Research On Knowledge Graph Visualization Based On Model Validation Analysis

In the global market competition,the lead time of designing and manufacturing products is decreasing,while the complexity of products is increasing.In order to reduce rework caused by model design defects,major mainstream CAD(Computer Aided Design)software manufacturers are vigorously developing and providing quality inspection tools to help companies effectively improve model quality and reduce product production costs.In the process of model verification,a large amount of detection information will be generated,which contains a large amount of semantic information during modeling,but these data don’t been well utilized to help designers improve repair efficiency.In this regard,this thesis applies the knowledge graph to the manufacturing industry,and uses the graph structure to visualize the fault model modeling process and fault information.This research can not only help engineers quickly solve faults,but also analyze and summarize the modeling process.In addition,it can also optimize the modeling method and play the role of drawing inferences from one instance.The main research work of this thesis is as follows:(1)Extract model verification data to build a domain knowledge graph.Firstly,extract and analyze all the information generated after the model verification,and design the knowledge extraction method according to the characteristics of the data.Refer to the general knowledge graph construction method and combine the data characteristics to build a model validation domain knowledge graph.(2)Research the visual layout algorithm according to the structural characteristics of the data.Data in the industrial field often contains a large amount of structured data.Based on this feature improve and optimize the general force-directed layout algorithm,as well as improve the interpretability of visual graphics.(3)Build a visualization system based on the Web.Further improve data analysis capabilities by customizing visualization functions and interactive behaviors to ensure visual results are intuitive and accurate.Visual analysis of model validation data can help engineers quickly locate and analyze fault causes,and analysis of topological relationships among feature information can help optimize the modeling process and improve model quality.This thesis proposes an improved visualization layout algorithm based on model verification data,and combines this algorithm to design and implement a visualization system.It also has a certain degree of universality for the visualization of general knowledge graphs,and provides a certain reference value for the subsequent application of knowledge graphs in the industrial field.