基于有机元素含量的猪粪热值预测模型

探讨利用猪粪有机元素含量预测猪粪热值的可行性,建立基于C、H、N、S、O等5种有机元素含量的高、低位热值预测模型.采集了160个猪粪样品,利用EA3000型元素分析仪和IKA-C2000型氧弹量热仪分别测定其C、H、N、S、O元素的含量和高、低位热值.相关分析表明,C、H、N这3种元素的含量与热值之间存在极显著相关,而S、O这2种元素含量与热值之间的相关不显著;利用120个猪粪样品的测定结果,采用多元线性回归分析建立的基于C、H、N含量的热值预测模型可以较准确地利用碳、氢、氮预测猪粪热值,高位热值和低位热值预测模型的决定系数R2分别为0.836和0.785.用40个样品对热值模型进行外部验证,高位热值和低位热值均方根误差RMSE分别为621.31、620.85 J/g,相对均方根误差RRMSE分别为3.74％和4.06％.结果表明,多元线性回归法建立的基于有机元素含量的猪粪热值预测模型的预测结果较准确,利用猪粪的元素含量预测其热值是可行的.

Construction calorific value model for pig manure based on organic element contents

To discuss the feasibility of using organic element contents of pig manure to predict calorific value,the paper built a prediction model of gross calorific value and net calorific value based on the contents of organicelements. 160 pig manure samples were collected, and their C,H,N,S,O element contents were measured by EA3000and gross calorific value and net calorific value were measured by IKA-C2000 bomb calorimeters. Simple correlationanalysis result showed that the contents of C,H,N had very significantly positive correlation with calorific value, whilethe contents of S and O had no significantly positive correlation with calorific value. 120 samples were selected tobuild the model, results indicated that the calorific value prediction model based on C,H,N element contents usingmultiple linear regression analysis could accurately use C,H,N to predict calorific value of pig manure. Thedetermination coefficient of the prediction model of gross calorific value and net calorific value were 0.836 and 0.785respectively. In the 40 samples used for the external validation, the root-mean-square error of gross calorific valuewas 621.31 J/g, and the relative root-mean-square error was 3.74%; the root-mean-square error of net calorificvalue was 620.85 J/g, and the relative root-mean-square error was 4.06%. It was concluded that the calorific valueprediction model of pig manure based on organic element contents by built using multiple linear regression analysismethod could predict the gross and net calorific values accurately.