Research On Intelligent Design System For Laser Welding Technology Of Aerodynamic Components

At present,there is a problem that technology knowledge such as technological design experience and design cases are not effectively collected and applied in the process of hot working technology design of aerodynamic component,which causes the aerodynamic component to still rely on a large number of trial-and-error type process tests to obtain better process parameters in the process of technological design,and this way makes the component size super poor and other problems stand out,which seriously affects the overall quality of the aerodynamic product,Therefore,it is important to study how to use the process knowledge generated in the technological design process to realize the intelligent design of hot working technology.Based on the above-mentioned issues this thesis carried out research on the intelligent design of laser welding technology for the aerodynamic component,relying on the laser welding technology knowledge base of the aerodynamic component,proposed a knowledge question and answer to complete the design of parameters,case reasoning for parameter correction of intelligent design methods.The main research content of this thesis is as follows:Knowledge Question-and-Answer(Q&A)method for aerodynamic components laser welding technology based on knowledge graph.In the technological design process a large number of technology knowledge and design experience is not effectively collected and applied,this thesis uses a top-down graph construction method to build a knowledge graph of the laser welding technology of aerodynamic components,based on the knowledge graph to complete the design of the knowledge question and answer method to achieve the collection and application of technology knowledge.For the technologist proposed multi-class process problem requirements,this thesis proposes a Ro BERTa-Bi GRU-MHA(RBGMA)laser welding technology problem intent recognition model to achieve accurate classification of technology problems.Through the laser welding technology knowledge problem corpus constructed in this thesis,a comparative experiment on model recognition accuracy is conducted.The results show that the recognition accuracy of RBGMA model is superior to the commonly used CNN,RNN,and Ro BERTA models.Case-Based Reasoning method for aerodynamic components laser welding technology based on gray relation algorithm.This thesis collects completed cases of laser welding technology design case of aerodynamic components,organizes the formation of a technological design case database,realizes the collection of case knowledge,and based on the basic principles of case-based reasoning,proposes a gray relation case-based reasoning method with weight correction through analytic hierarchy process,builds a case-based reasoning model with welding defects as the target,helps designers adjust the parameters of design cases according to the form of case defects obtained through reasoning,and achieve reuse of case knowledge.Development of an intelligent design system for laser welding technology in aerodynamic components.Based on the aforementioned key technologies,this thesis utilizes the laser welding knowledge graph and process design case database of aerodynamic components as the data foundation,adopts Java programming language,and develops a framework based on Spring Boot and Vue.It implements an interactive intelligent design system for laser welding technology,with knowledge Q&A and casebased reasoning as the core business.It provides a knowledge service platform for laser welding technology design for aerodynamic components for technology designers,Improved the intelligence and informatization level of aerodynamic components in the laser welding technology design process.