Statistical numerical dosimetry
Abstract
——————————————————-
The aim of the project consists in developing, testing and validating a new methodology for numerical dosimetry. It is based on a non intrusive approach of the stochastic finite element method that allows to deal with the variability of the exposure situations. This method enables to make a statistical analysis that can be used to quantify the uncertainty of the exposure indicators.
——————————————————-
Partners
Ampère, Ecole Centrale de Lyon
UMR CNRS 5005
ICJ, University Claude Bernard Lyon 1
UMR CNRS 5208
XLIM, University of Limoges
UMR CNRS 6172
Team project NACHOS
INRIA
http://www-sop.inria.fr/nachos
Contact
Philippe Billoux
UMR CNRS 5005 – Ampère, Ecole Centrale de Lyon
Duration
36 months
Summary
——————————————————-
1) Introduction
ln parallel to biological studies, it is required to ensure that exposure of workers and general public to electromagnetic fields are below defined thresholds. Numerical computation allows to obtain accurate results but those results are strongly dependent on the human morphology and the sharpness of the model. It appears judicious to obtain, for a given category of individuals, SAR values which are representative in terms of mean value and standard deviation. The proposed approach reaches this aim as it enables to deduce a statistical analysis of the effects of the exposure.
2) State of the art
The numerical computation is CPU time consuming and requires a large memory. That’s why almost no investigation has been performed yet on the variability of the situation of exposure. However, France Telecom R&D recent works show the interest in statistically analyzing the effects of this exposure.
3) Scientific methodology
The project will be performed into three steps:
– Definition of a test case to numerically characterizes the interaction child head – mobile phone. It will be use in the following steps of the project. The results of the three numerical approaches developed by each partner will be faced on this test case and compared with existing data reported in literature for similar situations of real exposure.
– Sensitivity analysis to the electromagnetic properties of the tissues: the objective is to know a priori the tissues that have the main influence on the maximal SAR in the head. It will be performed using numerical experimental design method.
– Uncertainty analysis due to the electromagnetic properties of the tissues: the stochastic finite element method will be developed and tested. As feasibility example, a statistical analysis will be drawn on the test case introducing variability in the electromagnetic properties of tissues but also, in a lesser extent, variability in the morphology.
A PhD student will be employed on the basis of the obtained funds. He will make the link between the implied teams and will work successively on each location (Ecully, Limoges and Sofia Antipolis).
4) Partnership, skills
The project is performed by researchers from 4 public laboratories: Ampère (UMR CNRS 5005 – Ecully), XLlM (UMR CNRS 6172 – Limoges), ICJ (UMR CNRS 5208 – Ecully) and pro-ject-team NACHOS (INRIA – Sophia Antipolis).
Implied teams are complementary firstly for their skills in electromagnetism and bioelectromagnetism and secondly for applied mathematics and numerical analysis. The methods for numerical dosimetry developed by each team and that will be used in the project are also complementary.