Enhanced day-ahead market operations: incorporating renewable energy in AC network-constrained unit commitment

The transition to sustainable energy systems has led to a growing reliance on renewable energy sources (RES), introducing significant uncertainty into electricity market operations. This thesis addresses the complex problem of AC Network-Constrained Stochastic Unit Commitment (AC-NCSUC), which integrates unit commitment, AC power flow constraints, and stochastic modelling of RES variability into a unified optimisation framework. The AC-NCSUC problem is formulated as a non-convex Stochastic Mixed-Integer Nonlinear Programming (MINLP) model, which poses computational challenges due to its non-convex and combinatorial nature. To overcome these limitations, this thesis proposes a novel Hierarchical Sequential Mixed-Integer Linear Programming (H-SMILP) methodology. The proposed approach decomposes the problem into three iterative phases: feasibility restoration, objective improvement, and nonlinear constraint enforcement, each designed to progressively enhance solution accuracy and feasibility. Moreover, this thesis provides an in-depth analysis of the problem, consisting of three sub-problems that have been extensively studied in the field of optimisation. A careful analysis is provided, focusing in particular on the problem of power flow constraints, including a study of different mathematical formulations as well as various relaxation and approximation techniques aimed at improving tractability and scalability. To address RES uncertainty, the model incorporates scenario-based stochastic programming, which captures temporal variability and enables robust planning under uncertainty. Numerical experiments on modified IEEE test cases demonstrate that H-SMILP delivers high-quality, near-feasible solutions with improved computational efficiency. The present study validates the effectiveness of H-SMILP in solving large-scale, uncertainty-aware power system optimisation problems. This thesis presents a practical and scalable tool for Transmission System Operators (TSOs) to enhance the reliability and efficiency of day-ahead electricity market operations in the context of high RES integration.