WIND TURBINES AND FARMS USING AI FOR ENERGY STORAGE

Abstract

The article provides a comprehensive overview of wind turbine technologies, including both onshore and offshore systems, focusing on aspects such as grid integration, trends in the levelized cost of energy (LCOE), and applications of artificial intelligence (AI) in areas like forecasting, diagnostics, control, and digital twins from 2015 to 2025. The importance of energy storage solutions, such as battery energy storage system (BESS), liquid air energy storage (LAES), compressed air energy storage (CAES), power-to-x/H₂ (P2X/H₂), and pumped storage, is also discussed in relation to enhancing system flexibility and increasing the market value of wind energy. The text identifies research gaps and priorities for research and development. Wind energy remains cost-competitive, with onshore wind continuing to be among the lowest cost options for new energy capacity (weighted LCOE ≈ $0.034/kWh globally). Cost variations are mainly influenced by capacity factor (CF), the weighted average cost of capital (WACC), and execution risks. The scale of capacity and maturity of the supply chain are improving, particularly for offshore wind (12–20+ MW/turbine). However, costs and timelines are sensitive to logistics, weather conditions, and the availability of installation vessels. High-voltage direct current (HVDC) technology is becoming the standard for exporting offshore power, and transitioning to meshed/multi-terminal systems require interoperability across different vendors (InterOPERA) and standardized FAT/SAT/HIL testing. Ensuring grid-forming capabilities and converter compatibility is essential for the stability of low-inertia power systems, necessitating harmonized testing profiles and certification requirements at the wind turbine generator (WTG)/wind power plant (WPP)/HVDC levels. Artificial intelligence is increasingly being utilized in operations, achieving optimal results by combining forecasting methods (long short-term memory networks (LSTM)/ gated recurrent units (GRU)/Transformers) with fault detection and diagnosis (FDD)/ remaining useful life (RUL) predictions and wake steering techniques. These results are measured not only by the change in annual energy production but also by their impact on service profiles and network constraints.

Current and evolving ENTSO-E requirements include fault ride-through capabilities (low-voltage ride-through (LVRT) and high-voltage ride-through (HVRT)), active power (P) and reactive power (Q) control (Volt-VAR/Volt-Watt), ramp-rate constraints, primary and secondary frequency control (FCR/FRR), as well as limits on harmonics, flicker, and distortion. Converter-based generation units must ensure low short-circuit power (weak grid) and during disturbances, highlighting the increasing significance of dynamic models (root-mean-square simulation (RMS)/electromagnetic transient simulation (EMT)) and compliance validation during both commissioning and operation.