Research On Robotic MAG Adaptive Welding And Planning Based On Laser Vision Sensor

There are a large number of thick plate workpiece welding need in the marine equipment,engineering machinery and bridge structure industry.Traditional manual welding corresponds to high labor intensity,harsh operating environment.Therefore,an alternative method to arc welding robots intelligent is needed.At present,most of welding robots are "teachand-reproduce" type robots.The pre-set path operation can not solve the welding conditions,such as assembly gap and mismatch which may cause bad welding quality.In this paper,a laser vision sensor for welding is developed for the application of arc welding robot in the process of plate welding,and a software communication platform between vision sensor and FANUC robot controller is designed.The result shows that the arc welding robot acquires better application of plate automation and intelligent level.First,the calibration of the visual sensor system is finished,including the calibration of the camera and the parameters of the camera and then based on the projection of the line structure,the paper takes the optical calibration.This paper gives the height and the laser stripes position relationship formula,and finally takes the robot hand-eye calibration.Then,the welding seam image is analyzed synthetically,and the characteristic parameters of the weld bevels are extracted by the steps of pretreatment,threshold segmentation,centerline extraction and straight fitting.Then,based on the principle of line structure optical calibration,the data are transferred to the robot in real time using the communication platform.Taking into account the process of real-time image data acquisition fluctuations,we use smoothing filtering to interfere the fluctuations.And it is proved that the algorithm can acquire data steadily and accurately under the condition of torch swing.Finally,the robot weld tracking process test and algorithm verification are carried out on the welding test plate of the gradient gap.In this paper,the welding parameters,including welding speed and torch swing,are calculated by analytical model,and the attitude adjustment method and control strategy of torch attitude in multi-pass welding process are established.Finally,this paper uses the workpiece thickness of 20 mm groove gradient to carry out multi-channel planning and adaptive welding,which can adjust the welding arm swing and welding speed.And the threedimensional topography of each layer of weld bead is visualized.The validity of the above-mentioned variable-gap adaptive programming method is verified from the theoretical level,which provides theoretical support and algorithm verification for similar welding plate.