生物质固体成型燃料热值和碳元素预测模型的建立

Development of Prediction Models for Heating Value and Carbon Content of Biomass Solid Fuel

建立生物质固体成型燃料热值和碳元素预测模型。利用秸秆类生物质固体成型燃料的测试数据,以工业分析中的水分、灰分、挥发分和固定碳含量指标为4个自变量,分别以低位热值、高位热值和碳元素分析为因变量,通过多元线性回归模型(MLR)方法建立多元线性回归预测模型。内部检验和外部检验说明3组模型在应用域范围内均具有理想的预测能力,拟合效果良好,其中高位热值预测模型的R^2和R^2_(prep)分别为0.900和0.730,与已有研究相比,相对残差范围减小为–2.59%~2.26%,可为工业用途的生物质固体成型燃料的热值和碳元素分析快速做出反映。

The prediction models for the heating value and carbon content of biomass solid fuel were established.Based on the measured data of straw biomass solid fuel,MLR method was used to establish multiple linear regression forecast models.The industrial analysis of moisture,ash content,volatile and fixed carbon content indexes were secleted as four independent variables,while low heating value,high heating value or carbon content were the dependent variable.The internal and external inspections of the three models demonstrated that all the models had ideal prediction ability and the error range decreased remarkably.The R2and R2prep values of prediction model of high heating value were0.900and0.730,respectively,while the decreased relative residual error range was-2.59%to2.26%.The research is capable of rapidly reflecting the heating value and elemental analysis of biomass solid fuel for industrial use.