(SATURN CDT) AI-Assisted Optimisation of Soluble-Boron-Free Small Modular Nuclear Fission Reactor Cores

About the Project

Saturn_Nuclear_CDT UoM_Nuclear

Abundant clean energy for all relies on advanced computational design. Small modular pressurised water reactors that avoid soluble boron can simplify coolant chemistry and reduce environmental impact. They also create a complex high dimensional and multi-objective core-design optimisation challenge: excess reactivity, shutdown margin, power peaking and cycle length must be simultaneously optimized through internally linked design variables such as fuel loading patterns, enrichment zoning, burnable absorber distributions and control rod movement strategy.

This industry-sponsored PhD, sponsored by Rolls-Royce SMR Ltd., will develop a modern optimisation framework for small modular reactor (SMR) core design, using state-of-the-art optimisation algorithms to support more systematic, tractable, and effective exploration of complex design problems than is possible with traditional approaches. Case studies will be used to demonstrate how modern optimisation methods can identify robust design solutions and quantify trade-offs between competing objectives.

The project will follow a staged route with academic and industrial support and supervision. First, the student will establish a conventional reactor-physics simulation route using lattice and nodal methods, with benchmark cases such as the PRATIC soluble-boron-free SMR benchmark and verification against high-fidelity Monte Carlo calculations where available. Second, the student will automate multi-objective optimisation studies to explore trade-offs between conflicting objectives. Third, the student will investigate reduced-order, surrogate-assisted and AI-accelerated methods, assessing where these methods can accelerate design exploration while retaining an auditable link to trusted high-fidelity models.

Project-specific training will be provided in lattice and nodal codes, Monte Carlo, Python and optimisation algorithms, supporting UK capability in advanced SMR core design. Prior experience with specialist reactor-physics codes is helpful but not required. The student will work within the University of Manchester research environment, with industrial engagement from Rolls-Royce SMR Ltd.

The ideal candidate will enjoy quantitative problem-solving, programming and computational modelling, and will be motivated to apply modern optimisation and AI-assisted methods to challenging nuclear reactor design problems.

Key research activities are expected to include:

• developing and verifying reactor-physics models for soluble-boron-free SMR core analysis using lattice, nodal and selected Monte Carlo methods;
• building automated Python-based workflows for generating and assessing candidate core designs;
• applying multi-objective optimisation methods to explore trade-offs between cycle length, reactivity control, shutdown margin, power peaking, fuel utilisation and operational flexibility;
• developing reduced-order, surrogate-assisted or AI-accelerated models to make design-space exploration more efficient;
• demonstrating the final optimisation framework on representative SMR core-design case studies relevant to industrial design challenges.