Accelerated Adaptive Backstepping Control Based on the FWNN for the Multiple PMSGs System with Chaotic Oscillations

This paper presents an accelerated adaptive backstepping control method based on the fuzzy wavelet neural network (FWNN) for the multiple permanent magnet synchronous generators (PMSGs) system with chaotic oscillation. Firstly, considering the influence of the feedback signals of the wind farm bus on each PMSG, a coupled mathematical model with multiple PMSGs is established. Secondly, through dynamic analysis, we find that there are chaotic oscillations in the power generation system with two PMSGs, and the dynamic response of the system is highly sensitive to the parameters and initial states. We integrate the tangent barrier function (TBF), speed function (SF), FWNN and second-order tracking differentiator (TD) into adaptive backstepping technology, and design an accelerated adaptive neural network (NN) backstepping controller. The FWNN improved by the pseudo-exponential function is used to estimate unknown functions. The SF and TBF are used to constrain tracking errors (TEs) and ensure the boundedness of state variables. The complex problem of derivation of virtual control inputs is solved by second-order TD. The stability of the closed-loop system is attested by the Lyapunov function. Our scheme not only accelerates the convergence speeds of TEs, but also effectively suppresses chaos and ensures the boundedness of all signals. Finally, a series of experimental simulations verify the practicability and robustness of the scheme.