Research On UAV Pose Control Of Autonomous Cleaning System Of Power Grid Insulator

An insulator plays an important role in the electric power transmission.Since the insulator is exposed to long-term dusty environment,the dust is easily adsorbed on its surface to form a contamination layer.The charged particles in the contamination layer will seriously affect the insulation of the insulator.To ensure the safety of electric power transmission,it is necessary to regularly clean insulator.Up to now,the ways to clean the insulator are limited by manual cleaning,manned helicopter cleaning,and manually controlled unmanned aerial vehicle(UAV)cleaning,etc.These methods are low operational efficiency,low safety,high operational difficulty,and incomplete.In order to efficiently,autonomously,safely and fully clean the insulator,this dissertation studies that single UAV autonomously focus and clean the insulator and multiUAV cooperatively and autonomously clean the insulator while taking the advantages of UAV which is portable,reliable and able to autonomously operate.The outdoor experiments illustrate that the algorithms are efficient.The main content is divided into four parts as following:First,the dissertation studies that single UAV autonomously focus the insulator based on deep learning.In order to train the constructed model,this dissertation autonomously generates and labels the dataset.A network model DroH2oNet is proposed based on the residual network.This model takes the image as the input and the UAV velocity command as the output.Based on the labeled dataset,the Adam algorithm is used to train the model.Then,this dissertation studies the design and implementation of control system for single UAV to autonomously clean the insulator.The system consists of two subsystems:the onboard subsystem and the ground subsystem.For the onboard subsystem,the dissertation transforms the study of single UAV to autonomously clean the insulator into the study of the trajectory tracking with external disturbance.An adaptive antidisturbance trajectory tracking controller is proposed.For the ground subsystem,this dissertation proposes an auxiliary cleaning water flow curve prediction method and implements it on the ground station.Further,it displays the real-time states of the UAV,the distance of obstacles around the UAV and the amount of remaining water in the water tank.Also,the function of controlling UAV to cleaning is realized.Third,the dissertation studies that multi-UAV cooperatively clean the insulator based on formation control.Based on distance dynamics,a distributed varyingcoefficient formation controller is proposed to achieve the desired formation.The controlled is designed for two UAVs used to clean the insulator and a virtual UAV which is the insulator.The desired position of third UAV used to clean the insulator and the desired yaw angle of three UAVs are calculated according to geometrical relationships.Then,a proportional controller is designed to control the UAVs to reach the desired position and desired yaw angle.Finally,in order to illustrate the good performance of the trained model and the theoretically designed controller in the outdoors,a platform is constructed for the UAV to clean the insulator.Based on it,experiments are done outdoors.Results show that the single UAV can effectively and autonomously focus and clean the insulator.In addition,multi-UAV can also effectively and fully clean the insulator.