260 Peaking Solar Wind EE CVR DG High Base Load 825 1,260 1,260 Peaking Solar Wind EE CVR DG Low Base Load 825 1,260 1,260 Peaking Solar Wind EE CVR DG High Carbon Base Load 825 1,260 1,260 Peaking Solar Wind EE CVR DG No Carbon Base Load 825 1,260 1,260 Peaking Solar Wind EE CVR DG Base Load=NGCC; Peaking=Frame CT, Aero CT; EE=Energy Efficiency; CVR=Conservation Voltage Reduction; DG=Distributed Generation The optimized plan results in Table 15 provide PSO with insight in developing its IRP and five-year action plan. For example, no peaking generation is required in any of the five commodity pricing scenarios throughout the planning period. Also, base-load generation is not needed until 2022, which aligns with the timeline for both the retirement of Southeastern Unit 2 and PSO s 520MW PPA with Exelon. Table 15 shows that each of the scenarios which account 99

113 2015 Integrated Resource Plan for a carbon tax (scenarios 1-4) call for new wind resources beginning in 2016, illustrating the value of wind s carbon-free energy. In the first three scenarios (Base, High, and Low commodity pricing) which utilize a $15 per metric ton CO 2 tax, 20MW of wind capacity was added (100MW nameplate). When this tax was increased to $25 per metric ton CO 2 in the High Carbon scenario the model called for 40MW of wind capacity (200MW nameplate). No wind was selected in the fifth scenario where no CO 2 tax was applied Optimization Modeling Results of Load Sensitivity Scenarios Table 16 illustrates the anticipated relationship between the forecasted load and the company s required resources. The High Load scenario calls for additional base-load generation a yearly earlier, in 2021, than in based load scenarios analyzed above. The Low Load scenario shows new base-load generation beginning in 2022, however the total capacity required over the planning period is less than in all other scenarios. Both scenarios call for 40MW (200MW nameplate) of wind in Table 16. Cumulative SPP Capacity Additions (MW) for Load Sensitivity Scenarios Load Scenarios High Load Base Load 435 1,260 1,260 1,260 Peaking Solar Wind EE CVR DG Low Load Base Load Peaking Solar Wind EE CVR DG Base Load=NGCC; Peaking=Frame CT, Aero CT; EE=Energy Efficiency; CVR=Conservation Voltage Reduction; DG=Distributed Generation 100

114 2015 Integrated Resource Plan 5.3 Selected Plan Each of the seven scenarios provides insight into a potential alternative mix of resources for the future. For PSO the optimized portfolio modeled for the base commodity pricing scenario is the Selected Plan. This plan was developed based on the following considerations: Minimizing revenue requirements (i.e. cost to customers) over the planning period, while meeting capacity obligations Optimizes the mix of generation to hedge short-term energy price volatility in the SPP Integrated Marketplace. Installing economical CVR and other incremental DSM. Adding renewable energy resources (wind and solar) in a cost effective manner. The cumulative capacity additions associated with the Selected Plan are shown below in Table 17. Table 17. Cumulative SPP Capacity Additions (MW) and Average Annual Energy Position (GWh) for Selected Plan Selected Plan Base Commodity, Base Load Base Load 825 1,260 1,260 Peaking Solar Wind EE CVR DG Base Load=NGCC; Peaking=Frame CT, Aero CT; EE=Energy Efficiency; CVR=Conservation Voltage Reduction; DG=Distributed Generation Energy Position (GWh) 2016 (5,222) 2024 (2,901) In conjunction with the Company s five-year action plan, the Selected Plan offers PSO significant flexibility should future conditions differ considerably from its assumptions. For example, as EE programs are implemented, PSO will gain insight into customer acceptance and develop additional hard data as to the impact these programs have on load growth. This will assist PSO in determining whether to expand program offerings, change incentive levels for programs, or target specific customer classes for the best results. If current long-term renewable costs assumptions change, PSO could either accelerate or delay the installation of renewable generation facilities. Changes to PSO s existing portfolio associated with this Selected Plan are described in greater detail in Section 6.1 of this report. 101

115 2015 Integrated Resource Plan Demand-Side Management (DSM) Results Energy Efficiency (EE) Results In the Selected Plan, incremental EE resources were selected throughout the planning period. Overall, both residential and commercial programs are providing 240GWh of energy savings and a peak capacity reduction of 29MW by the end of the planning period. The programs providing the majority of the savings are Commercial Lighting, Residential Lighting and Residential Thermal Shell programs. Figure 22, illustrates the EE savings from existing and incremental programs relative to sales. Figure 22 also shows that under the Selected Plan 1% of forecasted energy use will be accounted for with EE programs by Figure 22. PSO EE and CVR Energy (GWh) for Existing and Incremental Programs Conservation Voltage Reduction (CVR) In the Selected Plan, 6 of the 9 available CVR tranches were ultimately selected by the model, resulting in a cumulative capacity reduction of 70MW. The first tranche of circuits was added in 2019, and a single additional tranche was added in in each subsequent year through The cumulative capacity additions to the Selected Plan from these tranches are depicted below in Table 18. PSO currently has programs in place to begin the implementation of CVR in The capacity associated with these programs is included in the load forecast and is not 102

116 2015 Integrated Resource Plan represented in Table 18. Figure 22, above, shows the impact of CVR on PSO s energy position over the planning period. Table 18. Cumulative CVR Additions in Selected Plan (MW) CVR Capacity (MW) Distributed Generation (DG) DG resources were not selected under any economic scenario during the planning period. Distributed rooftop solar was included as a resource based on historical additions within PSO s service territory. Figure 23 below, illustrates the embedded rooftop solar as well as the forecasted distributed rooftop solar additions that were trended from the installation history. Figure 23. PSO Cumulative Rooftop Solar Additions/Projections CO 2 Emissions Figure 24, below, offers a long-term view of the PSO total company projected CO 2 emissions under a mass-based view for the Selected Plan portfolio. Such projected emission levels are identified as of the interim (2022 through 2029) as well as final (2030 and 103

