基于主成分分析的BP神经网络模型在铁矿产地溯源中的应用

doi:10.13228/j.boyuan.issn1000-7571.011378

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (1096 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要铁矿在国家工业生产中占有重要地位,不同产地的铁矿在品质、有害元素含量上存在差异,为了提高海关对进口铁矿是否存在瞒报产地、以次充好等潜在风险的识别预警能力,尝试借助人工智能的方法来解决。以澳大利亚、巴基斯坦、巴西、南非、乌克兰和印度6个国家进口的共1 072批铁矿样品为研究样本,依据国家标准、行业标准对Hg、F、Cl、S、全铁(TFe)、SiO₂、Al₂O₃、P、H₂O、烧失量(L.O.I)、K₂O、TiO₂、MgO、CaO和Mn等15个指标进行检测。通过主成分分析提取了5个有效的主成分,累计方差贡献率为77.481%。以主成分分析提取的5个主要因子作为神经网络的输入指标,6个不同原产地作为输出指标,随机选取1 072组检测数据中的80%作为训练集,用以建立分析模型;剩下的20%作为验证集,用以验证模型的预测准确度,验证集识别准确率为100%,能够实现对6个国家铁矿原产地的准确识别。该模型的建立将直接应用于进口铁矿的产地鉴别,支撑进口铁矿的风险监管,维护贸易秩序,保障贸易便利。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	王虹
	魏秉炎
	王昊云

关键词 ：铁矿, 主成分分析, 产地溯源, BP神经网络

Abstract：Iron ore occupied an important position in national industrial production. The quality and harmful elements content varied for the iron ores from different origins. In order to improve the recognition and early warning ability of customs in potential risks of imported iron ores such as concealing the origin and selling as shoddy, the artificial intelligence was used to solve these problems. Total 1 072 batches of imported iron ore samples from Australia, Pakistan, Brazil, South Africa, Ukraine and India were selected as the research objects. Fifteen indexes including Hg, F, Cl, S, total iron (TFe), SiO₂, Al₂O₃, P, H₂O, L.O.I, K₂O, TiO₂, MgO, CaO and Mn were detected according to the national standard and industrial standard methods. Five effective principal components were extracted by principal component analysis. The cumulative variance contribution rate was 77.481%. Five main factors extracted in principal component analysis were used as the input indexes of neural network, and six different origins were used as the output indexes. 80% of the testing data in 1 072 groups were randomly selected as the training set to establish the analysis model. The rest data were used as validation set to verify the prediction accuracy of model. It was found that the recognition accuracy rate was 100%, which realized the accurate origin recognition of iron ores from six nations. The establishment of this model could be directly used for the origin identification of imported iron ores to regulate the risk of imported iron ores and secure the trade order and facilitation.

Key words： iron ore principal component analysis origin traceability BP neural network

收稿日期: 2020-12-30

ZTFLH:	TP183
	TF521

基金资助:天津海关课题(2019THK005)

作者简介: 王虹(1970—),女,高级工程师,大学本科,主要从事矿产品检测及相关课题研究工作;E-mail:wangh01@tjciq.gov.cn

引用本文:

王虹, 魏秉炎, 王昊云. 基于主成分分析的BP神经网络模型在铁矿产地溯源中的应用[J]. 冶金分析, 2021, 41(9): 11-17. WANG Hong, WEI Bingyan, WANG Haoyun. Application of BP neural network model based on principal component analysis in tracing the origin of iron ore. , 2021, 41(9): 11-17.

链接本文:

http://yjfx.chinamet.cn/CN/10.13228/j.boyuan.issn1000-7571.011378 或 http://yjfx.chinamet.cn/CN/Y2021/V41/I9/11

