Construction Of Knowledge Graph In Well Control For Testing Oil And Gas Domain

In recent years,with the rapid development of artificial intelligence,computer has been applied and developed in many related fields,and information technology has been widely applied in various professional fields.The knowledge system of many disciplines in their own fields is also growing,and the demand for integrated management of massive data is increasing.The knowledge map can satisfy the practical needs of structured storage,query and management of the large amount of knowledge in the professional field,and help to understand the related field knowledge more comprehensively.Currently,the related research of knowledge map is not satisfied with the general field,and has continued to expand to the professional field.In each field,knowledge map of specific fields has been gradually established,including the judicial field,the mathematical field,the agricultural field and many other professional fields.For the basic knowledge of oilfield theory,it also needs intelligent and modern methods to manage and learn.The operation conditions in the field of well control for oil and gas testing are more and more stringent,the oil field installations are more and more complex,and the technical knowledge involved is also increasing.However,at present,these knowledge contents are scattered,lack of effective management,which is not conducive to the learning of knowledge by relevant technical personnel,and it is difficult to bring into full play the value of well control data,and even more difficult to form a system.In order to solve the above problems,this paper constructed of knowledge graph in well control for testing oil and gas domain to better manage the data and make the knowledge of well control for testing oil and gas data for play a better role.The main work and contributions of this paper are as follows:(1)Obtain and clean data related to the well control for testing oil and gas domain.The multi-source corpus is constructed to obtain semi-structured data and unstructured text data of well control knowledge in the well control for testing oil and gas domain,and complete the construction of multi-source corpus.(2)The named entity recognition in well control for testing oil and gas domain is carried out,and the related methods are systematically studied.Bert-Bil STM-CRF deep learning model is used to complete the task,and the entities in well control for testing oil and gas domain are extracted from the data center.Based on the deep learning model,the active learning method is introduced to solve the problem of data annotation,and better named entity recognition effect and higher named entity recognition accuracy are achieved on the basis of a small amount of labeled data.(3)The entity relationship extraction in the domain of well control for testing oil and gas is carried out,and the corresponding methods are systematically studied.A deep learning model of BERT-Bi GRU-Attention is used to obtain the triple set of entity relationship for well control for testing oil and gas.The results of triple set are sorted out in the data file,and added to Neo4 j diagram database by batch insertion method to complete the construction of knowledge map in the field of well control for testing oil and gas.(4)Design and implementation of well control knowledge map system for oil and gas testing.This paper uses B/S architecture and C# programming language to develop the knowledge map system in the field of well control for oil and gas testing.The system includes entity recognition,relationship recognition,knowledge query and model incremental training.By visualizing the results of identifying entities and relationships in the field of well control for oil and gas testing,knowledge can be analyzed more intuitively.Model incremental training can gradually improve the knowledge map.