LIBS combined with SG-SPXY spectral data pre-processing for cement raw meal composition analysis

Yongqi Cai; Xiaoyu Ma; Bing Huang; Rongfeng Zhang; Xiaohong Wang; Xiaohong Wang

doi:10.1364/AO.505255

Applied Optics
Vol. 63,
Issue 6,
pp. A24-A31
(2024)
•https://doi.org/10.1364/AO.505255

LIBS combined with SG-SPXY spectral data pre-processing for cement raw meal composition analysis

Yongqi Cai, Xiaoyu Ma, Bing Huang, Rongfeng Zhang, and Xiaohong Wang

Open Access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Yongqi Cai, Xiaoyu Ma, Bing Huang, Rongfeng Zhang, and Xiaohong Wang, "LIBS combined with SG-SPXY spectral data pre-processing for cement raw meal composition analysis," Appl. Opt. 63, A24-A31 (2024)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Check for updates

Related Topics
Table of Contents Category
- Lasers, Optical Amplifiers, and Laser Optics
Optics & Photonics Topics
?

The topics in this list come from the Optics and Photonics Topics applied to this article.

About this Article
History
- Original Manuscript: September 8, 2023
- Revised Manuscript: December 9, 2023
- Manuscript Accepted: January 5, 2024
- Published: January 23, 2024
Virtual Issues
Applied Optics Radiography, Applied Optics, and Data Science 2023 (2024)

Abstract

Rapid testing of cement raw meal plays a crucial role in the cement production process, so there is an urgent need for a fast and accurate testing method. In this paper, a method based on the Savitzky-Golay (SG) smoothing and sample set partitioning based on joint $x\text{-}y$ distance (SPXY) spectral data pre-processing is proposed to improve the accuracy of the laser-induced breakdown spectroscopy (LIBS) technique for quantitative analysis of cement raw meal components. Firstly, the spectral data is denoised by SG smoothing, which effectively reduces the noise and baseline variations in the spectra. Then, the denoised data is divided into sample sets by combining the SPXY sample division method, which improves the efficiency of data analysis. Finally, the delineated data set is modeled for quantitative analysis by a back-propagation (BP) neural network. Compared to the modeling effect of the four oxide contents of CaO, ${\rm{Si}}{{\rm{O}}_2}$, ${\rm{A}}{{\rm{l}}_2}{{\rm{O}}_3}$, and ${\rm{F}}{{\rm{e}}_2}{{\rm{O}}_3}$ in the Hold-Out method, the correlation coefficient (R) was improved by 26%, 10%, 17%, and 4%, respectively. The root mean square error (RMSE) was reduced by 47%, 33%, 43%, and 21%, respectively. The mean absolute percentage error (MAPE) was reduced by 63%, 60%, 36%, and 51%, respectively. The results show that there is a significant improvement in the model effect, which can effectively improve the accuracy of quantitative analysis of cement raw meal composition by LIBS. This is of great significance for the real-time detection of cement raw meal composition analysis.

Full Article | PDF Article

More Like This

Accurate determination of plasma temperature and electron density using a reference target: one-point calibration LIBS elemental analysis of alloy samples

Huiling Zeng, Runhua Li, and Yuqi Chen
Appl. Opt. 63(2) 338-344 (2024)

Meat species identification accuracy improvement using sample set portioning based on joint x–y distance and laser-induced breakdown spectroscopy

Lianbo Guo, Weinan Zheng, Feng Chen, Weiliang Wang, Deng Zhang, Zhenlin Hu, and Yanwu Chu
Appl. Opt. 60(20) 5826-5831 (2021)

Surface hardness determination of laser cladding using laser-induced breakdown spectroscopy and machine learning (PLSR, CNN, ResNet, and DRSN)

Jiacheng Yang, Linghua Kong, and Hongji Ye
Appl. Opt. 63(10) 2509-2517 (2024)

Previous Article Next Article

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.

Fig. 1. LIBS experimental setup.

Download Full Size | PDF

Fig. 2. LIBS spectral information for cement raw materials.

Download Full Size | PDF

Fig. 3. Spectral data of initial cement raw meal.

