Thesis subjects
Filter by criteria

DES: Thesis subject SL-DES-21-0032

Environment and pollution / Earth and environmental sciences
TITLE English Français

Artificial Neural Networks contribution to karst flood forecasting on the CEA Cadarache site


The topic of the proposed PhD addresses groundwater rise flooding forecasting. Such phenomena represents an external risk that has to be taken into account in the frame of the INB design and during each safety check (as described by the ASN n°13 guide). Several events of this nature occurred in the Cadarache CEA site, leading to water ingresses and to temporary stoppage of some of the facilities. Assessing and anticipating this risk is thus a major stake for the CEA.

The main purpose of the thesis will be to develop an artificial neural network (ANN) model aiming at forecasting piezometric levels of a groundwater table using available data (groundwater levels and meteorological data). The ANN model will challenge the forecasts obtained using an existing semi-physical model (MECK) in the frame of the use of the “PERISCOP” tool (operational monitoring and flood risk assessment platform).

The PhD thesis work will be following these steps:

• Synthesis of the literature about neural networks forecasting.

• Selection of a few piezometers representing a type of behaviour in the aquifer using geostatistical methods, correlation and spectral analysis or CEA employee’s expertise.

• Implementation of forecasting models for the selected piezometers.

• Definition of a regionalization method based on previous research on the topic [2,7].

• Comparison of the results with the ones obtained using the semi-physical model (MECK) implemented in the PERISCOP tool.

During the thesis, the candidate will evolve in a pluridisciplinary and dynamic team so that he/she can acquire or deepen the following skills:

• Project management

• Redaction skills

• Spoken communication

• Written and spoken English (scientific publications and conferences)

About the specific technical skills:

• Data processing and analysis

• Statistical methods

• Meteorology and karst hydrogeology

• Experience in artificial intelligence and complex systems modelling

• Experience in natural hazards management: prevention and forecasting

• Programming and development (VBA and/or R and/or python)

These skills may be valued for applications to research positions such as CNRS or IRD, or in public institutions such as BRGM or IRSTEA. The operational purpose of the research could also be an added value for positions in the private sector.

Département de Technologie Nucléaire
Service Mesures et modélisation des Transferts et des Accidents graves
Laboratoire de Modélisation des Transferts dans l’Environnement
Place: Cadarache
Start date of the thesis: 01/09/2021

Sebastien MORILHAT  

CEA Cadarache
Bât 727 - Pièce 108

Phone number: +33 6 20 43 66 52

GAIA (Biodiversité, Agriculture, Alimentation, Environnement, Terre, Eau)


IMT Mines Alès
6 avenue de Clavières
30100 ALES