Global Sensitivity Analysis for Nuclear District Heating Reactors: What the Latest Research Reveals

Small Modular Reactors (SMRs) are rapidly gaining attention as a clean, stable, and cost-effective option for district heating. Yet many of their critical design assumptions—heat exchanger geometry, pressure vessel sizing, temperature levels, and cost factors—carry substantial uncertainties. Understanding how these uncertainties interact is essential before any real-world deployment.

A new peer-reviewed study in Energy applies global sensitivity analysis (GSA) and SimDec scenario decomposition to a nuclear district heating reactor for the first time, offering a uniquely transparent look into how uncertainties shape both economic performance and optimal design.

Why this study matters

District heating faces a major decarbonization challenge. Biomass is limited, heat pumps rely on volatile electricity prices, and gas is exposed to geopolitical risk. Nuclear heat-only SMRs could offer:

• stable long-term heat pricing
• low emissions
• near-100% conversion efficiency
• siting near cities due to passive safety

But major design decisions—primary water inlet temperature, heat exchanger sizing, reactor pressure vessel cost scaling—must withstand uncertainty.

This is where global sensitivity analysis becomes indispensable.

What the researchers did

The study integrates three components:

1. Techno-economic optimization of a natural-circulation SMR primary heat exchanger
2. Monte Carlo simulation of technical and economic uncertainties
3. SimDec global sensitivity analysis to explain how uncertainties shape outcomes

Figure 2 in the article (p. 3) shows the methodological workflow: OAT screening ➝ optimization ➝ Monte Carlo ➝ SimDec decomposition.

The approach evaluates both:

• how uncertainties affect Levelized Cost of Heat (LCOH)
• how they reshape the optimized design itself (tube length, inlet temperature, vessel sizing)

Key results

1. Pressure vessel and containment costs dominate sensitivity

As shown in Table 3 (p. 7), the reactor pressure vessel (RPV) and containment vessel (CV) cost factors have the highest combined sensitivity indices—far larger than thermohydraulic parameters.

This means cost structure matters more than physics for economic feasibility.

2. Optimal temperatures are higher than expected

Earlier studies assumed ~150°C primary inlet temperatures.
This analysis shows the cost-optimal range is closer to 160–170°C, driven by size and efficiency trade-offs.

SimDec visualizations in Fig. 6–7 (pp. 8–9) show how temperature shifts depending on RPV/CV cost corrections and design constraints.

3. Thermohydraulic uncertainties have surprisingly low impact

Except for core pressure drop, most thermal uncertainties barely influence LCOH or optimized design (p. 8).

This is unusual and suggests the SMR configuration is structurally robust to hydraulic variations.

4. Tube length is strongly shaped by interactions

Figure 8 (p. 10) shows striking multi-modal patterns in tube length due to baffle-number constraints.
A great example of SimDec revealing nonlinear, discrete interactions that standard Sobol indices alone might miss.

5. Nuclear district heating remains competitive

Figure 10 (p. 11) compares SMR heat production to biomass, gas, electric boilers, and heat pumps.

Conclusion: SMRs are cost-competitive, especially because they are least sensitive to fuel and electricity price shocks.

Why SimDec matters here

This is one of the first studies where GSA is applied on top of an optimization model, not before or after.
It shows:

• which uncertainties actually matter
• how the optimizer reacts to these uncertainties
• where engineering effort should focus

This is a prime example of why variance-based sensitivity indices and scenario decomposition are essential for modern engineering design.

Reference

Saari, J., Kozlova, M., Suikkanen, H., Semyagina, E., Hyvärinen, J., & Yeomans, J. S. (2024). Global sensitivity analysis of nuclear district heating reactor primary heat exchanger and pressure vessel optimization. Energy, 312, 133393. https://doi.org/10.1016/j.energy.2024.133393

‍