What We Publish

Stylized least-cost analysis of flexible nuclear power in deeply decarbonized electricity systems considering wind and solar resources worldwide

Lei Duan, Robert Petroski, Lowell Wood, and Ken Caldeira · Nature Energy 7, 260-269 · 2022

Key finding. In least-cost greenfield systems built on 42 countries' hourly weather and demand, the emissions-reduction level at which flexible nuclear enters is set mainly by the annual mean wind capacity factor — countries near 0.4 need very deep decarbonization before nuclear competes, while those at or below 0.25 take it early — and lowering nuclear capital cost from US$6,317 to US$4,000 per kilowatt-electric cuts the cost of a 100 per cent emissions-reduction system by 15 to 25 per cent in every country studied.

Four world maps in two rows under a 99 per cent emissions-reduction scenario. The top row is nuclear at the near-current cost of US$6,317 per kilowatt-electric and the bottom row at US$4,000; the left column shows optimized nuclear reactor capacity in kilowatts per person on an orange scale from 0 to 1.5, and the right column shows storage in hours of mean demand on a purple scale over the same range. Regions outside the 42 studied are stippled. At the higher nuclear cost most countries take little nuclear and a lot of storage; at the lower cost nuclear capacity is high almost everywhere and storage correspondingly low.
Lowering the capital cost of nuclear turns storage into generation. At the near-current cost (top row) least-cost systems lean on storage; at US$4,000 per kilowatt-electric (bottom row) they build nuclear across nearly every country and need far less storage to stay reliable. Figure 5 from Duan et al. (2022), Nature Energy 7, 260-269. Reproduced under author reuse rights. Extracted from the published PDF and resized for web display.

What question did this research address?

Solar and wind are now cheap, but systems relying on them get expensive as emissions constraints tighten past about 80 per cent, because filling multi-day gaps requires overbuilding capacity that then goes largely unused. Firm low-carbon generation could relieve that, and nuclear is the main candidate.

Proposed advanced reactor designs claim both lower capital cost and, through molten salt thermal energy storage placed between reactor and generator, the ability to follow load rather than run flat out. Whether such a plant would complement variable renewables or displace them was unclear.

This paper asked, for 42 country-level regions with their own weather and demand, at what emissions-reduction level flexible nuclear enters a least-cost system, what determines that threshold, and what the thermal storage is worth.

What did we find?

The Macro Energy Model was run at hourly resolution as a greenfield least-cost optimization, with solar, wind, natural gas with and without carbon capture, battery storage, and nuclear with thermal energy storage, over emissions-reduction constraints from zero to 100 per cent.

At near-current nuclear cost — the US Energy Information Administration figure of US$6,317 per kilowatt-electric — solar and wind win at moderate emissions reductions, and nuclear enters only in deeply decarbonized systems where the overbuild and curtailment required to fill renewable gaps become the dominant cost.

At US$4,000 per kilowatt-electric the entry point moves dramatically and by different amounts in different places. Nuclear enters at 84 per cent emissions reduction for German conditions at the higher cost but at 50 per cent at the lower cost; for Brazil the corresponding figures are 80 per cent and 6 per cent.

What sets the threshold is wind, not solar and not the shape of demand. Countries with annual mean wind capacity factors near 0.4 — the United States, China, Chile — require high emissions-reduction goals before nuclear competes, while countries at or below 0.25 — India, Japan, South Korea — bring it in at modest reductions or immediately.

Thermal storage is cheap relative to what it does. It contributes on average 11 per cent of system cost at the higher nuclear cost and 7 per cent at the lower, and removing it raises system cost by as much as 15 per cent under deep decarbonization. It is used mostly on intra-daily timescales, with longer gaps met by additional generating capacity.

Firm generation substitutes for storage rather than adding to it. At 99 per cent decarbonization without nuclear, the least-cost system builds battery storage averaging 3.9 hours of mean demand across countries; with nuclear it builds battery plus thermal storage averaging 0.9 hours.

Whether thermal storage helps or hurts renewables depends on the location. For China and South Africa it mainly absorbs wind and solar variability and lets more of them in; for the United States and Australia it mainly absorbs demand variability and lets more nuclear in.

Building order matters. A system that reaches 99 per cent decarbonization starting from a 50 per cent-reduction fleet already in place costs up to 12 per cent more than one optimized freely, because early solar and wind capital is long-lived and displaces cheaper later options.

The analysis is deliberately stylized — no pre-existing capacity, single-year optimization, fixed technology costs, a 7 per cent discount rate, and no representation of markets, politics, or existing infrastructure. Its purpose is to identify which countries merit detailed study, not to forecast what any of them will build.

Why does it matter?

It replaces a global argument about nuclear with a country-level criterion. The question of whether nuclear belongs in a decarbonized grid has a different answer in India than in the United States, and the paper identifies which physical variable — wind capacity factor — decides it.

It quantifies what cost reduction buys. A drop from US$6,317 to US$4,000 per kilowatt-electric is not a marginal improvement in nuclear's position; it lowers whole-system cost by 15 to 25 per cent at full decarbonization and moves nuclear's entry point down by tens of percentage points of emissions reduction.

The lock-in result is a warning about sequencing rather than about technology. Capacity built to meet a near-term target is still standing when the target tightens, so a least-cost path to 50 per cent is not a segment of the least-cost path to 99 per cent.

Citation

Lei Duan, Robert Petroski, Lowell Wood, and Ken Caldeira (2022). Stylized least-cost analysis of flexible nuclear power in deeply decarbonized electricity systems considering wind and solar resources worldwide. Nature Energy 7, 260-269.

Related