This is the third in a series of questions that highlights RFF’s Expert Forum on EPA’s Clean Power Plan.
RFF asks the experts: Is it possible for existing natural gas power plants to increase average utilization by 70 percent (building block #2) and, if so, at what cost?
With a boom in natural gas production in the United States, many view this fuel source as a possible “bridge” to a low-carbon future. In fact, building block #2 of EPA’s Clean Power Plan assumes that states can increase the average utilization of existing natural gas power plants to 70 percent, substantially higher than current rates. RFF asked the experts about the practicality and cost-effectiveness of this building block. Do such opportunities exist, and at what cost? Would an increase in the utilization of existing gas facilities render them unavailable to balance the intermittent supply from renewable energy generation?
Is it possible for existing natural gas power plants to increase average utilization by 70 percent (building block #2) and, if so, at what cost?
—Dallas Burtraw, Darius Gaskins Senior Fellow, Resources for the Future (See full response.)
“If you address this from the technical perspective, there is no question that the natural gas combined cycles that EPA focuses on are fully capable achieving and maintaining a 70 percent capacity factor. To achieve this, however, four issues will need to be addressed with careful modeling and planning.”
—Robert Hilton, Vice President, Power Technologies for Government Affairs, Alstom Power Inc. (See full response.)
“A more gradual phase-in of building block #2 over time would avoid this reliability-driven rush to new gas-fired generation, . . . allow a more thoughtful transition to increasingly competitive renewables and other clean energy resources, . . . and lower CO2 emissions from the power sector as a whole at a lower cost.”
—Steve Corneli, Senior Vice President, Policy and Strategy, NRG Energy (See full response.)
Dallas Burtraw
Darius Gaskins Senior Fellow, Resources for the Future
Recent history suggests a dramatic change in capacity factor is very possible, but prior experience does not provide evidence that it would be sufficient to achieve the building block target of 70 percent on average for all gas plants. This seems much more plausible for relatively newer plants. Of course, EPA’s proposed Clean Power Plan does not require this outcome for all plants or for any specific plant; rather it has to be sufficiently plausible to serve as a justification for the stringency of the proposal. The actual outcome could involve other changes in the electricity system, including investment in new facilities.
The national average capacity factor (utilization) for natural gas combined cycle plants (NGCCs) increased from 40 percent to 51 percent from 2008 to 2012, falling somewhat in 2013, and it is widely believed that additional shifting from coal to gas represents the lowest-cost option for reducing emissions from the power sector. In some regions, the change has been even more dramatic. For example, generation from natural gas in Pennsylvania increased by 58 percent between 2010 and 2012. EPA’s Clean Power Plan suggests that existing NGCC units could increase their capacity factors to 70 percent on average.
To determine whether this is plausible, one can look at recent performance. Capacity factors across the gas fleet show a pattern of variation by vintage. Capacity factors of the oldest plants are a lot lower than for all other plants and are fairly insensitive to relative fuel prices. The price of gas fell by half from 2008 to 2012 but, between those years, the median capacity factor of the oldest group only increased from 20 to 25 percent.
By comparison, between 2008 and 2012, the median capacity factor of the group of plants installed between 2001 and 2003 increased from about 32 to 46 percent. There is less heterogeneity among the newer plants, but few existing plants currently operate at 70 percent utilization. Econometric analysis in progress by my colleagues Joshua Linn and Lucija Muehlenbachs finds that the capacity factor of larger and newer gas plants responds more to gas prices than the capacity factor of smaller and older gas plants.
Another feature of natural gas generation is the variation in use by time of day. Gas turbines are expected to run during times of peak electricity demand, but NGCC units are expected to run more evenly on a daily basis. Nonetheless, over the past decade the capacity factor for NGCC units varies by a factor of two between times when it is least and most in use, suggesting considerable room for continued utilization over all hours.
The significance of these findings is two-fold. On the one hand, the fact that the newer part of the fleet is more responsive to changes in gas prices suggests that greater utilization is plausible on a fairly widespread basis, as EPA has asserted. On the other hand, gas plants that went online prior to the early 1990s now account for a much lower share of total generation than before. Because these plants do not tend to be responsive to changes in gas prices, the cost of increasing the utilization of these plants would be high—perhaps due to the characteristics of the plants, their location in the electricity system, and the availability of gas supply or proximity to low-cost coal units. In this vein, greatly expanding the utilization of gas units at older plants could be more costly than what is revealed by recent trends for newer plants.
EPA should consider carefully whether incentives should be given for construction of new facilities by including them in the calculation of the demonstrated emissions rates. Scenario modeling at RFF suggests that the expansion of natural gas use might be most likely to occur through investment in new units, especially if the emissions rate for new gas units is below the target emissions rate. EPA’s Clean Power Plan is ambiguous to some degree about the treatment of new plants. However, in the preamble, it appears that the emissions rate calculation is to be based only on existing units. In this case, there would be no additional incentive for new gas.
Robert Hilton
Vice President, Power Technologies for Government Affairs, Alstom Power Inc.
