Improving LIBS Quantification by Combining Domain Factors and Multilayer Perceptron Method
CUI Jia-cheng1, SONG Wei-ran1, YAO Wei-li2, JI Jian-xun1, HOU Zong-yu1, 3, WANG Zhe1, 3*
1. Department of Energy and Power Engineering, Tsinghua University; State Key Lab of Power Systems; Institute for Carbon Neutrality; International Joint Laboratory on Low Carbon Clean Energy Innovation, Beijing 100084, China
2. China National Coal Group Corporation, Beijing 100120, China
3. Shanxi Research Institute for Clean Energy, Tsinghua University, Taiyuan 030032, China
Abstract：Laser-induced breakdown spectroscopy (LIBS) is an emerging atomic spectroscopy technique with promising applications in coal analysis but is limited by its relatively low quantification performance. Various machine learning methods have been applied in coal analysis on LIBS to improve its quantitative performance in recent years. However, most of these machine-learning models were established purely based on statistics. They ignored the physical rules involved in the quantification, resulting in reduced robustness, application range, and a lack of model interpretability. This work proposed a physics-statistics combined regression method based on the dominant factor (DF) and multilayer perception (MLP), called DF-MLP, to incorporate spectral domain knowledge into machine learning. The new proposed method built a physical-based dominant model to predictelement concentration with the characteristic lines selected with spectral knowledge and correct the residual errors using MLP. DF-MLP combines the dominant factor model and residual error correction using the MLP method can utilize the domain knowledge to improve model robustness and interpretability without reducing complexity. DF-MLP was compared with normal MLP, dominant factor partial least squares regression (DF-PLSR), dominant factor support vector regression (DF-SVR), and other baseline methods, and optimal results were obtained. Compared with normal MLP, the proposed method reduces root mean squared error of prediction (RMSEP) by 13.21%, 14.54%, and 21.77% for carbon, ash, and volatile, respectively. Compared with DF-SVR, the proposed method reduces RMSEP by 14.75%, 23.13%, and 5.99%, respectively. We further discussed the impact of different modeling patterns in the dominant factor method. The experimental results showed that combining domain knowledge with machine learning methods was a feasible approach to improve the performance of LIBS quantification.
崔佳诚，宋惟然，姚蔚利，姬建训，侯宗余，王 哲. 结合主导因素和多层感知机提高LIBS定量化性能[J]. 光谱学与光谱分析, 2025, 45(04): 1022-1027.
CUI Jia-cheng, SONG Wei-ran, YAO Wei-li, JI Jian-xun, HOU Zong-yu, WANG Zhe. Improving LIBS Quantification by Combining Domain Factors and Multilayer Perceptron Method. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2025, 45(04): 1022-1027.
[1] Brech F, Cross I. Appl. Spectros., 1962, 16: 59.
[2] Guo L B, Zhang D, Sun L X, et al. Frontiers of Physics, 2021, 16(2): 22500.
[3] Wang Z, Afgan M S, Gu W, et al. TrAC Trends in Analytical Chemistry, 2021, 143: 116385.
[4] Zhang D, Zhang H, Zhao Y, et al. Applied Spectroscopy Reviews, 2022, 57(2): 89.
[5] Cisewski J, Snyder E, Hannig J, et al. Journal of Chemometrics, 2012, 26(5): 143.
[6] Ding Y, Yan F, Yang G, et al. Analytical Methods, 2018, 10(9): 1074.
[7] Sun C, Tian Y, Gao L, et al. Scientific Reports, 2019, 9: 11363.
[8] Li L N, Liu X F, Xu W M, et al. Spectrochimica Acta Part B: Atomic Spectroscopy, 2020, 169: 105850.
[9] D'Andrea E, Pagnotta S, Grifoni E, et al. Applied Physics B: Lasers and Optics, 2015, 118(3): 353.
[10] Song W, Afgan M S, Yun Y H, et al. Expert Systems with Applications, 2022, 205: 117756.
[11] Wang Z, Feng J, Li L, et al. Journal of Analytical Atomic Spectrometry, 2011, 26(11): 2289.
[12] Song W, Hou Z, Gu W, et al. Spectrochimica Acta Part B: Atomic Spectroscopy, 2022, 195: 106490.
[13] Dong M, Wei L, Lu J, et al. Journal of Analytical Atomic Spectrometry, 2019, 34(3): 480.
[14] Yang P, Li X, Nie Z. Optics Express, 2021, 29(13): 20687.
[15] Yue Z, Sun C, Chen F, et al. Biomedical Optics Express, 2021, 12(5): 2559.
[16] Zhang X, Zhang F, Kung H T, et al. International Journal of Environmental Research and Public Health, 2018, 15(11): 2390.
[17] DONG Mei-rong, WEI Li-ping, ZHANG Yong-sheng, et al(董美蓉，韦丽萍，张勇升，等). Proceedings of the Chinese Society of Electrical Engineering(中国电机工程学报), 2020, 40(20): 6617.