[1] 生态环境部.2021年将不再受理和审批固体废物进口相关申请[J].资源再生(Resource Recycling),2020(6):2.
[2] 王静静,房芳,周晓明,等.基于矿物元素含量的葡萄干产地溯源[J].新疆农业科学(Xinjiang Agricultural Sciences),2020,57(1):69-77.
[3] 王宁丽,于培良,赵立春,等.不同产地海地瓜活性物质检测及降糖活性评价[J].中国海洋药物,2019,38(2):11-16.
WANG Ningli,YU Peiliang,ZHAO Lichun,et al.Determination of active substances and evaluation of hypoglycemic activity in acaudinamolpadioides from different producing areas[J].Chinese Journal of Marine Drugs,2019,38(2):11-16.
[4] 贺媛媛,孙倩倩,郭波莉,等.矿质元素指纹分析在野葛产地溯源中的应用[C]//中国食品科学技术学会第十六届年会暨第十届中美食品业高层论坛.武汉:中国食品科学技术学会,2019.
[5] 杨健,周利,胡玲,等.ICP-MS法分析不同产地何首乌中40种无机元素[J].世界中医药,2019,14(11):2819-2828.
YANG Jian,ZHOU Li,HU Ling,et al.Quantitative analysis of 40 inorganic elements in Polygoni Multiflori Radix from different origins by ICP-MS[J].World Chinese Medicine,2019,14(11):2819-2828.
[6] 任春生,付冉冉,王艳,等.SPSS软件对铁矿中微量元素含量的统计分析[J].化学分析计量,2014,23(2):80-84,93.
REN Chunsheng,FU Ranran,WANG Yan,et al.Statistical analysis of trace elements in iron ores by SPSS Software[J].Chemical Analysis and Meterage,2014,23(2):80-84,93.
[7] 陈健,刘文中,陈萍.SPSS在淮南矿区煤中微量元素研究中的应用[J].选煤技术(Goal Preparation Technology),2009(2):46-50,81.
[8] 魏红兵,刘青山,孙世明,等.矿物组分“数字指纹”模型的建立和应用[J].工业技术经济,2010(11):107-111.
WEI Hongbing,LIU Qingshan,SUN Shiming,et al.Establishment and application of mineral constituents "digital fingerprint" model[J].Industrial Technology & Economy,2010(11):107-111.
[9] 武素茹,谷松海,宋义,等.进口铁矿产地鉴别模型的建立[J].计算机与应用化学,2014,31(12):1543-1546.
WU Suru,GU Songhai,SONG Yi,et al.Establishment of origin identification model of import iron ores[J].Computers and Applied Chemistry,2014,31(12):1543-1546.
[10] 张博,闵红,刘曙,等.X射线荧光光谱结合判别分析识别进口铁矿石产地及品牌[J].光谱学与光谱分析,2020,40(8):2640-2646.
ZHANG Bo,MIN Hong,LIU Shu,et al.X-ray fluorescence spectroscopy combined with discriminant analysis to identify imported iron ore origin and brand[J].Spectroscopy and Spectral Analysis,2020(8):2640-2646.
[11] 张开春,吴丽萍,姚军,等.X荧光谱与人工神经网络相结合对陶片产地识别的研究[J].核技术,2006(11):56-60.
ZHANG Kaichun,WU Liping,YAO Jun,et al.Study of recognition of production areas for ceramic fragments by X-fluorescence spectrum combined with artificial neural network[J].Nuclear Techniques,2006(11):56-60.
[12] 陈永明,林萍,何勇.基于遗传算法的近红外光谱橄榄油产地鉴别方法研究[J].光谱学与光谱分析,2009,29(3):671-674.
CHEN Yongming,LIN Ping,HE Yong.Study on discrimination of producing area of olive oil using near infrared spectra based on genetic algorithms[J].Spectroscopy and Spectral Analysis,2009,29(3):671-674.
[13] Giulio Binetti,Coco L D,Ragone R,et al.Cultivar classification of apulian olive oils: use of artificial neural networks for comparing NMR, NIR and merceological data[J].Food Chemistry,2017,219:131-138.
[14] Khajehzadeh N,Haavisto O,Koresaar L.On-stream mineral identification of tailing slurries of an iron ore concentrator using data fusion of LIBS, reflectance spectroscopy and XRF measurement techniques[J].Minerals Engineering,2017,113:83-94.
[15] 张立明.人工神经网络的模型及其应用[M].上海:复旦大学出版社,1993:46.
[16] 吕砚山,赵正琦.BP神经网络的优化及应用研究[J].北京化工大学学报(自然科学版),2001(1):67-69.
LÜ Yanshan,ZHAO Zhengqi.Optimization and application research of BP neural network[J].Journal of Beijing University of Chemical Technology(Natural Science Edition),2001(1):67-69.