A Two-Layer Hybrid Optimization Approach for Large-Scale Offshore Wind Farm Collector System Planning

Abstract

Constructing large-scale offshore wind farms (OWFs) has become the main direction of utilizing wind power to help realize the energy transformation. Traditionally, the planning of the OWF collector system would rely on either heuristic or deterministic optimization algorithms, which, respectively, suffer from unstable outputs and a lack of freedom in searching for a globally optimal solution. This article innovatively designs a hybrid optimization approach combining algorithms in these two categories to achieve a balance between improved economic efficiency and stable outputs. The whole design consists of two layers of hybrid optimizations. The outer layer is to partition wind turbines (WTs) into groups, where each group is allocated with an offshore substation with the optimized location for power collection and transmission. This partitioning and locating optimization is solved through a combination of the deterministic fuzzy C-means clustering method and the genetic algorithm (GA). The inner layer is to arrange optimal connections using proper cable ratings among WTs within each group, and GA is properly integrated into the deterministic two-phase Clark and Wright's saving algorithm to solve the problem. The collector system planning, in this article, concerns both the investment cost and the long-term power-loss cost. The former consists of the networks of internal medium voltage and the external high voltage, which collect the power from WTs and transmit it to the onshore grid. The proposed design is tested on a benchmark OWF collector system, and the test result verifies its achievements in higher economic efficiency with stable outputs.