Development of an optimized algorithm for the scheduling of NPP valve maintenance operations
This study focuses on the development of an optimized algorithm to match the workload (volume of maintenance operation in hours) with the manpower distribution (number of workers mobilized) during valve maintenance operations for all nuclear power plant (NPP) outages (shutdown period of the reactor...
| Author: | |
|---|---|
| Format: | master thesis |
| Publication Date: | 2025 |
| Country: | España |
| Institution: | Universitat Politècnica de Catalunya (UPC) |
| Repository: | UPCommons. Portal del coneixement obert de la UPC |
| Language: | English |
| OAI Identifier: | oai:upcommons.upc.edu:2117/428353 |
| Online Access: | https://hdl.handle.net/2117/428353 |
| Access Level: | Open access |
| Keyword: | Nuclear power plants -- França -- Maintenance and repair Algorithms -- Mathematical models Centrals nuclears -- França -- Manteniment i reparació Algorismes -- Models matemàtics Àrees temàtiques de la UPC::Energies::Energia nuclear::Centrals nuclears Àrees temàtiques de la UPC::Economia i organització d'empreses |
| Summary: | This study focuses on the development of an optimized algorithm to match the workload (volume of maintenance operation in hours) with the manpower distribution (number of workers mobilized) during valve maintenance operations for all nuclear power plant (NPP) outages (shutdown period of the reactor for the refuelling of the core where maintenance operations of it are conducted) in France. Two different models are presented, tested and compared. The f irst is a static model, originally used by EDF in previous years, in which the manpower distribution is predetermined by fixed parameters. This model is optimized to improve its accuracy. Thesecondisadynamicmodelinwhichthedistributionofmanpowerismodifiedaccordingto the duration and phases of the outage. This dynamic model is built using a new methodology based on key milestones of NPP outages. The construction and optimization of these models is based on the analysis of real outage planning data from the French nuclear fleet. Through extensive data processing, correlations in the outage planning of French NPPs are identified and analyzed. These findings are translated into mathematical models to create algorithms capable of predicting outage outcomes years in advance. Theresultsshowsignificantimprovementsinpredictionaccuracywhentheoriginalstaticmodel is optimizedcomparedtoitsinitialperformance. Furthermore,thenewdynamicmodeldelivers even better results, demonstrating its ability to adapt to unconventional outages. Finally, the development of an algorithm based on keyoutagemilestones to align workload and workforce distribution has shown promising results, providing a more accurate representation of reality. This algorithm can also be adapted to other areas of NPP maintenance operations by adjusting its parameters using historical data of the said area. |
|---|