...
Université catholique de Louvain

Optimization of computational dosimetry of workers in dismantling and decommissioning using machine learning

Opslaan als favoriet E-mailalert

Over de werkgever

The Université catholique de Louvain (UCLouvain) is internationally recognized for research and teaching quality in many different fields of expert...

De pagina van de werkgever bekijken

Nearly half of the world’s nuclear power reactors are approaching – or have already reached –their operational lifetime. Following the permanent shutdown of a nuclear facility, a plan is prepared describing in detail the decommissioning strategy, how and when each facility within the plant will be safely dismantled, and how radiation protection of workers and public is ensured. Radiation personal dosimetry is fundamental to ensure radiological protection of workers throughout all dismantling and decommissioning (D&D) operations. Typically, a legal dosimeter and an electronic personnel dosimeter are used. However, dosimeters are often not sufficiently accurate for a comprehensive monitoring of decommissioning activities [1]. Physical dosimeters are subject to high uncertainties, which can lead to both highly conservative dose estimates, introducing unnecessary limiting factors to decommission activities, but also to over or underestimation of doses to some body regions, especially in presence of highly inhomogeneous radiation fields. To improve the assessment of individual doses addressing external exposure of workers, SCK CEN has driven the development of innovative next-generation tools that effectively overcome the current limitations of personal dosimetry by switching from physical dosimeters to purely computational-based dose assessments. During the PODIUM project, an online dosimetry system was developed for monitoring of medical staff in interventional radiology and in workplaces with neutron radiation fields. In the context of the D&D application, it is the goal to extend and specialize this personal dosimetry system to fit the requirements specific to the decommissioning activities. One of these requirements is to provide real-time feedback to workers on their dose. The online feedback to workers in D&D is crucial to optimize their doses and for the application of the ALARA principle through training and visualization of the radiation.

In this PODIUM approach, the movement of the workers including their position and posture relative to the radiation source is monitored using RGB/depth cameras. The posture of the worker’s body is then translated to the realistic anthropomorphic flexible (RAF) phantom [2] and converted to a suitable format of MC simulation. When combined with source characterization, simulation of different organ doses can be performed. Thanks to the recent development in accelerated computing, Massively Parallel Processing (MPP) on CPUs or GPUs can dramatically reduce computing time [3], but this technique is still quite expensive. Especially in Monte-Carlo simulation of radiation transport for dosimetry, MPP or GPU computing still does not allow real-time dose calculation. We believe that recent advances in ML algorithms and their real-time performance implementations can benefit our personal dosimetry tool in D&D activities. This will be novel research that will allow us to solve the current limitations of the online dosimetry system.

The main objective of the PhD is the investigation of different ML algorithms (decision trees, neural network, among others) that fit the data which will be generated during the D&D project of different irradiation geometries. These data are organ doses of interests and/or effective dose calculated using MC simulations for synthesized scenarios of workers doing activities in different radiation fields specific to D&D. The results of the MC simulations will be provided to the PhD candidate for the training of the ML model. The resulted ML model should predict, with a certain accuracy, organ doses in different situations encountered during the handling and dismantling of radioactive components. The expected outcomes of the PhD are an implementation of the model in a software, and the results of the validation of this model in selected test cases. The model is also expected to be interfaced with the ALARA application that will be developed as part of an D&D research project.

Estimated duration: 4 years

Expert group RDA- Research in dosimetric applications

SCK CEN Mentor: Pasquale Lombardo, plombard@sckcen.be

SCK CEN Co-mentor: Mahmoud Abdelrahman, mabdelra@sckcen.be

Promotor: John A. Lee, UCLouvain, IREC/MIRO & ICTEAM/ELEN, john.lee@uclouvain.be

ADDITIONAL INFORMATION

Selection process

The position and funding are conditioned upon a first selection and interview with the mentors and promotor, followed by a defence of the project by the selected candidate in front of the SCKCEN scientific council and external experts.

REQUIREMENTS

Offer Requirements

REQUIRED EDUCATION LEVEL

Computer science: Master Degree or equivalent

Information science: Master Degree or equivalent

Engineering: Master Degree or equivalent

REQUIRED LANGUAGES

ENGLISH: Good

Skills/Qualifications

Background in data science, artificial intelligence, or machine learning.

A string interest for physics is an asset.

Nu solliciteren

Persoonlijke gegevens*
Je cv en bijlagen uploaden

*Door te solliciteren op een functie op Academic Positions ga je akkoord met onze algemene voorwaarden en ons privacybeleid.

Informatie over de vacature

Functienaam
Optimization of computational dosimetry of workers in dismantling and decommissioning using machine learning
Locatie
Place de l'Université 1 Louvain-la-Neuve, België
Gepubliceerd
2022-01-13
Uiterste sollicitatiedatum
2022-02-12
Soort functie
PhD
Opslaan als favoriet E-mailalert

Meer vacatures bij deze werkgever

Over de werkgever

The Université catholique de Louvain (UCLouvain) is internationally recognized for research and teaching quality in many different fields of expert...

De pagina van de werkgever bekijken

Vergelijkbare vacatures

...
Become an IB Examiner for Swedish A: Literature International Baccalaureate® (IB) 1 maand geleden
...
Become an IB Examiner for Croatian A: Literature International Baccalaureate® (IB) 1 maand geleden
...
Become an IB Examiner for Danish A: Literature International Baccalaureate® (IB) 1 maand geleden
Meer vacatures

Relevante stories

...
A Linguist’s View of the International Criminal Court University of Jyväskylä Leestijd: 4 min
...
Bridging Adolescence and Society with Neuroscience Erasmus Universiteit Rotterdam Leestijd: 4 min
...
The Power of Ultrafast Lasers Advanced Research Center for Nanolithography ARCNL Leestijd: 4 min
...
Going Nuclear to Save Time University of Jyväskylä Leestijd: 5 min
Meer stories