If you address this from the technical perspective, there is no question that the natural gas combined cycles (NGCC) that EPA focuses on are fully capable achieving and maintaining a 70 percent capacity factor. The gas turbines were designed for this level of operation, as base loaded, and the heat recovery steam generators would operate much better with less maintenance if operated as base loaded rather than cycling. There is ample support and service available to bring less efficient machines into competitive operation and to maintain the existing fleet at best operating condition.
However, the real issue lies in different sectors of the market. As noted by EPA, NGCCs currently average approximately 46 percent capacity factor in the most recently available data. The primary reasons for this level of operation are as follows:
- Economic or merit order dispatch: Even with today’s historically moderate to low natural gas pricing, not all NGCCs can compete in all markets based on bidding into competitive markets. EPA proposes that states will fix this situation by forcibly adjusting dispatch order.
- Over-capacity in certain regions of the Federal Energy Regulatory Commission, particularly regulated states: In many regions, there still exists excess capacity, largely owing to the fact that electricity demand has not fully recovered to pre-recession levels. With much of the NGCC capacity comprised of independent power producers or merchant plants, we see preferential dispatch for the regulated resources, thus leaving merchant plants as peak suppliers.
- Available fuel resources: Distribution infrastructure for natural gas still remains problematic in certain regions, so supplying NGCCs with adequate gas year round will cause some NGCCs to run on liquid fuels or shut down, reducing capacity factors.
- Consideration of plant locations and grid requirements for geographical area power demand: A few units may be located in areas where the demand is insufficient to reach these proposed capacity factors.
Coal plants have traditionally provided the base load with capacity factors of 65 to 73 percent, but presumably in EPA’s scenario this will be significantly reduced. Nuclear will remain in the 90 percent level for those units that are maintained in the fleet. For natural gas to assume the base load required to meet EPA’s scenario, the four issues will need to be addressed with careful modeling and planning. Assuming states follow EPA’s suggestion of tilting the market to an environmental dispatch, it would seem likely this will cause some increase in the price of electricity.
Finally, this further raises the issue of providing backup for renewable power intermittency. It would seem most likely this will be covered by simple cycle units that can quickly enter the market and, unless they exceed 219,000 megawatt hours of operation, will remain exempt from regulation. This may put further strain on the natural gas delivery system. Otherwise, it will be logical that backup power will come from a portion of the existing coal fleet that has been reduced in capacity factor in favor of NGCCs. However, with careful modeling and planning, even at a 70 percent capacity factor, there can be adequate head room for NGCCs to still provide significant backup to renewables, depending on demand.
Steve Corneli
Senior Vice President, Policy and Strategy, NRG Energy
The most noteworthy aspect of EPA’s building block #2 may well be not what it is, but when it is applied. It is clear from the arithmetic in EPA’s technical support documents that EPA assumes full redispatch to the maximum assumed level in 2020 and each year thereafter, and that the proposed rule’s interim goals are in fact derived by averaging together the annual targets based on this assumption (see Appendix 1 of EPA’s goal computation technical support document). This, together with the proposed rule’s requirement that the interim goal be met on average in the first 10 years, would require most of the redispatch-related emissions reductions to be achieved in 2020.
Such “overnight.” large scale redispatch raises a number of daunting problems in the real world, which are inadequately addressed in EPA’s analyses. In states with a large number of existing combined cycle units and existing coal plants, such a sudden major redispatch has the potential to create resource adequacy challenges by stranding assets needed for reliability, increasing consumer costs, and, in all likelihood, leading to unnecessarily large amounts of new natural gas power plant investment and the long term lock-in of their CO2 emissions.
Consider, for example, a hypothetical state that has approximately one-third of its power supply from coal plants and one-half of its power supply from combined cycle plants. Under typical recent capacity factors for combined cycle gas (44 percent) and coal plants (65 percent), building block #2 reductions would require coal plants in the state, on average, to reduce their generation output by 60 percent. Adding in the other building blocks will require additional generation reductions from coal plants, in some cases to 80 percent or more. Reducing coal plant generation output from existing plants by such large amounts overnight raises two immediate problems:
- The potential to render many baseload plants—which are only economic to operate at relatively high levels of output—unable to recover their fixed operations and maintenance costs, leading to immediate economic mothballing or retirement.
- 2. State policies that inadvertently drive even more plants into economic retirement. For example, a uniform carbon price that would be sufficient to achieve the redispatch of 60 percent or 80 percent of a state’s coal plant output would likely severely challenge the economics of all of its coal plants.
Going back to our hypothetical example, the sudden retirement of most or all coal plants in a state where they make up one-third of the supply stack clearly will threaten resource adequacy (unless the state enjoys a 33 percent or greater reserve margin) and risk a major resource adequacy crisis in states with much smaller reserve margins. Such a crisis would almost certainly be met by deployment of new natural gas plants—whose dispatchable capacity counts toward reserve requirements and whose CO2 emissions simply don’t count under the proposed rule.
By contrast, a more gradual phase-in of building block #2 over time would avoid this reliability-driven rush to new gas-fired generation. Such a phase-in should also allow a more thoughtful transition to increasingly competitive renewables and other clean energy resources, less emergency-inspired new gas plants, and lower CO2 emissions from the power sector as a whole at a lower cost.
