The Department of Energy (DOE) has varying degrees of responsibility over the sites identified for study by the National Renewable Energy Laboratory (NREL). Many are in active use (notably the 17 national laboratories), while others are not in active use. Many of these inactive sites are under DOE jurisdiction either for remediation of contamination or for continued monitoring post-remediation, while others have more complicated ownership issues and are deemed not under DOE control.
The available information about the DOE properties comprised critical input into the NREL and Colorado School of Mines (CSM) analyses. Some of the site characterizations were more descriptive than others based on the ready availability of information. For each site, DOE provided a location point in latitude and longitude, and the site acreage, along with limited details about the site ownership and jurisdiction within DOE. For many of the 55 sites that were analyzed in greater depth, some additional site-specific characteristics were provided to inform the analysis. However, no comprehensive information about the site topography or geography was obtained by NREL to inform the analysis. In the absence of this information, NREL and CSM assumed in their analysis of techno-economic potential that the sites were circles with area equal to the site acreage, centered at the latitude and longitude point. The resolution of the resource maps varied. The properties were also not fully characterized for infrastructure needs to develop energy resources on-site for on-site use or for sale to regional wholesale electricity markets or to local utilities.
DOE identified 148 sites for study by NREL.1 From that list, they performed an initial screening on the feasibility of adding an energy project at the site. The preliminary assessment considered land ownership, available acreage, and compatibility with existing uses. Subsequently, DOE reduced the list to 55 sites for further study. A limited further analysis was then conducted on those sites among the 55 that they considered the most promising. A review of the analysis is the subject of this chapter.
NREL and CSM were each funded with $150,000 for analyses of renewable energy resource development potential and fossil energy and uranium resource development potential, respectively. An additional $50,000 was given NREL for project management and report compilation.
The committee appreciates the fact that NREL and CSM took on a difficult project with very limited and, frankly, insufficient funding. The limited funding available constrained the methodology and approach used to frame the analysis, imposing limitations on the rigor of the analysis and subsequent results and conclusions. The committee undertook its review of the analysis with knowledge of the constraints imposed by the limited funding. The committee identifies some of the more salient shortcomings of the methodology and approach resulting from this limitation throughout this chapter and
1 Alicen Kandt, “DOE Large-Scale Power Production Study: May Meeting with NAS,” presentation to the Committee, May 20, 2015.
again in its findings and conclusions. Other methods and approaches to the analysis, suggested by the committee and referenced in this chapter, might have reasonably led to different conclusions. Engaging DOE site personnel early in the study effort to collect data and information to more completely characterize a site and its surrounding infrastructure, as well as interviews with potential developers to shed light on site-specific considerations and costs for energy resource development, might have better informed the analysis.
NREL examined a number of technologies for generation of renewable electricity for export off-site. The analysis used the NREL Renewable Energy Optimization (REopt) tool to calculate a levelized cost of electricity (LCOE) for each resource, other than geothermal, on each of the 55 sites. Data inputs to the model included site area and latitude and longitude, renewable energy resource data available from other studies or estimated for immediate purposes, technology effectiveness and cost assumptions, available transmission capacity excluding costs, and relevant government policies and availability of financial incentives for renewable resource development. The REopt tool uses the input data to determine system size, system cost, electricity output, and operations and maintenance cost, which are then amortized over the project life to calculate the LCOE. This provided an initial estimate of techno-economic potential of the resource (Kandt et al., 2016, p. 6). The inputs to and assumptions embedded within the model are further described in Appendix C.
