变量重要性-反向传播人工神经网络辅助激光诱导击穿光谱测定铁矿石中硅、 铝、 钙和镁含量

刘曙; 金悦; 苏飘; 闵红; 安雅睿; 吴晓红

doi:doi:10.3964/j.issn.1000-0593(2023)10-3132-11

光谱学与光谱分析, 2023, 43 (10): 3132, 网络出版: 2024-01-11

变量重要性-反向传播人工神经网络辅助激光诱导击穿光谱测定铁矿石中硅、铝、钙和镁含量

Determination of Calcium, Magnesium, Aluminium and Silicon Content in Iron Ore Using Laser-Induced Breakdown Spectroscopy Assisted by Variable Importance-Back Propagation Artificial Neural Networks

论文大纲

刘曙 ¹金悦 ²苏飘 ²闵红 ¹安雅睿 ³吴晓红 ¹

作者单位

¹ 上海海关工业品与原材料检测技术中心, 上海 200135

² 上海海关工业品与原材料检测技术中心, 上海 200135上海理工大学材料与化学学院, 上海 200093

³ 上海理工大学材料与化学学院, 上海 200093

摘要

快速准确测定铁矿石中的硅、铝、钙、镁含量对铁矿石质量评价具有重要作用。受制于多变量分析方法过拟合现象以及不同种类样品基体效应, 使用激光诱导击穿光谱（LIBS）准确测定铁矿石中硅、铝、钙、镁含量仍然是当前存在的挑战。采用变量重要性-反向传播人工神经网络（VI-BP-ANN）辅助LIBS定量分析铁矿石中硅（以SiO2计）、铝（以Al2O3计）、钙（以CaO计）和镁（以MgO计）的含量。在这项研究中, 收集了12种244批铁矿石代表性样品的LIBS光谱, 优化了光谱预处理方法, 使用随机森林（RF）对LIBS光谱特征的重要性进行了测量, 使用袋外(OOB)误差优化RF模型参数, 变量重要性阈值用于优化BP-ANN校准模型的输入变量。变量重要性阈值和神经元数量通过五折交叉验证（5-CV）的测定系数（R2）和均方根误差（RMSE）进行优化。结果显示测试样本SiO2、 Al2O3、 CaO和MgO含量预测均方根误差（RMSEP）分别为0.377 2 wt%、 0.133 9 wt%、 0.059 2 wt%和0.141 1 wt%, R2分别为0.970 1、 0.955 4、 0.987 1、 0.997 5。相比于使用相同的预处理方法作为PLS、 SVM、 RF和BP-ANN四种模型的输入, VI-BP-ANN在校准集和预测集都显示出出色的预测能力。结果表明LIBS与VI-BP-ANN的结合有潜力在实际应用中实现铁矿石硅、铝、钙、镁含量的快速准确预测。

Abstract

The rapid and accurate determination of calcium, magnesium, aluminium and silicon content in iron ore plays an important role in iron ore quality assessment. The accurate determination of calcium (CaO), magnesium (MgO), aluminium (Al2O3) and silicon (SiO2) in iron ore using laser-induced breakdown spectroscopy (LIBS) remains a challenge due to the overfitting of multivariate analysis methods and matrix effects between different types of samples. In this paper, variable importance-back propagation artificial neural network (VI-BP-ANN) assisted LIBS was used for the first time to quantify the content of SiO2, Al2O3, CaO and MgO in iron ore. In this study, LIBS spectra of 12 representative samples of 244 batches of iron ore were collected, spectral pre-processing methods were optimised, the importance of LIBS spectral features was measured using random forest (RF), RF model parameters were optimised using out-of-bag (OOB) errors, and variable importance thresholds were used to optimise the input variables for the BP-ANN calibration model. The variable importance thresholds and the number of neurons were optimised by five-fold cross-validation (5-CV) of the coefficient of determination (R2) and root mean square error (RMSE). The results showed root mean square error of prediction (RMSEP) for the SiO2, Al2O3, CaO, MgO content of the test samples were 0.372 3 wt%, 0.129 8 wt%, 0.052 4 wt% and 0.149 0 wt% respectively, with R2 of 0.977 1, 0.950 4, 0.987 8 and 0.997 7, respectively. Compared to using the same preprocessing method as input to the three PLS, SVM and RF models, the VI-BP- ANN model showed excellent performance in both the calibration dataset and prediction dataset. The results indicate that the combination of LIBS and VI-BP-ANN has the potential to achieve fast and accurate prediction of calcium, magnesium, aluminium and silicon content of iron ore in practical application.

参考文献

[1] Wang P, Li N, Yan C, et al. Analytical Methods, 2019, 11(27): 3419.

[2] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.13—2007 Iron Ores-Determination of Calcium and Magnesium Content-EGTA-CyDTA Titrimetric Method(GB/T 6730.13—2007 铁矿石钙和镁含量的测定 EGTA-CyDTA滴定法), 2007.

[3] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.11—2007 Iron Ores-Determination of Aluminium Content-EDTA Titrimetric Method (GB/T 6730.11—2007 铁矿石铝含量的测定 EDTA滴定法), 2007.

