Font Size: a A A

The Research On Adaptive Multipass Welding Technology Of Medium And Thick Plate

Posted on:2024-06-14Degree:MasterType:Thesis
Country:ChinaCandidate:X P WeiFull Text:PDF
GTID:2531307094456304Subject:Materials Processing Engineering
Abstract/Summary:PDF Full Text Request
With the increasing requirement of automation and intelligence in modern manufacturing industry,the application demand of multi-layer and multi-pass robot welding is increasing day by day.At present,the planning methods of multi-pass welding parameters of robot are mostly offline planning.The process parameters and torch trajectories of all welds are uniformly planned based on the groove size before welding,and no adjustment is made during welding.However,there are still some differences between these calculation models and actual weld forming.Due to the continuous superposition of multi-pass welds,welding deformation,process parameter fluctuation and other reasons,the errors between the off-line programming welding path shaping and the actual shaping after welding will accumulate continuously.In addition,if the gap and size of the weld groove change,it is also easy to produce welding defects.In order to solve the above problems,this paper establishes the intelligent welding system of multi-pass robot welding,which is composed of line laser sensing system,host computer expert system,industrial robot and welding system.The linear laser sensing system is used to obtain groove clearance and guide the robot to track the weld trajectory.The host computer expert system receives the groove size information obtained by laser sensor scanning,matches the qualified process parameters from the database according to the model rules in the knowledge base,and realizes the pass planning,parameter planning and trajectory planning of variable clearance multilayer multi-pass welding.Finally,the instructions are transmitted to industrial robots and welding equipment to complete the welding,and the planning results are displayed in the human-computer interface.Firstly,through the orthogonal process test,the weld formation data under different process parameters are obtained.The positive and negative prediction models of welding process parameters and weld size were established by BP neural network and iterative correction algorithm.Provide data support for subsequent multi-pass welding parameters and trajectory planning.The welds of multipass welding can be divided into three types: unilateral binding type,bilateral binding type and three-side binding type.A seam stacking model of multi-pass welding is established.The relationship between seam size and seam spacing is explained by mathematical geometry model.Secondly,in view of the existing in the traditional method of multi-layer welding multichannel planning calculation model for cumulative error,easy to produce welding defects problem.A layered planning method for multilayer and multipass robot welding of thick plate is proposed.The welding layer of multilayer and multipass welding can be divided into three types: bottom welding,filling welding and cover welding.Among them,the bottom welding adopts the principle of self-defined heat input;Filling welding adopts the principle of smooth and efficient;Cover welding adopts the principle of forming control first.According to the bottom gap of each layer scanned before welding,the reasonable number of welding passes and welding parameters of the current layer to be welded are planned adaptively.Through the actual welding experiment,the weld surface forming is smooth,the weld has no internal defects,and the planning effect is good.Finally,aiming at the deformation of the heating Angle in the welding process of the workpiece,the parameters of the planning are modified by scanning the groove Angle.For the welds with variable groove clearance,the filling strategy of n welds with fixed parameters without swing and 1 weld with variable parameters swing welding was adopted.Through theoretical calculation,the range of groove clearance variation that can be adapted under different process Windows is obtained.V-groove sample with thickness of 28 mm and variable gap welding experiment with clearance of 4.5-8mm were used to verify the planning strategy.The welding parameters obtained by planning can be adjusted adaptively according to the change of groove clearance,and the weld is beautiful in shape and good in filling effect.
Keywords/Search Tags:Robot, Multipass welding, Neural network, Adaptive programming
PDF Full Text Request
Related items