Apply now »

Best-estimate plus uncertainty analysis of accidental scenarios in a Lead Fast Reactor

INTRODUCTION

SCK CEN has recently launched research on an advanced Small Modular Reactor (SMR), specifically a lead cooled Fast Reactor or SMR-LFR (https://www.sckcen.be/en/smr-lfr) . Cooled by lead, the SMR-LFR could become an indispensable element in a sustainable energy security supply. In this perspective, together with a robust, international consortium, SCK CEN aims to build a Belgian demonstration model of the very first SMR-LFR by 2035, called LEANDREA. (https://www.sckcen.be/en/news/leading-nuclear-european-organizations-unite-develop-eagles-300-next-generation-lead-cooled-small-modular-reactor)

SCOPE OF THE WORK

One mandatory step to obtain the nuclear license for construction of the reactor is the approval by the Belgian regulator of the Preliminary Safety Assessment Report, also containing the safety analyses in accidental conditions.

The safety analyses are requested to be run with best-estimate simulation tools, supported by conservative assumptions on input parameters and boundary conditions or by best-estimate input parameters supplemented by uncertainty quantification (Best-Estimate Plus Uncertainty – BEPU), if difficult to justify a conservative envelope.

The object of the thesis is to develop an optimized methodology for BEPU analysis and apply it to the Unprotected Loss of Flow accident and other scenarios made available by SCK CEN. The RELAP5-3D model of the reactor will be coupled to the open-source softwares RAVEN (https://github.com/idaholab/raven), and, according to the progress of the work, DAKOTA (Sandia National Laboratories, https://dakota.sandia.gov/about-dakota/ ), widely used for uncertainty quantification applications in the nuclear sector (see for instance [1] and [2]). The work includes a global sensitivity study aimed to show the dependency of the safety analysis related acceptance criteria on the simulation model’s input parameters and embedded physical models affected by the uncertainty.

Other accidental scenarios made available by SCK CEN can be investigated depending on the availability of time.

DURATION OF THE WORK

The foreseen duration of the MSc thesis work is 6 months.

ORGANIZATION OF WORK

The work is organized in parallel tasks as described in the following sections.

I. Performance of uncertainty quantification

1. Practice with RAVEN and DAKOTA functionalities by following the available online tutorial and test cases.
2. Program RAVEN and DAKOTA to run uncertainty quantification and sensitivity studies.
3. Define the uncertainty distribution of the input parameters.
4. Run and analyze the results of the uncertainty quantification of representative accidental scenarios of the reactor, ranking the parameters by importance, according to their impact on the acceptance criteria.
5. Conclude on the need for R&D to reduce uncertainties for the most influencing parameters/models
6. Conclude on the potentiality of the two softwares in support of decision-making process (depending on the work progress)

II. Development of an optimized methodology for uncertainty quantification and sensitivity analysis

In parallel with the work in section I, develop an optimized methodology for the uncertainty quantification for safety analysis with RELAP5-3D. Latin Hypercube Sampling for uncertainty propagation and Sobol global sensitivity analysis will be tested at first. The work is subdivided in the following tasks:

1. Literature overview on uncertainty quantification method and global sensitivity analysis
2. Proposal of methodologies to be tested
3. Selection of an optimized methodology after the feedback of the application described in section I.

REFERENCES

1. V. Narcisi et al. : Uncertainty Quantification method for RELAP5-3D using RAVEN and application on NACIE experiments.
2. F. Mascari et al. : MELCOR – DAKOTA coupling for uncertainty analyses in the SNAP environment/architecture - Nuclear Engineering and Design - Volume 421, May 2024

Apply now »