Towards comprehensive multiphase flow modelling for nuclear reactor thermal hydraulics

In any nuclear reactor, ensuring that the nuclear fuel always remains properly cooled is the main achievement of the thermal hydraulic design, which thus has utmost impact on the safety and the performance of the plant. Often, this thermal hydraulic design and the plant safety assessment rely on computational models that, by providing a mathematical representation of the physical system, predict the fluid dynamic behaviour of the coolant and the rate of heat transfer in the system. In a nuclear plant, in normal operating conditions or in accident scenarios that require emergency cooling, this often requires solving gas-liquid multiphase flow problems. Unfortunately, although computational tools of any degree of complexity are now available, modelling and computation of gas-liquid multiphase flows is still mainly limited to well-defined flow conditions and/or entirely based on empiricism. The aim of this fellowship is to develop an advanced computational model that overcomes these limitations and goes well-beyond currently available capabilities. At the present time, different techniques reach good accuracy in distinct and well-defined flow conditions, but none has been successful in modelling the entire spectrum of gas-liquid multiphase flows without a priori knowledge of the flow regime. This strongly limits the applicability of available models to flows that are of industrial interest, since these rarely exhibit the same well-characterized and defined flow features. In this project, by means of novel numerical techniques, advanced modelling methods will be coupled in the same computational model and selectively applied based on suitability to the local flow conditions. This will ensure accuracy and unprecedented applicability to multiphase gas-liquid flows, avoiding limiting assumptions but at the same time unrealistic computational requirements.

In the nuclear sector, such a model will provide leading edge modelling and simulation capabilities, underpinning improved operation of the current reactor fleet and design and assessment of future plants. Confident predictions will inform the reactor design and the assessment of safety limits, reducing empiricism and conservatism. In addition, the number of costly experiments will be limited to a smaller number of model-driven tests. Reactors that are safer and produce electricity at a cheaper price and with a reduced waste footprint will underpin Government's plan for between 16 GW and 75 GW of new nuclear generation capacity by 2050. This new capacity will be essential to ensure a secure, sustainable and low-carbon energy future to the UK and respect the legally binding commitment to reduce carbon emission by 2050 of at least 80% with respect to 1990.

In addition, the work will have wider application outside the nuclear sector in the optimization of the design and operation of the numerous industrial equipment exploiting gas-liquid multiphase flows across all branches of engineering (e.g. enhanced mixing by bubbles in bubble columns, fluid dispersion and mass transfer in separation equipment, two/three phase flow streams in extraction, treatment and transportation of oil and gas). At the same time, the fine resolution of spatial and temporal scales as well as of the majority of the interfacial details will allow more fundamental studies to be made. These will shed new light on the many aspects of multiphase flows that still miss thorough understanding, which negatively affects the design and operation of multiphase equipment. The project will benefit from close collaboration with esteemed academics within the UK and overseas (Massachusetts Institute of Technology and North Carolina State University) and industrial leaders in the development of computational products for the nuclear industry and in the analysis and assessment of nuclear reactor thermal hydraulics (Siemens Industry Software Ltd and Frazer-Nash Consultancy).

The main target of the present fellowship is the modelling capabilities available to the nuclear industry through commercial software packages (e.g. STAR-CCM+) or proprietary models. Recipients include reactor designers (e.g. Rolls-Royce, EDF Energy), regulating bodies (e.g. the Office for Nuclear Regulation) or engineers involved in the analysis and safety assessment of nuclear plants (e.g. Frazer-Nash Consultancy and Wood).
In view of the complexities involved, an all flow-regime multiphase computational model might require decades to reach industrial maturity. However, the organization of the fellowship in work-packages is aimed at delivering impact on different time frames:
- Short-term (3-5 years). Improvements to averaged multi-fluid models (WP 1) that are already best-practice in industry. Timescale can be expected to be comparable to the duration of the fellowship. Impact will be provided on a similar timescale also by availability of high-quality experimental data and high-fidelity simulation results (WP 4) that can support model development and validation.
- Medium term (5-10 years). Improvements to models that track large interfaces, but still model interface transfer processes (WP 3). These models are currently under development and not yet exploited in industry, therefore impact and exploitation on a medium time frame can be reasonably expected.
- Long term (10-20 years). All flow-regime model where large interfaces, and all interface transfers, are fully-resolved. Mainly under conception, these models have the highest potential but might require decades to reach the necessary maturity to be fully-exploited in industry.

Designers will benefit from the availability of more advanced numerical tools that can help improving thermal hydraulic design of reactors, and the level of safety and economic competitiveness of future plants. Using the EPR reactor to be built in Hinkley Point C power station as a reference, an increase of 1% in electricity output (driven by higher efficiency or reduced uncertainty in the safety margins) equals to 350 GWh/year. This translates to around £ 17.5M per year at the current electricity market price. In addition, reliable computational tools can reduce the number of costly experiments to a smaller number of model-driven tests. Examples of the cost of experimental programs are provided by the estimated cost of a UK national thermal hydraulic facility, expected of the order of some tens of million pound sterlings (NIRAB-75-10, 2016), not accounting for operating costs or specific experimental setups.
At the same time, impact is expected on regulators and safety assessors responsible for the safety evaluation of reactors. Accurate predictive models will enable more thorough and less uncertain evaluations of safety margins, potentially limiting conservatism and redundancy of safety systems and benefiting economic viability of reactors, while ensuring the necessary levels of safety are always guaranteed. To put safety in context, continuously increasing estimation for the costs of the Fukushima disaster has been recently set by the Japanese government to $ 187 billion.
Safer plants will improve public acceptability of nuclear energy. In addition, plants that are cheaper to build will reduce investment costs and ensure availability of electricity at more affordable prices, reducing at the same time the CO2 footprint from the energy sector.

Large impact is expected on the developers (e.g. Siemens Industry Software Ltd and ANSYS) of computational fluid dynamic software. By implementing the more advanced models developed in their codes, these companies can make them available to all their users, reaching a number of recipients much larger than the nuclear industry.

Lastly, fellowship will impact on the development of multidisciplinary skills in thermal hydraulics and multiphase flows. In nuclear, these are much needed for United Kingdom's future energy plan to be successful.