煤种成浆性的人工神经网络预测模型研究Mathematical model based on artificial neural network for coal slurrying prediction
李艳昌,杨雨濛,刘建忠
LI Yanchang,YANG Yumeng,LIU Jianzhong
摘要(Abstract)：
为更好地预测煤的成浆性,以大量煤种成浆浓度试验数据为基础,建立了3个输出因子的神经网络成浆浓度预测模型,模型采用L-M算法,对输入数据进行数据预处理,最后对比分析了神经网络预测模型与回归分析模型的预测结果。结果表明,以A_d、哈氏可磨性指数HGI和氧含量O为输入因子的模型预测结果平均绝对误差为0.63%,以M_(ad)、HGI和O为输入因子的模型预测结果平均绝对误差为0.60%,以M_(ad)、HGI和氧碳比O/C为输入因子的模型预测结果平均绝对误差为0.40%,3种组合的模型结果均小于回归分析模型的平均绝对误差1.15%。因此神经网络模型比回归分析模型有更好的预测能力,其中以M_(ad)、HGI和O/C为输入因子的神经网络模型预测结果最好。
In order to improve prediction accuracy,based on experimental data of coal slurrying,BP neural network model with three input factors was set up for slurry concentration prediction. The BP neural networks' algorithm was Levenberg- Marquardt algorithm. The input data was treated in order to get accurate results. The A_d,HGI,O input factors neural network model's mean absolute errors was 0. 63%,the M_(ad),HGI,O model's mean absolute error was 0. 60% and the M_(ad),HGI,O/C model's mean absolute error was 0. 40%,but the exist regression model's mean absolute error was 1. 15%,so the neural network models were effective in predicting the slurrying,and the M_(ad),HGI,O / C model was the best among the three prediction models.
关键词(KeyWords)：
煤;成浆性;神经网络;L-M算法
coal;slurryability;neural network;Levenberg-Marquardt algorithm
基金项目(Foundation): 国家重点基础研究发展计划(973计划)资助项目(2010CB227001)
作者(Author):
李艳昌,杨雨濛,刘建忠
LI Yanchang,YANG Yumeng,LIU Jianzhong
DOI: 10.13226/j.issn.1006-6772.2016.01.002
参考文献(References)：
- [1]张荣曾.水煤浆制浆技术[M].北京:科学出版社,1996.
- [2]陈训刚.微光波对水煤浆的成浆性和气化性的影响[D].杭州:浙江大学,2008.
- [3]李珊珊.精细水煤浆的颗粒孔隙-成浆及燃烧机理研究[D].杭州:浙江大学,2006.
- [4]王睿坤,刘建忠,胡亚轩,等.水煤浆掺混湿污泥对浆体成浆特性的影响[J].煤炭学报,2010,35(S1):199-204.Wang Ruikun,Liu Jianzhong,Hu Yaxuan,et al.Influence of wet sludge on the slurrying properties of coal-water slurries[J].Journal of China Coal Society,2010,35(S1):199-204.
- [5]朱雪丹.粒度级配对神府煤成浆性能的影响[J].洁净煤技术,2014,20(4):1-3,6.Zhu Xuedan.Influence of grading size on slurryability of Shenfu coal[J].Clean Coal Technology,2014,20(4):1-3,6.
- [6]Atesok G,Boylu F,Sirkeci A A,et al.The effect of coal properties on the viscosity of coal-water slurries[J].Fuel,2002,81(14):1855-1858.
- [7]Kaushal K Tiwari,Sibendra K Basu,Kumaresh C Bit,et al.Highconcentration coal-water slurry from Indian coals using newly developed additives[J].Fuel Processing Technology,2003,85(1):31-42.
- [8]Mishra S K,Senapati P K,Panda D.Rheological behavior of coalwater slurry[J].Energy Sources,2002,24(2):159-167.
- [9]杨建刚.人工神经网络实用教程[M].杭州:浙江大学出版社,2001.
- [10]Martin T Hagan,Mohammad B Memhaj.Training feedforward network with the marquardt algorithm[J].IEEE Transactions on Neural Networks,1994,5(6):989-993.
- [11]殷春根,骆仲泱,倪明江,等.煤的工业分析至元素分析的BP神经网络预测模型[J].燃料化学学报,1999,27(5):408-414.Yin Chungen,Luo Zhongyang,Ni Mingjiang,et al.Relationship between ultimate analysis of any coal and its proximate analysis data[J].Journal of Fuel Chemistry and Technology,1999,27(5):408-414.
- [12]Li Yongxin,Li Baoqing.Study on the ultrasonic irradiation of coal water slurry[J].Fuel,2000,79(3/4):235-241.
- [13]于敦喜,徐明厚,刘小伟,等.燃煤可磨性指数的人工神经网络预测[J].煤炭技术,2003,22(9):91-93.Yu Dunxi,Xu Minghou,Liu Xiaowei,et al.Forecast of artificial neural network for burn coal grindaiblity index[J].Coal Technology,2003,22(9):91-93.