Decoupled Prediction Method for Water Pollutant Concentration Based on SPA-SVR Using Ultraviolet Spectroscopy

Jiguang Jiang; Lei Shi; Chengzhi Su; Chuan Chang; Xiaotian Li; Xiaolong Hou; Aixin Tian

doi:10.3788/LOP220700

Journals >Laser & Optoelectronics Progress >Volume 60 >Issue 7 >Page 0730004 > Article

Laser & Optoelectronics Progress
Vol. 60, Issue 7, 0730004 (2023)

Decoupled Prediction Method for Water Pollutant Concentration Based on SPA-SVR Using Ultraviolet Spectroscopy

Jiguang Jiang^1、**, Lei Shi¹, Chengzhi Su^2、*, Chuan Chang¹, Xiaotian Li¹, Xiaolong Hou¹, and Aixin Tian¹

Author Affiliations

¹College of Mechanical and Electrical Engineering, Changchun University of Science and Technology, Changchun 130022, Jilin, China

²Artificial Intelligence Research Institute, Changchun University of Science and Technology, Changchun 130022, Jilin, China

show less

DOI: 10.3788/LOP220700 Cite this Article Set citation alerts

Jiguang Jiang, Lei Shi, Chengzhi Su, Chuan Chang, Xiaotian Li, Xiaolong Hou, Aixin Tian. Decoupled Prediction Method for Water Pollutant Concentration Based on SPA-SVR Using Ultraviolet Spectroscopy[J]. Laser & Optoelectronics Progress, 2023, 60(7): 0730004 Copy Citation Text

show less

Fig. 1. Flow chart of coupling parameters detection model

Download full size

Fig. 2. Ultraviolet spectrum detection system

Download full size

Fig. 3. UV absorption spectra of standard solution. (a) COD; (b) turbidity

Download full size

Fig. 4. UV absorption spectra of mixed solution. (a) COD+turbidity interference; (b) turbidity+COD interference

Download full size

Fig. 5. Selection of characteristic wavelength. (a) COD; (b) turbidity

Download full size

Fig. 6. Selection of characteristic wavelength of COD and turbidity. (a) 230.4 nm; (b) 249.2 nm; (c) 272.8 nm; (d) 295.3 nm; (e) 347 nm; (f) 374.6 nm

Download full size

Fig. 7. Three dimensional characteristic point diagram of ultraviolet spectrum

Download full size

Fig. 8. Comparison of coupling interference results. (a) COD+turbidity interference; (b) turbidity+COD interference

Download full size

Fig. 9. Comparison of coupling analytical prediction results. (a) COD; (b) turbidity

Download full size

Sample	COD /mg·L^-1	Turbidity /NTU	Absorbance
Sample	COD /mg·L^-1	Turbidity /NTU	$λ$ _230.4	$λ$ _249.2	$λ$ _272.8	$λ$ _295.3	…	$λ$ _374.6
1	5	0	0.415	0.147	0.069	0.020	…	0.001
2	0	5	0.120	0.092	0.089	0.076	…	0.003
3	5	5	0.400	0.211	0.139	0.090	…	0.004
$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$	$⋮$
108	125	50	2.108	2.047	1.384	0.710		0.252
109	150	50	2.122	2.159	1.503	0.748	…	0.265
110	150	60	2.157	2.394	1.712	0.763	…	0.384

Table 1. UV spectral characteristic dataset

Sample	COD /（mg·L^-1）		Prediction error COD /%	Turbidity /NTU		Prediction error Turbidity /%
Sample	True value	Predicted value	Prediction error COD /%	True value	Predicted value	Prediction error Turbidity /%
1	9.60	9.46	1.46	0.40	0.39	2.50
2	17.50	17.72	1.26	3.00	3.05	1.60
3	10.60	10.24	3.40	0.40	0.41	2.50
4	16.40	16.42	0.12	0.50	0.51	4.00
5	25.30	24.84	1.82	12.00	11.79	1.75
6	8.50	8.62	1.41	6.30	6.42	1.90
7	21.30	21.33	0.14	21.20	20.87	1.56
8	72.80	71.97	1.14	34.80	35.20	1.15
9	43.70	44.60	2.02	19.60	19.57	0.15

Table 2. Verification results of actual water sample model

Prediction index	R²	RMSE /（mg·L^-1）	RSD /%
COD	0.998	0.343	1.50
Turbidity	0.998	0.125	2.43

Table 3. SVR model prediction performance index

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information