top of page

PUSH Report Released

This is the penultimate draft of our report to the Sloan Foundation.

PUSHing for Storage: A Case for Repurposing Decommissioned Mines for Pumped Underground Storage Hydro (PUSH) in the United States. October 2021.

Roman Sidortsov, Shardul Tiwari, Timothy Scarlett, Ana Dyreson, and David Watkins

Technical Report submitted to the Alfred P. Sloan Foundation in response to their grant (G-2018-11305) “Enhancing electrical grid and community resilience through repurposing decommissioned mines into underground pumped storage facilities.”

Final report will be posted to this page when it is ready for public download.

Want a teaser? The report provides lots of gritty details that explain why we can say:

  1. Our study case study in Negaunee, Michigan, showed that the mine could be ultra-long duration storage, providing continuous power to 30,000 people for 3.5 months--at a profit--once it is built.

  2. Based upon how it was designed, if the Mather B was a hydropower station, it would be among the largest such facilities in the world. It is a battery, however, not a generator!

  3. Our team's geospatial analysis shows about 1,000 possible sites for grid-scale PUSH facilities in the United States, with enough energy capacity to meet the projected national needs for storage.

  4. Perhaps more importantly, it is in the public's interest to solve these problems--energy storage, mine reclamation, rural economic development, energy justice-- so we can use energy transition as an opportunity to bring a sustainable economic resources to communities that have been abandoned by mining, as they struggle with reclamation and revitalization.

Download PDF • 5.93MB

Summary of the Findings

This technical report is based on a study funded by the Alfred P. Sloan Foundation “Enhancing electrical grid and community resilience through repurposing decommissioned mines into underground pumped storage facilities.” Between January 2019 and October 2021, we completed a landscaping study that examined the technical, economic, legal, regulatory, water quality, social, and community engagement opportunities and barriers for repurposing decommissioned mines into pumped underground storage hydro (PUSH) facilities. The study focused on Mather B, a decommissioned iron ore mine in Negaunee, Michigan, United States while also extrapolating from those results to consider the applicability of PUSH on a national scale.

PUSH is a type of closed-loop pumped storage hydro (PSH) technology where the upper reservoir is located either at or below the ground’s surface, while the lower reservoir and turbomachinery are built entirely underground. This closed-loop PSH application is capable of providing essential grid services while also mitigating and remediating environmental damage caused by past mining activities and offering sustainable economic development opportunities for post-mining communities. These attributes of PUSH can mitigate the complexities of the licensing and permitting process and improve the economic feasibility of PUSH facilities. As noted in the literature, a PUSH facility can be developed using mature technological systems—materials and machinery—used by conventional PSH. Several studies vetted technical feasibility of PUSH; however, developers should expect site-specific engineering challenges. Dealing with potentially contaminated mine water during the dewatering and operating stages is likely to be one of such challenges. For this reason, we conducted water quality testing, which did not indicate major water quality concerns with potential outflows from the upper portion of the Mather B site.

Community outreach; historic records from local archives, including documents like maps of the surface and underground workings and other company records; as well as oral histories of mining operations and community life all proved to be instrumental in our analysis of Mather B dimensions, structural integrity, soil and water contamination, property rights, design considerations and other issues. Based on our analyses, we made several high- and low-volume reservoir and head estimates, which translated into the following five fully and partially subterranean PUSH designs:


High Volume Estimate (m3)

Low Volume Estimate (m3)

Maximum Head (m)

Scenario 1: Surface pond to Levels 11-12




Scenario 2: Levels 2-4 to Levels 11-12




Scenario 3: Level 6-8 to Levels 11-12




Scenario 4: Surface to levels 7-12




Scenario 5: Shaft only




Using these designs, we calculated the nameplate and energy storage capacities for several scenarios under three models. The first model calculated the maximum installed nameplate capacity with seven-hour discharge time for a partially and completely subterranean PUSH facilities using the high and low volume estimates of the available upper- and lower-reservoirs. The main objective of this model is to demonstrate the potential for a system requiring a high-power output. We do not believe that this model contains realistic options for the case study location because it is very unlikely that both the current market conditions and the transmission and distribution infrastructure will accommodate facilities with these nameplate capacities. The second model simulates daily energy storage scenarios that are based on the site characteristics noted above. This model presents the most realistic design options that can be developed at the Mather B site. The third model extends the second model to explore the potential of PUSH as a solution for long-term energy storage. The third model’s scenarios have the largest storage capacities for a completely and partially subterranean PUSH facility. Although this model is similar to the first model, it appears to be more practical because it is not constrained by the limitations of the first model and it is not likely to require the construction of additional shafts. The maximum power and energy capacities are 295 MW and 1,666 MWh under the daily energy storage model and 73MW and 52,188 MWh for the long-term energy storage model.