Download Full Size | PDF

Fig. 4. Spectral data of cement raw meal after SG smoothing.

Download Full Size | PDF

Fig. 5. Two-dimensional visualization of the number of samples and oxide concentration levels in the training set.

Download Full Size | PDF

Fig. 6. Correlation coefficients for training models based on the Hold-Out method.

Download Full Size | PDF

Fig. 7. Comparison of predicted and actual values of training models based on the Hold-Out method.

Download Full Size | PDF

Fig. 8. Correlation coefficients for training models based on the SG-SPXY method.

Download Full Size | PDF

Fig. 9. Comparison of predicted and actual values of training models based on the SG-SPXY method.

Download Full Size | PDF

Tables (3)

Table 1. Concentration Distribution of the Four Main Oxides in 32 Samples

View Table | View all tables in this article

Table 2. Division of Training and Test Sets of Cement Raw Material Samples

View Table | View all tables in this article

Table 3. Modeling Results

View Table | View all tables in this article

Equations (13)

Equations on this page are rendered with MathJax. Learn more.

\bar{I_{i}} = \frac{1}{110} \sum_{n = 1}^{110} I_{i, n},

A_{n} = \sum_{i = 1}^{M} {(I_{i, n} - \bar{I_{i}})}^{2},

\bar{A} = \frac{1}{110} \sum_{n = 1}^{110} A_{n},

X = [x_{1}, x_{2}, x_{3}, \dots, x_{n}],

W = {[w_{1}, w_{2}, w_{3}, \dots, w_{n}]}^{T},

{n e t}_{i} = \sum_{j = 1}^{n} w_{ij} x_{j} + (- 1 * θ) = X W,

y_{i} = f ({n e t}_{i}) = f (X W) .

d_{x} (p, q) = \sqrt{\sum_{j = 1}^{J} {[x_{p} (j) - x_{q} (j)]}^{2}}; p, q \in [1, N],

d_{y} (p, q) = \sqrt{{(y_{p} - y_{q})}^{2}} = | y_{p} - y_{q} |; p, q \in [1, N],

d_{xy} (p, q) = \frac{d_{x} (p, q)}{max d_{x} (p, q)} + \frac{d_{y} (p, q)}{max d_{y} (p, q)}; p, q \in [1, N] .

R = \frac{\sum_{i = 1}^{n} [(x_{i} - \bar{x}) (y_{i} - \bar{y})]}{\sqrt{\sum_{i = 1}^{n} {(x_{i} - \bar{x})}^{2} \cdot \sum_{i = 1}^{n} {(y_{i} - \bar{y})}^{2}}},

R M S E = \sqrt{\frac{\sum_{i = 1}^{n} {(s_{i} - r_{i})}^{2}}{n}},

M A P E = \frac{\sum_{i = 1}^{n} (\frac{| r_{i} - s_{i} |}{r_{i}} \cdot 100 %)}{n} .

Sample Number	CaO/%	$S i O_{2}$ /%	$F e_{2} O_{3}$ /%	$A l_{2} O_{3}$ /%	Sample Number	CaO/%	SiO₂/%	Fe₂O₃/%	Al₂O₃/%
1	41.79	2.96	2.96	3.33	17	53.59	24.1	3.2	6.93
2	45.1	2.42	2.42	2.24	18	54.14	23.78	3.19	6.56
3	39.22	2.15	2.15	3.83	19	55.5	23.3	3.05	6.33
4	38.84	0.49	0.49	3.22	20	57.08	23.13	3	6.1
5	46	10.5	1.58	2.06	21	58.14	22.8	2.91	5.82
6	45.39	10.66	1.63	2.2	22	59.44	22.56	2.83	5.55
7	44.63	11.31	1.76	2.44	23	60.63	22.27	2.74	5.32
8	43.85	11.92	1.98	2.64	24	61.5	22.02	2.69	5.09
9	43.17	12.47	2.09	2.84	25	62.65	21.81	2.58	4.78
10	41.91	13.56	2.46	3.14	26	64.01	21.45	2.49	4.46
11	41.16	14	2.51	3.33	27	65.78	22.07	3.22	4.56
12	40.44	14.52	2.64	3.48	28	57.22	21.36	2.78	6.5
13	39.37	15.33	2.84	3.73	29	62.23	20.41	3.2	4.68
14	38.3	16.23	3.17	3.96	30	42.4	13.76	2.13	3.12
15	37.61	16.71	3.16	4.16	31	47.57	32.67	3.54	6.52
16	51.18	25.49	3.31	6.96	32	55.21	24.78	3	7.24