The REopt tool by its nature ignores other site-specific characteristics that can expand or limit energy resource development potential. LCOE as defined and used in the analysis is a necessary input to determining techno-economic potential, but as characterized by NREL, it is not a sufficient criterion for selecting resources or technologies to site on a particular parcel of land. The committee suggests that using LCOE as the primary criterion for screening or down-selecting technologies for future analysis is misleading and limiting. Many other widely used methods are available, including the following: more rigorous site analysis (see, for example, Walker, 2009); discounted cash flow analysis (see, for example, Brown and Campbell, 2003); and utility-scale integrated resource planning and dispatch models (see, for example, Mai, 2013). Supplementing REopt results with qualitative analysis for screening might well have led to different technologies being selected and at different sites. A narrow focus on techno-economic potential ignores the practical realities of actually developing energy resources on a particular site
The output of the REopt model was reported for each resource on each site, resulting in creation of a database with 243 possible projects, consisting of 55 photovoltaic (PV), 20 concentrating solar power (CSP), 54 wind, 8 land-fill gas, 54 waste-to-energy (WTE), and 52 biomass projects. The database of REopt outputs included individual project’s projected electric generation capacity, system installed cost, annual operations and maintenance costs, land area required, and the LCOE. The database created is included in the NREL/CSM report (Kandt et al., 2016).
FINDING 2.1. The database created for the NREL/CSM analysis, as a repository of REopt output, has limited utility owing to the construct of the analysis and overreliance on LCOE for screening. As such, the committee finds the database to be of limited use for identifying specific projects at specific sites for development. The information may be of interest to potential energy project developers who may be interested in studying DOE properties further for energy resource development.
Following the initial LCOE screening, a more thorough analysis was conducted for certain sites identified by NREL as having higher priority. Seventeen projects were chosen (Table 2-1), including the
TABLE 2-1 Overview of the Possible Projects Examined in the Barriers and Opportunities Analysis
|Technology||Site Name||Area (acres)||Resource Available (MW)||System Capacity (MW)||Levelized Cost of Electricity ($/MWh)|
|Photovoltaics||Nevada National Security Site||775,680||Unlimited||100.0||$82|
|Los Alamos National Laboratory||28,000||Unlimited||62.8||$82|
|Shirley Basin South, Wyoming, Disposal Site||1,527||Unlimited||50.9||$46|
|Biomass||Separations Process Research Unit||200||82.1||82.1||$91|
|Fermi National Accelerator Laboratory||6,811||187.1||100.0||$97|
|Landfill gas||Grand Junction||360||6.8||6.8||$81|
|National Energy Technology Laboratory (Pennsylvania)||63||2.5||2.5||$86|
|Kansas City Plant (Bannister Road)||120||2.5||2.5||$91|
|Waste to energy||Bonneville Power Administration Ross Complex||250||138.3||100.0||âˆ’$25|
|Argonne National Laboratory||1,700||487.9||100.0||$5|
|Concentrating solar power||Nevada National Security Site||775,680||Unlimited||50.0||$200|
|Los Alamos National Laboratory||28,000||Unlimited||50.0||$210|
|Geothermal||Shoal, Nevada, Site||2,560||-||-||-|
|Lakeview, Oregon, Disposal Site||40||-||-||-|
|Nevada National Security Site||775,680||-||-||-|
|Central Nevada Test Area, Nevada, Site||2,560||-||-||-|
two or three projects with lowest LCOE for each resource.2 This was done so that each type of resource would be represented in the more detailed analysis. Unfortunately, this procedure may have eliminated from consideration favorable projects that might otherwise have been “next in line” had the top-two sites (lowest LCOE) for that resource been disqualified due to the other factors (e.g., on-site land constraints, developer risk, etc.). Such sites may have lacked favorable LCOE and would have already been dropped in the first screening phase. (Geothermal energy is an exception, and the analysis used the size of the resource as the screen for further evaluation.)
The analysis also constrained the combination of projects on any one site to some extent. For example, developing solar PV and wind at the same site was not considered, nor was limiting the development potential to the lowest LCOE resource at a given site. A resource with a lower LCOE might have been excluded from development at a given site if that type of resource was already being considered at another site for the more detailed screening. NREL included geothermal projects on four sites that were identified as top opportunities for further analysis among the 17 listed in Table 2-1. These
17 projects on the 14 sites were then examined in a market barriers and opportunities analysis framework that took into account the following: possible barriers and opportunities in site ownership and control, off-takers for the power generated, permitting, and first-order economics. The resulting estimates of renewable energy potential estimates represent approximations of technical, economic, and achievable potential simply defined for the 17 projects examined under the barriers and opportunities analysis.
