基于神经网络模型的锅炉汽包水位监测

为解决目前锅炉就地水位计测量的汽包水位与真实值均存在较大误差的问题，对锅炉就地水位计的传热机理做了分析，建立了基本传热模型，实现了就地水位计的温度压力补偿，并通过神经网络训练得到了在不同汽包压力和水位计表面温度分布情况下的补偿值，进行了机组运行参数及水位计表面温度对水位计误差值产生影响的研究，从而解决了就地水位计指示值偏低的问题。现场运行情况表明，通过神经网络训练的计算水位值更加合理，为汽包水位的准确测量提供了新的途径。

Monitoring Boiler Drum Water Level Based on Neural Network Model

In order to solve the problem of obvious error of measurement for boiler drum water level by local water level meter, the mechanism of heat conduction of local level meter is analyzed. The basic heat conduction model is established to implement temperature and pressure compensation for local level meter. Through neural network training, the compensating values under different pressure of boiler drum and temperature distribution on surface of level meter are derived. The influence of unit operating parameters and surface temperature of level meter on measurement errors is studied. Thus the problem of lower indication obtained from local level meter is solved. The practical operation shows that the water level calculated by neural network training is more reasonable, this offers new approach to accurate water level measurement of boiler drum.