Research On Robot Intelligent Welding Seam Tracking Method And System Control For Complex Working Conditions

After the "Made in China 2025" was proposed,the "teach-back" robot welding technology has been widely used in the production process of marine engineering,ships,rail transit,pressure boilers,nuclear power and automobiles.With the development of these industries,the requirements for welding manufacturing technology gradually tend to be unmanned,intelligent,high-precision and high-quality,especially in the production environment that is dangerous and difficult for humans to contact for a long time.However,various complex factors such as arc light,spatter,smoke,workpiece structure,workpiece clamping accuracy,workpiece machining error,welding thermal deformation,weld groove type,weld misalignment and weld gap can cause changes in the weld trajectory.It brings great difficulties to the current "teach-back" robot welding technology to achieve highdemand welding manufacturing in various industries.In order to overcome this situation,an intelligent welding technology that can realize robot automatic tracking welding and improve forming quality under various complex welding conditions is urgently needed.For this reason,this paper researches and develops a robot intelligent welding seam tracking method and system control technology based on vision and arc sensing,and conducts welding seam tracking and forming quality real-time control for the robot gas metal arc welding(GMAW)process.The research content of this technology mainly includes the design of robot intelligent welding tracking system,image processing algorithm,welding seam start-stop point detection algorithm,welding seam tracking algorithm,adaptive swing tracking welding forming and arc length control.In order to realize automatic welding under complex working conditions,a robot intelligent welding tracking system is developed.In this system,a structured light vision sensing system and an arc sensing system are designed in order to detect the weld groove information and welding parameters during the welding process.In order to integrate multiintelligent sensor information and realize the guidance of robot welding,based on the MFC platform,this paper uses multi-thread technology to develop the welding seam tracking software,which realizes image acquisition,welding current acquisition,data communication,image processing,sensor information fusion,and robot intelligent welding control.In order to realize the accurate extraction of groove information such as weld feature points and weld gaps,a set of strong robust image processing algorithms is proposed.The algorithm overcomes the interference of arc radiation,metal spatter,metal vapor and surface reflection,and realizes the accurate extraction of weld seam feature point and weld gap.The core steps of the algorithm include weld feature extraction initialization,image preprocessing,stripe center point extraction,center line fitting,stripe segmentation,abnormal feature point correction and weld feature region update.In order to test the performance of the proposed algorithm,detailed welding experiments were carried out on the feature extraction performance of the algorithm under complex conditions such as strong reflective surface,weld gap,weld misalignment,arc radiation,spatter and smoke.The experimental results show that the proposed algorithm can achieve sub-pixel precision extraction of weld feature points and high-precision detection of weld gaps,and has good robustness.In order to realize the automatic alignment of the welding initial point and the automatic end of the welding tracking process,an algorithm for identifying the welding start-stop point is proposed.The algorithm uses the changing characteristics of the laser stripes at the end of the workpiece to identify the welding start-stop points.In order to test the effectiveness and performance of the algorithm,the performance of the algorithm under different workpiece poses is tested in detail.The experimental results show that the proposed algorithm is effective,and the identification accuracy of the welding start point and end point is not more than 0.24 mm and 0.26 mm,respectively,at a search speed of1200mm/min.In order to overcome the influence of lead distance,welding thermal deformation and mechanical vibration on the welding seam tracking control,a welding trajectory planning algorithm is proposed.In this algorithm,the proposed algorithm overcomes the interference of vision sensor advance distance and mechanical vibration,eliminates the phenomenon of positive and negative alternation of welding seam trajectory deviation,and maintains the stability of the welding process.Combined with the designed position-given welding seam tracking controller,the algorithm realizes the real-time high-precision control of the welding trajectory.In order to verify the effectiveness and performance of the proposed algorithm,the tracking welding performance is comprehensively tested for various conditions such as surface reflection,groove,medium and high welding speed,complex curve weld,joint type and so on.The Welding test results show that the proposed algorithm achieves accurate seam tracking welding,and the average value of the absolute seam tracking error in the X and Z axis directions does not exceed 0.28 mm and 0.19 mm at a GMAW welding speed of 1200mm/min.In order to realize the swing tracking welding of the weld with variable gap,an Eyein-Hand adaptive swing welding tracking algorithm is proposed.The adaptive swing welding tracking algorithm realizes the planning of the swing center trajectory through the sliding curve fitting algorithm,and adaptively adjusts the swing parameters in real time according to the size of the weld gap.In order to ensure the welding quality in the welding process,an arc length control model is proposed.The arc length control model transforms the fuzzy change of arc length into coordinate compensation through coordinate transformation,and performs fuzzy quantitative fusion with the three-dimensional coordinates detected by visual sensing to realize the control of welding arc length in the welding process.The arc length control model converts the fuzzy change of arc length into coordinate compensation through coordinate transformation,and performs fuzzy quantitative fusion with the three-dimensional coordinates detected by the vision sensor to realize the control of welding arc length in the welding process.In order to verify the effectiveness of the proposed method and test its performance,detailed tests are carried out on the swing welding tracking performance of the curve weld,the adaptive swing welding forming of the weld with variable gap,and the welding current stability with arc length control.The experimental results show that the proposed algorithm can achieve accurate swing tracking welding,and the average welding seam tracking error does not exceed0.33 mm.The proposed adaptive swing tracking algorithm can improve the adaptability of swing welding forming to the weld gap.The proposed arc length control method realizes the stability control of welding parameters and guaranteeing the stability of welding quality.