Abstract
This thesis addresses the integration of Variable and Renewable Energies (VRE) in the Occitanie region of France, a pertinent issue in the broader context of the global energy transition. The study revolves around the critical question of managing power system flexibility in the face of increased VRE integration.A hybrid deep-learning forecasting model is developed, combining time-series decomposition with Convolutional LSTM networks, to predict power system flexibility requirements. Subsequently, a cost-optimization model is introduced to explore flexibility sourcing from a spectrum of technologies. Last, a scenario analysis is conducted on future Occitanie power system flexibility. Using the cost-optimization model, we explore how can flexibility management can be performed under varying conditions.The findings provide valuable insights into power system flexibility, emphasizing the strategic importance of technologies like hydropower, grid, and storage. The thesis contributes significantly to the field of energy system management and flexibility, offering both theoretical frameworks and practical methodologies that are relevant to current challenges and future developments in renewable energy integration.