数据驱动的工业过程运行监控与自优化研究展望

现代工业过程向大规模、连续化、集成化方向发展，有必要对生产全流程运行的决策、协同控制、底层控制进行有效监控，也是当前国际控制领域的研究热点。本文首先分析了工业过程全流程运行监控的内涵与行业现状；其次，阐述了基于模型的控制系统故障诊断与容错控制方法，以及数据驱动的异常工况诊断与自愈控制方法的研究现状，并指明了信息物理系统（Cyber-physical systems, CPS）时代智能安全运行监控与自优化的发展机遇；最后，论述了工业过程运行监控与自优化研究的新方向和最新进展，包括：1) 数据驱动的决策、协同控制、底层控制多层面联合监控；2) 基于机理、数据、知识多源动态信息融合的异常工况诊断；3) 专家知识与控制手段相结合的协同层自愈控制；4) 数据驱动的运行动态性能分析与自优化；5) 支撑运行监控与自优化系统的实现技术。

With the development towards large-scale, continuous, and integrated modern industrial processes, it is essential to effectively monitor the plant-wide operations that cover decision, cooperative control, and base-level control. The operation monitoring has recently become an active area of research both in academia and industry. Firstly, the demanding work of operation monitoring and the current status in industrial area are analyzed in this paper. Secondly, the existing methods on model-based fault diagnosis and fault-tolerant control, and data driven abnormal situation diagnosis and self healing control are reviewed, while the opportunities are analyzed under cyber physical systems (CPS) circumstances. Finally, future research directions and recent progresses on the topic of operation monitoring and self-optimization of industrial processes are discussed, including: 1) data-driven multi-level comprehensive monitoring of decision, cooperative control, and base-level control; 2) multi-source dynamic information based abnormal situation diagnosis that combines first principles, process data, and expert knowledge; 3) cooperative self-healing control which combines expert knowledge and control strategy; 4) data driven dynamic performance analysis of process operation and self-optimization; and 5) technologies that implement operation monitoring and self-optimization system.