FINDING 2.2. While LCOE can be a useful screening metric, it is just one of several metrics that should be used for identifying a list of sites for energy development. Many other widely used methods are available, including the following: more rigorous site analysis, discounted cash flow analysis, and utility-scale integrated resource planning and dispatch models. The use of LCOE as the sole criterion most likely led to a different list of top sites for renewable resource development than would have resulted if the sites were more completely characterized and screened using more relevant criteria, such as those used in the market opportunities analyses.
The input assumptions for the LCOE analysis strongly affect the ranking. The results could have been very different, had the study considered the following: the cost of environmental remediation in order to develop sites safely, costs of bringing electricity to market (e.g., cost of transmission and necessary distribution system upgrades), and a fuller suite of complementary technologies (e.g., renewable energy with storage on-site).
The information resulting from the analysis in the NREL report is insufficiently robust to completely address the goal of the overall project to identify energy resource potential on DOE sites. Thus, to emphasize what was previously stated, LCOE should be only one of several criteria in site screening and should not be used as the primary means to rank sites for development. Some consideration of other data—essentially the value of information (i.e., the value to an investor of any information that would improve his/her decision whether to make a potential investment)—should have been taken. As the committee discussed with potential developers, these other sources of information were more appropriate for decision-making.
FINDING 2.3. Rankings presented in the NREL report are not a firm indication of the attractiveness of any particular resource or site for development. Even sites that might have the best overall economics might be too cumbersome to develop, due to restrictions related to national security or environmental issues, and/or might present risks that the market would not be willing to bear.
The REopt model, while useful for some purposes, is not well suited to be the primary screening mechanism for potential sites for renewable energy development and use. NREL presented the results of some sensitivity analyses it conducted using REopt, varying certain key assumptions to assess the impact on LCOE, but the committee believes that, given the fundamental misfit of the model to purpose of the study, the results were not particularly insightful and did not change the results of the screening analysis. In addition to being potentially inaccurate, the model results were reported to at least three significant figures, making the analysis appear highly precise when an order-of-magnitude calculation might have been more appropriate for ranking.
The sensitivity analysis returned a range of LCOE for the candidate projects. Considering such ranges earlier in the screening process might have revealed some interesting insights that might then have influenced the down-selection of resource types at a given site for further analysis. Important data and information, such as more specific site characterizations and proximity of a site to surrounding energy infrastructure and to load centers, even if available, would not have fit easily into the REopt tool. The committee believes that the inability of the tool to accept more complete site-specific data is a significant flaw in the study design. The committee believes these omissions resulted in an inaccurate assessment of a site’s energy resource potential.
In addition, ignoring real-world concerns of developers in screening and selecting resources and sites for development could lead to very different conclusions. A potentially fatal shortcoming of the analysis was down-selecting resources and sites prior to soliciting input from potential developers. Such input could have provided guidance in determining the most appropriate approach and method for developing screening criteria—without total reliance on one model that focused on only a single criterion. Further, a number of DOE sites are actually developing renewable resources on their sites. Early discussions with these sites on criteria that they were using to make project decisions would have also better informed the project participants. The committee could not establish whether the DOE sites that had already developed such resources would have been selected using LCOE as the sole criterion for resource and site selection.
FINDING 2.4. Interviews with potential developers may have helped NREL and CSM analysts to view the sites through the practical lens of the market.
