基于BP神经网络的供热机组低压缸质量流量在线监测方法

Online Monitoring Method of Mass Flow in Heating Units'LP-CylinderBased on BP Neural Network

汽轮机低压缸质量流量的实时监测对极寒地区大型供热机组具有重要意义,但却存在难以在现场直接测量的问题。提出了一种基于BP神经网络的供热机组低压缸质量流量在线监测方法。以机组负荷、主蒸汽流量、主蒸汽压力和中压缸排汽压力作为输入,以低压缸质量流量为输出,建立BP神经网络计算模型,利用插值方法获取样本数据。最后,基于PI数据平台开发了汽轮机低压缸质量流量的实时监测系统,并在2台600MW供热机组上进行了实际应用测试。结果表明,该系统给出的常运行工况监测值与经验值是相吻合的,对极寒温度下机组低负荷、大流量供热运行时的空冷岛防冻预警具有一定的指导意义。

Thereal-time monitoring of the mass-flow in turbine's LP-cylinder is of great significance for large heating units in extremely cold regions.However,it has some problem such as we cannot measure this parameter directly.Soin this paper,an online monitoring method of mass flow in heating units'LP-cylinder based on BP neural network is proposed.Take unit load,main steam flow,main steam pressure and intermediate pressure cylinder exhaust pressure as input,and the low pressure cylinder mass flow as output to establish a BP neural network calculation model.The interpolation method is used to obtain sample data.Finally,based on the PI data platform,the real-time monitoring system of the mass flow inturbines'LP-cylinder was developed and applied at two 600MW heating units.The result have verified that the monitoring system's estimation valueof several typical operating conditions is consistent with their empirical value,which has certain guiding significance to the air-cooling island's anti-freezing warningof heating units operating with low-load and high-flow extraction steam under extremely cold temperature.