[4] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.9—2016 Iron Ores—Determination of Silicon Content—The Silicomolybdic Blue Spectrophotometric Method Reduced by Ammonium Ferrous Sulfate (GB/T 6730.9—2016 铁矿石硅含量的测定硫酸亚铁铵还原-硅钼蓝分光光度法), 2016.

[5] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.14—2017 Iron Ores—Determination of Calcium Content—Flame atomic Absorption Spectrometric Method (GB/T 6730.14—2017 铁矿石钙含量的测定火焰原子吸收光谱法), 2017.

[6] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.74—2017 Iron Ores—Determination of Magnesium Content—Flame Atomic Absorption Spectrometric Method (GB/T 6730.74—2017 铁矿石镁含量的测定火焰原子吸收光谱法), 2017.

[7] State Administration for Market Regulation and Standardization Administration of China(国家市场监督管理总局, 中国国家标准化管理委员会). GB/T 6730.56—2019 Iron Ores—Determination of Aluminum Content—Flame Atomic Absorption Spectrometric Method (GB/T 6730.56—2019 铁矿石铝含量的测定火焰原子吸收光谱法), 2019.

[8] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.63—2006 Iron Ores—Determination of Aluminum, Calcium, Magnesium, Manganese, Phosphorus, Silicon and Titanium Content—Inductively Coupled Plasma Atomic Emission Spectrometric Method (GB/T 6730.63—2006 铁矿石铝、钙、镁、锰、磷、硅和钛含量的测定电感耦合等离子体发射光谱法), 2006.

[9] General Administration of Quality Supervision, Inspection and Quarantine of the Peoples Republic of China and Standardization Administration of China(中华人民共和国国家质量监督检验检疫总局, 中国国家标准化管理委员会). GB/T 6730.62—2005 Iron Ores—Determination of Calcium, Silicon, Manganese, Titanium, Phosphorus Magnesium, Aluminium and Barium Content Wavelength Dispersive X-ray Fluorescence Spectrometric Method (GB/T 6730.62—2005 铁矿石钙、硅、镁、钛、磷、锰、铝和钡含量的测定波长色散X射线荧光光谱法), 2005.

[10] Chen T, Zhang T, Li H. TrAC Trends in Analytical Chemistry, 2020, 133: 116113.

[11] YANG Ya-wen, YAN Cheng-lin, XU Ding, et al(杨雅雯, 严承琳, 徐鼎, 等). Metallurgical Analysis(冶金分析), 2020, 40(12): 14.

[12] Wang Z, Afgan M S, Gu W, et al. TrAC Trends in Analytical Chemistry, 2021, 143: 116385.

[13] Grant K J, Paul G L, ONeill J A. Applied Spectroscopy, 1991, 45(4): 701.

[14] Death D L, Cunningham A P, Pollard L J. Spectrochimica Acta Part B: Atomic Spectroscopy, 2008, 63(7): 763.

[15] Hao Z Q, Li C M, Shen M, et al. Optics Express, 2015, 23(6): 7795.

[16] Guo Y M, Guo L B, Hao Z Q, et al. Journal of Analytical Atomic Spectrometry, 2018, 33(8): 1330.

[17] ZHAO Wen-ya, MIN Hong, LIU Shu, et al(赵文雅, 闵红, 刘曙, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2021, 41(7): 1998.

[18] Wang J, Shi M, Zheng P, et al. Journal of Applied Spectroscopy, 2018, 85(1): 190.

[19] Sun C, Tian Y, Gao L, et al. Scientific Reports, 2019, 9: 11363.

[20] Yang Y, Li C, Liu S, et al. Analytical Methods, 2020, 12(10): 1316.

[21] Liu K, Tian D, Xu H, et al. Analytical Methods, 2019, 11(37): 4769.

[22] Tang H, Zhang T, Yang X, et al. Analytical Methods, 2015, 7(21): 9171.

[23] Liu K, Tian D, Wang H, et al. Analytical Methods, 2019, 11(9): 1174.

刘曙, 金悦, 苏飘, 闵红, 安雅睿, 吴晓红. 变量重要性-反向传播人工神经网络辅助激光诱导击穿光谱测定铁矿石中硅、铝、钙和镁含量[J]. 光谱学与光谱分析, 2023, 43(10): 3132. LIU Shu, JIN Yue, SU Piao, MIN Hong, AN Ya-rui, WU Xiao-hong. Determination of Calcium, Magnesium, Aluminium and Silicon Content in Iron Ore Using Laser-Induced Breakdown Spectroscopy Assisted by Variable Importance-Back Propagation Artificial Neural Networks[J]. Spectroscopy and Spectral Analysis, 2023, 43(10): 3132.

变量重要性-反向传播人工神经网络辅助激光诱导击穿光谱测定铁矿石中硅、铝、钙和镁含量

关于本站 Cookie 的使用提示

全站搜索

变量重要性-反向传播人工神经网络辅助激光诱导击穿光谱测定铁矿石中硅、 铝、 钙和镁含量

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索

变量重要性-反向传播人工神经网络辅助激光诱导击穿光谱测定铁矿石中硅、铝、钙和镁含量