Scholarlink

<p>Floating offshore wind farms (FOWFs) provide access to deepwater wind resources beyond the reach of fixed-bottom technology. However, large-scale deployment of mooring systems remains constrained by high cost, environmental impact, and operational complexity, accounting for up to 30% of total capital expenditure. Although conventional radial layouts are easy to certify, they require numerous anchors and heavy materials, increasing cost and disturbing the seabed. Shared-anchor mooring configurations can achieve substantial material and anchor savings, yet they introduce nonlinear dynamic coupling and fatigue risks that existing design frameworks rarely address. This paper presents a prototype open-source Digital Twin (DT) framework for shared mooring arrays, combining multifidelity physics-based simulations (OpenFAST, MoorPy, FAST. Farm) with AI-driven forecasting. Based on 84 years of hindcast ocean data, long short-term memory (LSTM) neural networks predict real-time tension and fatigue accumulation. The framework demonstrates feasibility through combined simulations, probabilistic extreme-value analysis, and scaled hardware-in-the-loop (HIL) tests. It provides array-level prognostic insight by integrating sensing, anomaly detection, and adaptive decision logic. All simulation code, data sets, and plotting scripts are released openly to ensure transparency and reproducibility. Applied to a 10-turbine case study at the São Miguel (Azores) site, the framework achieved anchor reductions of approximately 40% and material savings of 34% relative to radial baselines, while maintaining API and DNV-compliant safety margins. The AI component provides actionable early warnings for preventive interventions under extreme sea states, extending anomaly lead times by several hours. The main contributions are: i) a prototype, open-source DT framework combining physics-based modeling and AI forecasting for shared moorings; ii) the development of a fatigue-aware anomaly detection and adaptive control scheme; and iii) quantitative evidence that deepwater costs, environmental footprint, and operational risk can all be reduced through intelligent shared-mooring design.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid Physics–AI Digital Twin Framework for Shared Mooring Systems in Deepwater Floating Offshore Wind Farms

  • 期刊论文

  • 发表日期：

    2026年3月31日

查看全文
  • Mahtab Shahin,
  • Ahmed Nagi Nasr,
  • Avleen Malhi,
  • Sanja Bauk,
  • Osiris Valdez Banda,
  • Pentti Kujala,
  • Ralf-Martin Soe,
  • Shan Wang

摘要

Floating offshore wind farms (FOWFs) provide access to deepwater wind resources beyond the reach of fixed-bottom technology. However, large-scale deployment of mooring systems remains constrained by high cost, environmental impact, and operational complexity, accounting for up to 30% of total capital expenditure. Although conventional radial layouts are easy to certify, they require numerous anchors and heavy materials, increasing cost and disturbing the seabed. Shared-anchor mooring configurations can achieve substantial material and anchor savings, yet they introduce nonlinear dynamic coupling and fatigue risks that existing design frameworks rarely address. This paper presents a prototype open-source Digital Twin (DT) framework for shared mooring arrays, combining multifidelity physics-based simulations (OpenFAST, MoorPy, FAST. Farm) with AI-driven forecasting. Based on 84 years of hindcast ocean data, long short-term memory (LSTM) neural networks predict real-time tension and fatigue accumulation. The framework demonstrates feasibility through combined simulations, probabilistic extreme-value analysis, and scaled hardware-in-the-loop (HIL) tests. It provides array-level prognostic insight by integrating sensing, anomaly detection, and adaptive decision logic. All simulation code, data sets, and plotting scripts are released openly to ensure transparency and reproducibility. Applied to a 10-turbine case study at the São Miguel (Azores) site, the framework achieved anchor reductions of approximately 40% and material savings of 34% relative to radial baselines, while maintaining API and DNV-compliant safety margins. The AI component provides actionable early warnings for preventive interventions under extreme sea states, extending anomaly lead times by several hours. The main contributions are: i) a prototype, open-source DT framework combining physics-based modeling and AI forecasting for shared moorings; ii) the development of a fatigue-aware anomaly detection and adaptive control scheme; and iii) quantitative evidence that deepwater costs, environmental footprint, and operational risk can all be reduced through intelligent shared-mooring design.