基于BP神经网络的汽轮机排汽焓在线计算方法

On-line Computation Method for Steam Turbine Exhaust Enthalpy Based on BP Neural Network

汽轮机排汽点处于湿蒸汽区,由于现场缺乏蒸汽湿度的测量手段,排汽焓值难以直接确定,其计算取值一直是火电机组热力系统热经济性在线分析诊断的难点之一。影响排汽焓的因素很多,它们与排汽焓之间属于典型的非线形关系。本文利用BP神经网络良好的非线形映射能力,以NK200-12.7/535/535型空冷机组为例,介绍了应用BP神经网络建立汽轮机排汽焓计算模型的方法,并成功应用于某电厂热经济性在线诊断系统中。

The steam exhaust point of a steam turbine is located in a wet steam zone. Due to a lack of steam wetness measuring instruments on site it is difficult to determine directly the enthalpy of turbine exhaust steam. The calculation of this enthalpy has long been one of the difficult issues in the on line analysis and diagnosis of thermal economics of a thermal power plant thermodynamic system. There exist numerous factors, which can affect the turbine exhaust enthalpy. The relationship between these factors and the exhaust steam enthalpy is a typical nonlinear one. With a NK200 12.7/535/535 type of air cooled plant serving as an example the authors have recommended a method for setting up a model for calculating the turbine exhaust enthalpy by using a BP neural network. The method has been successfully put into use on an on line diagnosis system of thermal economics for a power plant.