Pneumatic Valve Position Control of Third-order LADRC Based on RBF Neural Network

Pneumatic control valves are used extensively in industrial processes. Due to the valve's nonlinear characteristics and the impact of time-varying disturbances, the PID control strategy struggles to achieve the desired levels of precision and robustness in control systems. This paper introduces a control scheme combining a third-order Linear Active Disturbance Rejection Controller (LADRC) with a Radial Basis Function (RBF) neural network. Firstly, the third-order transfer function model of the pneumatic control valve is established. Then a third-order LADRC is designed, and the Jacobian information of the valve position is recognized online by RBF neural network to realize the online tuning of the control parameters. Finally, the results of simulation and experimental show that the proposed method boasts excellent anti-interference capability and superior control precision.