Research On Welding Robot Technology In Tube Wall Tube Plate

In the waste heat boiler industry,there are a large number of tube wall tubes welded on the inner surface of the cylinder.This is a welding method in which the process parameters and the attitude of the torch are complicated and difficult to realize.The welding worker must pass through the 300 x 400mm manhole to enter the interior of the narrow boiler barrel and maintain the bent or semi-twisted position to complete the welding operation.Therefore,the welding workers are labor-intensive,technically demanding,and the space is relatively closed to harm the body.This makes the waste heat boiler industry-related talent gaps large and urgently need to improve the automation level.The main diff-iculties in welding the inner tube wall of the cylinder are as follows:1.It belongs to all-position welding.The weld bead is distributed on the inner wall of the boiler barrel.It also includes flat welding,vertical welding,overhead welding and transition welding on one workpiece.Need to adjust according to the weld position;Second,the weld is a girth weld,the torch's attitude needs timely adjustment;Third,the welding robot is difficult to imitate the experience of skilled welding workers.In this paper,we focus on the three technical difficulties in the automation of the tube wall welding in the cylinder body,and deeply study the digital collection of experience of the welding workers,the experience digital division of the welding area segmentation,and the key technologies of torch gun gesture recognition.An intelligent identification method is proposed.According to the principle that the direction of gravitational acceleration is always vertical and the MEMS accelerometer is used to obtain the absolute posture of the welding torch relative to the boiler cylinder.According to the MEMS data,the relative posture of the welding torch and the girth weld is obtained.The pitch angle and working angle of the welding torch are calculated by combining the data of the two sensors,and feedback data is provided for the adjustment of the welding torch attitude.Finally,a machine learning correlation algorithm is used to combine the welding attitude with the welding process to realize the function of intelligent adjustment of the welding process with the welding attitude.A manual color-set method for manual welding operation is proposed.At present,the quality of operation of the most advanced robot for all-position welding cannot reach the operational level of a primary welder.The fundamental reason is that the machine is difficult to perform the empirical adjustment of the welding attitude and welding process according to the real-time working conditions.Mount the MEMS and gauge sensors in the designated positions on the torch.When a skilled welder carries out a special welding torch to perform the operation,the sensor feeds back the attitude data of the welding torch in real time,and calculates the empirical adjustment data of the welding torch to the welding torch according to the attitude recognition method of the all-position torch.Finally,these data are clustered and integrated,and the experience program is formed as the control basis of the welding robot,and the purpose of integrating the human experience into the welding robot can be achieved.In the intelligent recognition accuracy test of welding gestures,20%of the original training samples were used as test data.The results show that the accuracy of recognition is 99.6094%and the recognition accuracy is high when the all-position welding zone is divided into 8 categories.