基于人工神经网络的锅炉烟气酸露点预测

Prediction of Flue Gas Acid Dew Point in Boilers with Artificial Neural Network Method

锅炉烟气酸露点的准确预测是实现锅炉节能和低温腐蚀防控的关键问题。鉴于现有的烟气酸露点理论模型或测试方法均存在一定的局限性,本文拟通过人工神经网络来实现锅炉烟气酸露点的预测。通过已有试验数据和理论公式计算值比较,本文建立的人工神经网络能够对案例锅炉的烟气酸露点及其波动范围实现准确预测,为该锅炉的排烟温度和节能装置控制提供精确指导,同时也为建立一个限定条件较少的烟气酸露点通用预测模型拓展思路。

Accurate prediction of the acid dew point temperature of boiler flue gas is a critical key to the boiler energy recovery and low temperature corrosion protection.Most of the existing widely used theoretical models or test methods for the acid dew point cannot act as a universal method due to their particular limitations.The present study try to develop an artificial neural network model to obtain good predictions for the acid dew point of the boiler exhaust gas.Compared with some test data and theoretical calculations,the neural network model can provide good predictions of the acid dew point as well as its fluctuation range for the studied boiler.These predictions can make guidance for an accurate control of the temperature of the boiler flue gas and energy recovery system,and also the present method maybe help to open up an idea for establishing an universal prediction model for the acid dew point of the boiler flue gas.