INVESTMENT OPPORTUNITIES AND TECHNOLOGY SELECTION: IPP VALUE PROPOSITION FOR ERCOT A WHITE PAPER BY WÄRTSILÄ

Transcription

1 INVESTMENT OPPORTUNITIES AND TECHNOLOGY SELECTION: IPP VALUE PROPOSITION FOR ERCOT A WHITE PAPER BY WÄRTSILÄ

2 INVESTMENT OPPORTUNITIES AND TECHNOLOGY SELECTION: IPP VALUE PROPOSITION FOR ERCOT INDEX Executive summary 1 Introduction 4 2 Business logic behind hedging 5 3 How does the heat rate call option work? 7 4 Evaluating power plant technology options for hedging Technical performance of Wärtsilä 18V50SG and GE 7FA Calculating the heat rate cell levels for Wärtsilä 18V50SG and GE 7FA Derating impacts maximum contracted capacity Calculating minimum requirement for option fee 13 5 Value of a heat rate call option in ERCOT Back-cast value for a heat rate call option in ERCOT 14 6 Financial feasibility of Wärtsilä 18V50SG and GE 7FA Day-Ahead Market participation only Day-Ahead + Real-Time Market optimization Day-Ahead + Real-Time + Ancillary Services Market optimization Conclusions of the back-cast analysis 22 7 Future sensitivity analysis Price curves for the year Fair value of a heat rate call option with the year 2015 data Financial analysis with the year Conclusions 26 APPEDIX A: Project Finance Calculation Tables for the Modeled Cases 27 1

3 EXECUTIVE SUMMARY: The Electric Reliability Council of Texas (ERCOT), which operates the electric grid for most of the state of Texas, has experienced steadily growing demand over the last several years. In a state where 1,000 new residents are added each day, ERCOT has set new demand records, particularly during Along with rising demand, recent generating unit retirements and the cancellation or postponement of several capacity projects has presented new challenges to maintaining adequate planning reserve margins. Studies indicated that ERCOT could see reserve margins drop below established reliability targets in the coming years. Prices in ERCOT s energy-only market had been unable to attract sufficient investment to ensure adequate capacity reserves, so the system-wide scarcity pricing cap was increased, reaching $9,000 per MWh in Adjustments to scarcity pricing have spurred investment activity in ERCOT, with 8000 MW of capacity primarily industrial gas turbines planned for development. Many of these projects are based on the GE 7FA.05 gas turbine. The IPP investment business case is to provide Day-Ahead Market heat rate call option backed by the 7FA.05 to a Load-Serving Entity. Back-cast analysis using historical price and temperature conditions during has shown that the fair market value of a heat rate call option based on a Wärtsilä 18V50SG internal combustion engine (ICE) power plant is higher than the 7FA.05 power plant. This is because the Wärtsilä 18V50SG has higher efficiency and less derating at high temperatures than the 7FA.05. However, the average value of both technology options is lower than the required option fee to cover the debt service and operations and maintenance costs, and does not justify investment purely based on hedging. To examine project feasibility and internal rate of return (IRR), Wärtsilä has conducted detailed dispatch and financial modeling for a GE 7FA.05 and a Wärtsilä 18V50SG power plant in the ERCOT Day-Ahead (DA), Real- Time (RT), and Ancillary Services (AS) Markets for The Wärtsilä 18V50SG power plant financially outperforms the GE 7FA.05 due to the flexibility of Wärtsilä ICE technology. The ability of the Wärtsilä 18V50SG to start up within five (5) minutes, reduce load to 20% of full capacity, and incur no maintenance penalties from frequent starts and stops allows the Wärtsilä plant to participate in ERCOT s Ancillary Services Market, offering a variety of ancillary service products as shown in Figure E.1. Figure E.1. The Wärtsilä 18V50SG generates more revenue than the GE 7FA.05 due to greater dispatchability in markets 2

4 Because of the longer startup time of the GE 7FA.05 and increased maintenance costs associated with frequent starts, the 7FA.05 project could only offer limited ancillary services. As a result, the Wärtsilä 18V50SG is able to achieve 19.3% equity IRR and a 10.3% project IRR offer limited amount of ancillary services or participate in the Real-time market while the GE 7FA.05 is only able to achieve a 9.2% equity IRR and a 6.4% project IRR (back-cast analysis for ). The IRR performance of the 7FA.05 and the Wärtsilä 18V50SG in different market participation cases are shown in Figures E.2 and E.3. Extending this analysis to the hypotetical year 2015, the Wärtsilä 1850VSG power plant achieves a 19.9% return on equity while the 7FA.05 is only able to reach 9.6% return on equity. As market conditions predict there will be higher price spikes in ERCOT in coming years, and increasing demand for hedging products which will drive up the hedge prices. The more valuable heat rate call option, ability to earn revenue for several markets, and higher IRR makes the Wärtsilä 18V50SG power plant a superior investment for IPPs compared to 7FA ,0% 10,0% 8,0% 6,0% 4,0% 2,0% 0,0% 10,3% 7,3% 6,4% 5,2% 5,9% 5,2% DA DA + RT DA + RT +AS Project IRR - 7FA.05 Project IRR - Wärtsilä 18V50SG Figure E.2. Project IRR in the different market participation cases 25,0% 20,0% 15,0% 10,0% 5,0% 0,0% 19,3% 11,6% 9,2% 5,9% 7,7% 5,9% DA DA + RT DA + RT +AS Equity IRR - 7FA.05 Equity IRR - Wärtsilä 18V50SG Figure E.3. Equity IRR in the different market participation cases 3