117 2015 Integrated Resource Plan beyond) implementation periods set forth in the Final CPP. These charts offer a summary depiction of PSO s trends versus a 2012 (Actual) baseline regarding CO 2 emissions that result from actions undertaken as part of this IRP process. Figure 24. PSO Projected 'Mass-Based" CO2 Emissions under the Selected Plan 5.4 Risk Analysis The Selected Plan Portfolio and a Combustion Turbine Substitution Portfolio were evaluated using a stochastic, or Monte Carlo, modeling technique where input variables are randomly selected from a universe of possible values, given certain standard deviation constraints and correlative relationships. This offers a powerful tool to test various plans over a distributed range of certain key variables. The output is a distribution of possible outcomes, 104

118 2015 Integrated Resource Plan providing insight as to the risk or probability of a higher cost (revenue requirement) relative to the expected outcome. A substitute portfolio consisting of NGCTs was modeled using base commodity pricing and base load to evaluate the relative Revenue Requirement at Risk (RRaR) as compared to the Selected Plan which substantially relies on the addition of NGCC. The key facet of this scenario is that in 2022 four NGCTs were installed in place of the NGCC facility. The four combustion turbines were modeled as two sets of two GE model 7EA turbines, rated at 180MW per set, for a total capacity value of 360MW. The capacity from these combustion turbines effectively displaced the need for a new NGCC facility in 2022, which was included in the Selected Plan. Both scenarios which were stochastically modeled called for repowering Northeastern Unit 4 in The intent of modeling another portfolio was to observe the difference in revenue requirements between the Combustion Turbine Substitution portfolio and that of the Selected Plan. Combustion turbines have lower build costs that a combined-cycle facility, however combined-cycle facilities are more efficient and therefor result in more favorable dispatch costs. This study included multiple risk iteration runs performed over the IRP study period with three key price variables (risk factors) being subjected to this stochastic-based risk analysis. The results take the form of a distribution of possible revenue requirement outcomes for each plan for the planning horizon. Table 19 shows the input variables or risk factors and the correlation coefficients. The range of values associated with the variable inputs is shown Figure 25 through Figure 27. Table 19. Risk Factors and Their Relationships Gas Power CO2 Gas Power CO2 1 Standard Deviation 19% 14.7% 43% This risk modeling effort only evaluates the variable cost factors which differ between plans. Each portfolio includes resources that produce energy which is sold into the SPP market (even EE resources are treated as resources that sell the energy they avoid into the market at $0 variable cost), then recognize a net profit based on the variable cost of energy produced 105

119 2015 Integrated Resource Plan compared to the hourly market price at the time of energy production. This net profit is the number that is reported in each simulation run, and the range of these net profits is used to determine risk. Figure 25. Variable Input Prices for Gas (nominal $/mmbtu) $/MWh Average Max Min Figure 26. Variable Input Prices for Power (nominal $/MWh) 106

120 2015 Integrated Resource Plan Figure 27. Variable Input Prices for CO 2 (nominal $/metric ton) Stochastic Modeling Process and Results For each portfolio, the differential in net profit between the median (50 th percentile) and 95th percentile results from the multiple runs was identified as RRaR. The 95 th percentile is a level of required revenue sufficiently high that it will be exceeded, assuming the given plan is adopted, only five percent of the time. Thus, it is 95% likely that those higher-ends of revenue requirements would not be exceeded. The larger the RRaR, the greater the likelihood that customers could be subjected to higher costs relative to the portfolio s mean or expected cost. Conversely, there is equal likelihood costs may be lower than the median value. These higher or lower costs are generally the result of the difference, or spread, between fuel prices and the resultant SPP market energy prices, or CO 2 costs. The greater that spread, the more margin is enjoyed by the Company and its customers. Figure 28 illustrates the difference in CPW for each portfolio from the lowest cost iteration to the highest cost iteration. 107

121 2015 Integrated Resource Plan Figure 28. PSO Portfolio Revenue Requirement at Risk (RRaR) ($000 s), Figure 28 shows that the RRaR for the Selected Plan is greater than that of the Combustion Turbine Substitution portfolio by $62M. On the surface it appears that the Combustion Turbine Substitution portfolio is the better choice. However, the revenue risk must be considered with respect to the overall cost of the portfolio. Table 20 shows the CPW of both the Selected Plan and CT Substitution scenario. Table 20. Cumulative Present Worth (CPW) of Selected Plan and Combustion Turbine Substitution Portfolio ($M) CPW ($M) Selected Plan 19,758 CT Substitution 19,930 Delta 172 The Selected Plan has a CPW equal to $172M less than that of the Combustion Turbine Substitution portfolio. When CPW is factored into the risk analysis it becomes apparent that spending an additional $172M to hedge against a potential risk of $62M would not be prudent. 108

122 2015 Integrated Resource Plan 6.0 Conclusions and Five-Year Action Plan 6.1 Plan Summary The optimization results of this IRP demonstrate that PSO, as a stand-alone entity in the SPP RTO, can serve customer needs over the planning period with additional base-load combined-cycle generation, wind and solar renewables, and DSM resources, including EE measures such as CVR. The following are summary highlights of the Selected Plan: PSO s Selected Plan Adds 50MW (nameplate) of utility-scale solar resources per year in , for a total of 200MW (nameplate) of utility-scale solar by the end of the planning period. Adds 100MW (nameplate) wind energy in Implements customer and grid energy efficiency programs, including Conservation Voltage Reduction, reducing energy requirements by 517GWh and capacity requirements by 99MW by Fills long-term needs through the addition of 390MW of natural gas combined-cycle generation by repowering Northeastern Unit 4, or equivalent in 2022, and an additional 870MW of natural gas combined-cycle resources in the years Fills short-term need of 266MW in 2021 through a short-term PPA. Anticipates retirement of Southwestern Units 1 (68MW) and 2 (79MW) in 2021 and 2023, respectively. Anticipates retirement of Weleetka Units 4, 5, and 6 (150MW total) in Note: The modeling for this IRP was conducted prior to the issuance of the EPA CPP final rule. Specific PSO capacity position changes over the 10-year planning period associated with the Selected Plan are shown for 2015 and 2024, in Figure 29 and Figure 30, respectively. 109

