Case Study 2: Small Modular Reactors — Physics Meets Economics

Introduction: The Promise and the Problem

For seventy years, the nuclear industry has followed a simple economic logic: build bigger. A 1,200 MWe reactor produces roughly four times the power of a 300 MWe reactor, but it does not cost four times as much to build, operate, or staff. This economy of scale has driven reactor sizes upward from the 60 MWe Shippingport reactor (1958) to the 1,650 MWe European Pressurized Reactor (EPR).

But the economy-of-scale argument assumes that larger plants can be built on time and on budget. Recent Western experience has shattered that assumption. The Vogtle Units 3 and 4 in Georgia, USA — the first new nuclear units built in the United States in 30 years — were originally estimated at $14 billion and came in at over $35 billion, seven years late. Flamanville 3 in France is 12 years late. Olkiluoto 3 in Finland was 14 years late.

Small Modular Reactors (SMRs) propose a fundamentally different approach: build smaller, build in a factory, deploy in modules. This case study examines whether the physics supports the economics.

The Physics Case for SMRs

Passive Safety from Size

The most compelling physics argument for SMRs is that smaller cores are inherently easier to cool passively. The reason is straightforward:

Decay heat scales with volume (proportional to the amount of fuel, hence to $R^3$), while heat loss scales with surface area ($R^2$). As a reactor shrinks, the surface-to-volume ratio increases as $1/R$, making it easier to remove decay heat by passive mechanisms (radiation, natural convection, conduction) without pumps or external power.

Quantitatively, consider the power density and the cooling requirement:

$$\frac{\text{Decay heat per unit surface area}}{\text{at shutdown}} \propto \frac{P_{\text{thermal}} \times 0.06}{4\pi R^2} \propto \frac{q_v \cdot R^3}{R^2} = q_v \cdot R$$

where $q_v$ is the volumetric power density (W/m$^3$). Smaller $R$ means less decay heat per unit surface area, which means passive cooling is more effective.

For a 77 MWe NuScale module (226 MW(th)):

• Decay heat at shutdown: ~14 MW
• Vessel surface area: ~70 m$^2$ (approximate)
• Heat flux: ~200 kW/m$^2$

Compare with a 1,100 MWe PWR (3,400 MW(th)):

• Decay heat at shutdown: ~204 MW
• Vessel surface area: ~180 m$^2$ (approximate)
• Heat flux: ~1,130 kW/m$^2$

The NuScale module has a heat flux roughly 5.6 times lower — well within the range that natural convection to a surrounding water pool can handle indefinitely.

Natural Circulation

Many SMR designs eliminate reactor coolant pumps entirely, relying on natural circulation to move coolant through the core. The driving force is the density difference between the hot (rising) and cold (descending) coolant:

$$\Delta P_{\text{buoyancy}} = g \int_0^H [\rho_{\text{cold}}(z) - \rho_{\text{hot}}(z)] \, dz$$

For a NuScale-type integral design with the steam generator located directly above the core inside the reactor pressure vessel, the natural circulation driving head is roughly:

$$\Delta P \approx g \cdot \Delta\rho \cdot H$$

where $\Delta\rho \approx 50\,\text{kg/m}^3$ (water density difference between cold leg at 260°C and hot leg at 300°C at 127 bar) and $H \approx 8\,\text{m}$ (height of the circulation loop):

$$\Delta P \approx 9.81 \times 50 \times 8 \approx 3,900\,\text{Pa}$$

This modest pressure head (0.039 bar) is sufficient to drive adequate coolant flow through the compact core because the flow resistance is also small (short fuel assemblies, low power density). Natural circulation eliminates a major class of potential failures: loss of forced circulation.

Integral Design: Eliminating Pipe Breaks

In a conventional PWR, the reactor pressure vessel is connected to the steam generators by large-diameter pipes (the "hot leg" and "cold leg"). A break in these pipes — a Large-Break Loss of Coolant Accident (LBLOCA) — is one of the most severe design-basis accidents.

In integral SMR designs (NuScale, mPower, SMART), the steam generator is inside the reactor pressure vessel. There are no large-diameter pipes that can break. The largest penetration is typically 2–3 inches, limiting the maximum leak rate to a level that passive makeup systems can easily compensate.

The Economics: Where Physics Meets the Real World

The Scaling Problem

The fundamental economic tension of SMRs is captured by the scaling exponent in the cost-capacity relationship:

$$C = C_0 \left(\frac{P}{P_0}\right)^n$$

where $C$ is the capital cost, $P$ is the capacity, and $n$ is the scaling exponent. For conventional nuclear plants built as one-of-a-kind projects, $n \approx 0.6$–$0.7$ — meaning that a plant twice as large costs only $2^{0.65} \approx 1.57$ times as much. This economy of scale strongly favors large plants.

For an SMR to compete, it must overcome this scaling disadvantage through:

1. Factory learning curves: Manufacturing multiple identical modules should reduce the cost of each successive unit. Learning curve theory predicts that each doubling of cumulative production reduces cost by a fixed percentage (the "learning rate"). A 10% learning rate means the 16th unit costs $(0.9)^4 = 0.66$ times the cost of the first unit.

2. Reduced construction risk: Factory-built modules can be quality-controlled and tested before shipment, reducing the on-site construction schedule and the cost overruns that plague large nuclear projects.

3. Simplified design: Passive safety systems eliminate pumps, diesel generators, and associated redundancy, reducing the sheer amount of equipment and piping.

