
KEYNOTE I
Title:
How to Take Care of Turbo Generators
Keynote Speaker:
Dr. Jean-Pierre Ducreux, Senior Research Engineer, Électricité de France (EDF), France
Abstract:
In France, nearly 87% of electricity is generated by nuclear and fossil fuel power plants with the remainder being produced by renewable energy sources (hydro, wind, etc.) Large-scale power plants are therefore essential to ensure optimal electricity production. To ensure the best possible availability, EDF has set up an important monitoring and maintenance plan for its generation facilities.
Turbine generators used in the production of electrical energy and driven by steam or gas turbines steam or combustion turbines (TAV and TAC) are subject, due to their operation, to significant stresses that can be the cause of some failures. These failures are mostly minor and have very little impact on the availability. Nevertheless, it can happen that they get worse and turn into a real fault that it is essential to monitor and diagnose.
It is in this context that a methodology of identification of faults of a turboalternator is set up. This detection methodology will be based mainly on the signal from a radial flux probe located in the air gap of the machine. There are many diagnostic methods for electrical machines. Some authors have shown that it is possible to establish a diagnosis via thresholds on local or global measurements. However, these methods are limited for the identification of the severity of faults. This has led EDF to develop rules for identifying defects of synchronous machines based on pattern recognition methods.
Therefore, rules were established that allow to determine the type of fault that characterizes the test data. It consists in performing a hierarchical classification to identify the kind of defects. This technique has been submitted to the data and gives excellent results since all the test data were assigned to the right type of defect.
To formalize the relationships between the faults and the number of candidate classes, a finite state automaton and a study on many possible scenarios was conducted. This approach improved all the results of the classification of the test data by removing some confusion areas together with some prototypes.
Speaker’s Biography:
To formalize the relationships between the faults and the number of candidate classes, a finite state automaton and a study on many possible scenarios was conducted. This approach improved all the results of the classification of the test data by removing some confusion areas together with some prototypes.
KEYNOTE II
Title:
Health Condition Monitoring Techniques - Oil Immersed Transformers
Keynote Speaker:
Dr. Seyed-Saeid MOOSAVI-ANCHEHPOLI, Research Engineer and Technical Lead, Technological Research Institute Railenium, Valenciennes, France
Abstract:
The transformer is an indispensable asset in railway network infrastructure and distribution network. They are also expensive and account for massive capital expenditure in the contemporary electrical network. Not only do they require huge fiscal investments, but the reliability and dependability of the electrical railway and distribution network depends primarily on their operational stability. It is, therefore, imperative that railway companies give priority to failure prevention and the sustenance of optimal operational status of their electrical network. Condition monitoring and asset management is therefore a key concern in all electrical utility providers, especially for rail and electrical infrastructure managers.
Over the recent years, transformer health indexing (HI) has become a popular tool for performing transformer health assessments on a larger fleet of transformers. HI is a tool that allows asset engineers/managers to make informed decisions by processing available data of the transformer and convert those into an overall “condition” score. This condition is usually based on “scores” and “weighting”, which are calculated from a set of algorithms designed to evaluate both field conditions, inspection results, on-site test results, etc. which permitted to estimate the remaining useful lifetime (RUL).
In this work, results of an industrial project in collaboration with SNCF Réseau related to the determination of traction transformer faults, fault diagnosis, HI condition monitoring and RUL estimation methods for the oil immersed transformers will be presented and several case studies will be discussed.
Speaker’s Biography:
Dr. Seyed-Saeid MOOSAVI-ANCHEHPOLI, obtained his PhD from Université de Technologie de Belfort-Montbéliard (UTBM), France in 2013. He was an Assistant Professor at UTBM in 2013-2014, and Université de Caen Normandie-France in 2019-2020. In 2015, he was Head of Signaling and Electrification System, Line 2 of Mashhad City Metro - Metro-system Co, Iran. From 2015 to 2019, he was Associate Professor at Amol University of Special Modern Technologies (AUSMT) in Iran. Since 2021, he is a Research Engineer at the Technological Research Institute Railenium, Valenciennes, France.
KEYNOTE III
Title:
Artificial Intelligence for Forecasting in Engineering Applications and Predictive Maintenance
Keynote Speaker:
Dr. Lama Itani, Education and Research Engineer, MathWorks, France
Abstract:
Industrial equipment failure can lead to costly downtime, outweighing the cost of replacing the equipment. Predictive maintenance aims to reduce this unplanned downtime by utilizing sensor data to anticipate necessary maintenance. However, acquiring data from physical equipment under typical fault conditions can be challenging, potentially leading to catastrophic failure or being too costly.
In this presentation we will navigate the solution of creating a digital twin for a triplex pump enabling the generation of sensor data under various fault conditions through simulation. We will then use Machine Learning to craft the predictive maintenance algorithm to help in recognizing which components in the pump are about to fail.
Speaker’s Biography:
Lama Itani holds a PhD in Mechanical Engineering from IFP Energies Nouvelles. Before joining MathWorks (makers of MATLAB and Simulink), she worked on topics related to energy optimization and signal processing for different OEMs. Today Lama is an Application Engineer in the Academia Group at MathWorks France.
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |