Modélisation et analyse technico-économique de la cavitation et de la fissuration d'une roue de turbine hydraulique

Abstract

In 2017, scheduling of Hydro-Quebec’s hydroelectric turbines was essentially determined according to their profitability. Nevertheless, the turbines most solicited degrade more rapidly and breakdown more often. In fact, the global cost of the complete lifecycle is not taken into consideration. This is why we propose to complete a techno-economic analysis of the turbine runners and to study maintenance strategies. In the context of this analysis and following a preliminary study previously completed by Hydro-Québec, two main degradation factors are identified : crack propagation and cavitation. Once the software tool used to complete the analysis has been presented, the two degradation factors are defined in two distinct chapters. The asset management tool chosen is VME, developed by EDF R&D teams, with whom IREQ often collaborates. Its graphical interface allows both technical and economic indicators to be included. Afterwards, the results are generated by a Monte-Carlo simulation. For crack propagation, first the physics of the phenomenon, the origin of the breakage and its impact on reliability are presented. The consequences of maintenance on the lifespan of the turbine runners are then explained via a simplified model. The most important aspect lies in the presentation of the digital model : it is composed of five sub-models, which are defined in detail prior to assembling them into a final version. Finally, this first chapter concludes with a verification step to confirm that the results are aligned with those expected. For cavitation, a brief literature review explains the physics of the phenomenon. A second part describes a cavitation model based on observations which are simple and true to the opinions of the experts consulted. This model allows physics still poorly understand to be linked to the costs created by cavitation. The last part describes how to integrate the proposed model in VME and introduces paths of evolution to reach a more precise model.

Date	19 Jun 2018
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Antoine Tahan (Supervisor) & Martin Gagnon (Co-supervisor)