This is the sixth in a series of questions that highlights RFF’s Expert Forum on EPA’s Clean Power Plan.
RFF asks the experts: how can coal power plants reduce emissions and be made more efficient—and at what cost (building block #1)?
The emissions rate targets assigned to states are built on four building blocks that EPA says represent best practice in reducing emissions throughout the electricity system. Building block #1 focuses on measures to make coal plants more efficient. EPA seeks comment on its proposed finding that the average heat rate for coal power plants (the heat content of fuel input per unit of electricity output) could be improved by 6 percent on average across the fleet. Do such opportunities exist, and at what cost? Would they lead to emissions reductions?
"The evidence suggests EPA’s finding is technically plausible and economically reasonable—given that the agency adopts a flexible approach to achieve compliance."
—Dallas Burtraw, Darius Gaskins Senior Fellow, Resources for the Future (See full response.)
"EPA’s conclusion that 4 percent improvement using “best practices” is possible at no or low cost was based upon an analysis [that]… did not consider any technical aspects of the facility, or if any characteristics changed over the 11 year period, including a change in operating mode… whether the facility conducted a retrofit …or if the facility changed coals."
—Anthony Licata, Partner, Licata Energy (See full response.)
Dallas Burtraw
Darius Gaskins Senior Fellow, Resources for the Future
One of the ways to reduce emissions as part of EPA’s proposed Clean Power Plan is to improve the operation of coal-fired power plants. EPA finds that the heat rate—the heat content of fuel input per unit of electricity output—at existing plants can be reduced by 6 percent on average across the fleet. This would correspond to an equivalent reduction in emissions rate. Is this goal attainable?
Empirical evidence from research at RFF suggests the following:
- EPA’s goal is technically plausible.
- The costs could be low, at least for modest heat rate improvements of 1 to 2 percent.
- However, the cost of a 6 percent improvement in heat rate (while holding utilization constant) could be significant at many plants.
- The flexibility of this proposal to average across plants and to shift generation to more efficient plants means that greater opportunities may be available at lower cost, compared to when plants are examined on an individual basis.
My colleagues and I examined data on the performance of coal plants in 2008, sorting plants by boiler type and other characteristics. Our results indicate that the average heat rate would be reduced by 5.5 percent—without changing the utilization of individual plants—if each plant improved to match the operating performance of the best 10 percent in its group. This indicates that EPA’s goal is technically plausible.
In its technical documents, EPA examines the operation of plants over many years and finds substantial variation in heat rate at individual plants on an hourly basis. This suggests that improvements are possible if that variation could be reduced. EPA suggests that 4 percent improvement could be achieved on average through changes in maintenance and operation at individual plants at low cost or no cost, and another 2 percent could be achieved through capital investments at individual plants.
We examined 25 years of operation of coal plants to see how their heat rates responded to changes in relative fuel prices while holding constant the utilization of the plants. Our findings support EPA’s claim that costs are low, at least for modest heat rate improvements of 1 to 2 percent. Larger changes are outside the range of variation observed very frequently, and it is difficult to extrapolate, but clearly the marginal costs would increase with increasingly stringent improvements.
If poor heat rates waste fuel, why would such opportunities for heat rate improvements exist across the fleet? There are multiple possible explanations, at least in the short run. These include automatic fuel adjustment clauses in some states that may reduce the incentive to save fuel costs, liquidity constraints in some systems that may make it difficult to fund investments, and the so-called new source review test—which is the threshold where investments to improve a plant’s operation could enable its greater utilization and thereby increased emissions of other pollutants. This test could trigger a requirement for tens of millions of dollars in expenditures to improve controls for other pollutants, a cost that may deter investments to improve heat rates.
Nonetheless, opportunities to reduce costs should not go unrealized in the long run. While the results from Linn et al. suggest the cost of a 6 percent improvement in heat rate (while holding utilization constant) could be significant at many plants, a central characteristic of the Clean Power Plan is that stringency is built on EPA’s findings of what is technically possible. EPA does not require specific measures to be taken. That is, a heat improvement of 6 percent on average in the operation of the coal fleet does not require a 6 percent improvement at any given plant. One way this is relevant is that large capital investments, which EPA suggests could lead to a 2 percent improvement in heat rate on average across the fleet, may result in improvements of 10 percent at an individual plant. For example, investments such as major overhaul of a plant’s turbine are expensive and would only be observed at some plants—but they would lead to substantial reductions in heat rate. The flexibility of this proposal to average across plants and to shift generation to more efficient plants means that greater opportunities may be available at lower cost, compared to when plants are examined on an individual basis.
The system-level costs of a flexible approach to achieve heat rate improvements have been estimated in a couple of papers that incorporate the engineering opportunities expected to be available at individual plants, with the opportunity to change operation of the system to achieve further improvements on average. These papers (Burtraw, Woerman, and Paul 2012; Burtraw and Woerman 2013) examined a set of investments that were less substantial than a major overhaul of the turbine, and identified that the marginal cost of emissions reductions achieved through an average heat rate improvement of 4 to 5 percent in the coal fleet would cost about $10 to $30 per ton of carbon dioxide reduced. This is somewhat greater than the cost suggested by EPA but still reasonable when measured against the social cost of carbon emissions, as estimated by the Interagency Working Group. These reductions are measured against a future baseline for 2020; measured against the heat rates in 2012, the reduction would be nearly another percent according to the modeling, approaching the 6 percent target identified in building block #1.
In conclusion, the evidence suggests EPA’s finding that a 6 percent reduction in heat rate from 2012 levels is technically plausible and economically reasonable—given that the agency adopts a flexible approach to achieve compliance.
Anthony Licata
Partner, Licata Energy
EPA’s proposal for building block #1 of the Clean Power Plant Rule (CPP) is based on improvements in the heat rate at existing coal-fired power plants. EPA concluded that up to a 6 percent heat rate improvement could be obtained at an average cost of $100 per kW. EPA stated that 4 percent of the heat rate improvement, on average, was achievable through no- or low-cost options it refers to as “best practices” and the remaining 2 percent improvement was achievable with some capital expenditure.
EPA projects 88,000 megawatts of installed coal capacity will be removed from service by 2020, largely from smaller, older, and less-efficient plants. Opportunities for the greatest heat rate improvements would come from these same less-efficient plants, a number of which may be able to achieve a 3 to 6 percent heat rate improvement, even with cost being considered. Most of the remaining plants are “flagship” units—larger capacity and supercritical plants—already have upgrades that may limit heat rate improvements to about 1 percent, meaning that other remaining plants may have to achieve upgrades greater than 10 percent, which is not practical.
There are concerns about EPA’s analysis. EPA’s conclusion that a 4 percent improvement using “best practices” is possible at no or low cost was based upon an analysis of more than 800 units, using what EPA describes as statistical process control methods. For each unit, the gross heat rate in a given hour over an 11-year period was compared against hourly ambient temperature data and hourly average load as a percent of maximum load. EPA calls this latter variable a “capacity factor”, although others consider a capacity factor as averaged over a much longer time. The degree to which the heat rate for the unit is consistent or inconsistent over the 11-year period for any given temperature and hourly average load is, according to EPA, a measure of how well or how poorly the facility is being controlled. A well-controlled unit should, EPA argues, have a very consistent heat rate under a given ambient temperature and average hourly load. To the extent that there is scatter in gross heat rate data for any given hourly load or temperature condition over this 11 year period, EPA considers that a sign that more consistent process control can improve heat rate. It appears that EPA meant that plants could be better controlled with neural networks. Many plants already have a neural network control system. An upgrade to a plant without such a system may see a heat rate improvement of 0.75 to 1.5 percent.
However, the statistical approach EPA used did not consider any technical aspects of the facility, or if any characteristics changed over the 11-year period, including a change in operating mode (switching from base load to cycling), whether the facility conducted a retrofit (installing low nitrous oxides burners, selective catalytic reduction installations, scrubbers, new burner or furnace management systems), or if the facility changed coals or any other characteristics of the fuel. EPA also did not compare units against one another or do any sort of subcategorization. Unfortunately, because there are many more factors that impact heat rate that are beyond the operator’s control than just ambient temperature and average hourly load, and because EPA did not factor changes to the plant over that 11-year period, EPA’s statistical approach likely mistook the effects of these other parameters and changes to the plant as indicators of poor process control.
Moreover, most power plants are already equipped with statistical process control systems that monitor thousands of plant parameters and are designed to optimize operation of the plant. EPA’s analysis that only looked at two parameters is much less reliable than the advanced process optimizers that are already installed at the plants.
Finally, there are few technological opportunities to achieve heat rate improvements at low or no cost. Other than tuning combustion systems and patching leaks in ductwork, low or no cost technologies would provide minimal heat rate improvements. All other technologies that would improve heat rate require capital investment. A paper published in July 2013 by members of the American Society of Mechanical Engineers finds that a 500 megawatt wall-fired coal boiler could achieve a 0.34 percent improvement in boiler efficiency by upgrading its fuel delivery system at a cost of nearly $14 million. A major steam turbine upgrade and rebuild can cost $30 million to $40 million. There appears to be significant cost of heat rate improvements that EPA did not consider in its evaluations of no- or low-cost options.
