How does hydrological inflow variability affect reservoir management under alternative electricity market designs within a stylized hydro-economic representation of the Norwegian power system?

Details

Supervisors

Willems, Bert

Faculty

Louvain School of Management

Degree label

Master [120] : ingénieur de gestion, à finalité spécialisée

Abstract

enNorway’s electricity system relies heavily on hydropower, reservoir storage and a transmission network with persistent internal constraints. This thesis examines how hydrological inflow variability affects reservoir management under two alternative market-design representations: a physically constrained nodal benchmark and a two-stage zonal design followed by nodal redispatch. In order to address this question, this thesis develops a stylized operational model of the Norwegian power system, based on a lossless DC optimal power flow formulation implemented in Python for Power System Analysis (PyPSA). The model includes reservoir hydropower, run-of-river generation, transmission constraints, spill and load shedding. The model is run over the year 2023 at hourly resolution. A rolling-horizon procedure is used to represent sequential reservoir decisions. Each optimization covers a 48-hour window, while only the first 24 hours are implemented. Reservoir storage levels are then carried forward as the initial conditions for the next window. The analysis is based on five hydrological inflow scenarios. These are constructed by scaling the 2023 inflow profile by -15%, -10%, 0%, +10% and +15%. The results suggest that hydrological conditions mainly affect the system’s available water margin. Market design changes the way this water margin is used. In the nodal benchmark, transmission constraints are considered directly, storage decisions and prices are determined. In the zonal market stage, internal scarcity signals are more clustered, which leads to deeper seasonal reservoir drawdowns. Redispatch then corrects part of this outcome and brings aggregate storage, spill and load-shedding results closer to the nodal benchmark, but this correction requires substantial daily and zone-specific adjustments. Taken together, these results show that reservoir management in a hydro-dominated system cannot be understood only by total inflow availability. The timing of water, storage constraints and the ability of the network to deliver electricity across space all influence how scarcity and surplus appear in the system