Research And Exploration Of Substation Intelligent Inspection Robot

Intelligent Today,substation,as an important part of the power system,is the core hub for ensuring the safe and stable supply of power quality.As the core link of smart grid,intelligent substation has attracted much attention.The use of intelligent inspection robots to build smart power stations is a research hotspot.The development of popular inspection robots is the driving force for promoting substation intelligence.According to this,this paper studies the two core parts of the power system and vision system of the inspection robot;adopts the universal embedded chip as the main control chip to facilitate mass production and wide-scale promotion;design the dual-sensing vision system to HD camera and The infrared light sensing component is a visual sensing component,and the two complement each other to improve the image acquisition accuracy.The motion control algorithm and the visual algorithm are designed to develop a patrol robot prototype.The specific research contents of this paper are as follows:First of all,in the case of reading many domestic and foreign literatures,after fully understanding the development status of domestic and international inspection robots,in order to further popularize the application of inspection robots in power systems,the concept of hierarchical management of substation robots was proposed to popularize robots.As the research content and research basis and significance of this topic,the key technologies of the popular inspection robot for substation are expounded.Secondly,the technical analysis of the inspection robot system is carried out.The functions of each part of the inspection robot are explained by modularization.The overall mechanical 2D and 3D models of the robot are designed and designed by Auto CAD software and Solidworks software.The overall layout was planned;the permanent magnet motor and the conventional motor were compared according to the spatial layout plan,combined with the existing experimental conditions,and finally a more mature scheme was selected for driving.Then,the software control system of the inspection robot is analyzed.Machine vision uses the Python programming language to build a shallow convolutional neural network model in Keras,and compares the recognition accuracy with the traditional neural network in digital recognition.The analysis verifies that the convolutional neural network is superior to the traditional neural network in digital recognition.Finally,based on the laboratory conditions,the prototype of the inspection robot was built,and the experimental platform of the inspection robot was built.The stability experiment of the robot indoor cruise experiment,positioning experiment and climbing performance was completed,and the PID control parameters were driven.The tuning determines the stability of the operation of each sensor module and human-computer interaction interface,and processes the experimental data.Experiments show that the robot designed in this paper can basically meet the requirements of independent tracing,remote control,real-time image acquisition,etc.,laying a solid theoretical and practical foundation for the development of popular inspection robots.