RBF neural network compensation-based adaptive control for lift-feedback system of ship fin stabilizers to improve anti-rolling effect

For the conventional fin stabilizer, its practical anti-rolling effect is difficult to reach the theoretical design. In order to solve this problem, the reasons of system error are analyzed, and then the uncertain factors of system are discussed. A new lift-feedback control system is designed, and the improved advantage is analyzed. For reducing the nonlinear error of feedback, a new fin-axis mechanism is designed, which can directly detect actual lift. However, lift is difficult to detect, which could be transformed into displacement easily measured through deduction of Timoshenko beam theory. Then quantitative relation is established between lift and displacement, and the main influencing factors are analyzed. The effectiveness of lift-feedback control system is verified by fin-axis strain simulation and PID control simulation. In view of the fin stabilizer system with uncertainties, the compensation of system is needed. The RBF neural network is applied to estimate the uncertainties. Lyapunov equation is adopted to satisfy the stability conditions. The adaptive law is designed to improve the system accuracy. RBF neural network can identify system model error on-line, which can guarantee the stability of closed-loop system. The simulations demonstrate the design of control strategy can effectively improve anti-rolling effect.