TODAY’S STUDY: A DEEPER DIVE INTO THE TEXAS ALL-OF-THE-ABOVE ENERGY FUTURE
Exploring Natural Gas and Renewables in ERCOT, Part III: The Role of Demand Response, Energy Efficiency, and Combined Heat & Power
Shavel, Fox-Penner, et al, May 29, 2014 (The Brattle Group via The Texas Clean Energy Coalition)
Introduction and Summary
In December 2013 we completed a series of simulations of the future of the ERCOT electricity marketplace through 2032 on behalf of the Texas Clean Energy Coalition (TCEC). The goal of these simulations was to show how renewable and natural gas-fired electricity sources would develop in the next two decades, and how this development would depend on gas prices, the existence of a capacity mechanism in ERCOT, national carbon policies, and other key market drivers. We found a wide range of 2032 outcomes, from scenarios with over 40% of ERCOT’s 2032 energy coming from renewables to other cases in which nearly all future capacity additions would be gas-fired.1
In order to simulate the future ERCOT grid we relied primarily on ERCOT’s own long-term planning and operating data, supplemented by selected additional items from our research. In the areas of electric power demand and sales, the role of demand response and energy efficiency, and the role of distributed sources of generation, we relied almost entirely on ERCOT’s base case forecasts and no sensitivity analysis. In short, our simulations were focused primarily on exercising the large-scale supply side of the marketplace under “reference case” demand-side assumptions.
In this report we expand our prior work to incorporate a more extensive set of demand-side scenarios. These scenarios, most of which would require state policy and/or market changes, provide a more complete answer to our original motivating question, i.e., what are the possible range of outcomes for the full set of electricity resource options in the future Texas power grid? What are the drivers of these futures, and how much impact do they have on electricity prices, greenhouse gas emissions, and other important factors? How much of the future depends on policy choices versus the inexorable tide of market forces?
Our new work employs an updated version of the unique modeling system employed in our 2013 work. This system -- described in more detail below -- combines a model that simulates the decisions of market-driven developers of a wide range of new electric resources with a model that simulates the minute-by-minute operation and control of the grid by ERCOT. By combining these two perspectives, our modeling system finds future trajectories that represent, for any given scenario, a realistic set of resources the market is willing to build and that can be integrated and managed by ERCOT to yield adequate and reliable power service.
B. Scenarios And Results
Our new results are drawn from a revised set of scenarios involving both key supply and demand-side drivers. The four scenarios we examine are:
• Phase III Reference. A new Reference case, with updated forecasted power sales and base case gas prices, as well as added CHP potential and a refined and expanded portfolio of DR programs;
• Enhanced Energy Efficiency. The Phase III Reference case with an added portfolio of cost-effective energy efficiency programs;
• Moderate Federal Carbon Policy. The Enhanced Energy Efficiency case with an added requirement that all coal-fired facilities capture and sequester 50% of their CO2 by 2025; and
• Strong Federal Carbon Policy. The Enhanced Energy Efficiency Case, but with (a) a rule that all coal-fired plants sequester 90% of their CO2 by 2025; (b) higher natural gas prices due to increased gas demand to replace coal units that cannot cost-effectively sequester 90%; and (c) lower renewable energy costs from more rapid deployment. The highlights of our findings include these results:
• Across all Phase III scenarios, natural gas and renewable additions dominate the supply picture, with gas providing both low-cost baseload energy and ancillary services that integrate wind and solar energy. The original forms of complementarity we have discussed in prior reports for the TCEC continue to occur, albeit in a more nuanced manner with the introduction of the increased options of EE, DR, and CHP.
• New large CHP installations at petrochemical facilities are very economical and the simulation indicates that the full potential of these opportunities will be realized by 2032 in all scenarios. However, the high capital costs and rapid required payback required of smaller CHP units prevent any further CHP adoptions in our scenarios.
