紫外吸收光谱结合SPA-ELM算法的水体磷酸根离子检测研究

doi:10.3964/j.issn.1000-0593(2024)01-0082-06

摘要
参考文献
相关文章 (15)

全文: PDF (4415 KB)
输出: BibTeX | EndNote (RIS)

摘要：在工业锅炉中随着水蒸气蒸发，大量的钙镁离子留在炉水中，如果不加处理，在水冷管中会形成水垢，造成爆管停炉。为了保障设备的安全运行，消除安全隐患，工业上通过维持水中一定含量的磷酸根离子来去除锅炉中的钙、镁水垢。传统的磷酸根离子检测技术主要有比色法、分光光度法、色谱法、电位法等，这些方法前期处理步骤较繁琐且耗时较长。光谱法是通过测定物质的吸收光谱并建立和浓度关系的数学模型，对物质浓度定量的一种分析方法。为了满足磷酸根离子快速、实时测量的要求，提出了一种基于紫外吸收光谱结合SPA-ELM算法快速测量磷酸根离子的方法。按照我国《工业锅炉水质GB/T 1576—2018》中所规定的进入热水锅炉前的水质参数要求，配置37份浓度范围在5~100 mg·L^-1磷酸根溶液，使用搭建的实验装置，采集紫外吸收光谱。使用SPXY将样品按照7∶3的比例随机划分训练集、测试集；使用Savitzky-Golay（S-G）滤波对数据预处理以提高光谱的信噪比；通过连续投影算法（SPA）压缩光谱数据，筛选出5个与磷酸根强相关的特征波长；使用极限学习机（ELM）将特征波长处的吸光度与样本浓度进行拟合，以决定系数R²和均方根误差RMSE作为模型评价指标，建立磷酸根离子的回归模型。采用所提出的建模方法所建立的模型训练集的R²与RMSE分别为0.997 2和1.301 5 mg·L^-1，测试集的R²与RMSE分别为0.999 5和0.517 4 mg·L^-1。为了验证所提出的SPA-ELM预测模型效果，另外建立了LASSO-ELM、PCA-ELM、SPA-PLS和SPA-SVR四种预测模型进行对比。实验结果表明，SPA-ELM建立的预测模型的R²和RMSE均优于其他四种预测模型，说明采用的特征选择方法和回归方法均为最优，能够对磷酸根浓度范围为5~100 mg·L^-1的水体进行准确预测，为水中磷酸根离子的检测提供了一种新的解决方法。

关键词：磷酸根离子；紫外吸收光谱；连续投影算法；极限学习机

Abstract：With the evaporation of water vapor in industrial boilers, a large amount of calcium and magnesium ions are left in the boiler water. If not treated, scale will form in the water-cooled tubes, causing tube explosion and boiler shutdown. In order to ensure the safe operation of the equipment and eliminate potential safety hazards, the calcium and magnesium scale in the boiler is removed by maintaining a certain amount of phosphate ions in the water.The traditional detection techniques for phosphate ions mainly include colorimetry, spectrophotometry, chromatography, potentiometry, etc. These methods have cumbersome and time-consuming preliminary processing steps. The spectroscopic method is an analytical method to quantify the concentration of a substance by measuring the absorption spectrum and establishing a mathematical model of the relationship between the concentration and the substance. A method for rapidly measuring phosphate ions based on ultraviolet absorption spectroscopy and the SPA-ELM algorithm was proposed. According to the water quality parameter requirements before entering the hot water boiler stipulated in “Industrial Boiler Water Quality GB/T 1576—2018”, 37phosphate ion solutions with the concentration range of 5~100 mg·L^-1 were prepared, and the UV absorption spectrum was collected using the established experimental equipment. The training and test sets were divided randomly according to the ratio of 7∶3 by SPXY. Data were preprocessed by Savitzky-Golay (S-G) filtering to improve the signal-to-noise ratio of the spectrum. The dimensionality of the spectrum was reduced by Successive Projection Algorithm(SPA). Five characteristic wavelengths strongly correlated with phosphate ionswere screened out. Finally, the Extreme Learning Machine(ELM) was used to fit the absorbance at the characteristic wavelength with the sample concentration, and the regression model of phosphate ions was established withR² and RMSE as the evaluation indexes of the model. TheR² and RMSE of the training set established by the method proposed in this paper are 0.997 2 and 1.301 5 mg·L^-1, and theR² and RMSE of the test set are 0.999 5 and 0.517 4 mg·L^-1. In order to verify the effect of the SPA-ELM prediction model proposed, four other prediction models, LASSO-ELM, PCA-ELM, SPA-PLS and SPA-SVR, were established for comparison. The experimental results show that theR² and RMSE of the prediction model established by SPA-ELM are better than those.It shows that both the feature selection and regression methods adopted in this paper are optimal. The modelling method adopted in this paper can accurately predict the water with phosphate concentration ranging from 5 to 100 mg·L^-1, which provides a new solution for detecting phosphate ions in water.