The NuScale CFPP Case

The NuScale Carbon Free Power Project (CFPP) at Idaho National Laboratory provides a sobering real-world test of SMR economics:

The project was canceled in November 2023 when the cost escalation made it uncompetitive with other clean energy options available to the participating utilities. The primary cost driver was not the reactor itself but the balance-of-plant, site preparation, and the reality that "first of a kind" engineering costs are high regardless of the reactor size.

What Would Make SMRs Competitive?

The economic analysis suggests that SMRs need:

1. Fleet deployment: A single SMR is expensive. The economic case requires dozens to hundreds of identical units to drive down costs through learning. South Korea's success with the APR-1400 (built at Barakah for ~$4,600/kWe) demonstrates that standardized design + continuous construction = low cost. SMRs aim to replicate this with factory production.

2. Regulatory efficiency: Current licensing processes were designed for large, one-of-a-kind plants. Each NRC review costs $100–300 million and takes 3–5 years. If the same process must be repeated for each SMR design, the costs are prohibitive. Type-certification (certify the design once, deploy many) is essential.

3. Co-siting advantages: SMRs may be most competitive not as standalone power plants but as replacements for retiring coal plants (using existing grid connections, cooling water, and workforce) or as industrial heat sources (providing 500–900°C heat for hydrogen production, desalination, or chemical processing — applications where renewables cannot easily compete).

Current SMR Projects: A Status Report (as of 2025)

Under Construction or Operating

In Advanced Licensing

What the Data Shows

As of 2025, exactly one SMR design is producing commercial electricity: the Chinese HTR-PM. Russia's floating nuclear power plant (Akademik Lomonosov, using two RITM-200-derived reactors) has been operating since 2020 but at submarine-derived scale. The Western SMR programs are 3–8 years from operation.

The HTR-PM is instructive. Its TRISO fuel — billions of tiny coated particles, each a self-contained micro-containment — makes meltdown physically impossible. Even at extreme temperatures (tested up to 1,620°C), fission product release from TRISO particles is negligible. This is not a safety system that can fail; it is a property of the fuel itself.

Analysis: Matching Physics to Markets

Where SMRs Have Physics Advantages

Where SMRs Face Headwinds

The HALEU Question

Several advanced SMR designs (TRISO-fueled HTGRs, fast reactors, micro-reactors) require High-Assay Low-Enriched Uranium (HALEU) — uranium enriched to 5–20% ${}^{235}\text{U}$, above the ~5% of conventional PWR fuel. As of 2025, there is no commercial HALEU supply chain outside Russia. The U.S. Department of Energy is funding centrifuge capacity at Centrus Energy in Piketon, Ohio, and other facilities, but a reliable Western HALEU supply is not expected before 2027–2028.

The physics of HALEU is straightforward — higher enrichment means higher $\eta$, enabling smaller cores, longer fuel cycles, and more flexible fuel forms. But the enrichment infrastructure is a bottleneck that cannot be solved by physics alone.

Conclusion: The Verdict Is Not In

Small Modular Reactors represent a genuine attempt to change the economic and safety paradigm of nuclear energy. The physics case is strong: smaller cores enable passive safety, natural circulation, and integral designs that eliminate entire categories of accidents. The engineering case is plausible: factory fabrication should reduce construction risk and cost.

But the economics remain unproven. No Western SMR has demonstrated competitive electricity costs. The NuScale CFPP cancellation shows that "first of a kind" costs are high even for simplified designs. The path to cost competitiveness requires fleet deployment — building not one but dozens of identical units — and fleet deployment requires confident customers, which requires demonstrated cost competitiveness. This is a classic chicken-and-egg problem.

The most likely path to resolution is the one already emerging: government co-funding of first-of-a-kind projects (TerraPower Natrium, GE-Hitachi BWRX-300), combined with niche market deployment where SMRs have unique advantages (process heat, remote sites, coal replacement). If the first few units perform well, the learning curve may bring costs down enough for broader deployment. If they do not, SMRs will remain a promising concept with unproven economics — a status that nuclear energy has, unfortunately, occupied before.

The physics, at least, is ready. The question is whether the institutions are.

Discussion Questions

1. Scaling vs. learning. The economy of scale favors large reactors ($n \approx 0.65$), while the learning curve favors mass production (10–20% cost reduction per doubling of production). At what production volume does a 300 MWe SMR at $8,000/kWe (first unit) become cheaper per MWh than a 1,200 MWe plant at $6,000/kWe? Set up the calculation using the learning curve formula $C_n = C_1 \times n^{-b}$ where $b = -\ln(\text{learning rate})/\ln 2$.

2. Passive safety quantification. A NuScale module can remove decay heat by natural convection to a surrounding pool indefinitely. Calculate the minimum pool volume needed to absorb 72 hours of decay heat from a 226 MW(th) reactor (assume an initial heat rate of 14 MW decreasing as $t^{-0.2}$, and the pool heats from 40°C to 100°C without boiling at 1 atm).

3. TRISO fuel and the elimination of meltdown. TRISO particles have been tested to 1,620°C without significant fission product release. The maximum fuel temperature in an HTGR loss-of-coolant accident (with no operator intervention) is calculated to be ~1,200°C. What is the safety margin in degrees? Why does the pebble-bed geometry (where each "pebble" contains ~15,000 TRISO particles in a graphite matrix) enhance passive safety compared to conventional fuel rods?

4. Coal-to-nuclear conversion. The U.S. has ~200 coal plants retiring by 2035, with existing grid connections, cooling water permits, and trained workforces. What are the physics, regulatory, and economic advantages of siting SMRs at these locations? What are the challenges?

5. The Russian monopoly problem. As of 2025, Russia is the only country with commercial operating experience in fast reactors (BN-600, BN-800) and the only significant supplier of HALEU. Discuss the geopolitical implications for Western SMR programs that depend on these technologies.