• Energy efficiency and demand response provide substantial opportunities to displace future capacity additions and lower overall electricity costs. Our program portfolio was designed to be moderate in size and use well-established approaches primarily driven by ERCOT energy prices. Nonetheless, 3 GW of new EE programs and around 2-4 GW of new DR programs are identified as economically achievable in ERCOT in our modeling scenarios. In total, this represents a 40% to 50% reduction in projected peak demand growth (depending on the carbon policy scenario).
• Real energy prices in Reference scenarios remain within the band of prices actually experienced in ERCOT between 2010 and 2012. The highest annual average price for a converged year is about $67/MWh ($2012) for the Strong Federal Carbon Policy scenario, which has higher gas prices. However, even this extreme scenario price is $3/MWh lower than its counterpart scenario in Phase II.
• Carbon emissions are slightly below all comparable Phase II scenarios. These lower emissions are the net effect of reduced sales (including from EE programs) and higher renewables penetration, offset by reduced retirements of inefficient older capacity.
C. Important Study Considerations
As in the prior phase of this research, we emphasize that these results are not intended to be our definitive prediction of the most likely future path for the Texas power marketplace nor our explicit policy recommendations. Instead, our goal is to illustrate the relative effect of important drivers, determine whether renewables and gas are likely to be complementary or competing, and explore the effects of a limited set of policy directions. Among other limitations, the set of generation resources, Demand Response (DR) options, and Energy Efficiency (EE) programs we include in our modeling system is by no means intended to be exhaustive. There are unquestionably other options that we could not include in our scenarios that will play a role in Texas’ energy future, whether in the form of new demand response options, new low-carbon generation technologies, or expanded forms of traditional power.
There are also important limitations and assumptions in this study that should be noted. First, we do not assume major technical breakthroughs in new energy technologies such as electricity storage, small nuclear reactors, or carbon capture and sequestration. Second, even drawing from the current and forecasted set of technologies we do not include every current option. For example, concentrating solar power plants are now in use in the Southwestern United States, but we do not include them as a resource option simply due to budget limitation. Third, our solar Photovoltaics (PV) resource option should be viewed as utility-scale solar, as we do not include state or federal policy changes that would accelerate distributed solar in Texas. Fourth, our model reflects much but not all of time variability of solar and wind power, thus slightly understating the integration resources needed for these additions. Finally, our modeling system does not formally incorporate risk aversion and fuel price uncertainty, which would reduce gas investment relative to wind, solar, and coal-fired options. These considerations are discussed further in the introduction to our Phase II study…
As we have found in all our work for TCEC, the relationship between natural gas and renewable generation is multifaceted, with substantial room for both to grow in nearly all futures. In Phase II, gas prices, renewable policies, and renewable cost reductions stood out as critical drivers of the mix between these two types of generation. In Phase III these findings continue to hold, but they are shaped by DR and EE which reduce overall sales and peak demand. In Phase III, capacity growth is a little more evenly divided between gas and renewables relative to Phase II.
A particularly good example of this is illustrated by the capacity additions in the Moderate Carbon scenario. As noted above, the abrupt loss of 3.5 GW of coal-fired capacity causes the market to immediately supply a similar amount of CCGT capacity and later build out higher levels of renewables, ultimately ending with a two-to-one ratio of renewable nameplate capacity additions to CCGT additions. Since wind capacity produces about half the energy per rated MW, the incremental energy contributions are more closely aligned.
Our results also show that expanded DR and energy efficiency options help reduce total energy costs and can be successful in ERCOT’s market-driven environment. By 2017 we identify between 450 and 760 MW of economic, achievable DR in ERCOT across our scenarios, a 20% to 30% increase over the existing ERCOT portfolio. By the end of our forecast horizon, the economic new DR grows to between 2.3 and 3.8 GW, a 90% to 150% increase over the current portfolio (it varies by scenario). This includes dynamic pricing, which is cost-effective in all scenarios and capable of providing more than 1 GW of peak reduction by 2032. In our reference case, the size of the total DR portfolio (existing and new) in 2032 is 6,350 MW, 7.8% of the projected system peak (the relative size of the DR portfolio exceeds the national average of existing DR across ISO/RTOs of around 6%). Of this total, new DR capacity is between 2.3 GW and 3.3 GW, depending on the carbon policy scenario. Combined with 3 GW of peak reduction in the expanded EE portfolio, this represents a 40% to 50% reduction in projected peak demand growth over the forecast horizon.