5 1. INTRODUCTION During 2011 the Electricity Reliability Council of Texas (ERCOT) experienced several significant weather events coupled with extensive unexpected generator outages that prompted emergency procedures. ERCOT has been carefully monitoring its system resource adequacy, recognizing a threat of declining reserve margin in upcoming years. These concerns have spurred consideration of a capacity mechanism (project 40000) in ERCOT to attract new investment. However, in 2014 the Public Utility Commission of Texas decided to continue with the energy-only market design. In energy-only markets, the only sources of revenues for generators are energy and ancillary services, while capacity is not rewarded, as in some other ISO markets. To attract new investments, there must be sufficient scarcity pricing events in this type of market design. ERCOT has worked on the market enhancements during the past few years, with the target to provide adequate scarcity price signals together with enhancements in Real-Time Market design (Operating Reserve Demand Curve, ORDC B+). A real-time price adder was increased from $3000 to $9000 per megawatt-hour (MWh) to reflect the value of spinning and non-spinning reserves based on the value of lost load (VOLL) and the loss of load probability. These enhancements, together with declining market reserve margin, have boosted investment activity in ERCOT. Currently, there is more than 8000 MW of new peaking gas capacity in the development pipeline, but only a couple hundred megawatts of capacity is actually under construction. A great majority of this 8000 MW planned capacity will be industrial gas turbines, specifically GE s 7FA.05 industrial gas turbine. There are several Independent Power Producers (IPPs) pursuing the 8000 MW capacity additions. Because IPPs are subject to the ups and downs of the electricity market, the investment business logic is to provide Day-Ahead Market heat rate call option to load serving entities. A load serving entity (LSE) is willing to purchase this type of insurance product, as without it they risk being in short position during a price spike. In return, the IPP requires a long-term heat rate call option to finance the project. Current market prices and forward curves do not indicate the need for new peaking capacity yet, and therefore IPPs have not managed to sell their heat rate call options to off-takers. Basically, all IPPs are selling the same product, which is the Day-Ahead heat rate call option. As the product is same, the off-taker is picking up the cheapest offer that it can find in the market. Consequently IPPs are developing low capital cost projects, and in this field the large GE 7FA.05 gas turbine seems to be the most economical solution. Wärtsilä has analyzed historical ERCOT market pricing and the performance of the GE 7FA.05 gas turbine plant compared with the Wärtsilä 18V50SG internal combustion engine (ICE) power plant under various market participation cases. This report will show the feasibility for both solutions, as well as more advanced business cases and value proposition for flexible peaking capacity. Based on this analysis, the cheapest capacity to construct might not be the winner for internal rate of return. By providing a more valuable product to offtakers, leading IPPs ensure financing before other IPPs and reach higher return on equity. 4

6 2. BUSINESS LOGIC BEHIND HEDGING The energy-only market design in Texas should create scarcity pricing events when the available reserve capacity reaches a low threshold. These pricing events are typically driven by hot weather or unexpected plant failures in the market. Generators, especially merchant plants, would like to see as many of these pricing events as possible, as they are able to cover their fixed costs and get return on investment during these events. These pricing events are particularly essential for peaking capacity that would not be able to cover fixed expenditures during normal operations, when the market price is set by the operating cost of the marginal unit needed to meet load. However, these scarcity pricing events cannot be predicted years ahead, which makes the financing of peaking plants challenging. Therefore, IPPs are looking for an off-taker willing to pay a fixed fee for the capacity. As these pricing events typically occur during the summer months, the system load is also close to its peak. This creates a risk for load serving entities (LSEs) that are obliged to serve their load in all conditions. If the load serving entity does not have adequate capacity or financial contracts to meet its load, it needs to buy the remaining energy from the market. The table below shows an illustrative example of the financial impact of a price spike from the LSE perspective for three cases: capacity shortage during the price spike, in balance, and a capacity surplus. The illustrative example in Table 2.2 shows the financial impact of multiple price spikes for a utility with a 200 MW capacity shortage. IMPACT OF MARKET PRICE Capacity shortage In balance Capacity surplus Market price [$/MWh] 9,000 9, 000 9,000 Customer load to be served [MW] Contracted capacity [MW] ,000 Capacity shortage/surplus [MW] (200) Financial impact ($1,800,000) - $1,800,000 Table 2.1. Illustrative example of a price spike from the LSE perspective IMPACT OF MARKET PRICE Market price [$/MWh] Capacity shortage Number of price spikes Financial impact Maximum price for hedge [$/kw/month] 1 Price spike 5 Price Spikes 10 Price Spikes 15 Price Spikes 9,000 9,000 9,000 9,000 (200) (200) (200) (200) (1,800,000) (9,000,000) (18,000,000) (27,000,000) Table 2.2. Illustrative example of financial impact of multiple price spikes The simplified example illustrates the financial impacts from severe price spikes, but it also provides a good basis for the hedge contract pricing. If the utility believes there will be only one scarcity price event for which it is 200 MW short, the maximum price it should pay for the hedge is $0.75/kW/month. With 15 price spikes, the maximum price for the hedge fee rises to $11.25/kW/month. The IPP can calculate the contract value to cover its debt service, fixed operations and maintenance expenses, and evaluates if the LSE is willing to pay 5

