Probabilistic Graphical Model-Based Operational Reliability-Centric Design of Offshore Wind Farm Feeder Layouts

The rapid expansion of offshore wind energy necessitates robust and cost-effective electrical collector system (ECS) designs that prioritize lifetime operational reliability. Traditional optimization approaches often simplify reliability considerations or fail to holistically integrate them with economic and technical constraints. This paper introduces a novel, two-stage optimization framework for offshore wind farm (OWF) ECS planning that systematically incorporates reliability. The first stage employs Mixed-Integer Linear Programming (MILP) to determine an optimal radial network topology, considering linearized reliability approximations and geographical constraints. The second stage enhances this design by strategically placing tie-lines using a Mixed-Integer Quadratically Constrained Program (MIQCP). This stage leverages a dynamic-aware adaptation of Multi-Source Multi-Terminal Network Reliability (MSMT-NR) assessment, with its inherent nonlinear equations successfully transformed into a solvable MIQCP form for loopy networks. A benchmark case study demonstrates the framework’s efficacy, illustrating how increasing the emphasis on reliability leads to more distributed and interconnected network topologies, effectively balancing investment costs against enhanced system resilience.