基于知识图谱的采煤机智能维护知识库构建

Construction of intelligent maintenance knowledge base for shearer based on knowledge graph

  • 摘要: 为了满足采煤机元部件级故障源精确定位需求,提升综采工作面开采效率和安全可控性,将知识图谱技术引入采煤机故障检修知识的动态建模、形式化存储及智能交互过程,构建了基于知识图谱的采煤机智能维护知识库。从采煤机的硬件拓扑、故障检修、传感监测等方面,规范了相关术语与命名结构;定义并抽取了采煤机智能维护知识图谱的实体、关系与属性,建立了包含整机、部件、子部件、元件和零件5类实体的硬件拓扑网络子图,包含故障类型、位置、现象、原因、解决方法5类实体的故障检修网络子图,包含传感器、监测位置2类实体的传感监测网络子图;通过实体消岐与共指消解过程合并硬件拓扑、故障检修和传感监测3个网络子图,从而形成采煤机智能维护知识网络图,并通过对网络节点之间的相互作用关系进行形式化描述来表达智能维护知识;采用Neo4j,Py2neo等技术,搭建了一个可实现动态交互的采煤机智能维护知识库原型系统,初步实现了故障信息检索、技术指导等功能。

     

    Abstract: In order to meet the requirement for precise location of fault sources at the meta-component level of shearers and improve mining efficiency and safety controllability of fully mechanized working faces, knowledge graph technology is introduced into the process of dynamic modeling, formal storage and intelligent interaction of shear fault maintenance knowledge. An intelligent maintenance knowledge base of shear based on knowledge graph is constructed. From the aspects of the shearer hardware topology, fault maintenance, sensor monitoring, related terms and naming structures are standardized. The entities, relationships and attributes of the shearer intelligent maintenance knowledge graph are defined and extracted, and a hardware topology sub-network diagram is established,which contains 5 types of entities, including whole machine, component, sub-component, element and part. And a fault maintenance sub-network diagram is established, which contains 5 types of entities, including fault type, location, phenomenon, cause and solution. A sensing monitoring sub-network diagram is established, which contains 2 types of entities, including sensor and monitoring location.Through the process of entity disambiguation and common reference resolution, 3 sub-network diagrams, including hardware topology, fault maintenance and sensor monitoring, are merged to form a network diagram of intelligent maintenance knowledge of coal shearers. And the interaction relationship between network nodes is formally described to express intelligent maintenance knowledge. Using Neo4j, Py2neo and other technologies, a prototype system of intelligent maintenance knowledge base of shearer that can realize dynamic interaction is built, and the functions of fault information retrieval and technical guidance are realized initially.

     

/

返回文章
返回