7 the required hedge fee. For example, if the IPP needs $5.0/kW/month for seven years to finance an investment, and the LSE believes that there will be 15 price spikes, there is a good chance that the IPP and the LSE will agree to a contract. The illustrative example clearly shows the severe financial impact of a price spike if the LSE is in short position during the price spike. Just one hour could cost $1,800,000 and will have negative impact on profitability. With the increase in the ERCOT scarcity price cap from $3,000/MWh to $9,000/MWh, the impact and probability of these severe price spikes have increased. Nobody knows exactly how many price spikes there will be next year or in future years, so estimating the impact of price spikes is difficult. The probability of price spikes will increase when the capacity margin decreases. IPPs would sell their project based on this uncertainty and the potential financial impact of scarcity pricing events. Ultimately, different LSEs have varied assumptions and sensitivities about the frequency and economic consequences of price spikes. 6

8 3. HOW DOES THE HEAT RATE CALL OPTION WORK? The most typical hedging product that IPPs are selling is the so called Day-Ahead heat rate call option. The call option means that the off-taker (the LSE in this case) has the right to use the contracted capacity, but it is not obliged to do so. For this right, the off-taker is willing to pay an upfront fee to the seller. The value of the call option can be only calculated afterwards; therefore call options are seen as a type of insurance product. For instance, if the LSE paid $3.75/kW/month for a 200 MW call option (based on the example in the previous table), but there were ten price spikes during the year, then the LSE actually made $9,000,000 profit with the call option (or did not lose this amount). On the other hand, the IPP could have made more profit with the higher call option price. Consequently, the price of hedge is based on the future expectations of prices. The heat rate call option consists of two components: the option fee and utilization fee. The option fee is paid upfront and is based on the same calculation principles shown in Table 2.2. From the IPP perspective, the role of the option fee is to cover the debt service and annual fixed expenses of the capacity investment. The utilization fee is based on the operational economics of the power plant that underlies the heat rate call option. The utilization fee is shown typically as MBtu/kWh and is based on the operational cost the power plant. The typical utilization fee for a 7FA.05 gas turbine is about 13,000 14,000 MBtu/kWh. If the off-taker calls the option, it will cover the operational expenses during the operation. We will show the calculation of the heat rate call option utilization fee in the next section. Typically, heat rate call options are financial contracts which mean that there is no physical obligation to deliver the energy, but the contract value is settled afterwards. The utilization fee is often called the strike price for the call options, so there is no need to actually call the option, but it is assumed that the contracted capacity will come online when the price is high enough. The following example shows the financial settlement of a typical heat rate call option in different situations. The example in Table 3.1 demonstrates the importance of plant availability during the price spike situations. When the plant is available, the option seller returns the difference between the market price (market heat rate) and the strike price to the off-taker and receives only the utilization fee. If the plant is not available during the price spike, the financial impact for the option seller is severe, as it only receives the utilization fee, but does not receive anything from the market. Summer derating 9,000 9,000 9, ,250,000 2,250,000 2,250,000 13,000 13,000 13,000 2,237,000 2,237,000 2,237,000 8,948 8,948 8, ,789,600 1,789,600 1,789,600 10,400 (1,789,600) (259,600) FINANCIAL SETTLEMENT Plant available Not available Market price [$/MWh] Gas price [$/MBtu} Implied market heat rate [Btu/kWh] Heat rate call level [Btu/kWh] Heat rate settlement [Btu/kWh] Heat rate settlement [$/MWh] Contracted capacity [MW] Capacity available [MW] Financial settlement to off-taker [$] Financial settlement to seller [$] Table 3.1. Heat rate call option financial settlement Typically, industrial gas turbines derate during the hot temperatures. This must be taken into account especially in Texas, where majority of the price spikes occur during the summer. The option seller can either carry the risk or sell only the derated capacity. 7

9 4. EVALUATING POWER PLANT TECHNOLOGY OPTIONS FOR HEDGING One of the most popular power plant solutions in Texas for hedging purposes is the GE 7FA.05 industrial gas turbine. There are several projects under development with 2-4 x 7FA.05 configuration to drive down the investment cost on $/kw basis. 7FA.05 is low capital cost capacity with high heat rate of about 10,000 Btu/ kwh. In the traditional hedging business case, the heat rate is not very important as the power plant is only operating a couple of hundred hours per year. However, in the following sections we will look closer into the performance of GE 7FA.05 industrial gas turbine compared to the Wärtsilä 18V50SG internal combustion engine (ICE) power plant. Using this analysis we can demonstrate new hedging business cases in which the Wärtsilä 18V50SG power plant enables to increase the rate of return. 4.1 Technical performance of Wärtsilä 18V50SG and GE 7FA.05 The operating characteristics of a Wärtsilä 18V50SG power plant consisting of 12 x 18.4 MW reciprocating engines and a GE 7FA.05 power plant were examined. Both plants have a nominal capacity of 220 MW. The operational characteristics of both plants are shown in Table 4.1. The modular plant architecture of the Wärtsilä 18V50SG power plant has significant value in the IPP business case. First, the shaft risk, which is the resource s probabilistic loss of load in the event of failure, is twelve times smaller than the 7FA.05 gas turbine. This is important in the hedging business case due to the importance of availability, as discussed in the previous section. Second, the modular design would allow several off-takers for the plant and more modular contract specifications. ICEs can be dispatched individually with the same heat rate as the whole plant. In other words, full efficiency is maintained even at part load. The efficiency, expressed as heat rate, of the Wärtsilä 18V50SG power plant is significantly better than the 7FA.05 power plant at full load and at minimum stable load. The Wärtsilä 18V50SG plant can operate at 20% minimum load which enables more revenue opportunities in the ancillary services markets. The higher efficiency also provides more operations in the energy market as the plant moves lower up in the merit than the 7FA.05. The big difference between 7FA.05 and Wärtsilä 18V50SG is startup operations and cost. Due to thermal stress during startup, the 7FA.05 incurs a maintenance impact every time the turbine is started. Several studies have estimated the start-up cost for 7FA.05, and we have used a start-up cost for this analysis based on published data, shown in Table 4.2. The start-up cost of the gas turbine is $15,000 per start, while Wärtsilä engines do not have a startup cost. Further, the number of starts does not affect the maintenance cost for Wärtsilä engines. The starting of a Wärtsilä 18V50SG engine uses compressed air for combustion, enabling virtually instantaneous start. In a gas turbine, the compressor must accelerate to reach firing speed and then selfsustaining speed. To prevent thermal stress, limits on airflow velocity and combustion temperature constrain how quickly a gas turbine can start and reach full load. The startup cost has a big impact on plant economics and dispatch, as the plant must be able to cover its startup cost when it operates in the market. The variable operations and maintenance (O&M) cost, or VOM, is calculated differently for 7FA.05 than for Wärtsilä 18V50SG. The 7FA.05 VOM of $0.9/MWh includes only consumables, whereas the major maintenance is covered through the startup cost. The $5.5/MWh VOM of Wärtsilä 18V50SG shown in Table 4.1 includes all major maintenances, as well as the consumables (lube oil, urea etc). Plant flexibility is another important factor that will impact availability and revenue opportunities. The Wärtsilä 18V50SG power plant can be synchronized to the grid in 30 seconds and reach full load in less than five (5) 8

