Pumped Storage Hydropower Supply Curves

Resource potential is often assessed in terms of geographic (or resource), technical, and economic potential—each of which represents a succession of additional complexity and input assumptions that leverage similar data and a common analysis flow.

The closed-loop portion of the PSH resource assessment uses high-resolution digital elevation models (30-meter resolution) to identify potential upper and lower reservoirs within the technology parameters specified by the NREL adaptation of the ANU model. Design specifications include minimum head height of 200 meters and a maximum head height of 750-meters, dam heights of 40, 60, 80, and 100 meters, and a maximum conveyance length between upper and lower reservoir of 12 times the head height. This yields a large set of potential reservoirs with many overlaps.

Once the reservoirs are identified, technical potential criteria are applied to refine the development areas. The criteria eliminate any reservoirs that intersect existing water bodies and waterways; glaciers and ice-covered areas, protected federal lands; urban areas; critical habitat areas; or reservoirs within 1,000 feet of a wetland. Optional criteria can eliminate reservoirs intersecting roads or farmland or allow reservoirs intersecting ephemeral streams.

The HydroLAKES dataset of reservoir locations and characteristics is used to find potential sites that pair new off-river reservoirs with existing reservoirs that could be along river systems—to incorporate existing reservoirs into the full set of potential PSH reservoirs.

To find potential sites that use open-pit mines as reservoirs, NREL researchers search for pit features using 10-meter-resolution digital elevation models, removing any pits smaller than 10,000 square meters in area and shallower than 1 meter in depth. Pits are then identified by searching within 1 kilometer of mine locations in the U.S. Geological Survey mine symbol dataset and visually inspecting the results to remove errors.

After applying technical potential criteria, the remaining upper and lower reservoirs are evaluated for capacity (upper and lower reservoir capacity must be within 10% of each other), the 12× distance criteria, and total paired system cost. Exceptions are made for existing reservoirs, where the size similarity criterion is not enforced, and for mine pits, where head heights as low as 100 meters are allowed. The paired reservoirs are then optimized by cost to develop a non-overlapping reservoir dataset for each combination of assumed storage duration (8, 10, or 12 hours in this dataset) and optional technical potential criteria. Total costs include hydropower site and transmission development. The site development costs are taken from NREL’s bottom-up component-level PSH cost model, except for open-pit mine sites, which are not assigned costs because of high uncertainty. Retrofit costs to use existing reservoirs are also uncertain, so costs for sites using existing reservoirs are approximated by excluding the cost of that reservoir and dam. In addition, transmission spurline costs are taken from NREL analysis supporting the ReEDS model. See ReEDS Model Documentation Version 2020, NREL Technical Report (2021).

Ultimately, this exercise results in a spatially resolved characterization of the technical potential quantity, quality, and cost of PSH resources, which can be sorted to represent a "supply curve" for a specific scenario. The figure below plots the supply curve of closed-loop PSH capital cost in dollars per kilowatt versus cumulative generating capacity in gigawatts for the contiguous United States for 8-, 10-, and 12-hour storage durations and the default assumptions for where to prohibit or exclude closed-loop PSH construction. This supply curve includes both closed-loop sites and sites that use existing reservoirs. Resource and cost data binned by cost ranges are also included in the NREL Annual Technology Baseline beginning in the 2022 data year.

The open-pit mine PSH site assessment, where pits are paired with potential off-river reservoirs, finds 15 candidate locations, as shown on the map below. These are mostly scattered throughout the western United States, with one site in Massachusetts. Many of these sites include active mining operations, so PSH development would compete with existing site use.

The latest (third-generation) dataset includes several updates and changes from the methodology described in the original (first-generation) report. Incorporating stakeholder participation and review, the updated data better reflects the range of technical possibilities for designing closed-loop PSH in the United States and supplements closed-loop opportunities with sites that can use existing reservoirs and open-pit mines as reservoirs. It also integrates updated geospatial data where possible and includes multiple scenarios for storage duration, dam height, and several site exclusion criteria. While some changes limit the range of system specifications to better capture realistic design constraints, multiple dam heights and a lower minimum head lead to many more sites being identified in the second-generation dataset, and expansions in the third-generation data incorporate potential low-cost development opportunities that are not included in the closed-loop data. The other key difference in the third-generation data is that costs are estimated using NREL’s bottom-up component-level cost model, whereas first-and second-generation costs are estimated using ANU’s pumped hydro energy storage cost model with costs adjusted to align with industry expectations published in the 2020 Grid Energy Storage Technology Cost and Performance Assessment.

*Sites utilizing open pit mines as reservoirs have a minimum head of 100 meters.

**The reservoir volume similarity criteria is not enforced when pairing existing reservoirs with potential off-river reservoirs.