论文部分内容阅读
生物质燃料的致密成型过程是一个影响因素多且具有高度非线性的复杂工艺过程,运用传统的建模方法很难建立出精准的预测模型。该文选用秸秆作为生物质原料,通过生产实验,采集样本数据,将粒度、含水率以及成型机压辊间隙作为输入,利用模糊理神经网络可识别、自适应及模糊信息处理于一体的优点,建立网络模型,实现对秸秆燃料的密度及致密成型过程中比能耗的预测。经过实验证明,模型的密度预测最大绝对误差为0.0142 g/cm~3,最大相对误差为1.13%,比能耗预测最大绝对误差为1.5604 kWh/t,最大相对误差为2.27%,该预测模型的建立,可实现预测,达到优化生产参数的目的。
The compact forming process of biomass fuels is a complex process with many factors and high nonlinearity. It is very difficult to establish a precise prediction model by using traditional modeling methods. This paper chooses straw as the raw material of biomass, through the production experiment, collecting the sample data, taking particle size, water content and gap of the forming machine as the input, using fuzzy neural network to recognize, adaptive and fuzzy information processing in one, The establishment of network model, to achieve the density of straw fuel and dense molding process than the specific energy consumption forecast. The experimental results show that the maximum absolute error of model prediction is 0.0142 g / cm ~ 3, the maximum relative error is 1.13%, the maximum absolute error of specific energy prediction is 1.5604 kWh / t, and the maximum relative error is 2.27% Establish, can realize the forecast, achieve the purpose of optimizing the production parameter.