10 minutes. The plant can be shut down completely in less than a minute, and does not require a minimum uptime nor down-time. While the 7FA.05 can start within 10 minutes, frequent starts come with a maintenance penalty which limits the plant dispatch, especially in the Real-Time Market. In addition, the 7FA.05 typically requires six hours minimum run-time when it is used to back-up a heat rate call option. There are technical limitations behind the long minimum run-time, but the main reason is the start-up cost, which will be explained in the next section. Output (ISO) Output (sea level, ISO temp) Efficiency (ISO) Minimum stable load Efficiency at minimum stable load Start-up time to full load 18V50SG 18.4 MW 12 x 18.4 = 221 MW 8,266 Btu/kWh 20 % Btu/kWh 5 min 7FA MW 227 MW 9,838 Btu/kWh 40 % Btu/kWh 15 min Start-up fuel Start-up cost (maintenance) VOM 0.3 MBtu/MW/start 1.5 MBtu/MW/start 0 USD/start 15,000 USD/start 5.5 USD/MWh 0.9 USD/MWh Table 4.1. The Wärtsilä 18V50SG outperforms the GE 7FA.05 in heat rate, startup time and startup economics Gas CT CONE Area VOM ($/MWh) LTSA ($/FFS) 1 1 Eastern Mid-Atlantic area 2 Southwest Mid-Atlantic area 3 Rest of RTO areas 4 Western Mid-Atlantic ,846 17,501 18,565 16,968 5 Dominion service territory ,887 Table FA.05 start-up cost and variable operation and maintenance cost (VOM) 2 1Long-term service agreement( LTSA) payments and major maintenance events depend on gas turbine operations typically measured through factored fired starts, or FFS. 2http://brattle.com/system/news/pdfs/000/000/196/original/Cost_of_New_Entry_Estimates_for_Combustion-Turbine_and_Combined-Cycle_Plants_in_PJM_Spees_et_al_Aug_24_11.pdf?

11 4.2 Calculating the heat rate call levels for Wärtsilä 18V50SG and GE 7FA.05 As was discussed earlier in this analysis, the Day-Ahead heat rate call option has a utilization fee or implied utilization heat rate. This heat rate defines the minimum strike level for the option contract. The lower the heat rate, the more valuable the option contract, as the better heat rate covers a wider price band in the market. The components that are included in the implied heat rate in this analysis are the following: Plant heat rate at full load at 95 Fahrenheit (summer temperature) Variable O&M cost (VOM) Start-up cost divided by minimum run-time The calculation of the implied heat rate for Wärtsilä 18V50SG and GE 7FA.05 is shown in Table 4.3. IMPLIED HEAT RATE Plant size [MW] Gas price [$/MBtu] Heat rate at 95 F [Btu/kWh] Variable O&M [$/MWh] Heat rate -Variable O&M [Btu/kWh] Start-up cost [$/start] Minimum run time [h] Start-up cost [$/MWh] Heat rate - Start-up cost [Btu/kWh] Implied heat rate for call option [Btu/kWh] W18V50SG , , ,282 GE 7FA , , ,778 13,050 Table 4.3. Wärtsilä 18V50SG has a lower implied heat rate than the 7FA.05 gas turbine The implied heat rate (and the minimum heat rate for call option) for 7FA.05 is 13,050 Btu/kWh and for Wärtsilä 18V50SG is 9,282 Btu/kWh. The difference between the implied heat rates means that an IPP with a Wärtsilä 18V50SG power plant could ask higher value for the heat rate option contract, as the lower implied heat rates enables lower strike price and therefore more valuable price impact mitigation. The value of this heat rate difference in the ERCOT market will be reviewed later in this analysis. 10

