Research On Motion Control Of Water Wall Climbing Robot Based On Neural Network

With the development of China’s economy and the growing urban population,demand for power supply has increased significantly.Fuel boiler is the energy and power conversion device of thermal power plant which burns various fuels in combustion chamber.The high temperature makes the liquid in the water wall pipe evaporate into gas,which is supplied to the steam turbine generator matched with the boiler for power generation.However,corrosion and wear of water wall tubes under high temperature conditions has always been a major issue in power system.The traditional method of manually detecting the thickness of water wall is adopted,which is inefficient and dangerous.Therefore,wall climbing robots have been wildly applied in power plant to test the thickness of water wall detection.Due to the particularity of wall climbing robot motion and the influence of working environment,its path tracking control has become a research difficulty.In view of the above background,this paper studies the motion control system based on the wall climbing robot independently developed in the laboratory.Consulting relevant references in domestic and abroad,By consulting the relevant literature at home and abroad and the description and functional requirements of the working environment of the water-cooled wall climbing robot,the wall adsorption mode,moving mode,driving mode and wall thickness detection mode of the wall climbing robot are studied to determine the rationality of the scheme.According to the actual working conditions and task requirements of water wall climbing robot,consult the relevant domestic and foreign literature.The wall adsorption mode,motion mode,driving mode and wall thickness detection mode of the wall climbing robot are studied in order to determine the rationality of the scheme.This paper selects the two-level multi CPU distributed control system of PC and MCU,which is based on the task requirements of the wall climbing robot.The PC is centered on the personal computer.The PC uses a computer interaction interface designed by Visual Studio 2010 platform to send and receive commands with the MCU by means of wireless serial communication.The hardware circuit of the MCU includes power,communication,camera,Ultrasonic Thickness Gauge,crawling control,etc.The MCU applies STM32F103RCT6 microprocessor to receive commands by the PC and carry out tasks.The reliability of the control system in hardware and the rationality of software are verified through experiments.Because the wall robot may offset deviate from the straight path in the work,the kinematics and dynamics model of the wall climbing robot is established,the kinematics and dynamics model of the wall climbing robot is established.The results show that the speed and deflection angle of the wall climbing robot can be adjusted by controlling the speed of the left and right wheels.This paper presents a Double Loop Control strategy of external path tracking control and internal speed control.Both controllers use PID Neural Network algorithm to design.Compared with the traditional PID control algorithm,the path tracking controller has more advantages in the accuracy and response speed of path tracking.The speed tracking controller is better for the dynamic response performance of the control system,with short adjustment time,less mistakes,smaller overshoot,strong anti-interference ability and preferable robustness.