A report released today by the North American Electric Reliability Corporation (NERC) found that the Clean Power Plan can be met with modest changes in the generation mix. The required changes, chiefly displacing coal generation with cleaner alternatives, are actually less aggressive than the generation mix change trajectory we have seen over the last decade as the cost of cleaner energy sources like renewable energy and natural gas has declined. In fact, the coal power plant retirements NERC projects between now and 2030 will be more than offset by the nearly 50 GW of new gas generating capacity that is already planned for addition between now and 2020. NERC finds that the Clean Power Plan would only increase coal plant retirements during 2016-2020 from an already planned 33.5 GW to between 41 and 43 GW, an increase of fewer than 10 GW.
However, even these achievable findings greatly overestimate the challenge of meeting the Clean Power Plan because NERC used unreasonably pessimistic assumptions about the contributions of renewable energy and energy efficiency. NERC forced arbitrary and unreasonable assumptions into its model that limited the deployment of energy efficiency and renewable energy, forcing the model to rely almost entirely on coal-to-gas shifting to meet the Clean Power Plan (CPP). Because renewable energy and energy efficiency can be deployed very quickly, a fact noted by NERC’s Report, they will likely play a far greater role in meeting the early CPP targets than NERC allowed them to. As a result, NERC’s concerns about the feasibility of meeting the early year CPP targets are likely to be exaggerated.
Flawed assumptions caused NERC to overstate the generation mix changes required to meet the Clean Power Plan.
NERC’s analysis arbitrarily and inaccurately barred large-scale use of the compliance options offered under EPA’s Building Blocks 1, 3, and 4, so NERC’s model was forced to almost entirely rely on Building Block 2, coal-to-gas shifting, to comply. This impact was compounded by the fact that using a resource with some emissions, like natural gas, requires a far greater shift in the resource mix than use of a zero-emission resource, like renewable energy or energy efficiency. Specifically, because a MWh of gas generation emits nearly half as much CO2 as a MWh of coal generation, nearly twice as many MWh of coal generation must be displaced to achieve the same level of emissions reductions versus if zero-emission compliance options were used instead.
Had NERC not arbitrarily limited renewable energy and energy efficiency, they could have played a far greater role in meeting the CPP targets for the early 2020s in particular, addressing NERC’s concerns about the feasibility of those targets. Renewable energy is already greatly exceeding EPA’s CPP targets, providing surplus emissions reductions that will reduce the obligation on other generation sectors to reduce emissions. Moreover, energy efficiency and renewable energy could have made significant contributions to meeting system capacity needs, further reducing any concerns identified by NERC.
Several of NERC’s most problematic assumptions – related to energy efficiency, renewable energy, and fossil plant retirements – are discussed below.
NERC arbitrarily limited wind and solar expansion to unrealistically low levels. NERC’s assumptions capped the amount of wind and solar energy that could be deployed by its economic model. These caps are well below recent trends in solar and wind energy deployment, so they greatly limit renewable deployment in the study and therefore greatly increase the burden on other compliance options, namely coal-to-gas generation shifting. Specifically, NERC’s assumptions limit wind energy’s growth to around 2 percent per year, while the actual U.S. wind industry annual growth rate since 2006 is more than ten times larger at 24.4. Similarly, solar deployment was limited to 12 GW between now and 2030, even though the solar industry reports that 20 GW are expected to come online in the next two years alone.
The NERC report does note that renewable energy can be added very quickly and therefore can play a key role in meeting early Clean Power Plan compliance targets. NERC’s report correctly notes that wind energy has the shortest timeline for bringing a project online, fast enough to fully deploy needed wind energy by the 2020 start date of the Clean Power Plan. Even the very short timeline listed for wind in the NERC report is likely pessimistic, as it misses the fact that wind developers already have a large pipeline of wind projects that are ready to be built, as much of the permitting and interconnection work has already been completed. Had NERC allowed the model to deploy renewable energy, it could have played a critical role in meeting the early year compliance targets, alleviating NERC’s concerns about the feasibility of those targets.
NERC used obsolete information about the reliability services capabilities of modern wind turbines. NERC’s report also contradicted other findings by its own technical experts by using outdated information on the ability of renewable resources to provide essential reliability services. Here are some of the recent statements that NERC and others have made about wind’s reliability:
- Contradicting NERC’s report, NERC recently noted, “This issue does not exist for utility-scale wind energy, which offers ride-through capabilities and other essential reliability services.”
- NERC stated “Modern wind turbine generators can meet equivalent technical performance requirements provided by conventional generation technologies with proper control strategies, system design, and implementation.”
