Posters

Abstract

Motivated by the urgent need to accelerate the transition towards a cleaner energy system, the penetration of Renewable Energy Sources (RES), namely wind and solar energy, has been increasing worldwide in recent years. Still, the inherent variability and unpredictability of RES power are posing challenges to an extensive and reliable integration of renewable energy. Several studies have already investigated how to mitigate the intermittency and unpredictability of wind power by means of various sources of flexibility, such as electrochemical storage, pumped storage hydropower and other emerging technologies (e.g., electrical vehicles and hydrogen). However, less attention is paid to run-of-river hydropower plants, even if they represent a significant share of the energy mix in several European countries, such as Italy, Switzerland and France. In this study, we consider a Virtual Power Plant (VPP) integrating a cascade of run-of-river hydropower plants (CRORHS) and wind farms. The energy management of the VPP is optimized from the point of view of an aggregator participating in the Day-Ahead electricity Market (DAM). Even if the storage capacity of run-of-river hydropower plants is very limited, the water reservoirs can be exploited to reduce the market penalties caused be the unpredictable nature of wind energy. The integration of the CRORHS in the VPP significantly increases the complexity of the problem, as it calls for a detailed modelling of the temporal dynamics, water propagation between consecutive assets in the cascade, hydroelectric safety constraints, environmental requirements and the technical constraints of each hydropower plant. We formulate a mixed-integer programming problem to model the participation in the DAM of a VPP acting as a price-tacker, under the aforementioned constraints, as well as market constraints. A two-level scenario-based stochastic architecture accounts for three sources of uncertainty, namely water inflow from the river tributaries, wind power generation and market prices. Based on day-ahead forecast, the first level of the proposed architecture employs scenario-based stochastic optimization to schedule the VPP resources on the DAM. Then, based on short-term forecast, the second level employs Model Predictive Control to tackle the VPP dispatch near to real-time. Realistic forecast scenarios are available in the framework of a collaboration with the French aggregator Compagnie Nationale du Rhône. We perform several simulations to investigate the ability of the CRORHS to improve the value of wind energy in the DAM. Results show the economic advantage provided by the aggregation, compared to an uncoordinated participation of the wind farms and the CRORHS in the DAM. Then, three scenarios of wind penetration in the VPP (small, moderate and high wind penetration) are simulated to determine at which point the flexibility provided by the run-of-river hydropower plants is exhausted. Finally, two sensitivity analyses assess: 1) the impact of the considered weather year on the observed results; 2) which of the CRORHS parameters, namely reserves capacity and turbines capacity, has the higher impact on the observed VPP market revenue.