123 2015 Integrated Resource Plan Figure PSO Nameplate Capacity Mix Figure PSO Nameplate Capacity Mix Specific PSO energy production changes over the 10-year planning period associated with the Selected Plan are shown for 2015 and 2024, in Figure 31 and Figure 32, respectively. 110

124 2015 Integrated Resource Plan Figure PSO Energy Mix Figure PSO Energy Mix Figure 29 through Figure 32 indicate that the Selected Plan would increase PSO s reliance on natural gas and renewable generation, add DSM resources, and reduce dependency on coal solidfuel based generation. Specifically, over the 10-year planning horizon the Company s capacity mix attributable to solid fuel-fired assets would decline from 18.7% to 8.5%, and natural gas assets increase from 57.8% to 62.8%. New solar assets will make up 3.1% of the capacity mix and 111

125 2015 Integrated Resource Plan wind assets increase to 18.8%. Demand-side resources (energy efficiency and CVR) are added to the mix at 2.9% of total nameplate capacity resources. PSO s energy output attributable to solid fuel generation decreases from 54.7% to 16.4% over the period of , while energy from natural gas resources increases from 13.2% to 46.9%. The Selected Plan introduces solar resources, attributing to 3.0% of total energy. Reliance on thermal PPA energy is decreases from 16.7% to 2.6%. Figure 33 shows PSO s annual SPP capacity position over the planning period, per the Selected Plan. Thermal PPA s replace the capacity and some of the energy from the retiring Northeastern Unit 4 in Some of these PPA s expire in 2021 and new natural gas combinedcycle generation is added, along with the repower of Northeastern Unit 4 into a natural gas combined-cycle. Figure 33. PSO Annual SPP Capacity Position (MW) Table 21 provides a summary of the Selected Plan which resulted from resource optimization modeling under the load and commodity pricing scenarios. 112

126 2015 Integrated Resource Plan Table 21. Summary of PSO Selected Plan Resource Additions from a Capacity (MW) Viewpoint 113

127 2015 Integrated Resource Plan PSO Five-Year Action Plan Steps to be taken by PSO in the near future as part of its Five-Year Action Plan include: 6.2 Conclusion 1. Continue the planning and regulatory actions necessary to implement economic energy efficiency programs in Oklahoma. 2. Conduct an RFP to explore opportunities to add wind generation in the near future to take advantage of the Federal Production Tax Credit. 3. Conduct an RFP to explore adding cost effective utility-scale solar resources. 4. Initiate the RFP process to evaluate PSO s options for replacing the existing PPAs when they expire in 2021 and Evaluate the greenhouse gas rules. Work with the Oklahoma Executive Branch, Oklahoma Department of Environmental Quality, and the Office of the Attorney General on Oklahoma s response to the EPA s greenhouse gas rule. 6. Be ready to adjust this Action Plan and future IRPs to reflect changing circumstances PSO s Selected Plan provides the Company with an increasingly diversified portfolio of supply- and demand-side resources which provides flexibility to adapt to future changes to the power market, technology, and environmental regulations. The addition of efficient natural gasfired generation along with increased renewables and demand-side management mitigates fuel price and environmental compliance risk. Inasmuch as there are many assumptions, each with its own degree of uncertainty, which had to be made in the course of resource portfolio evaluations, material changes in these assumptions could result in modifications. The action plan presented in this IRP is sufficiently flexible to accommodate possible changes in key parameters, including load growth, environmental compliance assumptions, fuel costs, and construction cost estimates, which may impact this IRP. By minimizing PSO s costs in the optimization process, the Company s model produced optimized portfolios with the lowest, reasonable impact on customers rates. 114

128 2015 Integrated Resource Plan Appendix Exhibit A: Demand Portfolio Filing Summary

129 The Demand Portfolio filing (Cause No. PUD ) which is pending approval with the Oklahoma Corporation Commission is summarized below. The plan addresses the requirements of OAC 165: This portfolio builds on the momentum of current Demand Portfolio (Cause No. PUD ). The filed Demand Portfolio can be viewed at The goals of PSO s Demand Portfolio are to minimize the long-term cost of utility service and avoid or delay the need for new generation, transmission and distribution investment consistent with OAC 165: (a) effective January 1, PSO s Demand Portfolio seeks to accomplish these goals by overcoming barriers that prevent residential and business customers from adopting energy efficient technologies. PSO also intends for one program to specifically leverage load management capability to reduce peak demand on the system, which should, all things being equal, decrease the amount of investment required to meet peak demand. PSO s Demand Portfolio of Programs Residential Home Weatherization High Performance Homes Energy Saving Products Education Conservation Voltage Reduction Behavioral Modification Commercial & Industrial High Performance Business Business Demand Response Conservation Voltage Reduction Behavioral Modification The table below summarizes the cost-effectiveness, budget and energy and demand savings of the proposed Demand Portfolio. PSO s Demand Portfolio Benefit Cost Tests, Budget & Savings Portfolio Year TRC Result UCT/PA CT Result Annual Program Costs ($ million) Energy (GWh) Savings Demand (MW) Savings Savings (% of Sales) (less opt-out) % % %

