Moteur de recherche d'offres d'emploi CEA

TRANSPAREMS - Particle transport in stochastic media with memory effects - Saclay (91)

Vacancy details

General information


The French Alternative Energies and Atomic Energy Commission (CEA) is a key player in research, development and innovation in four main areas :
• defence and security,
• nuclear energy (fission and fusion),
• technological research for industry,
• fundamental research in the physical sciences and life sciences.

Drawing on its widely acknowledged expertise, and thanks to its 16000 technicians, engineers, researchers and staff, the CEA actively participates in collaborative projects with a large number of academic and industrial partners.

The CEA is established in ten centers spread throughout France



Description de l'unité

SERMA is the reactor physics and applied mathematics division at CEA, whose main R&D activities concern code development, validation and application for the numerical simulation of nuclear systems, in the domain of radiation shielding, reactor physics and depletion, criticality-safety and nuclear instrumentation.
Situated within SERMA, the LTSD laboratory (Stochastic Transport and nuclear Data Lab) is in particular devoted to the development of the current generation (TRIPOLI-4) and next generation (TRIPOLI-5) Monte Carlo code for particle transport.

Position description


Neutronics and reactors physics



Job title

TRANSPAREMS - Particle transport in stochastic media with memory effects - Saclay (91)


In order to model the transport of particles in multi-scale disordered media, two approaches are possible. The first is based on the generation of a very large number of random geometries, to estimate the averages of the quantities of interest. This "reference" approach requires extremely long calculation times. The second approach is based on the derivation of so-called “effective” transport equations, taking into account in a condensed way the effects of disorder during a single calculation. The reliability of these models, fast but approximate, must be established with respect to the reference solutions. In a previous project we have investigated the "reference" solutions for random media. The TRANSPAREMS (Stochastic TRANsport of PARTicles with Spatial Memory Effects) project aims to design a new class of "effective" transport models by developing innovative Monte-Carlo methods to take into account the spatial memory effects induced by the disorder on the particle trajectories.

Contract duration (months)


Job description

Within the framework of the GEOSTOH project (2021 – 2022), we have achieved a major breakthrough in the processing of the "reference" approach to the simulation of particle transport in Markovian random media. Based on the theory of Poisson tessellations, we were able to first establish a rigorous mathematical framework and then develop the simulation tools necessary to sample a set of three-dimensional Markovian media with spatial gradients. This made it possible to calculate for the first time “reference” solutions for this very important class of random media. Until now, for reasons of algorithmic complexity and computational cost, these simulations were almost exclusively limited to dimension d=1. Recent results have allowed sampling in dimension d=3 for the case of Markovian random media, subject to a hypothesis of spatial homogeneity. In the GEOSTOH project, the homogeneity hypothesis was relaxed for three-dimensional media: the ability to take into account non-homogeneous statistical properties, and in particular spatial gradients, turns out to be essential in the modeling of realistic disordered structures, such as the stratification of materials due to gravity or the complex interfaces of multiphase mixtures (Rayleigh-Taylor instabilities).

In the TRANSPAREMS project, we propose to extend the advances of GEOSTOH to "effective" transport models: we will develop new Monte-Carlo methods, inspired by the Chord Length Samplig (CLS) approach, making it possible to treat the effects disorder on the trajectories of the particles in a condensed way, within a single transport calculation.
In the case of spatially homogeneous Markovian media, the CLS method has recently been developed and tested in dimension d=3. In general, the CLS method shows deviations from the reference solutions, due to its inability to take into account spatial correlations. However, it has been shown that it is possible to significantly improve the precision of the CLS method by introducing spatial memory effects: instead of regenerating the material pseudo-interfaces each time the particles move, the particles will remember the position of the last interfaces crossed, and this during several consecutive movements. Several models with memory have been proposed, including Poisson-Box Sampling (PBS), or Local Realization Preserving (LRP). By increasing the number of interfaces that each particle remembers, the accuracy of these “effective” methods with spatial memory increases considerably (at the expense of computation time).
The central idea of the TRANSPAREMS project is to generalize “effective” models of the PBS or LRP type (with memory) to the case of inhomogeneous Markovian media, with spatial gradients. This would allow to have fast effective methods, offering more fidelity thanks to the possibility of taking into account the spatial heterogeneities, and with an increased precision thanks to the inclusion of the memory effects.

Methods / Means

Monte Carlo methods and code development

Applicant Profile

PhD in mathematics, physics or engineering

Taste for numerical simulation


Position location



Job location

France, Ile-de-France, Essonne (91)



Candidate criteria


  • French (Fluent)
  • English (Fluent)

Prepared diploma

Bac+8 - Doctor of philosophy (PhD)

Recommended training

Physics, Mathematics, Computer science


Position start date