These results also show several subtle tradeoffs between DR and other resources. For example, slower load growth encourages older base load resources to stick around, which is economical, but slows the turnover of the fleet. This has a two-edged effect on fuel efficiency and emissions, slowing the growth of new CCGTs but generally increasing the growth of wind power but not solar. Thus, DR in the ERCOT system is more complementary to wind than to either solar or gas. Of course, the most important feature of DR is that it saves customers’ money by deferring plant construction, while still reducing emissions overall.
Conversely, strong carbon policies reduce the need for DR by necessitating fleet turnover. However, even here we see that slower growth, DR, and EE as well as higher gas prices induce about twice as much wind development as CCGT additions when coal plant capacity is removed.
If one assumes that the U.S. will adopt a strong carbon policy in the long term but not in the near term, the logical evolution of these resources might be to emphasize DR and energy efficiency now. Since DR does not have large long-term capital servicing requirements, few costs would be “stranded” when a stronger climate policy triggered a larger fleet transition than current policies allow.
The combined effects of lower load forecasts, DR, EE and CHP have slightly reduced average customer bills and greenhouse gas emissions. The combined effects of higher gas prices, lower load growth, enhanced DR and CHP installations lower CO2 emissions about 4% by 2032 versus the Phase II Reference Case, or 143 million metric tons. This is the equivalent of closing one 600 MW coal plant for 30 years. New EE programs further reduce CO2 by 10 MMT, the equivalent of one year’s emissions from an 800 MW coal plant. Table VII-8 shows the major air emissions under all Phase III scenarios.
Average wholesale electricity market prices have risen across the board in our Phase III scenarios due to higher natural gas prices, partially offset by the price-reducing effects of DR, CHP, and EE. The net effect remains slightly higher prices for the Phase III scenarios versus their phase II counterparts. These results are shown in Figure VII-12, which shows all Phase II average price results (lighter bars) alongside Phase III scenarios (darker bars). In inflation-adjusted terms, prices in the Reference scenarios remain within the band observed between 2010 and 2012, from a low of about $42/MWh to a high of about $67/MWh under the strong carbon rule. Importantly, the inclusion of EE, DR, and CHP in the Phase III scenario reduces the higher-priced carbon rule scenarios, as what would otherwise have been. In Phase II, 2032 prices in the carbon rule cases topped out at almost $70/MWh, $3/MWh more than the same scenario in Phase III.
We began our ERCOT simulations for TCEC to examine tradeoffs and complementarities between gas-fired and renewable generation sources. With the results of this Phase, we integrate energy efficiency, combined heat and power, and demand response into ERCOT’s future. This larger set of options provides a richer set of tradeoffs and complements. Demand response and energy efficiency lower the need for new supply-side resources of any type in a cost-effective manner. Until carbon rules are in effect or other policies change, gas CCGT additions dominate supply-side additions until the late 2020s, when unsubsidized renewables begin to compete on price alone. Following this period – and especially following carbon rules - renewable capacity additions become the majority, but substantial gas investment continues to be made to integrate renewables into the system.
As with any modeling and market simulation, the specific results we have calculated are subject to many assumptions. Moreover, the range of results across our scenarios show that different price and policy futures will have an overwhelming influence on the future of ERCOT’s grid. Amidst this huge range of specific outcomes, the sum total of our results indicate that both gas and renewables are likely to be developed in substantial amounts together in the Texas markets, with gas prices, carbon and renewables policies, and renewables price reductions servings as the most important drivers. In addition, our results show that expanded energy efficiency and demand response programs are economical for Texas energy customers across nearly every realistic future for the ERCOT market.