- Regarding voltage and reactive power control, NERC has noted “As variable resources, such as wind power facilities, constitute a larger proportion of the total generation on a system, these resources may provide voltage regulation and reactive power control capabilities comparable to that of conventional generation. Further, wind plants may provide dynamic and static reactive power support as well as voltage control in order to contribute to power system reliability.”
- NERC recently noted that “However, by causing conventional generators to have their output dispatched down, wind and solar generation can increase generator headroom and, therefore, the amount of total frequency response being provided.” Moreover, NERC’s previous analysis found that many coal plants are not currently providing sustained frequency response. In addition, NREL has documented that “Wind power can act in an equal or superior manner to conventional generation when providing active power control, supporting the system frequency response and improving reliability.”
A longer summary of wind’s contributions to reliability services is included in the table at the bottom of this post, and more documentation of those and other points is available here.
Unrealistically low energy efficiency gains. NERC assumes that energy efficiency can reduce electricity demand growth by 0.5 percent per year. NERC’s assumption is only one-third of the contribution that has been cost-effectively achieved by states that have implemented energy efficiency policies, as documented in EPA’s CPP proposal. As noted in the Brattle Group’s February 2015 review of EPA’s Clean Power Plan, EPA’s estimates are themselves likely conservative, in part because they rely only on energy efficiency improvements stemming from utility programs, which currently represent roughly only 50 percent of all energy efficiency spending. In addition to utility-sponsored EE programs, states and power companies can draw on many other forms of energy efficiency savings to comply with state goals, including private sector investments in EE, savings from improved building codes and appliance standards, and transmission and distribution upgrades.
A worst-case estimate of coal plant retirements. Even NERC’s achievable level of estimated coal power plant retirements is unrealistically high, driven primarily by the arbitrary limit the model placed on the use of the renewable energy and energy efficiency Building Blocks, as described above. NERC’s model was therefore forced to rely almost exclusively on shifting generation from coal to gas generation.
However, it should be kept in mind that nothing in the CPP dictates that any coal power plant must retire, so any estimate of coal plant retirements is at best guesswork. Unlike capacity-focused environmental rules that specify pollution-reducing upgrades that are required at power plants, the Clean Power Plan is an energy-focused rule that will be met almost entirely through changes in how frequently emitting power plants operate. The Clean Power Plan only requires states to meet a certain emissions target, and states are provided total freedom to decide how to reach that target.
In many cases, the optimal compliance portfolio for a state or utility is likely to include paying the relatively small fixed cost to keep a coal plant operational so that it can run for a limited number of hours to meet local or regional reliability needs. Doing so will have minimal impact on Clean Power Plan compliance, as the emissions associated with operating during a small number of hours are quite small and most of the plant’s generation will be displaced by low-carbon resources.
Speculating about generator retirement decisions is an inherently uncertain exercise, as it involves weighing a number of unit-specific characteristics, local and regional reliability needs, plant owner expectations about future market conditions and policies, and the combined impact of different policies. That speculation is even more uncertain when attempting to assess the retirement impact of environmental rules that do not impose unit-specific restrictions, but rather allow a nearly infinite range of compliance options that can be achieved by changing how power plants are dispatched.
NERC’s estimate ignores the flexibility the Clean Power Plan affords states by assuming compliance is possible through only coal plant retirements. In vertically integrated markets, utilities will not retire capacity needed to ensure reliability. In areas participating in ISO/RTO markets, capacity payments and reliability-must-run contracts likewise ensure that power plants will not retire if reliability is at risk. In both areas, a number of other solutions are available, such as converting coal power plants to burn gas at the same site using the same equipment, or even converting coal power plants to synchronous condensers to provide local voltage support.
Grid operators and utilities are likely to employ long-standing practices to reduce capacity factors in response to market forces and emissions targets without retiring plants that are needed to maintain reliability for some period of time. Today, many coal plants already run at very low capacity factors, in some cases because they are needed for local or regional reliability needs and the fixed cost of keeping such a plant open is relatively low. The current national average coal plant capacity factor is 60 percent, and many plants fall well below that level. If needed for local or regional reliability needs, coal power plants will continue to run at lower capacity factors unless and until grid upgrades are completed or new lower-carbon power sources displace them and meet those needs. Thus, NERC’s prediction for coal plant retirements should be viewed as a worst-case estimate that will not occur if those plants are needed for reliability.
