矿山生产机电设备是机械、电气、液压、控制等多形式系统的复杂耦合结构,工作过程中环境和工况条件变化多样,缺乏有效的技术手段解决矿山设备运行健康状态的实时感知问题。 

借助物联网、工业互联网、人工智能和大数据挖掘技术,研究矿山设备状态知识建模与在线诊断方法,将传统的“人-机”交互监控模式提升为“传感—机器认知—机器决策”的智能化监控模式。 

分析了基于物联网的矿山设备状态感知系统架构,定义了多源信息感知层、边缘智能层、大数据分析层和数据与知识共享迁移层的 4 个层次的作用,提出了设备状态知识共享与迁移模式;结合本体语义、置信规则库和数字孪生技术,设计了面向矿山机械设备系统状态知识建模的信息描述、知识表示、决策融合方法,提出了面向矿山设备运行全过程的实时感知、演化分析与智能交互的“虚实融合”感知模型,实现虚、实系统运行过程的“精准映射、信息对偶、融合交互、协同演进”。 

分析了数据驱动的矿山机械设备状态诊断方法研究现状、技术架构、存在问题和研究趋势,提出结合数字孪生、深度学习、迁移学习等方法,构建机理模型、经验知识与数据深层特征相融合的矿山设备状态诊断模式,研究矿山设备状态大数据分析与应用技术,研发矿山设备状态诊断与全生命周期管理等智能化应用服务系统。 形成了矿山设备运行信息感知、知识建模与状态在线判识方法体系,以实现矿山机械设备故障状态自诊断、早期隐患预知维护、智能调度与协同管控,为矿山生产智能化、无人化提供技术支撑。

Electromechanical equipment for mine production is a complex coupling structure of mechanical,electrical,hydraulic,and control systems. During the working process,the environ ment and working conditions are varied. Currently there is a lack of effective technologies used for the real-time perception of health status of mining equipment. The Internet of Things,Industrial Internet,artificial intelligence,and big data mining technologies are used to study the modeling methods of mining equipment state knowledge and online diagnostic technologies. The traditional “ human- machine” interactive monitoring mode will be promoted to the “sensor-machine cognition-machine decision” intelligent monitoring mode. The architecture of mining equipment state awareness system under Industrial Internet is analyzed,and the roles of the four layers of multi-source information perception layer,edge intelligence layer,big data a- nalysis layer and data and knowledge sharing migration layer are defined,and the equipment state knowledge sharing and migration mode is proposed. Information description,knowledge representation,and decision fusion methods for the state knowledge modeling of mining equipment systems are designed using ontology semantics,confidence rules,and digital twin technologies. A “virtual-real fusion” perception model for real-time perception,evolutionary analysis,and intelligent interaction for the entire process of mine equipment operation is proposed to achieve “precise mapping,information duality,fusion interaction,and collaborative evolution” of virtual and real system operation processes. The research status,technology architecture,existing problems and research trends of data-driven mining equipment condi- tion diagnosis methods are analyzed,the combination of digital twins,deep learning,transfer learning and other meth- ods are proposed to build the mining equipment status diagnosis mode with the integration of a mechanism model,empirical knowledge and deep features,the big data analysis of mining equipment status and application technology are studied,and the mining equipment status diagnosis and comprehensive Intelligent application service system such as life cycle management,and a big data platform framework for mining equipment condition diagnosis are constructed. Mine equipment operation information perception,knowledge modeling and status online identification method system is formed to provide guidance for the self-diagnosis of mining equipment fault status,early maintenance of hidden dagers,intelligent scheduling and collaborative management and control,and to provide technology for intelligent and unmanned mining support.