The budget limitations were significant for CSM, as they were for NREL. CSM’s analysis, however, did not rely solely on the use of a modeling framework. Instead, CSM analysts examined DOE sites and compared their locations to maps of natural gas, oil, and uranium resources obtained primarily from U.S. Geological Survey (USGS) sources. Sites’ proximities to resource claims or active resource development were used as a criterion to rank sites more highly than those without nearby activity. This assumption rests on the availability of nearby infrastructure for energy development on said sites. No comprehensive quantification of available resources was undertaken; rather, a list of top sites was identified on the basis of proximity to actual resource activities. For its analysis, CSM did not screen all 148 DOE sites originally furnished to NREL; rather, it started with the 55 sites identified by DOE following NREL’s initial screen. The potential for development of fossil resources on the 93 sites that were discarded was not evaluated.
CSM’s analysis was not a technical assessment of the resource potential. A more rigorous and comprehensive analysis was outside the scope of work, given the limited funding of the study. Also, despite coal being a resource specifically addressed in the committee’s statement of task, it was not addressed in CSM’s statement of work, due in part to limited funding and time to conduct the analysis. This omission is noted further in the committee’s evaluation of the coal resource.
The lack of resource quantification is in contrast to the NREL study of renewable resources. The resource maps that were displayed in the CSM presentation were useful, but included nothing more than sets of information that could be obtained rather easily by others and, thus, provided no study-specific information. In fact, a similar approach was used by a geologist from USGS who appeared before the committee to provide the committee with a list of possible sites for coal resource development.3
FINDING 2.5. The analysis by CSM, while very superficial and based on regional resource maps rather than site characterizations, does provide a sense of which locations of the 55 that were analyzed might hold potential for oil, gas, and uranium resource development.
The committee believes that the basic design of the NREL/CSM study was flawed. Early in the process, there were missed opportunities to learn about DOE or other federal sites that already have developed energy resources and to engage with the energy industry development community. While REopt may be a useful tool for estimating LCOE and economic potential, it does not appear suitable for
3 Peter Warwick, U.S. Geological Survey, “Review of Coal and Geologic Carbon Dioxide Storage Resources Underlying DOE Lands,” presentation to the committee, May 21, 2015.
this analysis. And, without a full accounting of costs (e.g., site remediation, infrastructure needs to deliver the resource, and others), LCOE would not even accurately reflect the true cost of development.
Developers believed there to be other issues of greater importance that were not considered in the NREL/CSM analysis. These would include the need for environmental impact statements (as required for development on federal lands) and interconnection proximity. The committee believes that a more thorough and thoughtful analysis of the sites might conclude that substantive environmental and interconnection issues would eliminate all of the top sites identified for development. Projects not included in the top 17 (Table 2-1) might be more appropriate for development if they could better meet environmental, interconnection, and other requirements while still offering an attractive LCOE. Since a resource’s LCOE must be compared against wholesale or marginal electricity prices to determine the economic viability of development, it is not possible to discern with any confidence that a resource would ever be developed at a particular site, even if technical potential exists.
FINDING 2.6. Developers identified environmental impact statements and proximity to infrastructure, such as electricity interconnects, as key criteria for suitability for energy resource development.
RECOMMENDATION 2.1. The identification of top sites for energy resource development should be guided by the most relevant criteria for development, as identified by developers or other practitioners, such as need for environmental studies and proximity to grid infrastructure.
Early interaction with the DOE sites themselves—those that have already developed renewable energy projects on their property—could have provided information very relevant to this study as well. For example, the committee heard from Pantex,4 Brookhaven National Laboratory (BNL),5 Lawrence Livermore National Laboratory,6 and the Savannah River Site7 about their activities in developing renewable resource technologies on their sites. With the exception of BNL, all of these projects were for the production of electricity for on-site use. In the case of BNL, this effort involved a power purchase agreement (PPA) with Long Island Power Authority (LIPA) for the sale of power from the solar project at BNL. The project was bid through a LIPA request for proposals and competitively awarded a PPA. Since any DOE site development for interconnection to external grids would require a PPA, this case study information could have been useful to the NREL analysts in developing their analytical approach and study design.
FINDING 2.7. Case studies of energy resource development already under way at DOE properties could have better informed the NREL analysis.