The Clean Power Plan is achievable
Emissions from the nation’s fossil fuel-fired power plants have already fallen 15 percent since 2005, putting the nation halfway towards the 2030 goals of the Clean Power Plan before it has even taken effect. Renewable energy in particular has grown dramatically in recent years, well in excess of the growth rate trajectory EPA envisioned in setting its Clean Power Plan targets. EPA’s Building Block 3 only calls for non-hydro renewable energy to provide 12 percent of U.S. electricity by 2030. The U.S. is already at 6.8 percent non-hydro renewables today, and wind and solar planned for the next two years alone will add several percentage points to that.
As shown below, the U.S. has already reached EPA’s 2020 target for renewable energy. The U.S. produced 281 million MWh of non-hydro renewable energy during 2014, exactly meeting EPA’s 2020 target for non-hydro renewable energy generation. Even under the conservative assumption that non-hydro renewable energy continues to expand at the linear growth rate it has experienced over the last 10 years, renewable generation will exceed EPA’s target by 1.1 billion MWh cumulatively over the 2020-2029 compliance timeframe. Because renewable energy growth compounds as the economics of wind and solar continue to improve, and because the growth trajectory for 2005-2014 predates much of the cost reduction-driven growth in wind and solar generation, this linear growth projection is likely to be very conservative.
By greatly exceeding EPA’s targets, renewable energy can reduce the compliance burden on the other Building Blocks and other parts of the electric sector, addressing any concerns about the feasibility of compliance. The 1.1 billion MWh of surplus renewable energy credits generated with linear growth over the 2020-2029 period will be worth more than 500 million metric tons of emissions reductions, 1/7th of the cumulative 2020-2029 emissions reductions required under the total Clean Power Plan relative to 2012 emissions levels. These surplus emission savings would greatly mitigate any concerns about the ability to reliably and cost-effectively achieve the other EPA Building Blocks.
Despite a number of flawed assumptions, NERC’s report found that the Clean Power Plan can be met with modest changes in the generation mix. These manageable changes, chiefly displacing coal generation with cleaner alternatives, are actually less aggressive than the emission reduction trajectory we are already on for economic reasons. Even though it found modest changes in the generation mix that are achievable, NERC’s analysis is likely to be very pessimistic because it arbitrarily limited energy efficiency and renewable energy growth. NERC’s use of two different analytic models does not shield itself from this criticism because similarly flawed assumptions were used in both models. Had NERC’s analysis allowed states to use the full range of compliance options provided by the Clean Power Plan, the predicted changes in the generation mix would likely have been even more modest and achievable.
NERC’s report ignores technological advances, noted in other NERC reports, that allow modern wind plants to provide grid reliability services as well as or better than conventional generators.
|– Excellent voltage and frequency ride-through per FERC Order 661A requirements– Power electronics electrically separate wind turbine generators from grid disturbances, providing them with much greater ability to remain online through disturbances
|– Many cannot match wind’s capabilities or meet Order 661A ride-through requirements
|Reactive and voltage control
|– Wind turbine power electronics provide reactive and voltage control equivalent to that of conventional generators– Power electronics can provide reactive power and voltage control even when the wind plant is not producing power– Because reactive needs are location-specific on grid, 661A approach of providing reactive in locations where it is needed is more efficient than blanket requirement
|-Conventional generation provides this service.
|Active power control
|– Can provide extremely fast response in seconds, far faster than conventional generation– Like other generators, wind will provide this response when it is economic to do so- Xcel Energy sometimes uses its wind plants to provide some or all of its frequency-responsive automatic generation control
|– Like wind, many baseload generators do not provide active power control for economic reasons, though they technically can
|–Adding wind can help system frequency response by causing conventional generation to be dispatched down-Wind can provide frequency response, but it is typically more costly for it to do so than for other resources as it requires curtailing wind generation in advance-A market-based solution would procure frequency response from the lowest cost resources
|– Changes in conventional generator operating procedures have greatly reduced frequency response– Only 70-75 percent of generators have governors that are capable of sustaining frequency response for more than one minute, and about half of conventional generators have controls that may withdraw sustained frequency response for economic reasons– “Only 30 percent of the units on-line provide primary frequency response. Two-thirds of the units that did respond exhibit withdrawal of primary frequency response.” So, “Only 10 percent of units on-line sustain primary frequency response.”- The cost of providing and sustaining frequency response is very low for a conventional generator, so a market-based solution would incentivize the needed frequency response at low cost
|–Can provide with no lost production by using power electronics and the inertia of the wind turbine rotor; this capability is commercially available but not widely deployed because there is no payment for any resource to provide this service
|-Conventional generation provides this service.
|Increases need for operating reserves, integration cost
|– Very small impact on total reserve need and integration cost
|-Contingency reserve needs and costs are quite large