Introduction
People
R&D topics
Projects
Teaching
Publications
Applications
Partners
Career
Video
Datasets

Follow us on Facebook LinkedIn Group

Atmosphere Identification for Protection of Population in Preparation for Accidental Releases - MARIONETTE

Principal investigator:

Prof. Dr. Juš Kocijan

Duration:

1.1.2025 – 31.12.2027

Funding:

ARIS - Slovenian Research and Innovation Agency, L2 - 60149 (B), co-financing from Krško Nuclear Power Plant







Abstract:

Project MARIONETTE discusses identifying the atmosphere’s ability to disperse for accidental radioactive pollutant release.

Several EU countries are investing in new nuclear power plants, and the EU classifies them as carbon-neutral energy sources due to their lack of CO2 emissions during electricity generation. Expert studies are underway in Slovenia to determine whether a new nuclear power plant should be established in Krško, where one (NEK) is already situated. NEK recently received a license extension for the next 20 years based on modernizations and security enhancements.

Project MARIONETTE aims to advance the understanding and speed up quantifying radionuclides’ dispersion in the ambient air in the event of accidental releases. Due to unknown accidental emissions, traditional physics-based dispersion models cannot be used quickly. Models are crucial for protecting the population against potential hazards associated with nuclear power plants.

The dispersion of accidental radioactive air pollution is a significant scientific concern, irrespective of one's stance on nuclear power generation. Scientific research on accidental dispersion is needed for reasons such as accidents or terrorist threats. Furthermore, the project is also applicable for accidental releases of toxic pollutants during chemical accidents, like a train accident in the USA and incidents in Slovenia covered by the SEVESO European Union Directive.

The typical research question across nuclear and chemical industries and terrorist activities concerns understanding how airborne pollutants spread during accidental releases. Rapid and accurate descriptions of atmospheric dispersion are needed, especially in challenging terrains. The limitations arising from the initial lack of information on the quantity and time distribution of accidental pollutant releases hinder the use of traditional dispersion models.

To address these challenges, the project proposes a combination of theoretical and experimental modelling, utilizing system identification and machine learning methods. This approach aims to accelerate, enhance and replace physics-based models used for atmospheric dispersion from unpredictable and unsteady sources like accidental releases.

We plan to develop:

  • A method for determining the atmosphere’s dispersion ability for arbitrary accidental emission of pollutants into the ambient air.
  • A surrogate dispersion model for simulating the nonlinear spatial dependence of concentrations on weather variables to considerably speed up computations.
  • A combined model consisting of a numerical weather model, whose outputs are used as physics-based or surrogate dispersion-model inputs and surrogate dispersion model to upgrade the existing dispersion models.
  • The validation of the developed methods in the Krško testbed comprising the realistic environment of the Krško basin.


Project workpackages:
  1. Development of the purpose-made Krško testbed. (Present level of realisation: 0%)
  2. Development of the benchmark toy model. (Present level of realisation: 0 %)
  3. Development of methods for identifying the atmosphere’s dispersion ability combining physics-based and data-driven models. (Present level of realisation: 0 %)
  4. Modelling of the dispersion and utilisation of the Krško testbed with realistic data for an accidental atmospheric release. (Present level of realisation: 0 %)


Project partners:


Selected publications:

To appear