Solver-Free Heuristics to Retrieve Feasible Points for Offshore Wind Farm Collection System

TL;DR

This paper introduces a series of solver-free heuristics for designing offshore wind farm collection systems, capable of quickly finding feasible solutions while respecting key constraints, and improving solver efficiency.

Contribution

It presents a novel multi-step heuristic workflow that handles typical constraints and repairs infeasible solutions, enhancing offshore wind farm network design.

Findings

Feasible solutions found within seconds.

Heuristics improve solver convergence times.

Workflow effectively manages real-world offshore wind farm constraints.

Abstract

A set of solver-free heuristics for the offshore wind collection system problem are presented. Currently, methods of this type are not able to cope with typical constraints, and most of their variations minimize only for accumulated cable length. The first algorithm is a two-steps decision process, where the output is the design of a tree network satisfying cable thermal limits constraints, but vulnerable to violate planarity constraints. Subsequently, a cable crossings repair heuristic is introduced in order to fix infeasible points from the first heuristic. Finally, a refining heuristic (negative cycle cancelling refining heuristic) takes over to improve feasible points. The latter iteratively swaps cables, intending to find cycles with negative costs that will lead to investment savings. The sequence of heuristics supports the most important restrictions of the problem. The applicability of the workflow is empirically demonstrated by means of a set of large-scale real-world offshore wind farms. The numerical results indicate that: (i) feasible points can be retrieved in computing times in order of seconds, and (ii) warm-starting can help solvers to converge significantly faster for problems with solution time in order of several hours.