Research On Multi Scale Information Retrieval In BIM Model Based On Natural Language Processing

With the deep application of BIM(building information modeling)technology in the field of construction,a large number of multidisciplinary data are integrated into BIM model,resulting in the problem of information overload.How to quickly and accurately obtain the multi-scale information of space,equipment and management data in BIM model is an urgent problem.Traditional keyword based information retrieval methods can’t correctly analyze the user’s query intention due to semantic problems,resulting in low retrieval accuracy and irrelevant return results,which can’t meet the needs of accurate multi-scale information query in BIM data,such as space,equipment and management data.Therefore,this paper proposes a multi-scale information retrieval method in BIM model based on natural language processing.Natural language processing technology can improve the ability of BIM retrieval system to correctly understand the user’s query intention,so as to make correct processing and ensure the accuracy of retrieval results.Knowledge Graph technology can associate the data information in BIM,form the semantic expression of data,and make the query results more relevant.Therefore,this paper proposes a multi-scale information retrieval method based on BIM Technology,natural language processing technology and knowledge Graph technology.The main research work is as follows.(1)The method of building knowledge Graph of BIM model automatically is proposed.Based on IFC Standard,the data in BIM model and the relationship between data are analyzed.Through the steps of ontology concept construction,entity extraction and relationship extraction,BIM model knowledge Graph with multiple object relationships(such as peer relationship,parent-child relationship,spatial relationship,etc.)and attribute relationships(such as material,length,etc.)is constructed.This method can realize the automatic construction of knowledge Graph of large BIM model with tens of thousands of nodes and relationships,and realize the semantic expression of building information.(2)A multi-scale information retrieval method in BIM model based on natural language processing is proposed.The natural language processing technology is used for word segmentation,semantic disambiguation,named entity recognition and syntactic analysis of multi-scale natural language query.The query statement is transformed into the form of logical expression,and the query intention is identified based on rules.Finally,it is mapped to different scale query templates to realize accurate query of multi-scale information in BIM model.(3)Combined with Web GL technology,neo4 j Graph database and other related technologies,this paper designs and develops BIM multi-scale information retrieval system.The system is applied to the actual project to verify the effectiveness of the retrieval method proposed in this study.The results of the retrieval method proposed in this study are more in line with the user’s query intention,and the user can efficiently and accurately query the multiscale information used in the decision-making process.