130 A brief description of each of the proposed Demand Portfolio s eight programs is provided. Home Weatherization PSO s Home Weatherization Program seeks to generate energy and demand savings for limited income residential customers through the installation of a wide range of cost effective weatherization and other measures in eligible dwellings. The purpose of the Home Weatherization Program is to provide PSO s limited income and hard-to-reach residential customers the financial assistance they need to make their homes more energy efficient, increase comfort levels, and reduce their utility bills. The objective of the Home Weatherization program is to focus on residential customers or communities with customers with total annual household incomes at or below 200% of federal poverty guidelines and households with annual incomes less than $45,000. Goals of the program are to: Achieve customer energy and cost savings. Educate customers on the benefits of continued or expanded energy efficiency and conservation efforts including, but not limited to, a change in customer usage habits through energy education. Supplement the resources of existing state, federal, nonprofit, and tribal housing authority weatherization programs to allow more eligible dwellings to be treated on an annual basis. This may include the replacement of non-energy specific elements that without replacement prevent eligible homes from receiving state, federal, nonprofit, and tribal housing funds, under the condition that specific energy efficiency and conservation measures, as deemed cost effective by PSO, are also installed at the residence. High Performance Homes PSO s proposed High Performance Homes Program seeks to generate energy and demand savings for residential customers through the promotion of comprehensive efficiency upgrades to building envelope measures and HVAC equipment for both ENERGY STAR new homes and retrofits. The purpose of the High Performance Homes Program is to provide PSO residential customers with inducements for increasing building envelope efficiencies and installing items such as high efficiency appliances and HVAC equipment. The High Performance Homes Program has four components: 1) In-house audits followed by direct inducements for envelope and equipment upgrades. PSO intends to continue providing the same program model currently offering through the Multiple Upgrades approach. 2) High Performance Homes will continue providing the Single Upgrade approach. The Single Upgrade component provides direct inducements to the customer through the service provider for qualifying upgrades. The Single Upgrade was

131 previously offered in the Energy Saving Products and Services program. Consolidated Single Upgrades with Multiple Upgrades under High Performance Homes will align like-measures and simplify overall portfolio administration for the program years. 3) The High Performance Homes will also offer an upstream, distributor-level inducement for efficient air conditioners and heat pumps. This is a new program component. 4) High Performance Homes will continue to offer an efficient New Homes program where builders are paid inducements for building to efficient standards. Energy Saving Products PSO s Energy Saving Products seeks to generate cost effective energy and demand savings for residential customers through the promotion of ENERGY STAR qualified products such as CFLs, and LEDs. The purpose of this service is to provide PSO residential customers inducements for purchasing energy saving products that meet high efficiency standards. This will, at a minimum be a continuance of the current lighting component of the existing Energy Saving Products and Services program offered by PSO. PSO s program budget includes funds for customer rebates and a food bank give-away component. A third party will be responsible for coordinating the purchase and delivery of the bulbs to local food banks. The Contractor will work with the local food bank to distribute CFLs to food panties within PSO s service territory. PSO proposes a mix of energy saving products for PSO s service territory. PSO may adjust rebate amounts when programs are implemented to ensure cost effectiveness, and to incentivize customer participation. Education PSO s proposed Education Program seeks to generate energy and demand savings while increasing awareness for PSO s 2016 to 2018 Demand Program offerings. In addition to educating customers about energy efficiency, the program will also be an effective marketing tool, and will direct customers to other energy conservation and demand reduction programs offered by PSO. PSO s program budget includes funds for direct delivery of energy education and conservation kits. PSO may choose to use a third party implementer to design educational materials, assemble and deliver kits, provide customer outreach, and other program processes. The education program has proposed three components as described below.

132 Schools Education Program - A lesson plan is provided to classroom teachers, which engages fifth grade students in learning about energy efficiency while also practicing mathematics and science. The students are then provided a take-home energy efficiency kit. Energy savings are achieved when these measures are installed in homes. Although the lessons encourage energy efficient behavior, the only impacts PSO will report will come from actual kit components that are installed at participants homes. Residential Conservation Kits - The residential conservation kits will provide customers with energy efficiency educational materials that help customers to make energy efficient choices. The materials range from tips on saving energy at home to fundamental concepts needed to identify and assess energy savings opportunities. Additional materials will highlight the opportunities for participating in PSO s energy efficiency and demand response programs with a special focus on PSO s new Power Hours program. The conservation kits will also include easy installation energy efficiency measures such as CFLs, LEDs, LED night lights, and filter tone alarms that encourage timely replacement of air conditioner air filters. Nonresidential Conservation Kits - The nonresidential conservation kits will provide commercial and industrial customers with energy efficiency educational materials that help customers to make energy efficient choices. The materials will discuss fundamental concepts that are needed to identify and assess energy savings opportunities. Additional materials will highlight the opportunities for participating in PSO s energy efficiency and demand response programs. In particular, the materials will promote PSO s High Performance Business program, with a special focus on PSO s offering of free or discounted facility audits, and PSO s small commercial direct install program component. The conservation kits will also include CFLs and LEDs. The residential and nonresidential conservation kits will be marketed to customers through numerous channels including mailers, bill stuffers, printed and online advertisements, and prompts on PSO s web site. The commercial kits may also be offered directly to customers by PSO representatives. Behavioral Program The behavioral program will be a holistic, stacked program with a conscious eye toward engaging multiple facets of decision making and behavior, most importantly, emotions, reason and social interaction in order to induce cost effective energy saving behavior in customers. Program roll-out will be January 2017, however work may commence prior to that date. This program will work closely with PSO s marketing firm and with PSO s existing Web Portal to engage and motivate customers to alter their energy use and to participate in other PSO offerings, lower their energy bills. The program will also seek to increase customer satisfaction by being engaging and interactive, fast and convenient,

133 personalized and customer friendly. Provide strategy and pricing for enrollment, retention of participants and associated metrics for service components including: Working with existing Web Portal - Deliver personalized energy usage information and offer simple energy-saving tips customized to each household. Electronic channels - Use campaigns, push notifications, electronic reports, social media, smart phone applications, and other tools to engage and encourage behavioral changes for energy efficiency. Going Green - Eliminate all paper reports or direct mail that is wasteful to the environment. Segmentation - Deliver the right messaging to different groups of customers to improve their customer experience. Rewards/Challenge option - Recognize customers contributions and successes toward reducing energy usage, lowering their bills and participating in other PSO services. High Performance Business PSO s commercial and industrial (C&I) program is a continuation of the existing program with proposed enhancements. This program seeks to achieve cost-effective reduction of peak demand and energy use by assisting C&I customers in generating energy savings through promotion of high efficiency electric end use products. The program provides a combine provision of financial inducements with access to technical expertise to maximize program penetration across the range of C&I customers. The program has additional goals: Increase customer awareness of applicable energy saving measures Achieve customer cost savings Increase the market share of commercial grade high efficiency technologies sold through market channels, Increase the installation rate of high efficiency technologies in C&I facilities by businesses that would not have done so absent the program The existing program allows for customers to self-sponsor or work with a trade alley. The proposed program has three programs: 1. prescriptive and custom, 2. direct install and behavioral, and 3. engineering support, ASHRAE audits, benchmarking, and master planning. The program will induce customers to make cost effective energy efficiency upgrades through offering direct install, rebates, customer education, and other promotional activities.