12 4.3 Derating impacts maximum contracted capacity Thermal power plants suffer derating at high temperatures, as the inlet air mass flow decreases. However, different technologies have varied performance and extent of derating. The derating curves for Wärtsilä 18V50SG and GE 7FA.05 are shown in Figure 4.1. It is evident from Figure 4.1 that the GE 7FA.05 derates significantly at high ambient temperatures, while the Wärtsilä 18V50SG plant only begins to slightly derate at 97 F. As was previously shown, derating can be costly to the IPP if the full contracted capacity is not available when the capacity is called. To estimate the impact of derating for the IPP business case we have analyzed the hourly temperature data in the Houston area from 2011 to 2014 to assess the impact of actual ambient temperatures on both technologies. It can be seen from the derating graphs (Figures 4.2 to 4.5) that 2014 was quite a mild year, with minimal derating for a Wärtsilä 18V50SG power plant. The summer of 2011 was very hot, as evidenced from the amount of derating in Figure 4.2. As weather cannot be forecasted years ahead, the IPP needs to decide the acceptable capacity derating level when signing the heat rate call option contract. In this analysis, we select a conservative strategy, and contract only the maximum available summer capacity for both plants during the hottest days: GE 7FA.05 derates several times to 87% of its maximum capacity. For a 227 MW plant this means that only 197 MW can be contracted. Wärtsilä 18V50SG derates to 97% of its maximum capacity. For a 221 MW plant this means that 214 MW can be contracted. 105% 100% Capacity output 95% 90% 85% 80% Temperature [F] V50SG 7FA.05 Figure 4.1. The Wärtsilä 18V50SG experiences less derating at high temperatures than the 7FA.05 gas turbine 11

13 Capacity available 105% 100% 95% 90% 85% 80% Date 7FA V50SG 2011 Figure 4.2. Available capacity in 2011 based on historical temperatures Capacity available 105% 100% 95% 90% 85% 80% Date 7FA V50SG 2012 Figure 4.3. Available capacity in 2012 based on historical temperatures Capacity available 105% 100% 95% 90% 85% 80% Date 7FA V50SG 2013 Figure 4.4. Available capacity in 2013 based on historical temperatures Figure 4.4. Available capacity in 2013 based on historical temperatures 105% Capacity available 100% 95% 90% 85% 80% Date 7FA V50SG 2014 Figure 4.5. Available capacity in 2014 based on historical temperatures 12

14 4.4 Calculating minimum requirement for option fee To calculate the minimum required option fee, we need to estimate the capital cost of GE 7FA.05 and Wärtsilä 18V50SG, the capital structure, cost of capital, fixed operations and maintenance cost, and construction period. Based on market intelligence and Wärtsilä s in-house expertise, we have estimated $500/kW overnight EPC cost for GE 7FA.05 and $700/kW for Wärtsilä 18V50SG. The financial assumptions used in the modeling for both plants are shown in Tables 4.4 and 4.5. ASSUMPTIONS W18V50SG GE 7FA.05 Capacity [MW] Project lifetime [years] Tax rate 37.5% 37.5% Interest rate 5% 5% Overnight EPC cost [$/kw] Owner's cost [$/kw] Construction period [months] Total investment cost [$ Mn] Equity share Debt share % 70% % 70% Loan term [years] Fixed O&M [$/kw] Table 4.4. Financial assumptions for Wärtsilä 18V50SG power plant The overall investment cost for the Wärtsilä 18V50SG plant is $42 million more than the 7FA.05 plant. This means that to receive financing, Wärtsilä 18V50SG requires higher heat rate call option fee than the 7FA.05. The need for fixed cash flow is highest during the first year of operation, since the debt interest is the highest. Consequently, we can estimate the minimum required option fee based on the first year fixed costs. The calculation for the required minimum option fee is shown in Table 4.5, which is based on the debt service, minimum debt-service-credit-ratio (DSCR) and O&M costs. REQUIRED OPTION FEE Debt interest [$] 5,964,090 4,545,105 Debt repayment [$] 6,117,015 4,661,646 Total debt service [$] Minimum DSCR 12,081, ,206, Requided income for debt service 14,497,326 11,048,101 Fixed Operations & Maintenance [$] 2,435,420 1,448,260 Required debt service + Fixed O&M [$] 16,932,746 12,496,361 Contracted capacity [MW] Required heat rate call option fee [$/kw/month] Table 4.5. Required option fee to finance the plant Due to the higher capital and fixed O&M costs, the required option fee is greater for the Wärtsilä 18V50SG power plant. The difference in required option fee between 7FA.05 and Wärtsilä 18V50SG is $1.3/kW/month which equates to approximately $3.3 million difference annually. In other words, the heat rate call option provided by Wärtsilä 18V50SG must be at least $3.3 million more valuable than the option provided by 7FA.05. It should be noted that the required option fees would only cover the fixed operating costs and debt service, but not return on equity W18V50SG GE 7FA.05

15 5 VALUE OF A HEAT RATE CALL OPTION IN ERCOT As discussed previously in this analysis, the value of a heat rate call option is based on the expectations of future prices and market participation. Consequently, there is always uncertainty over the real value or so called fair value of the option. Currently, there are many peaking projects under development, but it is often difficult to sell a heat rate call option at a price that would justify investment in new generation. A majority of the market players are not willing to pay the asking price for a heat rate call option, and current forward option prices are way below to incentivize new capacity construction. To analyze the value of a heat rate option for a GE 7FA.05 power plant and Wärtsilä 18V50SG plant in the ERCOT market, and to better understand the role of market prices on the heat rate call option fair value, we have performed a back-cast analysis for the years Back-cast value for a heat rate call option in ERCOT To analyze the value of heat rate call options in ERCOT during we have modeled the financial settlement for technology-specific call options against hourly Day-Ahead market prices. In the analysis, the implied heat rate varies from year to year according the average natural gas price. Based on the years , the average minimum implied heat rate for the GE 7FA.05 would be 13,828 Btu/kWh, while the Wärtsilä 18V50SG power plant achieves an implied heat rate of 9,395 Btu/kWh. The results for Wärtsilä 18V50SG and GE 7FA.05 are shown in Tables 5.1. and 5.2. BACK-CAST VALUE OF HEAT RATE CALL OPTION IN ERCOT - W18V50SG Contracted capacity [MW] Average gas price [$/MBtu] Heat rate at 95 F [Btu/kWh] Variable O&M [$/MWh] Start-up cost [$/start] Minimum run time [h] Implied heat rate for call option [Btu/kWh] Hours when price over implied heat rate [h] Option called [times] Option called [hours] Heat rate call option fair value [$] Heat rate call option fair value [$/kw/month] W18V50SG W18V50SG W18V50SG W18V50SG W18V50SG Average ,282 8,282 8,282 8,282 8, ,286 9,737 9,355 9,201 9,395 2,692 2,906 2,590 2,527 2, ,692 2,906 2,590 2,527 2,679 28,825,712 9,715,036 6,751,048 10,192,327 13,871, Table 5.1. Value of heat rate call option in ERCOT for Wärtsilä 18V50SG based on back-cast analysis 14

