Optimizing hydro power plant maintenance planning
The School of Engineering FHNW has developed a method to support operators of AXPO hydro power plants.
Objectives
Identifying the best maintenance scenarios for hydro power plants with the help of linear programming.
Background
Operating hydro power plants in the most efficient way is crucial for economic reasons and to guarantee electricity supply security. The planning complexity of individual plant maintenance optimization arises from diverse configurations of generators, pumps, bypasses and reservoirs and large seasonal variability of hydrological flows and electricity prices. Revenue loss minimization for maintenance windows is a computationally challenging task for which the standard Stochastic Dynamic Programming approaches are limiting in terms of computation time.
Results
The Institute of Data Science FHNW successfully develops an analytical solution from scratch to assess maintenance scenarios of individual hydro power plants. Leveraging a linear programming approach the implemented model reaches predictive accuracy comparable to Stochastic Dynamic Programming for a majority of plants and scenarios. Given the dramatically reduced computational costs our solution is well suited for an interactive planning tool.
Information
Client | AXPO |
Execution | FHNW Institute for Data Science |
Duration | 6 months |
Team | Dr. Michael Graber, Prof. Dr. Daniel Perruchoud, Simon Beck |
Contact
Head of FHNW Institute for Data Science