134 Key design updates of the program include: Include a direct install component for facilities Pay down the cost of facility audits for a limited number of facilities Pay a share of LEED certification expenses on qualifying new construction and existing building projects. Business Demand Response The Business Demand Response (BDR) program is the continuation of an existing program. The program is targeted to commercial and industrial customers served by PSO capable of reducing demand on short notice. After receiving a curtailment message from PSO, facility operators shed electric load in many different ways such as shutting down motors, pumps, compressors, air conditioning equipment, and lighting. Some customers are even able to maintain normal production levels while temporarily reducing electric demand. The primary goal of this program is to achieve cost-effective reduction in demand during critical summer peak capacity periods, and to maintain and improve overall system reliability. Additional goals of the program may be to achieve customer energy and cost savings and educate customers on the benefits of continued or expanded energy efficiency and conservation efforts. The target market includes all commercial and industrial customers. Monetary inducements will be paid to customers willing to allow PSO to install devices capable of interrupting specific electrical end use load(s) within a customer facility. PSO anticipates additional active participation from both the commercial and industrial sectors. BDR is a vital peak demand reduction program. PSO has gained valuable experience with this program. Educating customers about the ways to save and reduce peak usage has been valuable to both PSO and customers. Like relationships with trade allies in other programs, relationships with BDR customers are sensitive, must be cultivated over time and very valuable. The need for peak demand reductions isn t consistent from yearto-year, but it is important for capacity planning and markets to maintain an active BDR program due to the ramp-up time in recruiting and educating customers. Capacity takes years to bring online, likewise, building a reliable BDR program also takes time. Conservation Voltage Reduction Conservation Voltage Reduction (CVR), sometimes referred to as Volt VAR Optimization (VVO), refers to the installation and maintenance of technology that optimizes and lowers delivered voltages creating energy savings for customers. The CVR program is a newly proposed energy saving program for PSO. However, PSO has proven experience with CVR in the Owasso pilot.

135 CVR uses a centralized intelligent control system to continuously monitor and automatically control substation and line voltage regulating devices to maintain the overall distribution line voltage within a narrower bandwidth than possible without CVR as shown in the Voltage Profile Examples figure below. The line devices primarily consist of multiple capacitor banks, voltage regulators and line voltage monitors that are strategically located along the entire length of the distribution line. Historically, these line devices operated independently based on predetermined and manually set control settings at specific locations. For example, voltages at a substation were set at the high end of the range to assure voltages would be above the minimum requirement at the end of the line. This operating scheme resulted in wider voltage variations along the distribution line within the upper and lower voltages typically incurred. CVR creates demand and energy savings and allows for more consistent voltage across the circuit while still remaining in compliance with minimum voltage levels. Voltage Profile Examples CVR is an energy efficiency measure and would not be installed on the distribution system for any other reason. As stated above, CVR technology controls voltage levels and lower voltages will result in the customer end-use equipment consuming less energy and yet providing the same services. The customer realizes savings from the reduced energy consumption as a result of the lower voltages. Customers on the circuits where CVR technology is installed will benefit from a reduction in kwh billed and all PSO customers will benefit from reduced peak demand. Initially, and specific to the portfolio, PSO will install CVR technology on 24 circuits in 2017 to be fully operational in Groups of circuits were chosen across PSO s south Tulsa territory based upon residential and small commercial loading which are more conducive to benefiting from CVR technology. Groups of circuits with higher usage and certain operational characteristics were chosen. PSO will operate the CVR technology all hours of the year with exception to operation needs and evaluation, measurement and verification testing needs.

136 Demand Portfolio Participants, kwh & kw Savings and Program Costs Public Service Company of Oklahoma Proposed Demand Portfolio Plan for Demand Portfolio Proposed Programs Customer Class Participants kwh Energy Savings kw Demand Savings Program Cost Home Weatherization Residential 1,900 3,344,823 1,129 $3,476,701 High Performance Homes Residential 6,585 6,159,578 3,365 $8,545,840 Education Residential 26,000 4,216, $2,127,000 Energy Saving Products Residential 208,256 23,065,590 2,670 $3,754,950 Business Demand Response Commercial & Industrial ,968 48,994 $3,116,200 High Performance Business Commercial & Industrial 2,987 43,481,363 7,366 $10,112,152 Behavioral Residential & Small Business $162,000 Conservation Voltage Reduction All $0 Portfolio Total 246,234 80,513,226 64,120 $31,294, Demand Portfolio Proposed Programs Customer Class Participants kwh Energy Savings kw Demand Savings Program Cost Home Weatherization Residential 1,940 3,450,442 1,156 $3,524,018 High Performance Homes Residential 6,560 6,105,304 3,395 $8,271,210 Education Residential 26,000 4,216, $2,113,500 Energy Saving Products Residential 202,587 22,505,674 2,605 $3,687,883 Business Demand Response Commercial & Industrial ,594 54,319 $3,296,900 High Performance Business Commercial & Industrial 3,007 44,251,630 7,439 $9,949,878 Behavioral Residential & Small Business 101,000 14,911,066 4,779 $1,223,500 Conservation Voltage Reduction All $29,600 Portfolio Total 341,654 95,712,615 74,289 $32,096, Demand Portfolio Proposed Programs Customer Class Participants kwh Energy Savings kw Demand Savings Program Cost Home Weatherization Residential 1,980 3,453,039 1,169 $3,612,347 High Performance Homes Residential 6,110 5,356,964 2,867 $7,252,780 Education Residential 26,000 4,216, $2,113,500 Energy Saving Products Residential 195,720 21,903,581 2,535 $3,880,597 Business Demand Response Commercial & Industrial ,221 59,644 $3,546,900 High Performance Business Commercial & Industrial 2,970 41,835,365 7,354 $8,988,956 Behavioral Residential & Small Business 152,000 22,899,137 7,280 $1,803,500 Conservation Voltage Reduction All 30,910 23,878,982 6,265 $1,722,646 Portfolio Total 416, ,842,195 87,711 $32,921,225 3-year Portfolio Total 300,068, ,807 $96,312,558 Notes: Participants are approximate expected customers. Energy & Demand values are net at the generator (with losses). In the 3-year Portfolio Total, Business Demand Response kw demand savings are not cumulative.

