Research On Properties Modeling And Knowledge Base Intelligent Design Of Magnetic Materials

Materials are based on multi-dimensional knowledge and complex engineering design process.They are widely used in the fields of aerospace,mechanical and electronic energy environment,information storage,biology and so on.The Aerospace Industry Group has long accumulated a wealth of practical knowledge and experience in the design of magnetic materials,but this knowledge is scattered throu ghout the project experience of text and material design experts.The design knowledge of enterprises has not been effectively summed up,collated and refined;Because of the complex and highly nonlinear relationship between the properties of magnetic materials and the composition of materials,the optimization of material formulation is an urgent problem in this group.This paper is based on the knowledge map of aerospace magnetic materials and the main contents are as follows:First of all,the key words of knowledge,professional terms preprocessing.In order to ensure the rationality and accuracy of the statistical data,the key words representing the development context of magnetic alloy materials are counted by Bicomb software,and the keywords whose frequency is greater than 7 are obtained.In order to ensure the reasonableness and accuracy of the statistical data,some similar keywords are merged.Construct high frequency keyword matrix.Based on high frequency keyword matrix,SPSS software is used to cluster analysis.The 30 high-frequency magnetic key words of magnetic alloys are divided into 4 categories,and the tree classification structure is applied to show the classification results.Secondly,the knowledge dimension model is designed,and the number of entities is determined by entity identification technology,fusion technology and link technology(material number,191 pieces,and attribute matching),and the knowledge map of magnetic material is constructed.According to the properties of magnetic materials,the magnetic materials are divided into seven dimensions: knowledge type dimension,design process dimension and product dimension.The semantic network of magnetic material knowledge is constructed by using the SNet L syntax description form of triple node-edge-node.Finally,the personalized recommendation of magnetic material products is realized by using Java and Python language combined with collaborative filtering algorithm on the basis of the constructed knowledge map.The integrated modeling of magnetic material,information and performance is realized by using BP neural network,and the results of network prediction are analyzed,and the evaluation system structure of material performance is set up for the optimization of the formula of magnetic material product.The product knowledge management of Dw-025 sintered Nd Fe B was carried out by using magnetic material knowledge map,and the material process was designed.The results show that the intellectual construction method of magnetic material knowledge base proposed in this paper can make full use of the existing data resources and realize the personalized recommendation of magnetic alloy material products.By using the semantic network of the composition and properties of magnetic materials on the knowledge map and the BP neural network algorithm,the model between material,information and magnetic properties is established,and the optimal selection of the properties and composition of magnetic materials is finally realized.In order to overcome the research difficulties for material experts caused by the weak basic theory of magnetic alloy selection at present,the design and application of aerospace magnetic materials in China are developed at a high speed.