Decomposition methods for the maintenance scheduling problem in hydropower systems
Preventive maintenance is essential for the reliable operation of hydroelectric systems. However, maintenance planning is a complex activity since outages of hydro-turbines for maintenance can impact hydropower operation costs. In addition, operation costs are also determined by physical laws governing water and electricity, uncertain water precipitations and interdependencies between short-term decisions and long-term operations. In this project, we propose to develop and implement mathematical optimization techniques for solving the maintenance-scheduling problem of the hydroelectric generators in Rio Tinto, a large aluminium producer in Québec. The solution of this problem could yield significant savings for the company as well as more efficient usage of hydroelectricity in Quebec.