A scalable interior-point framework for stochastic N-1 secure pre-dispatch in large-scale power systems

The secure and economic operation of large-scale power systems under uncertainty requires optimization frameworks capable of jointly addressing demand variability and reliability constraints. This paper considers a two-stage stochastic pre-dispatch formulation that explicitly incorporates the N-1 security criterion, allowing pre-dispatch decisions to remain feasible under both stochastic demand realizations and single-component outages. The resulting large-scale optimization problem is addressed through a structure exploiting primal-dual interior-point approach that leverages the block-sparse architecture induced by scenarios and contingencies. By combining a Schur-complement based decomposition with parallel solution of scenario contingency subproblems, the proposed framework achieves scalable and robust performance. A lightweight adaptive predictor-corrector mechanism is incorporated to enhance numerical stability in ill-conditioned regimes without increasing algorithmic complexity. Computational experiments on benchmark and large scale transmission systems demonstrate that the proposed approach reliably handles hundreds of scenarios and contingencies while preserving feasibility and predictable scaling behavior. Overall, the results show that stochastic N-1 secure pre-dispatch can be addressed within a computationally viable framework when appropriate structural exploitation is employed, providing a practical bridge between stochastic optimization models and security-constrained power system operation.