褐煤配入高熔点煤灰的灰熔点预测模型研究

测定了不同比例的褐煤与高熔点煤灰的混灰在弱还原气氛下的灰熔点,并且采用BP神经网络模型对灰熔点与灰成分及其组合参数之间的关系进行预测。结果表明:3种低灰熔点褐煤的灰熔融特性可以通过配入高熔点煤灰显著提高,混灰的灰熔点变化与配比间呈现非线性变化规律,灰熔点上升趋势总体可分为‘前段快速上升后段平缓’和‘前段快速上升中间段平缓后又上升’2种类型,配入灰熔点更高的高熔点煤灰对提高褐煤灰熔融温度效果不一定更优;使用摩尔分数作为基准,输入层包含8个灰成分参数和5个组合参数(硅值、酸值、碱值、白云石比率和R250)的BP神经网络模型对混灰熔点的预测优于仅包含8个灰成分参数的输入层预测模型,且该模型可对混合灰熔点的预测效果较好。

Ash fusion temperature prediction model of lignite blending high melting point coal ash

In order to study the effects of blending high melting point coal ash on the ash fusion temperature of lignite,the ash fusion characteristics of lignite mixing with different anthracite ash were measured,and the BP neural network was used to predict the mixing coal ash fusion temperature based on its chemical composition and combined parameters.The results indicate that three kinds of low ash melting point of lignite can be increased by blending the anthracite ash,and the changing trends are not linear with the proportions of adding ash.The rising of ash melting point can be divided into two types:'first rising and then leveling off' and 'first rising,then keeping unchanged and finally rising again′.There may be not better effect to add the ash with much higher ash fusion temperatures for improving the ash melting temperature of lignite.BP neural network can obtain a better performance when the input layer is consisted of 8 chemical composition parameters and 5 combined parameters(silica value,acid value,base value,dolomite ratio and R250) based on molar fraction and it can be a useful tool to predict the mixing ash fusion temperature.