Hybrid ANFIS-OTC strategy for MPPT operation and mechanical vibration reduction in offshore wind turbines

Abstract

This paper presents a hybrid control strategy for offshore wind turbines that combines an Adaptive Neuro-Fuzzy Inference System (ANFIS) with conventional Optimal Torque Control (OTC). The main objective is to maintain maximum power point tracking (MPPT) efficiency while minimizing mechanical vibrations at the tower-top structure. While standard OTC extracts aerodynamic power effectively, it ignores structural loads. To address this limitation, an ANFIS-based adaptive gain scheduler is developed, where the controller is trained offline using a dataset generated from a multi-policy heuristic control framework that combines the OTC law with vibration-sensitive modifications. The ANFIS uses generator speed and tower-top acceleration as inputs to estimate an adaptive gain factor that modulates the OTC torque reference. In real-time operation, this gain scales the conventional OTC torque command, enabling a trade-off between MPPT performance and structural vibration mitigation. OpenFAST simulations demonstrate that the proposed controller successfully reduces mechanical load fluctuations while maintaining high MPPT performance compared to the traditional approach.