Key words：Phosphate ions; UV absorption spectrum; Successive projection algorithm; Extreme learning machine

收稿日期: 2022-12-02 修订日期: 2023-04-28

中图分类号:

O433.4

基金资助: 国家自然科学基金项目(61805030)，重庆市基础与前沿技术研究专项(cstc2020jcyj-msxmX0147)，重庆市教委科技项目(KJQN202000640, KJZD-M202200602)资助

通讯作者: 王金梅 E-mail: wangjm@cqupt.edu.cn

作者简介: 郑培超，1980年生，重庆邮电大学光电工程学院教授 e-mail: zhengpc@cqupt.edu.cn

引用本文:

郑培超，尹义同，王金梅，周椿棪，张莉，曾金锐，吕强. 紫外吸收光谱结合SPA-ELM算法的水体磷酸根离子检测研究[J]. 光谱学与光谱分析, 2024, 44(01): 82-87.
ZHENG Pei-chao, YIN Yi-tong, WANG Jin-mei, ZHOU Chun-yan, ZHANG Li, ZENG Jin-rui, LÜ Qiang. Study on the Method of Detecting Phosphate Ions in Water Based on Ultraviolet Absorption Spectrum Combined With SPA-ELM Algorithm. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2024, 44(01): 82-87.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2024)01-0082-06 或 https://www.gpxygpfx.com/CN/Y2024/V44/I01/82

[1] Tomasz Switkowski, Sylwester Kalisz, Joanna Wnorowska. Results in Engineering, 2022, 13: 100350.
[2] Ardy Husaini, Putra Yudhistira Perdana, Anggoro Adimas Dwi, et al. Heliyon, 2021, 7(2): e06151.
[3] Liu Shijie, Wu Haomin, Zhao Qinxin, et al. Engineering Failure Analysis, 2022, 142: 106847.
[4] van Puijenbroek P J T M, Beusen A H W, Bouwman A F. Journal of Environmental Management, 2019, 231: 446.
[5] Zhu Xiangyu, Ma Jian. TrAC Trends in Analytical Chemistry, 2020, 127: 115908.
[6] Cao Pingping, Zhu Yuanyang, Zhao Wenzhu, et al. Water, 2019, 11(11): 2339.
[7] Akhter Fowzia, Siddiquei H R, Alahi Md Eshrat E, et al. Sensors and Actuators A: Physical, 2021, 330: 112861.
[8] Mohammad S El-Shahawi, Hassan Alwael. Microchemical Journal, 2019, 149: 103987.
[9] Wei Hong, Pan Dawei, Zhou Zhengwen, et al. Ecotoxicology and Environmental Safety, 2021, 221: 112444.
[10] Chen Xiaowei, Yin Gaofang, Zhao Nanjing, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2021, 244: 118827.
[11] Yu Yinshan, Qu Yan. Optik, 2020, 217: 164915.
[12] Marie-Nolle Pons, Aziz Assaad, Chafatayne Oucacha, et al. Ecohydrology & Hydrobiology, 2017, 17(1): 46.
[13] Hugh L Kennedy. Digital Signal Processing, 2020, 102: 102743.
[14] ZHENG Pei-chao, ZHAO Wei-neng, WANG Jin-mei, et al(郑培超, 赵伟能, 王金梅, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2021, 41(1): 136.
[15] Gabriela Krepper, Florencia Romeo, David Douglas de Sousa Fernandes, et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2018, 189: 300.
[16] Yaseen Z M, Sulaiman S O, Deo R C, et al. Journal of Hydrology, 2019, 569: 387.
[17] Maciel Daniel Andrade, Barbosa Claudio Clemente Faria, Novo Evlyn Márcia Leão de Moraes, et al. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 182: 134.
[18] Lekomtsev Aleksander, Keykhosravi Amin, Moghaddam Mehdi Bahari, et al. Petroleum, 2022, 8(3): 424.
[19] Deeksha Chandola, Harsh Gupta, Vinay Anand Tikkiwal, et al. Procedia Computer Science, 2020, 167: 626.
[20] Spangenberg Martin, Bryant James I, Gibson Sam J, et al. Scientific Reports, 2021, 11(1): 3682.