Fuzzy-type 2 fractional fault tolerant adaptive controller for wind turbine based on adaptive RBF neural network observer

Nowadays, wind turbines have become one of the most significant sources of clean electricity generation because of their renewability and being free. The wind turbine control techniques must be robust to both effective wind speed fluctuations and modeling uncertainty. In the presence of fault effects, the control system should decrease the fault effects of the control process on the system structure and the generated power characteristics by maintaining an acceptable level of performance. This paper presents a novel adaptive controller with type 2 fuzzy fault tolerance based on the Radial Basis Functions (RBF) adaptive neural network observer that guarantees the system reliability in the condition that the speed sensor and generator torque bias are faulty. Utilizing fast nonsingular terminal sliding surface make high speed tracking and remove the singularity problem associated with conventional terminal sliding mode control. Also utilizing the fractional order operator in sliding surface make more degree of freedom and also by utilizing of long-term memory characteristic of fractional operator, more stability and robustness will be made and also the chattering effect will be decreased. Subsequently, fuzzy logic type 2 is employed to derive the switching control law, ensuring stability and eliminating chattering phenomenon. Additionally, an adaptive radial basis functions is utilized to estimate faults and uncertainties in the system. For better performance and fair comparison between the proposed method and the other methods, all the coefficients for all controllers is adjusted with ant colony optimization. The results of the suggested method are compared to those of a classic adaptive sliding mode controller. The tracking error for the proposed method in under normal conditions (without fault) has been approximately 0.04. The performance of the proposed method under two faulty conditions scenario, compared to ASMC, has shown an average superiority of 2.5 times under the application of 0.8 fault and a superiority of 1.5 times under the application of 1.2 fault. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.