137 Exhibit B: 2015 Fuel Supply Portfolio and Risk Management Plan 2015 Integrated Resource Plan

138 RECEIVED Public Utility Division 5/14/2015 Public Service Company of Oklahoma 2015 Fuel Supply Portfolio and Risk Management Plan May 15, 2015

139 RECEIVED Public Utility Division 5/14/2015 Table of Contents I. INTRODUCTION... 1 A. FUEL PLANNING OBJECTIVES... 3 B. RESOURCES & CAPABILITIES Generation Purchased Power Wind Energy...7 C. PRIOR PERIOD RESULTS... 7 D. GAS-ELECTRIC HARMONIZATION 14 II. ENERGY AND SUPPLY FORECAST A. DEMAND FORECAST B. FUEL SUPPLY AND ENERGY FORECAST Fuel Energy...18 III. PROCUREMENT STRATEGY A. BACKGROUND B. PROCUREMENT PLAN IV. RISK MANAGEMENT A. HEDGING B. RESOURCE DIVERSITY AND OPTIMIZATION C. CONTRACT PROVISIONS D. CUSTOMER PROGRAMS E. COST OF ENERGY V. SUMMARY... 30

140 RECEIVED Public Utility Division 5/14/2015 I. Introduction Organized in Oklahoma in 1913, Public Service Company of Oklahoma ( PSO or the Company ) is a vertically integrated utility engaged in the generation, transmission, and distribution of electric power to approximately 542,000 retail customers in eastern and southwestern Oklahoma, and in supplying and marketing electric power at wholesale to other electricity utility companies, municipalities, rural electric cooperatives and other market participants. As of December 31, 2014, PSO had 1,133 employees. Under Order No , Cause No. PUD , PSO provides this Fuel Supply Portfolio and Risk Management Plan (Plan) on an annual basis. This document serves to support PSO s plans to provide reliable, adequate, and flexible fuel supplies for its customers at the lowest reasonable cost. It also explains the strategy for the purchase of energy and capacity to meet multiple reliability, economic, and environmental objectives. As an integrated public utility, PSO holds franchises and/or other rights to provide electric service in various municipalities and regions in its service territory. PSO owns 4,436 MW of generating capacity, which it uses to serve its retail and other customers. Exhibit 1 illustrates the approximate boundaries of PSO s service territory (in green) and the location of its generation resources. The red circles represent PSO s coal units and the purple circles indicate the location of natural gas generation units. PSO also engages in the purchase of renewable energy, such as wind, and capacity to further diversify its portfolio of resources. Purchased Power Agreements ( PPAs ) for capacity and firm energy that are entered into by PSO also utilize Oklahoma resources. For example, PSO currently purchases power through a long-term PPA from Green Country, located in Jenks, Oklahoma. 1

141 RECEIVED Public Utility Division 5/14/2015 Exhibit 1: Map of PSO in Oklahoma PSO is a member of the Southwest Power Pool ( SPP ), a Regional Transmission Organization ( RTO ) that is mandated by the Federal Energy Regulatory Commission ( FERC ) to provide reliable supplies of power, adequate transmission infrastructure, and competitive wholesale prices of electricity. SPP s wholesale power market, known as the Integrated Marketplace ( IM ) opened on March 1, 2014, and consists of Day Ahead, Real-time, and Ancillary Service markets. SPP has provided market participants with financial instruments, Auction Revenue Rights ( ARR s) and Transmission Congestion Rights ( TCR s), to hedge against transmission congestion that may occur between generating resources and load. The IM has resulted in significant changes for unit commitment, fuel procurement, unit dispatch, operating reserve procurement, transmission congestion management, balancing authority operations and power settlements. PSO s participation in, and impacts from, the SPP IM will be discussed throughout this plan. As mentioned above, the new SPP IM has significantly changed how PSO and its sister operating company in the SPP, Southwestern Electric Power Company ( SWEPCO ) generating units are committed and dispatched. Accordingly, PSO and SWEPCO proposed changes to the Operating Agreement which were necessary to align it with SPP s operations under the new SPP IM. The proposed changes to the Operating Agreement were accepted by the FERC on January 8, 2014, with an effective date, concurrent with the SPP IM launch, but conditioned upon 2

