| The braided composite material is formed by infiltrating and curing the fiber braided structure fabric,and the braiding quality of the braided preform has a great influence on the material properties of the composite material.Braided preforms are successfully used in various fields due to their superior comprehensive performance.However,because the current braiding process still has certain shortcomings,it will lead to problems such as errors in braiding parameters,low braiding efficiency and quality inspection efficiency,and more human participation in the processing process will cause harm to the human body and high labor costs.Therefore,this paper mainly studies the planning of robot traction mandrel trajectory and machine vision for braiding angle detection to improve the efficiency and accuracy of braiding and braiding angle detection.This paper mainly takes carbon fiber braided preforms as the research object.Aiming at the problems of difficult design of processing technology,unstable fabric surface quality,low efficiency of braiding angle detection and inability to automate during the braiding process of mandrel with special-shaped space structure,a robot automatic traction and the trajectory planning method of real-time control of the speed of the core mold is used to braid the preform,and the braiding angle of the formed fabric is detected by machine vision.In the trajectory planning process,first discretize the spatially shaped mandrel,then obtain the position and attitude data of the robot tool center point through the homogeneous coordinate transformation algorithm,and calculate the robot traction speed and time through the braiding point dynamic compensation algorithm,to realize the automatic compensation of the braiding angle;finally,the braiding angle of the fabric surface is detected by the discrete angle mean gray algorithm.Firstly,aiming at the detection of spatially shaped tubular fabrics and the mandrel traction planning,the requirements for the traction and visual inspection process of the preform mandrel were established,and the hardware selection of the machine vision inspection system was carried out according to the structure of the knitting machine and the characteristics of the braiding process.type,and the fixture design and installation scheme verification are carried out through simulation software.Secondly,the planning problem of trajectory and speed of robot pulling mandrel in braiding process is studied.Establish the kinematics mathematical model of the braided traction mandrel robot,and derive the kinematics forward and reverse solutions of the braided traction robot.The trajectory planning of the robot end effector is carried out by simplifying the braiding model,and the traction speed and time are adjusted in combination with the dynamic compensation algorithm of the braiding point.Realize automatic compensation of braiding angle.Then,the problem of braiding angle visual detection is analyzed,and a discrete angle mean gray-scale algorithm for calculating the average braiding angle is proposed.The main steps of braiding angle image processing mainly include image preprocessing,edge detection,two-dimensional discrete Fourier transform and other steps;the realization of the algorithm is carried out by combining knowledge and the image processing flow is verified by combining MATLAB programming and the feasibility of the discrete angle mean gray algorithm.Finally,a visual simulation platform was built based on the ROS system,and the simulation of the mandrel traction trajectory process was carried out;the braiding of space-shaped tubular preforms was carried out through experiments.The experimental results show that the planning method can effectively reduce the preform.The braiding angle error of the formed body is within ±2°,which can improve the quality of the preform in actual production;finally,the braiding angle of different fabric surface images is measured by the discrete angle mean gray algorithm proposed in this paper,and the average braiding angle detection is obtained.The error is 0.2403°,which significantly improves the accuracy and stability of braiding angle detection. |