16 BACK-CAST VALUE OF HEAT RATE CALL OPTION IN ERCOT - 7FA.05 GE 7FA.05 GE 7FA.05 GE 7FA.05 GE 7FA.05 GE 7FA Average Contracted capacity [MW] Average gas price [$/MBtu] Heat rate at 95 F [Btu/kWh] 10,047 10,047 10,047 10,047 10,047 Variable O&M [$/MWh] Start-up cost [$/start] 15,000 15,000 15,000 15,000 15,000 Minimum run time [h] Implied heat rate for call option [Btu/kWh] 13,457 14,990 13,693 13,171 13,828 Hours when price over implied heat rate [h] Option called [times] Option called [hours] Heat rate call option fair value [$] 21,087,006 4,627,501 1,481,314 3,486,195 7,670,504 Heat rate call option fair value [$/kw/month] Table 5.2. Value of heat rate call option in ERCOT for GE 7FA.05 based on back-cast analysis The GE 7FA.05 heat rate call option is called on average 95 times per year and an average 593 operating hours during a four-year period. The implied market heat rate is above the 7FA.05 heat rate strike level 670 hours per year on average, it is not economical to call the option about 10 percent of the time because the 7FA.05 is not able to cover its start-up costs. As the Wärtsilä 18V50SG power plant does not have start-up costs and the minimum run-time is less than an hour, the Wärtsilä 18V50SG can respond economically every time the plant is called. Because of the lower implied heat rate of Wärtsilä 18V50SG, the option is called more often an average 2679 hours per year, which is over 2000 hours more than the 7FA.05 heat rate call option. The Wärtsilä 18V50SG heat rate call option is called more often and can mitigate the impact of price spikes more efficiently than the 7FA.05 call option, so it should be also more valuable to the off-taker. Based on the back-cast analysis over , the fair value for Wärtsilä 18V50SG heat rate call option on average would have been $5.40/kW/month, while the fair value for 7FA.05 call option would have been $3.24/kW/month. The Wärtsilä 18V50SG heat rate call option is therefore $2.14/kW/month more valuable than the 7FA.05 plant call option. Based on the calculation of required option fees for 7FA.05 and Wärtsilä 18V50SG shown in Table 4.5., the required minimum option fee for Wärtsilä 18V50SG is $1.3/kW/month due to higher capital expenditures. As the value of Wärtsilä 18V50SG heat rate call option is $2.14/kW/month more than the 7FA.05 call option, it can be concluded that the market value of the Wärtsilä 18V50SG heat rate call option sufficiently covers the difference in higher capital expenditures by 1.55 times. As a result, the Wärtsilä 18V50SG power plant provides more valuable heat rate call option for the off-taker than the 7FA.05 power plant. However, the average value of the heat rate call option for both technology alternatives is lower than the required minimum option fee to cover the debt service and fixed O&M costs. For 7FA.05 the deficit is $2.04/ kw/month, and for Wärtsilä 18V50SG the deficit is $1.19/kW/month. Only in 2011 would the fair value of heat rate call option for both plants been high enough to provide adequate income. As was stated before, the value of heat rate call option is based on the expectations on the future prices, and predicting prices accurately years ahead is impossible. Following the summer of 2011, there could have been several contracts that would have justified investments in new hedge capacity as similar pricing spikes in future years presented risks to LSEs. In the next section, we analyze this type of scenario, where an off-taker signed a heat rate call option with 7FA.05 power plant or a Wärtsilä 18V50SG power plant with the minimum required option fee in early 2011 and called the option according the analysis shown in Table 5.1. and Table

