Research On Reasoning Mechanism And Parameter Optimization Method Of Robot Welding Process

At present,robot has been widely used in the field of welding,and is also moving towards a higher degree of automation and intelligence.However,automation and intelligence are the means for welding robot to realize efficient processing,and welding technology,as the hub of product design and manufacturing,has a very important impact on welding efficiency,welding cost and welding quality.Intelligent planning of welding process and industrial robot are of great 'significance to realize higher degree of welding automation and intelligence.Based on process reasoning and process optimization and other related technologies,this paper studies the reasoning mechanism and parameter optimization method of robot welding process to improve the intelligent planning degree of robot welding process.In this paper,a general scheme of welding process reasoning and optimization is designed to meet the needs of robot welding process reasoning and optimization.By determining the knowledge representation model,process reasoning strategy,process optimization strategy and process evaluation strategy of robot welding process,the overall technical route of robot welding process reasoning and optimization was finally formed.Then the representability of knowledge of robot welding process is studied.Aiming at the instance knowledge,rule knowledge and evaluation knowledge in the welding process knowledge,consideration with the characteristics of the reasoning technology used,a robot welding process knowledge representation model based on ontology,generative rule was constructed.In terms of process reasoning,this paper proposes a welding process reasoning mechanism that combining Case-based reasoning(CBR)and belief rule-base inference methodology using the evidential reasoning approach(RIMER).In the process of welding process reasoning,the K-nearest neighbor algorithm is firstly used to retrieve similar cases in the case database.The case-based reasoning will be ended if the case can meet the actual needs after case modification;otherwise,the RIMER is carried out to reason the finally welding process.In the reasoning process based on RIMER,firstly,the knowledge representation scheme under uncertainty is determined,which includes other knowledge representation parameters such as attributes and rule weights,and then a confidence sub-rule base based on layered confidence structure is proposed.Then the evidential reasoning algorithm is used to implement the corresponding reasoning in the rule base.Finally,an example is given to verify the proposed process reasoning method,and the results show that the method can meet the needs of process reasoning.Finally,in order to optimize the reasoning results of welding process and effectively guide the actual welding,a welding parameters optimization method based on Stacking model fusion and particle swarm optimization is proposed.The prediction model of welding process performance was constructed through fusion of multiple base prediction models and secondary prediction models.In the optimization process,the prediction model is introduced into the particle swarm optimization algorithm to optimize the welding process parameters and obtain the optimal welding process parameters.The results show that the prediction accuracy of the model after model fusion is better than that of a single base predictor.The optimum process parameters obtained after optimization can obtain the welding bead with good performance.The proposed optimization method can effectively optimize the process reasoning results and guide the actual welding processing.