Social Hierarchy-based Distributed Economic Model Predictive Control of Floating Offshore Wind Farms

TL;DR

This paper presents a social hierarchy-based distributed economic model predictive control approach for floating offshore wind farms, utilizing neural networks and turbine repositioning to maximize power output.

Contribution

It introduces a novel control algorithm combining social hierarchy concepts with neural network-based predictions for real-time turbine repositioning in floating wind farms.

Findings

20% increase in energy production with YITuR

Neural network predictions enable fast dynamic optimization

Performance decreases with increased wind variability

Abstract

This paper implements a recently developed social hierarchy-based distributed economic model predictive control (DEMPC) algorithm in floating offshore wind farms for the purpose of power maximization. The controller achieves this objective using the concept of yaw and induction-based turbine repositioning (YITuR), which minimizes the overlap areas between adjacent floating wind turbine rotors in real-time to minimize the wake effect. Floating wind farm dynamics and performance are predicted numerically using FOWFSim-Dyn. To ensure fast decision-making by the DEMPC algorithm, feed-forward neural networks are used to estimate floating wind turbine dynamics during the process of dynamic optimization. For simulated wind farms with layouts ranging from 1-by-2 to 1-by-5, an increase of 20% in energy production is predicted when using YITuR instead of greedy operation. Increased variability in wind speed and direction is also studied and is shown to diminish controller performance due to rising errors in neural network predictions.