基于软测量技术的热力系统凝给水系统流量状态监测方法研究

Research on Flow Status Monitoring Method of Condensate Feedwater System in Thermal Power Plant Based on Soft Sensing Technology

摘要: 对热力系统凝给水系统流量的实时预测在系统状态监测、故障诊断方面有着重要意义。本文提出了融合机理模型与数据驱动的软测量方法，进而实现凝给水系统的流量实时预测。通过利用泵、阀门的设计参数以及传感器采集的实时数据，首先构建了凝给水泵、阀门和管道系统的流量实时计算模型，并在实时仿真平台中进行了模型构建，进一步开发实时通讯接口，进行了误差分析和评估。结果表明，该软测量模型在给水调节阀、给水泵以及凝水泵等关键流路的流量实时预测，平均流量误差控制在2.5%以内，所建立的软测量模型可以较好的预测凝给水系统流量。

Abstract: Real-time prediction of condensate feedwater flow in thermal power plants is of great significance for system status monitoring and fault diagnosis. This paper proposes a soft sensor method that integrates a mechanistic model and data-driven approaches to achieve real-time flow prediction of the condensate feedwater system. By utilizing the design parameters of pumps and valves, as well as real-time data collected by sensors, a real-time flow calculation model for the condensate feedwater pumps, valves, and piping system was first constructed. The model was then built in a real-time simulation platform, and a real-time communication interface was further developed, followed by error analysis and evaluation. The results show that the soft sensor model can effectively predict the flow of the condensate feedwater system in key flow paths such as feedwater regulating valves, feedwater pumps, and condensate pumps, with an average flow error controlled within 2.5%.