HPMC

High performance Monte Carlo reactor core analysis

 Partecipanti

Mappa

 Word cloud

Esplora la "nuvola delle parole (Word Cloud) per avere un'idea di massima del progetto.

 Obiettivo del progetto (Objective)

'Design and safety analysis of nuclear reactors is based on extensive use of computer codes for the coupled calculation of time-dependend neutron transport, thermal-hydraulics and burnup. State-of-the-art methods use deterministic techniques to solve the neutronics equations, which require various approximations for a full core: a limited number of energy groups, application of diffusion theory instead of transport theory, homogenization of fuel cells and fuel assemblies, pin power reconstruction, etc. These approximations can be overcome by the stochastic Monte Carlo method for neutron transport. However, coupling with thermal-hydraulics codes, long-time time dependence and application to full reactor cores for detailed (pin-by-pin) power density distribution is only at its infancy. The project aims at developing and demonstrating the application of full core Monte Carlo calculation for time-dependent safety analysis with thermal-hydraulic feedback and burnup using high performance computing. Although Monte Carlo calculations are very suitable for parallel execution, full core integrated problems require ultimate efficiency in parallel execution of the Monte Carlo calculation itself and complete optimisation of all coupling mechanisms when run on a supercomputer with large numbers of processors. The project will provide the general tools for reference calculations, applicable to different reactor types, to test the accuracy of current and future deterministic analysis methods.'

Introduzione (Teaser)

Nuclear reactor safety calculations require good estimates of their dynamic behaviour that are usually done with deterministic codes to solve the simplified problem. An EU project has proposed a Monte Carlo method without approximations.

Descrizione progetto (Article)

The Monte Carlo method is a basic tool in particle transport problems that is well suited for tasks requiring detailed modelling of geometry and physics.

It has been used in analyses of nuclear reactor behaviour for decades, but the applications have mainly been restricted by computer capacity.

The 'High performance Monte Carlo reactor core analysis' (http://www.fp7-hpmc.eu/ (HPMC)) project is extending the use of the method beyond steady-state problems such as shielding.

Researchers found that it is feasible to analyse dynamic behaviour within reasonable computational time.

Take for instance, the maximum power or temperature in an accident scenario.

Nuclear reactor modelling is a complicated task that combines detailed description of neutron transport and coolant flow through the reactor core.

Starting from the interactions between neutrons and the target nuclei, the intermediate step is the so-called lattice calculation, in which the geometry is modelled at the fuel rods assembly level.

These were used as input parameters for a 3D reactor simulator, which yields the reactor response under different operating conditions.

Using the Monte Carlo N-particle (MCNP) code, researchers were able to provide the spatial power distribution in the reactor, given information about temperature and coolant densities.

The power distribution was then input to thermal-hydraulic calculations to produce more accurate estimates of temperature and coolant densities.

These were fed back into the Monte Carlo calculations to update the power distribution.The inclusion of the thermal-hydraulic feedback also took Monte Carlo burn-up calculations several steps further.

Burn-up calculations by the Serpent code provide a picture of changes in irradiated nuclear fuel.

This allows the study of nuclear systems over long time periods ( typically, the fuel cycles of a nuclear power reactor.

These have been successfully tested for different reactor core geometries.

Throughout the project, special emphasis is being placed on finding ways to implement simple, efficient and fast solutions that will help speed up Monte Carlo calculations.

When the Monte Carlo calculations are sufficiently fast, they will be more commonly applied by designers and operators on complex nuclear reactor problems.