142 RECEIVED Public Utility Division 5/14/2015 removing text from a section of the agreement. The requested change was made and submitted on January 24, The FERC accepted the compliance filing in a March 25, 2014 Order. Additionally, the System Integration Agreement ( SIA ) has provided for integration and coordination of the power supply resources of the American Electric Power ( AEP ) East and AEP West Operating Companies. The SIA functions as an agreement in addition to, not in substitution of, the Operating Agreement. The SIA has provided for the distribution of power supply costs and benefits between two zones, the East Zone and the West Zone, corresponding to the AEP East and AEP West Operating Companies, while the Operating Agreement continues to govern the distribution of costs and benefits among SWEPCO and PSO. However, on May 1, 2014, the transmission reservation between the AEP East and AEP West Operating companies expired and as mentioned previously, the SPP IM became effective on March 1, Since both the East and West operating companies now operate in RTO wholesale energy markets, and without sufficient economic benefits to offset the cost of transmission services, the decision was made to eliminate those provisions of the SIA that were no longer necessary. Accordingly revisions to the SIA eliminating the power integration and coordination, with an effective date of June 1, 2014, were accepted by FERC on June 3, 2014 in FERC docket ER PSO provides retail service in Oklahoma at bundled rates approved by the Oklahoma Corporation Commission ( OCC ). PSO s rates are set on a cost-of-service basis. Fuel and purchased energy costs are recovered through the Fuel Cost Adjustment Factor. The factor is generally adjusted annually and is based upon forecasted fuel and purchased energy costs. Over and under-collections of fuel and purchased energy costs for prior periods are returned to or removed from customers in the year following when the new annual factors are established. A. Fuel Planning Objectives PSO s Plan is designed to ensure that sufficient quantities and types of fuel and power are available to safely and reliably meet customer needs under dynamic conditions, while striving to provide the over-all lowest reasonable delivered cost. In other words, PSO s fuel and purchased power procurement is first and foremost focused on ensuring that all the electric power that its customers want is available when they want it, and at the lowest reasonable delivered cost. 3

143 RECEIVED Public Utility Division 5/14/2015 B. Resources & Capabilities 1. Generation PSO s generating fleet is composed of both coal power plants and natural gas power plants, as summarized in Table 1. Plant Name Table 1: Plant Capacity Fuel Type Net Maximum Capacity (MW) Comanche Natural Gas 266 Riverside Natural Gas 1060 Southwestern Natural Gas 632 Tulsa Natural Gas 318 Weleetka Natural Gas 198 Northeastern, Units 1 and 2 Natural Gas 923 Northeastern, Units 3 and 4* Coal 937 Oklaunion** Coal 102 Total 4,436 * Northeastern, Unit 4 is scheduled to retire by April 16, ** Capacity at Oklaunion represents the PSO share. In addition, the steam generating units at Riverside can also use fuel oil to generate electricity. PSO maintains a limited quantity of fuel oil at the Riverside units as an emergency back-up fuel supply. The Riverside Plant is also connected to a pipeline capable of delivering fuel oil. PSO can also use natural gas to run Northeastern 3 and 4 at partial load in the event of coal curtailments or coal-related equipment outages. Comanche, Weleetka, Northeastern Unit 1, Riverside Units 3 and 4, and Southwestern Units 4 and 5 are each connected to one pipeline system. Northeastern Units 2-4, Riverside Units 1 and 2, Southwestern Units 1-3, and Tulsa Units 2 and 4 are each connected to two pipeline systems. These multiple natural gas pipeline connections provide the Company with access to reliable, flexible, and competitively priced natural gas supplies. The natural gas pipeline interconnections to each of PSO s natural gas plants are shown in Exhibit 2 on the following page. 4

144 RECEIVED Public Utility Division 5/14/2015 Exhibit 2: Existing Natural Gas Pipeline Interconnections to PSO PUBLIC SERVICE COMPANY OF OKLAHOMA NATURAL GAS FIRED POWER STATIONS PIPELINE SYSTEMS PSO RIVERSIDE PSO COMANCHE PSO NORTHEASTERN* PSO SOUTHWESTERN PSO TULSA PSO WELEETKA Enable OK OGT Enable OK and OGT provide low pressure service to the Northeastern Plant Site which serves the generation needs for Unit 2, duct burner gas to Unit 1, and startup or emergency generation replacement fuel for Coal Units 3 and 4. Similarly, the Northeastern 3 and 4 coal units have access to two competing rail carriers, Union Pacific ( UP ) and Burlington Northern Santa Fe ( BNSF ), for coal deliveries. Currently, UP provides coal deliveries to the Northeastern power plant with a shipping distance of approximately 1,000 miles. The BNSF railroad provides deliveries of coal to the Oklaunion power plant with a shipping distance of approximately 1,100 miles. The location of PSO s coal generating plants relative to the coal supply region and rail transportation routes is provided in Exhibit 3. As part of Oklahoma Regional Haze State Implementation Plan and for PSO to be in compliance with the Mercury and Air Toxics Standard, Northeastern Unit 4 is scheduled to retire by April 16, Purchased power agreements have been secured to replace the lost capacity. Northeastern Unit 3 is being retrofitted with Dry Sorbent Injection ( DSI ), Activated Carbon Injection ( ACI ), and a Fabric Filter Baghouse and will continue to operate. 5

145 RECEIVED Public Utility Division 5/14/2015 Exhibit 3: Location of PSO Coal Plants relative to the PRB* coal supply region *Powder River Basin 2. Purchased Power PSO s purchased power plan has been significantly impacted by the start of SPP s IM and the changes to the Operating Agreement that were made in recognition of the changing operating environment. PSO no longer engages in the purchase and sale of economic energy with SWEPCO. However, the start of the SPP IM has provided additional opportunities for AEPSC, on behalf of PSO, to engage in purchase power opportunities for the benefit of PSO s customers. Additionally, AEPSC s activities, on behalf of PSO, have enabled it to develop strong relationships within the SPP and across other regions. As a result, AEPSC continues to directly engage in trading with a variety of third-party market participants that buy and sell OTC electricity in the SPP. AEPSC s Commercial Operations employees leverage a broad cross section of operations and market knowledge to optimize the PSO system. The realization of the customer benefits from the SPP IM, including the optimization of Off System Sales (OSS) margins, will depend in large part on the continuing activities of PSO, through AEPSC s Commercial Operations organization, to optimize the use of available resources, working in the new IM framework established by SPP. 6

