| Wear plate is a kind of metal plate with wear-resistance,which is one of the main components of the pumped storage unit in hydropower plants.Its function is to resist the scouring and wear of high-speed water flow during the operation of the pumped storage unit,which can prolong the service life of the unit and reduce maintenance costs.The wear plate is connected to the base plate substrate by welding with a large number of plug welding holes.Currently,the welding of plug weld holes in wear plate still uses manual welding,which has problems such as high risk of manual operation,and unstable welding quality.In addition,it requires high technical requirements for welding operator.Therefore,an automated welding technology is needed to address these problems.This article focuses on the automated welding of plug welding holes on wear plates.Firstly this study investigated the welding problem of a single plug weld hole in a wear plate and proposed a "layered swinging" welding strategy suitable for welding robots.Then,addressing the problem of positioning weld holes of the whole wear plate,a porous positioning method based on drawing information was proposed,with an initial positioning accuracy of(27)4mm.To improve the positioning accuracy of the plug weld hole,a visual sensing system based on structured light was constructed,and the structure of the sensor was designed according to the actual situation.Moreover,the hardware of the system was analyzed and selected.Secondly,the model of the sensor based on structured light was established,and the mapping process from two-dimensional(2d)pixel coordinates to three-dimensional(3d)world coordinates was analyzed in detail.The calibration method was described in detail and the accuracy was verified.And,The error of the structure optical sensing model was≤0.5 mm.Based on OpenCV image processing library,an image processing algorithm was designed to locate the center of the top circle of the plug hole of an antiwear plate by utilizing the deformation of structured light at the edges of the plug hole.Start by setting four region of interest(ROI)to reduce redundant information in the image.Then,a new algorithm for extracting the center line of structured light,Gauss-extremum algorithm,is proposed,which refined the structured light.Next,The RDP algorithm and the least squares method is used to segment and fit the centerline.Lastly,the pixel coordinates of the feature points are mapped to the world coordinates of the robot by using the structured light vision sensing model,and the center position and radius of the top circle of plug weld hole are calculated by using the spatial circle fitting method.Experimental results show that the error of x and y direction is(27)0.6 mm,and the error of z direction is(27)1.1mm,and the recognition speed is 20 Hz.Finally,an experimental platform was built based on the Kuka robot,and the Automatic control software was independently developed to experimentally verify the research content of this paper.The experimental results showed that the proposed solution in this paper can achieve stable welding in different environments.The welding time of each plug weld hole is within 14 minutes with stable welding process.Compared to manual welding,the proposed method offers higher welding quality,simplified operation,reduced technical requirements,and minimized operational risks for operators.To a certain extent,the automation of plug welding for wearresistant plates has been achieved. |
