Sensitivity-Based Sequential Linear Approximation for Solving Multi-Period Security-Constrained Optimal Power Flow in Real-Time

Efficient methods for solving close to real-time multi-period security-constrained optimal power flow (MPSCOPF) problems are crucial for transmission system operators to compute redispatch actions that ensure system security in renewable-dominated power systems. This paper proposes a novel tractable methodology with acceptable accuracy. It involves solving sequential linear approximations of the problem with improved quality within an adaptive trust region. The linearization of the nonlinear network constraints is performed using sensitivities derived from the Jacobian matrix around the operating points at hand in each time period. The performance of the proposed methodology is evaluated on a 60-bus Nordic test system and benchmarked against the exact nonlinear model and an existing popular linear model from the literature. Numerical results show that the proposed method outperforms competing alternatives by achieving significantly better accuracy in resolving security violations, while maintaining a comparable computational time—making it well suited for real-time applications with mild violations of security constraints.