应用现代信息技术和人工智能实施设备诊断工程，逐步实现状态维修和预知维修，是大型流程工业企业降低生产成本的重要途径之一。概要介绍为实现这一目标所开发的机电装备实时监测网络和人工智能诊断技术。简要介绍了基于Ethernet和FDDI开发、应用于石化企业的机、泵群实时监测网络；首次提出了黑灰白集合筛选法，在一次原因分析法和故障机理及其识别特征研究基础上，应用此方法开发的基于黑灰白集合筛选法的机械故障诊断专家系统，用于工程实践取得了满意的结果。

Using modern information technology and artificial intelligence to achieve the condition based maintenance and predictive maintenance is one of the important ways to reduce the production cost in the process industries. The real-time monitoring network and artificial intelligent diagnosis technology for mechanical-electric plant was outlined in this paper. The Ethernet and FDDI based real-time monitoring network developed for compressors and pumps in petrochemical plants was introduced briefly. The black-gray-white gathering diagnosis method was given for the first time on the bases of approach to fault mechanism and distinctive symptoms. The mechanical fault diagnosis expert system based on black-gray-white gathering distinguishing sieve method developed in this work yields satisfactory results in the engineering practice.