Development And Implementation Of Digital Rapid Design System For Shelter

As a mobile studio,the shelter has good flexibility and adaptability,which is often used in engineering operations,medical rescue and other fields.With the continuous development of market demand,the traditional design method of shelter has been difficult to support the rapid production and manufacturing of shelter.Aiming at the low information resource utilization rate of historical design cases in current shelter manufacturing enterprises,this thesis proposes a case retrieval method for shelter design to improve the reuse degree of design cases.Aiming at the low design efficiency of the shelter structure in enterprises,parametric configurable variant design technology is used to improve the design efficiency of the shelter structure.Aiming at the low efficiency of drawing number application in enterprises,an intelligent drawing number application technology based on machine learning is proposed to improve the efficiency of drawing number application.The main research work of this thesis is as follows:(1)On the basis of overall analysis of shelter design resources,basic information resources such as historical shelter design case database,drawing number application information database and three-dimensional model database are established to provide basic support for the development of digital shelter design system.(2)Aiming at the low utilization rate of design case resources in enterprises,this thesis proposes a retrieval technology of design case of shelter based on knowledge graph fusion case-based reasoning technology(CBR).A top-down approach based on ontology is adopted to construct the knowledge graph of design cases oriented to shelter.The Jaccard similarity measurement method is used to calculate the relational similarity of cases,and the attribute similarity of corresponding nodes is calculated by combining the CBR.In addition,weight allocation is carried out for different types of nodes in the knowledge graph by AHP combined with the similarity deviation sum method.Finally,several historical cases similar to the target design shelter and corresponding three-dimensional models are obtained according to the comprehensive similarity ranking.(3)Aiming at the problems of low efficiency and low model reuse rate of the current enterprise shelter structure design,a parameterized variant design method based on configurable structure tree was proposed,and combined with configuration linked list,the borrowed configuration model was discriminated to complete model reuse.Through the enterprise history design case and the standardization of product structure to establish a three-dimensional model database facing to the shelter.The product configurable structure tree is used to extend the configuration module in the model base to the similar shelter model,which makes the configuration combination more diverse and improves the flexibility of shelter design and the reuse of model resources.(4)Aiming at the problems of low efficiency of drawing number application for enterprise product parts,a batch drawing number application technology based on machine learning is proposed.After pre-processing the drawing number application record,the K-means ++ algorithm is used to cluster the data,and then KNN algorithm is used to conduct preliminary coding for the parts.For the "same name and different number" parts,the deep learning method based on MVCNN is used to encode the parts,and finally the serial number is added to complete the drawing number application.(5)Combined with the above content,through the secondary development of Solid Works 2020 to complete the construction of the digital system,the realization of the shelter design case retrieval,rapid structural design,drawing number batch application and other functions,finally,a shelter design order as an example,to verify the practicability of the system.