Our economic feasibility analysis is built upon capital cost estimates derived from 436 data points for PSH and PUSH facilities published from 31 different sources. As a result of our analysis, we estimate the capital cost of building a Mather B PUSH facility at 1.34 million $/MW. A PUSH facility owner could recover the capital cost through several different pathways (or with a combination thereof). We modeled one of such pathways – participation in a wholesale electricity market. Our analysis used the daily price data for 2018-2020 from the following three the Midcontinent Independent System Operator’s (MISO) nodes closest to Mather B: UPPC.INTEGRATED, a load zone which is UPPC.WARDEN in local resource Zone 2 of MISO and a MICHIGAN.HUB. We calculated the average peak and off-peak price for last three years for both the designated peak period (between 11 pm to 5 am) and actual peak period in each of the target nodes. These calculations yielded average peak value price ratios (PVPRs) 1:4 and 1:6 for the designated and actual peak hour analyses. We then used the five PUSH facility scenarios from the daily energy storage model and determined that the actual peak hour model results in higher revenues for a PUSH facility. This demonstrated the need to conduct a full revenue optimization study for each PUSH facility. We also determined significant increases in profitability with the use of current tax incentives and lower discount rates (DR), from 10% and 8% to 5%. In contrast, reducing the financial lifetime of a PUSH project from 80 to 50 years had an insignificant impact on the profitability of a PUSH facility.

All of the facility scenarios posted losses under PVPRs around 1:4 and 1:6. However, these PVPRs are not reflective of the PVPRs that a PUSH facility operator would use to optimize profitability. Higher PVPRs are currently seen in U.S. energy markets including the MISO and they are projected to go higher with further influx of intermittent renewable energy (RE). For this reason, we tested the impact of 1:3, 1:4, and 1:5 PVPRs on the profitability of two of PUSH facility scenarios in our sensitivity analysis. Both PUSH facilities start operating profitably with the PVPR 1:4 and higher under 10% and 8% DRs. With 5% DR, the facilities becomes profitable with the PVPR 1:3. To complete our economic feasibility analysis, we determined that all our PUSH facility scenarios under the daily and long-term storage models can fully participate in the ancillary services market. As with our estimates of volume and head, we were very cautious and conservative when we extrapolated from existing data points to assess the impacts of financial incentives, revenue sources, and operational lifespan projections.

We used the data from the United States Geological Survey’s (USGS) Mineral Resource Database System (MRDS) to locate nationwide mine sites suitable for PUSH. The MRDS includes data on hundreds of thousands of historic and currently operating mines. We filtered the results based on several criteria and limited the analysis only to metallic mines and those mine facilities with descriptions that would allow adaptive reuse. The filtering identified 968 mines that could potentially host PUSH facilities across 15 different states. As a result of this analysis, we determined that despite being the single most comprehensive database of mining in the United States, the MRDS is missing all or some information about potentially important mines such as depth and size.

Keeping these data shortcomings in mind, Western United States appears to be a particularly attractive region for PUSH development because of the abundance of potential sites, plentiful wind and solar resources, existing intermittent RE facilities, population density, and load centers. The Upper Midwest is another attractive region for potential PUSH development, where most sites are concentrated in Michigan’s Upper Peninsula. Based on the criteria used to evaluate potential PUSH sites, with the exception of New York state, Eastern Seaboard states do not have decommissioned or active metallic mines that could be adapted for this purpose.

Combining the geospatial and techno-economic analyses, we determined that the United States has 285 GW and 137 GW of maximum cumulative national power capacities for partially underground and fully underground facilities of PUSH for daily storage. These capacities exceed the National Renewable Energy Lab (NREL) projection of 120 GW of energy storage needed to achieve a generation mix consisting of 80% renewable resources by 2050. The cumulative maximum storage capacities are 564,441 GWh for partially and 271,040 GWh for fully underground facilities. These capacities are promising because they equal to 14 % and 7% of all electricity generated in the United States in 2020. We further determined that PUSH could also be a long-term (seasonal) storage solution with the maximum cumulative power and energy capacities of 8.7 GW and 8,010 GWh based four pumping/discharge cycles a year.

The study in this report represents the most detailed consideration of PUSH application yet conducted in the United States. Our report uses existing data to build very cautious and reasoned estimates at each analytical stage. The resulting case study, when considered on the national scale, demonstrates that PUSH could be a transformative technological application in the United States, where investments required to solve problems for the national energy system could simultaneously solve problems for post mining communities throughout the country.


bottom of page