146 RECEIVED Public Utility Division 5/14/ Wind Energy PSO has made a significant commitment to wind generation through long-term purchased power agreements. PSO s wind contracts, like PSO s longer-term power purchases in general, were procured through competitive Request for Proposal ( RFP ) solicitations. Wind energy provides PSO s customers with price certainty well into the future and a hedge against many future environmental compliance requirements related to fossil-fired generation. C. Prior Period Results Summary PSO s generating plants combined with purchased power, offer a diverse fleet to the SPP IM. Table 2 below offers a comparison of the total generation resource mix in 2013 and Table 2: Resource Percentage Comparison Generation Resource (MWh Basis) Natural Gas Coal Fuel Oil <0.01 <0.01 Purchased Power Wind Energy In 2014, the Company had a reasonable expectation that the fuel mix experienced in 2013 would change to include a lesser share of natural gas in As expected, the most significant year-on-year change illustrated in Table 2 was the decrease in natural gas, which was offset by an increase in generation from purchased power. Wind and coal generation saw a small increase in 2014 compared to The impact to PSO s fuel costs as shown in Table 3 are the result of a decrease in natural gas generation and a corresponding increase in purchased power as referenced in Table 2, as well as the implementation of the plan detailed in the 2014 Plan. 7

147 RECEIVED Public Utility Division 5/14/2015 Table 3: Fossil Fuel Consumed During 2014 Fossil Fuel Type MMBtu % MMBtu Fuel Expense $ $/MMBtu Natural Gas 31,982, % $163,514, Coal 72,750, % $122,729, Fuel Oil* 35, % $585, Total 104,768, % $286,829, *Total fuel oil consumed for generation includes a monthly allocation of the annual pipeline lease payment for the fuel oil pipeline connected to Riverside. Exhibit 4 illustrates, on a monthly basis, how fuel costs at PSO generation resources and purchased power contributed to overall costs on a $/MWh basis. Exhibit 4: 2014 Average Monthly Cost of Selected Resources in $/MWh $/MWh Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Coal Generation Natural Gas Generation Purchased Power Wind Purchase Average 8

148 RECEIVED Public Utility Division 5/14/2015 PSO s fossil fuel mix (based on MMBtus consumed), as reflected in Table 3, continues to provide benefits to customers in the SPP IM. During 2014, PSO s average monthly delivered cost of natural gas varied between a low of $3.97 per MMBtu in October 2014 to a high of $8.10 per MMBtu in February As a result of PSO s fuel mix, PSO s average delivered cost of fossil fuel varied within a lower and tighter range, between a low of $2.32 per MMBtu in December 2014 to a high of $4.00 per MMBtu in February Coal Procurement Summary As a low-cost base load power source, coal units are generally dispatched when available before other sources of generation. To illustrate the significance of base load coal generation to PSO, the coal units at Northeastern and Oklaunion represent approximately 23.4 percent of PSO s generating capacity, but supplied 69.4 percent of PSO s generation in Northeastern Units 3 & 4 purchased approximately 3.97 million tons of coal in 2014 while Oklaunion purchased roughly 2.11 million tons (total plant basis) during the same period. Exhibit 5 summarizes the contracts used by PSO to acquire coal in Beginning in 2013, PSO began purchasing ultra-low sulfur PRB coal to replace the coal currently in stockpile with a lower sulfur coal capable of meeting the SO2 requirements in PSO s Environmental Compliance Plan. This compliance plan required a 0.65 lb. SO 2 /MMBtu emission rate at Northeastern Units 3 and 4 beginning on January 1, 2014 and then an emission rate of 0.60 lb. SO 2 /MMBtu beginning on December 31, Turning over and replacing the stockpile with ultra-low sulfur PRB coal throughout 2014 was necessary to comply with new SO2 emission limits. Shipments of coal from the Powder River Basin (PRB) to the Northeastern and Oklaunion plants during 2014 were made pursuant to transportation arrangements with UP and BNSF, respectively. 9

149 RECEIVED Public Utility Division 5/14/2015 Exhibit 5: List of Coal Contracts in Effect in 2014 Northeastern Generation Station Vendor Agreement Number Tons Purchased Alpha M1 725,723 Peabody COALSALES, LLC M1 122 Peabody COALSALES, LLC M2 2,000,000 Peabody COALSALES, LLC M3 1,003,374 Peabody COALSALES, LLC M2 241,307 Oklaunion Generation Station (Total Plant Basis) Vendor Agreement Number Tons Purchased Alpha Coal Sales Co., LLC M1 591,065 Alpha Coal Sales Co., LLC M2 131,484 Indiana Michigan Power Company ,439 NRG Power Marketing, LLC ,494 Peabody COALSALES, LLC M3 14,607 Peabody COALSALES, LLC M3 218,011 Peabody COALSALES, LLC M2 43,481 TraxEnergy, Inc TraxEnergy, Inc Natural Gas Procurement Summary PSO s natural gas generating units were brought on-line and taken off-line on relatively short notice and the actual unit loading and resulting natural gas demand was highly variable. Exhibit 6 below illustrates how PSO utilized multiple transaction types for purchasing gas to meet the varying demands for natural gas generation. The chart indicates a range of total purchases per day from a low of 5,238 MMBtus in June 2014 to a high of 277,014 MMBtus in July With the launch of the SPP IM, PSO s Long Term/Annual purchase volumes diminished to provide flexibility for purchases in the marketplace, as seen in Exhibit 6 below. 10

150 RECEIVED Public Utility Division 5/14/2015 Exhibit 6: PSO 2014 Natural Gas Purchases by Transaction Type 300, , , , ,000 50,000 0 January February March April May MMBtu June July August September October November December Month Long Term/Annual Short Term/Monthly Spot To take advantage of favorable conditions and further mitigate risk, PSO also used different pricing mechanisms in Approximately 88 percent of purchases for the year were at an index price. Conversely, as shown below, about 12 percent were on a fixed price basis. PSO used monthly index prices to effectively lock in prices on a monthly basis and provide greater price stability compared to daily indexes. The figure below illustrates the correlation between price and percentage of each pricing mechanism used for all natural gas purchases. Pricing Mechanism % of NG Purchases Wtd. Avg. Price Fixed 11.65% $5.47 Daily Index 74.18% $4.40 Monthly Index 14.16% $4.56 To transport natural gas supplies to PSO gas plants as necessary, transportation contracts with Enable Oklahoma Intrastate Transmission, LLC (Enable OK, formerly Enogex) and ONEOK Gas Transportation, LLC (ONEOK or OGT) were used. PSO uses a mix of firm 11