Application of the DEPS AI methodology to optimizing the nuclear fuel cycle as implemented in ANICCA

Introduction

As deep decarbonisation requires a massive expansion of electricity production to replace fossil energy, the interest in nuclear energy is reviving worldwide. Nuclear fission, the physical reaction powering today’s nuclear reactors, produces no greenhouse gas emissions (GHG). When assessing nuclear energy’s entire lifecycle, emissions range from 6 to 10 g[CO2eq]/kWh, comparable to hydropower and wind, and far lower than gas-powered systems (20 times less) and coal-powered systems (30 times less). This makes nuclear energy a powerful contributor to mitigating climate change. Beyond its low-carbon emissions, nuclear reactors provide substantial baseload electricity and direct heat, supporting diverse applications in industry, district heating, hydrogen production, and water desalination.

Advanced nuclear systems, with improved operating temperature, size, and fuel efficiency, are expected to play a key role in decarbonising heavy industrial processes. Current and advanced reactors could also enable the large-scale production of low-carbon hydrogen, expanding clean energy solutions.

In a world where clean energy demands are rapidly rising, nuclear energy – a safe, dispatchable, and low-carbon source – will remain essential to achieving a sustainable, net-zero future.

Public perception of nuclear energy varies widely within countries. Although its lifecycle GHG emissions are comparable to or even lower than those of renewable energy technologies, some countries still debate its environmental sustainability. Common reservations often centre on natural resource utilisation and radioactive waste management. The extremely long-lasting radiotoxicity of the waste generated by current reactors, which spans far beyond a typical public understanding of time, often amplifies unease and fear.

However, current nuclear reactors operating in an “open fuel cycle” do not optimise the use of natural resources and generate highly radiotoxic waste, as spent fuel still contains most of the original uranium and additional heavier elements responsible for its high radiotoxicity. Extensive research has thus focused on maximising uranium utilisation and reducing waste burden. While industrial-scale recycling of uranium and plutonium has been demonstrated, today’s recycling operations remain limited and still isolate small amounts of long-lived, high-level waste (HLW) that requires disposal.

Spent nuclear fuel from light-water reactors (LWRs) typically consists mainly of uranium (~95 % of the total mass inventory), plutonium (~1%) and small percentages of fission products (~4%) and minor actinides (~0.1%), like neptunium, americium and curium.

Scientific consensus supports innovative fuel cycle options: fast-neutron reactors allow for recycling uranium and plutonium multiple times, which would significantly enhance the use of natural resources and the management of radioactive waste. Recycling in fast-neutron systems would then be a game changer for uranium-based nuclear fission, potentially increasing the available energy of this source by a factor of 100.

Full recycling of all actinides entails partitioning to isolate minor actinides followed by their transmutation in dedicated burners (such as ADS) to convert them into lighter and less radioactive elements, reducing the long-term radiotoxicity of ultimate waste – while producing energy. This approach could drastically shorten the time required for the HLW radiotoxicity to decline to natural uranium ore levels – from hundreds of thousand years to mere centuries.

Research into Partitioning and Transmutation (P&T) of minor actinides has been conducted for decades. This work has identified the key technical components and processes necessary for a comprehensive full recycling strategy, which can be integrated into national policies. However, further demonstration is required to confirm the performance and cost-effectiveness of these technologies.

Advancing towards full recycling simulation

Key aspects of nuclear fuel cycles as illustrated in the figure below can nowadays be analyzed by the use of computer simulation tools, which have emerged as versatile and powerful tools in many institutions around the world for helping policy makers addressing suitable solutions concerning the different fuel cycle components, which range from the front-end ones (i.e. mining, conversion, enrichment and manufacturing) all the way to the back-end players, which are the ones that would make nuclear energy sustainable The Belgian Nuclear Research Centre (SCK CEN) has developed over the past few years the ANICCA simulation platform (Advanced Nuclear Inventory Cycle Code”). This code has been designed to be a flexible and reliable tool for scenario studies, ranging its different modelling stages from uranium mining to the final disposal (FD) of the spent fuel (SF) or high-level waste (HLW) from reprocessing.

The objective of this MSc thesis is to contribute for the further development of ANICCA by implementing in the future the AI methodology based on the Direct Environment and Policy Search (DEPS) developed at ULg for allowing more intuitive and rapid scenario analysis taking into account technical constraints of the various components of tf the fuel cycles (rector technologies, dedicated burners, reprocessing and fuel fabrication capacities and performances) as well as policy and financial constraints. Parametric uncertainty quantification studies for different technology options and parameters of the reference advanced Belgian fuel cycle scenario will be carried out. This work will provide an improvement of the code from the numerical approach and introducing policy constraints for the Belgian or regional (EU) cases aiming then at proposing possible evolution trajectories for the nuclear industry and policy makers according to constraints from physics, economics, industry, politics etc.

Objectives 

This project envisages a duration of 80 days of workload where the following tasks must be achieved:

• familiarization with fuel cycle calculations and ANICCA code;
• literature review: definition of uncertainty sources;
• identification of the needed information for implementing the DEPS methodology in ANICCA;
• quantification analysis with ANICCA code and the DEPS modified version;
• and, production of scientific outputs.