State-wise Economic Viability of Long-Duration Energy Storage Systems in the United States

TL;DR

This study develops a state-by-state framework to assess the economic viability of 100-hour long-duration energy storage (LDES) in replacing conventional firm generators across the contiguous United States. It uses a two-stage optimization (baseline and opportunity-value maximization) on 2050 energy matrices from Cambium, ReEDS, ATB, and AEO data to compute state-specific viability costs, defined as the maximum avoided cost per unit of LDES capacity. The results show a wide range of viability costs (including negative values) and indicate that under the DOE's $1{,}100$/kW target only four states are viable for 100-h LDES, with broader viability emerging under lower targets (nine under $500$/kW and seventeen under $300$/kW). The analysis also quantifies substantial capacity requirements for replacing firm generation (e.g., ~$646$ GW minimum for 43 positive-viability states, rising to ~$1{,}009$ GW when maximizing viability costs), underscoring state-specific system characteristics and large-scale deployment needs for future grid transition pathways.

Abstract

Long-duration energy storage (LDES) assets can be fundamental resources for the next-generation power systems. However, LDES technologies are still immature and their future technology costs remain highly uncertain. In this context, we perform in this paper an extensive study to estimate the maximum LDES technology costs (which we define as viability costs) under which LDES systems would be economically viable in each state of the contiguous U.S. according to their characteristics. Our results indicate that only 4 states (out of 48) would be able to remove firm conventional generation supported by LDES systems without increasing their total system costs under the current US-DOE cost target of 1,100 US$/kW for multi-day LDES. In addition, we find that states with the highest LDES viability costs have in general low participation of thermal generation, a high share of wind generation, and higher thermal-related fixed operation and maintenance (FO&M) costs.

Figures (7)

• Figure 1: LDES maximum viability cost and power capacity distribution across U.S. states in 2050. (a) Maximum viability cost of 100-hour LDES ($/kW) for each U.S. state, considering the power capacity that maximizes the viability cost. (b) Histogram of the LDES maximum viability cost ($/kW) across states with positive values. (c) Histogram of the LDES power capacity (GW) that maximizes the viability cost in states with positive values.
• Figure 2: The viability costs of LDES below which these technologies will be economically viable for the system in 2050, for: (a) Colorado, (b) Florida, (c) Iowa, and (d) Michigan.
• Figure 3: (a) U.S. total system capacity by technology in 2050: baseline model vs. opportunity value model with and without LDES. The first bar represents the total system capacity when considering gas and coal power plants with no inclusion of LDES. The second bar shows the scenario where gas and coal power plants are replaced by a combination of LDES, SDES, and intermittent energy sources, with LDES capacity set to the amount that maximizes the viability cost. The third bar represents a scenario where gas and coal power plants are also replaced by SDES and intermittent energy sources, but without the inclusion of LDES. The three bars on the right show the thermal replacement. The first bar indicates the amount of gas and coal capacity in the baseline, while the next two show the net increase in other technologies required to replace the output previously provided by gas and coal. (b) Same analysis as in (a), but for cost values instead of capacity. Both figures include data from 43 states, excluding values those where the maximum viability cost is negative.
• Figure 4: Box plots comparing key system metrics for U.S. states categorized by viability costs. Figure (a) compares the top 10 and bottom 10 states in terms of viability cost, and figure (b) extends to the top 24 and bottom 24, encompassing all states. Metrics include thermal generation participation, thermal capacity utilization, intermittent capacity factors, thermal FO&M costs, and solar and wind generation in the baseline model. The right-side panel presents the distribution of the $\alpha$ ratio for each case.
• Figure 5: Comparison of seasonal SoC differences: (a) Spring and (b) Winter. During spring, the colors are more intense, indicating higher charging levels, whereas in winter, the colors are lighter, reflecting greater discharging.