The committee suspects that assumptions used in the LCOE analysis might lead to incorrect conclusions regarding the technology types selected for detailed analysis. For example, in earlier versions of the analysis reported to the committee, NREL found solar PV technology to be best suited for development based on LCOE, but later analyses pointed to waste-to-energy technology as being the most cost effective from a LCOE perspective. This result may have been driven by assumptions of extremely generous tipping fees, coupled with a disregard of competition from the waste disposal industry. Clearly,
4 Mark Padilla, Nuclear National Security Administration, and Kevin Long, Consolidated Nuclear Security, LLC, “Pantex Renewable Energy Project,” presentation to the committee, November 13, 2014.
5 Patrick Looney, Brookhaven National Laboratory, “Long Island Solar Farm,” presentation to the committee, November 13, 2014.
6 Michael Brown, U.S. Department of Energy, “Livermore Site Solar Project,” presentation to the committee, November 13, 2014.
7 James DeMass, U.S. Department of Energy, “Biomass Cogeneration Facility: Savannah River Site; Aiken, SC,” presentation to the committee, November 13, 2014.
changes in the input assumptions led to very different rankings over the course of the project, and changes in either or both of these waste disposal assumptions could remove all of the waste-to-energy sites from the top list. In addition, other factors, such as water availability, need for additional site remediation, and electricity commodity prices, do not appear to have been taken into account.
Some of the NREL assumptions limited the efficacy of analyzing certain renewable energy technologies through the use of REopt, possibly providing misleading results. For example, the analysis limited all renewable energy projects to a generation capacity of 100 megawatts (MW) (see the section, “Comments on NREL Analysis for Wind,” in Chapter 0). Since many sites have very large land areas, siting wind projects of larger capacity might be very feasible.
Almost all of the CSP projects under development in the southwestern part of the United States include thermal energy storage (TES) to allow efficient use of their thermal generation capacity for continued production of electricity during peak evening usage periods. The information in the section of the NREL/CSM report concerning technology characterization needs to be clearer concerning storage.
The NREL study used reasonably accurate information as to solar prices at the time of the analysis. However, rapidly reducing prices of renewable technologies, particularly for photovoltaics, should have been incorporated into the LCOE model as they became available.
In summary, the committee believes that the NREL analysis provides a somewhat useful database of information about renewable energy resource potential on DOE properties. However, the committee is not confident that the analysis has identified or can identify the most attractive renewable energy development prospects based on the data, assumptions, and approach used to conduct the analysis. The CSM analysis provided some information on top prospective sites for oil, gas, and uranium extraction. In the case of uranium, however, the DOE Office of Legacy Management was not able to provide the committee with information on the Uranium Leasing Program and their management of these additional locations in the United States.
RECOMMENDATION 2.2. To provide a more useful ranking of sites for developers, the Department of Energy (DOE) should adopt a more robust approach featuring early outreach to developers, use of screening criteria other than levelized cost of electricity, and use of lessons learned at other DOE and similar federal sites.
Brown, H.F., and R.P.C. Campbell. 2003. Benefit-Cost Analysis: Financial and Economic Appraisal Using Spreadsheets. Cambridge, U.K.: Cambridge University Press.
Kandt, A., E. Elgqvist, D. Gagne, M. Hillesheim, A. Walker, J. King, J. Boak, J. Washington, and C. Sharp. 2016. Large-Scale Power Production Potential on U.S. Department of Energy Lands. Technical Report NREL/TP-7A40-64355. Golden, Colo.: National Renewable Energy Laboratory. June.
Mai, T., E. Drury, K. Eurek, N. Bodington, A. Lopez, and A. Perry. 2013. Resource Planning Model: An Integrated Resource Planning and Dispatch Tool for Regional Electric Systems. NREL/TP-6A20-56723. Golden, Colo.: National Renewable Energy Laboratory.
Walker, A. 2009. Renewable Energy Analysis for Department of Energy Savannah River Site. Golden, Colo.: National Renewable Energy Laboratory.