17 6 FINANCIAL FEASIBILITY OF WÄRTSILÄ 18V50SG AND GE 7FA.05 The analysis described in the previous section showed that ERCOT Day-Ahead Market prices over the period were too low to justify investment in GE 7FA.05 or Wärtsilä 18V50SG power plants purely based on hedging purposes. If a heat rate call option with the required price was purchased, there was a premium paid as the anticipated market risk did not materialize. This does not mean that the decision to sign a heat rate call option was wrong, as it provided insurance against potential price spikes in future years. In our next part of the analysis, we assume the following investment situation: Two different IPPs and two different off-takers: The first IPP has signed a heat rate call option with a 7FA.05 power plant backing up the contract prior to 2011 and received the required minimum option fee of $5.29/kW/month for 197 MW. The second IPP has signed a heat rate call option with a Wärtsilä 18V50SG power plant backing up the contract prior to 2011 and received the required minimum option fee of $6.59/kW/month for 214 MW. The first IPP builds a 227 MW 7FA.05 power plant, while the second IPP builds a 221 MW 18V50SG power plant. Both plants come online 1/1/2011. Both plants are dispatched against the Houston HUB Day-Ahead prices and the heat rate call option is financially settled against the HUB prices. As the target of the analysis is to compare the competitiveness and feasibility of 7FA.05 and Wärtsilä 18V50SG power plants, the financial settlement of the call option needs to be done on the same basis. If the off-taker of 7FA.05 call option pays $5.29/kW/month for the capacity, even though the fair value of the option is only $3.24/kW/month, then the premium paid by the off-taker is $2.05/kW/month. For the Wärtsilä 18V50SG call option, the premium is only $1.19/kW/month. We assume the same rationality in the market and between market participants. We also assume that the IPP with the Wärtsilä 18V50SG could receive the same margin from the option contract. The off-taker could analyze the difference in value between the 7FA.05 and Wärtsilä 18V50SG and draw the conclusion that the market-based value of Wärtsilä 18V50SG heat rate call option is $2.16/kW/month more valuable (see Tables 5.1 and 5.2.). Therefore, a utility that is willing to pay $5.29/kW/month for the 7FA.05 heat rate call option would be willing to pay $7.45/kW/ month ($ $2.16) for the Wärtsilä 18V50SG heat rate call option. The financial model consists of several modeling steps. The modeling approach for both plants is shown in Figure 6.1. The technical inputs for the Wärtsilä 18V50SG and GE 7FA.05 are listed in Table 4.1. The market inputs are gathered from ERCOT s webpage and the daily gas prices from the U.S. Energy Information Administration (EIA). 3 The site modeled is close to Houston, with an elevation at approximately sea level. Hourly temperature data is used to calculate the derating of the capacity and the impact on the heat rate (at hourly resolution). Weather data was gathered from the National Centers for Environmental Information webpage. 4 Technical, market and site inputs are used in the modeling platform to optimize plant dispatch against market prices. The market modeling platform runs with five-minute granularity and can optimize plant operation over Day-Ahead, Real-Time and Ancillary Services markets. The outputs of this dispatch optimization against 3http:// 4http:// 16

18 Technical input Modeling platform Market model outputs Financial model Heat rates (GTPro) Part load heat rates Variable O&M cost Capacity derating Start-up times Start-up costs Market input Day-ahead prices Real-Time Market prices Ancillary Services prices Daily gas price Excel tool for backcast analysis 5 minutes granularity to analyze the Real-Time Market Optimize the operation of asset against the LMP and offer capacity optimally between energy and Ancillary Service Assumes that capacity fits into the merit if it is feasible to operate the plant Takes into account temperature derating on hourly level Running hours and number starts 5 minute dispatch profile Revenue per market Operating costs (fuel, VOM, start-up costs) Gross margin from the market operations Investment input Investment cost Fixed Operation and Maintenance costs Capital structure Cost of capital Project feasibility Project IRR Equity IRR Cash flows Site input Option model Elevation Price node information Hourly temperature data Option fee income Option settlement Figure 6.1. Financial feasibility modeling steps market prices along with investment inputs and option fee income are then used to evaluate project feasibility. The financial model is a cash flow model over 30 years that provides metrics relevant to project feasibility, such as project and equity internal rate of return (IRR). The dispatch for Wärtsilä 18V50SG and GE 7FA.05 power plants is modeled over the years , and the average dispatch is used as an input for future feasibility analysis. Average option fee and option settlement data for are similarly used in the model. In other words, we assume that the average economic outcome of years will take place over next 30 years. An IPP that sells a heat rate call option for full capacity with the minimum implied heat rate at fair market value would be able to cover the fixed operations and maintenance costs and debt service. However, that IPP would not be able to generate return on equity if the plant is used only in the Day-Ahead market. When the market heat rate is above the strike level (implied heat rate) of the option contract, the plant would operate but the option seller receives only the utilization fee. The difference between the market price and the strike price is returned to the off-taker. Consequently, to generate additional income, the IPP must operate in the Ancillary Services or Real-Time markets. The following case studies further clarify this dilemma and show the importance of other market cash flows. 17

19 6.1 Day-Ahead Market participation only In this case study, both plants (Wärtsilä 18V50SG and GE 7FA.05) are used in the Day-Ahead Market only during Based on the results of the dispatch model, the operating profiles over these years differ for the two technologies. Both plants are offered into the market at their short-run marginal cost, while taking into account daily gas prices and the impact of ambient temperature on heat rate. On average, the 7FA.05 plant would operate 1123 hours per year with 101 annual starts. The Wärtsilä 18V50SG plant would operate on average 2628 hours with 432 starts over twice as many hours. The average operating time per start for the Wärtsilä 18V50SG plant is 6.1 hours, while the 7FA.05 is 11.1 hours. The higher average operating hours per start for the 7FA.05 plant is due to start-up cost, which makes the short operating pulses (less than 6 hours) uneconomical in most cases. However, the Wärtsilä 18V50SG plant would operate pulses as short as one hour, whenever the market price is above its short-run marginal cost. There is a slight difference between the actual operating hours and modeled option called hours (tables 5.1 and 5.2.), because in the dispatch calculation daily gas price was used, while in the option calculation we used average yearly gas price to define the implied heat rate. The annual operating hours, number of starts and average operating profiles for are shown for Wärtsilä 18V50SG and GE 7FA.05 power plants in Figure 6.2. Figure 6.2. Operational profiles of Wärtsilä 18V50SG and GE 7FA.05 in the Day-Ahead Market only case The operating hours are high for both plants, as both plants offer their capacity to the market without any margin. If for example a margin of $6.0/MWh was added on top of the short-run marginal cost, the number of operating hours would decrease without major impact on the gross margin. Revenues from the Day-Ahead Market are based on the hourly market prices, and operating cost is based on dispatch profiles and daily gas prices. The project financing calculations for the Day-Ahead Market only case is provided in Appendix A, Table A.1 for Wärtsilä 18V50SG and Table A.2 for GE 7FA.05. The 221 MW Wärtsilä 18V50SG plant is able to provide 7.7% IRR for equity and 5.9% IRR for the project. The 227 MW GE 7FA.05 project can reach only 5.9% IRR for equity and 5.2% for the whole project. Consequently, a project with Wärtsilä 18V50SG technology is able to provide higher return on investment than the 7FA.05. Someone could argue the higher IRR is the reason behind the higher heat rate call option fee for the Wärtsilä 18V50SG plant, but this is not true. The outcome of the heat rate call option settlement (option fee option settlement) is equal for both plants, since they receive the same premium on the contract as discussed earlier in this analysis. 18

