RBF-FO-PID control method of transmission line fault diagnosis robot under wind load vibration

Aiming at the problem that the transmission line fault Diagnosis robot is susceptible to wind load during operation, which leads to the robot’s deflection, poor stability and low quality of operational images, a wind vibration suppression control strategy for the transmission line fault diagnosis robot based on RBF-FOPID (Radial Basis Function-Fractional Order Proportional Integral Derivative) is proposed. The structure utilizes a momentum wheel device based on angular momentum conservation to achieve wind load balancing, and a RBF-FOPID control method to achieve momentum wheel wind load vibration suppression. Firstly, the wind load model under the working height of the robot is established by Kaimal wind spectrum and harmonic superposition method combined with the robot size data, and the dynamics model of the momentum wheel and the robot is established by Lagrangian method to obtain the dynamics model of the equilibrium system under the wind load, based on which, an analysis of the influence of wind load on the robot’s pose is conducted to establish a coupled relationship model among wind force, robot deflection angle, and momentum wheel driving torque. Secondly, a closed-loop control system for robotic wind load suppression control was designed based on Fractional Order Proportional Integral Derivative (FOPID) and the parameters of FOPID are optimized using Radial Basis Function (RBF) neural networks. Finally, a co-simulation platform was established using Matlab-Adams, and after parameter optimization of the controller, the optimized control strategy was implemented in the actual system. Different wind loads were applied to the robot by varying wind velocities, and real-time online control was implemented using RBF-FOPID to validate the effectiveness of the proposed vibration suppression strategy. Meanwhile, the robot field line experiments are completed and the proposed hybrid RBF-FOPID controller is verified to have faster response speed and better robustness by comparing the control performance of different algorithms on real mechanical systems.