Intelligent State-Constrained Control for Servo Valves via Neural Network-Based Real-Time Compensation

Rotary direct-drive servo valves (RDDSVs) have gained significant attention in high-performance electro-hydraulic servo systems due to their compact structure, rapid dynamic response, and high power density. However, improving the transient performance and steady-state accuracy of RDDSVs remains a challenge, primarily owing to inherent strong nonlinearities and disturbances characterized by high-frequency fluctuations and unmodeled uncertainties. To address these issues, this paper proposes an intelligent state-constrained control strategy with neural network-based real-time compensation for RDDSVs. Specifically, a nonlinear constraint function is introduced to directly restrict the range of state variables, thereby enhancing the system’s transient response. Subsequently, the universal approximation property of adaptive neural networks is exploited to estimate unmodeled disturbances, which significantly improves steady-state precision. Furthermore, nonlinear filtering technology is integrated to mitigate the computational burden on the controller while enhancing overall robustness. The stability of the closed-loop system is rigorously proven using Lyapunov theory. Finally, comparative simulations are carefully conducted to apply different control algorithms. The results validate the effectiveness and superiority of the proposed control algorithm.