20 6.2 Day-Ahead + Real-Time Market optimization The overall IRRs for both plants are quite low in the Day-Ahead Market only, and would probably not attract investments even though the financing has been secured through the heat rate call options. As the heat rate call options were sold with the cost-based implied heat rate, the plants are not able to make any additional money in the Day-Ahead Market, even though there would be more price spikes. In the case of additional price spikes, the revenue from energy production would increase, but simultaneously the financial settlement of the heat rate call option would increase. However, there are possibilities to get additional gross margin from other markets than the traditional Day-Ahead Market. The flexible Wärtsilä 18V50SG power plant could operate in the Real-Time Market in two different ways. First, the plant could be started to take advantage of Real-Time Market price spikes. Second, the plant could be shut down and the Day-Ahead commitment fulfilled through the Real-Time Market purchases when the price is below the short-run operating cost of the plant. When including the Real-Time Market into the analysis, the operational profile for the Wärtsilä 18V50SG power plant changes compared to participation in only the Day-Ahead Market. The 7FA.05 power plant has almost the same operational profile for both cases, as shown in Figure 6.3. Figure 6.3. Operational profiles of Wärtsilä 18V50SG and GE 7FA.05 in Day-Ahead + Real-Time Markets case On average the Wärtsilä 18V50SG plant operates almost 1000 hours less than in the Day-Ahead Market only case. The reason behind this change is the market procurement. If the price in the Real-Time Market is lower than the operating cost of the plant, it does not make sense to start the plant, but rather to fulfill the commitment through market procurement. This type of flexible behavior is not possible for the 7FA.05 plant due to its operational characteristics. Imagine a situation where the plant operator decides not to start the 7FA.05 plant as Real-Time prices are low, and then suddenly the price for next five-minute interval is at $9000/ MWh and the 7FA.05 plant misses the price spike. Not starting the plant is a very costly and risky strategy. As the Wärtsilä 18V50SG power plant is able to reach full output in less than five minutes and maintenance is not impacted by the number of starts, chasing price spikes in the Real-Time Market is possible for the Wärtsilä plant. The project finance calculations for the Wärtsilä 18V50SG and GE 7FA.05 plants are provided in Appendix A, Tables A.3 and A.4, respectively. The project economics for 7FA.05 are the same whether participating in the Day-Ahead Market or the Day- Ahead + Real-Time Markets, as only very small adjustments could be made to the operational profile of the 19

21 plant in the Real-Time Market. The project IRR is 5.2% and the Equity IRR is 5.9% for the 7FA.05 plant. For the Wärtsilä 18V50SG plant there is a very positive impact from the Real-Time Market operations. The project IRR is 7.3% and the Equity IRR 11.6%, which are acceptable numbers. Real-Time Market participation also improves the debt service capability of Wärtsilä 18V50SG plant, and the minimum DSCR is 1.5 which is above the minimum accepted level. Let s look a bit closer where the additional income for Wärtsilä 18V50SG plant comes from in this case. The revenue is about $3 million higher than in the Day-Ahead only case. This means that the plant is started in the Real-Time Market and is able to capture $3 million additional revenue from the market. The big change is in the operating cost of the plant which is about $7 million lower than in the Day-Ahead case. The lower operating costs are realized due to market procurement; when the plant is not operated in the Real-Time Market because prices are low, the Day-Ahead commitment is met by market procurement. 6.3 Day-Ahead + Real-Time + Ancillary Services Market optimization The Ancillary Services market is cleared in co-optimization with energy in ERCOT s Day-Ahead Market. As there is not a secondary market for ancillary services in the Real-Time Market, the ancillary service commitment is binding. Therefore, a plant which has sold ancillary services in Day-Ahead must also deliver the product in Real-Time. This market feature reduces the role of Real-Time market for the Wärtsilä 18V50SG power plant if it decides to sell ancillary services in ERCOT. There are currently four different ancillary services in ERCOT: Regulation Up, Regulation Down, Responsive Reserve and Supplemental. The first three products require that the resource providing the service must be synchronized to the grid, but the Supplemental service can be provided by non-synchronous generation. The Regulation Down and Responsive Reserve products require that the resource is operating at part-load to sell the head room for ancillary services. The minimum stable load for the Wärtsilä 18V50SG plant is 20%, so it is able to sell 80% of its capacity for upward ancillary services products. The minimum stable load for the 7FA.05 plant is 40%, so the headroom for ancillary services is 60%. While the plant is providing regulation down service, it would operate at full load and sell the capacity down to minimum stable load for the service. Figure 6.4. Operational profiles of Wärtsilä 18V50SG and GE 7FA.05 in the Day-Ahead + Real-Time + Ancillary Services Markets case 20