基于BP神经网络的配煤灰熔点预测研究

Investigation on Prediction of Melting Point of Mixed Coal Ash based on BP Neural Network

对伊泰煤与九洲煤进行了配煤灰熔点研究,结果表明使用原煤灰熔点进行加权平均得到的混煤灰熔点与化验值偏差较大,最高相差100℃左右,配煤灰熔点不能使用原煤灰熔点简单加权计算。根据混煤化验得到的灰成分再代入公式进行计算,得出的结果与化验值偏差较小,但适用范围对SiO_(2)与A1_(2)O_(3)含量有一定要求。研究发现基于BP神经网络的灰熔点预测对于配煤灰熔点的获取在工程应用中也是一种可行的方法,对于灰熔点的预测能够控制在±20℃以内,仅需要事先测定煤灰化学成分并以此计算各项复合系数,即可预测灰熔点。适用范围广,对指导电厂选煤、配煤、用煤以及燃烧器安全设计、锅炉机组的安全运行都极为有益。

The melting point of blended ash of Yitai coal and Jiuzhou coal is studied,results show that the melting point of mixed coal ash obtained by using the original melting point of coal ash for weighted average has a large deviation from the test value,the maximum difference is about 100℃,and the melting point of mixed coal ash cannot be calculated simply by using the original melting point of coal ash.According to the ash composition obtained from the mixed coal test and then substituted into the formula for calculation,the obtained result has a small deviation from the test value,but the application scope has certain requirements for the content of SiO_(2) and Al_(2)O_(3).It is found that the prediction of ash melting point based on BP neural network is also a feasible method for obtaining the melting point of mixed ash in engineering applications.The prediction of ash melting point can be controlled within±20℃.The ash melting point can be predicted only by measuring the chemical composition of coal ash in advance and calculating various composite coefficients based on it.It is very useful to guide the coal preparation,blending and use of coal in power plants,the safety design of burners and the safe operation of boiler units.