杉木砂光粉尘爆炸最小点火能预测模型  

The prediction model of minimum ignition energy for wood dust explosions

作　　者：李时维 喻孜 颜振东 刘海良 周捍东 LI Shiwei;YU Zi;YAN Zhendong;LIU Hailiang;ZHOU Handong(College of Materials Science and Engineering,Nanjing Forestry University,Nanjing 210037,China;Jiangsu Senmao Bamboo and Wood Industry Co.Ltd.,Yixing 214231,China)

机构地区：[1]南京林业大学材料科学与工程学院,南京210037 [2]江苏森茂竹木业有限公司,宜兴214231

出　　处：《林业工程学报》2025年第2期60-66,共7页Journal of Forestry Engineering

基　　金：国家重点研发计划(2022YFD2200705)。

摘　　要：旨在开发一个基于XGBoost(极端梯度提升)算法的杉木砂光粉最小点火能(MIE)预测模型,采用XGBoost技术对多个影响因素如粉尘质量浓度、含水率、喷射压力和比表面积径等进行综合分析,构建杉木砂光粉MIE的预测模型,并使用R^(2)、MAE、RMSE、MAD和MAPE 5个评价指标来综合评价模型的整体性能和该模型对各因素的权重。结果表明:在杉木砂光粉MIE预测模型研究中,此模型在训练集和测试集上的R^(2)分别为0.99961和0.96905,展现了高度的预测准确性。对于MIE上限的预测,模型在训练集和测试集的R^(2)分别为0.99971和0.98638,进一步证实了其有效性。在误差分析中,MAE、RMSE、MAD和MAPE均表现出模型在训练集上的高预测精度,且研究中较低的MAPE值表明模型预测与实际值之间的百分比误差较小,说明模型在泛化能力上表现良好。使用XGBoost模型对各因素权重分析表明,粉尘质量浓度是对MIE预测影响最大的因素。通过XGBoost模型的应用,不仅为杉木砂光粉MIE的预测提供了新视角,同时也为木材加工行业促进生产过程的安全管理提供了一种有效的风险评估工具。The study employed the XGBoost algorithm to develop the prediction model for the minimum ignition energy(MIE)of sanding dust of Chinese fir wood.XGBoost is a powerful machine learning technique that combines multiple weak learners to create a strong learner,effectively capturing complex relationships among variables.The model considered four key influencing factors,i.e.,dust mass concentration,moisture content,blowing pressure,and specific surface area diameter.These factors were used as input variables,while the MIE served as the target variable.The dataset was split into training and test sets to evaluate the model s performance and generalization ability.The results showed that the XGBoost model developed in this study demonstrated remarkable performance in predicting the MIE of sanded Chinese fir wood dust.The model s effectiveness was validated through a comprehensive evaluation using five key indicators,i.e.,coefficient of determination(R^(2)),mean absolute error(MAE),root mean square error(RMSE),mean absolute deviation(MAD),and mean absolute percentage error(MAPE).The model achieved an impressive R^(2)value of 0.99961 on the training set and 0.96905 on the test set,indicating that it had outstanding ability to capture the complex relationships between the input variables and the target variable.The model s robustness was further confirmed by its excellent performance in predicting the upper limit of MIE,with R^(2)values of 0.99971 and 0.98638 on the training and test sets,respectively.Error analysis using MAE,RMSE,MAD and MAPE demonstrated the model s high prediction accuracy.The low values of these metrics on the training set indicated that the model s predictions closely align with the actual MIE values.The low MAPE value suggested that the model s predictions deviated from the actual values by a small percentage,confirming its reliability and consistency across the range of MIE values.The XGBoost model also provided insights into the relative importance of the influencing factors.The weight analysis rev

关 键 词：杉木粉尘 最小点火能 预测模型 XGBoost 误差分析 

分 类 号：X932[环境科学与工程—安全科学]