The Construction And Application Of Knowledge Graph For Construction Organization And Management With Railway Tunnel Drilling And Blasting Method

As the main battlefield of our country’s railway construction gradually shifts to the central and western regions,the proportion of long tunnels under complex geographical and geological conditions is increasing,and the drilling and blasting method of railway tunnel construction in complex and dangerous environments faces long construction periods,complicated procedures,and construction organization.There are many management factors and intertwined relationships.The phenomenon of “information overload” brought about by the massive spatiotemporal data of multi-platform,multi-professional,multi-sensor,and multi-source heterogeneity is obvious,and the data acquisition ability and knowledge application ability are extremely inconsistent and matching problems are prominent.As a result,it is difficult to meet the outstanding problems such as the refinement,integration,and intelligent management needs of "smart railways" in the new era.This paper introduces theories and methods related to knowledge graph,and applies it to the construction scene of railway tunnel drilling and blasting method.The introduction of knowledge graph can effectively integrate multi-source and multi-professional domain knowledge.Combined with the path ranking algorithms of the graph structure,the reasoning and prediction of the material demand relationship in the construction process can be further realized,and the prediction results can realize the incremental supplement to the knowledge graph to a certain extent and help the construction organization and management personnel to assist the management and decision-making of the project progress.The main research contents of this paper include:(1)Construct an integrated expression model of multi-element coupling and correlation of railway tunnel drilling and blasting construction scene.The system analyzes the "six-inone" management target elements based on standardized management in the construction process,the five key influencing factors of the "man-machine-material law ring" covered by the management target elements,and the minimum construction cycle focusing on drilling and blasting construction.The coupling relationship between the elements "three elements",and the key influencing factors of construction are used as the starting point to expand the classification of sub-elements and the elaboration of the corresponding semantic relationship,so as to construct a unified semantic modeling expression of multi-element data.(2)A method for constructing knowledge graph of railway tunnel drilling and blasting construction scene domain is proposed.Taking the constructed relational model as the ontology,analyzing the characteristics of the construction scene of the railway tunnel drilling and blasting method,combining the knowledge and experience of domain experts,referring to the construction method of the open domain knowledge graph,and adopting the combination of "bottom-up" and "top-down" By means of knowledge extraction,association fusion,storage and expression,etc.,a knowledge graph of the construction scene of railway tunnel drilling and blasting method is constructed.(3)This paper proposes a knowledge graph path ranking relationship reasoning prediction method for railway tunnel construction material demand prediction.Based on the constructed domain knowledge graph,using the path sorting algorithm,the path between multiple types of users and construction material nodes is used as a feature,and then it is judged whether the path between the two concerned entities is sufficient to support the limited relationship of material requirements,and realize Quantitative forecasting of material requirements for multiple types of users in the construction process.(4)Build a prototype system and carry out related experiments.Taking the construction scene of the Kangding No.2 tunnel exit work area as an example,relevant experiments are carried out to verify the effectiveness of the method in this paper.The experimental results show that the knowledge graph constructed in this paper can show a variety of relationships among multi-element entities,and the inference and prediction results can assist managers to make management decisions and supplement the existing knowledge graph to a certain extent.