Oxide	Subset	Number	Max/%	Min/%
CaO	Train set	24	65.78	37.61
CaO	Test set	8	53.59	38.3
$S i O_{2}$	Train set	24	32.67	10.5
$S i O_{2}$	Test set	8	23.78	10.66
$A l_{2} O_{3}$	Train set	24	7.24	2.06
$A l_{2} O_{3}$	Test set	8	6.93	2.84
$F e_{2} O_{3}$	Train set	24	3.57	0.49
$F e_{2} O_{3}$	Test set	8	3.22	2.09

Oxides	Models	R	RMSE	MAPE/%
CaO	Hold-Out	0.72	0.43	24.18
CaO	SG-SPXY	0.91	0.23	8.94
$S i O_{2}$	Hold-Out	0.89	0.18	17.8
$S i O_{2}$	SG-SPXY	0.98	0.12	7.15
$A l_{2} O_{3}$	Hold-Out	0.77	0.46	28.62
$A l_{2} O_{3}$	SG-SPXY	0.90	0.26	18.23
$F e_{2} O_{3}$	Hold-Out	0.83	0.24	20.05
$F e_{2} O_{3}$	SG-SPXY	0.86	0.19	9.82

Sample Number	CaO/%	$S i O_{2}$ /%	$F e_{2} O_{3}$ /%	$A l_{2} O_{3}$ /%	Sample Number	CaO/%	SiO₂/%	Fe₂O₃/%	Al₂O₃/%
1	41.79	2.96	2.96	3.33	17	53.59	24.1	3.2	6.93
2	45.1	2.42	2.42	2.24	18	54.14	23.78	3.19	6.56
3	39.22	2.15	2.15	3.83	19	55.5	23.3	3.05	6.33
4	38.84	0.49	0.49	3.22	20	57.08	23.13	3	6.1
5	46	10.5	1.58	2.06	21	58.14	22.8	2.91	5.82
6	45.39	10.66	1.63	2.2	22	59.44	22.56	2.83	5.55
7	44.63	11.31	1.76	2.44	23	60.63	22.27	2.74	5.32
8	43.85	11.92	1.98	2.64	24	61.5	22.02	2.69	5.09
9	43.17	12.47	2.09	2.84	25	62.65	21.81	2.58	4.78
10	41.91	13.56	2.46	3.14	26	64.01	21.45	2.49	4.46
11	41.16	14	2.51	3.33	27	65.78	22.07	3.22	4.56
12	40.44	14.52	2.64	3.48	28	57.22	21.36	2.78	6.5
13	39.37	15.33	2.84	3.73	29	62.23	20.41	3.2	4.68
14	38.3	16.23	3.17	3.96	30	42.4	13.76	2.13	3.12
15	37.61	16.71	3.16	4.16	31	47.57	32.67	3.54	6.52
16	51.18	25.49	3.31	6.96	32	55.21	24.78	3	7.24

Oxide	Subset	Number	Max/%	Min/%
CaO	Train set	24	65.78	37.61
CaO	Test set	8	53.59	38.3
$S i O_{2}$	Train set	24	32.67	10.5
$S i O_{2}$	Test set	8	23.78	10.66
$A l_{2} O_{3}$	Train set	24	7.24	2.06
$A l_{2} O_{3}$	Test set	8	6.93	2.84
$F e_{2} O_{3}$	Train set	24	3.57	0.49
$F e_{2} O_{3}$	Test set	8	3.22	2.09

Abstract

Data availability

Cited By

Figures (9)

Tables (3)

Equations (13)

Applied Optics