基于深度信念网络和极限学习机的SO2浓度检测

摘要/Abstract

摘要： 使用差分吸收光谱技术(Differential optical absorption spectroscopy, DOAS)进行工业在线气体检测,在气体浓度较低时,其光谱吸收不明显,
信噪比较低,通过传统方法来对工业气体浓度进行反演,预测结果难以满足工业应用具体要求。针对SO$_2$气体的差分吸收光谱特点,
采用氚灯作为光源,采集189.73$\sim$644 nm波段内的标准浓度SO$_2$的吸收光谱高维数据,选取吸收光谱数据并进行预处理,然后
利用训练集数据建立深度信念网络模型进行低维特征提取。在此基础上,利用训练数据的低维嵌入特征构建极限学习机反演模型,
实现SO$_2$气体浓度计算,并对该模型进行了有效性测试,从而得到一种更加精确的SO$_2$气体浓度在线检测方法。

关键词: 气体浓度检测；SO2, 差分吸收光谱技术, 深度信念网络, 极限学习机

Abstract: Differential optical absorption spectroscopy (DOAS) is widely used for online gas detection in industry. However,
when the concentration of industrial gas is low, the spectral absorption is not obvious and the SNR is
very low. So if the inversion of industrial gas concentration is carried out by using the traditional
methods, it is very difficult to meet the requirements of industrial application. According to the
differential absorption spectra of SO$_2$, tritium lamp is used as the light source to collect the
high-dimensional data of absorption spectra in 189.73$\sim$644 nm band. And after selecting and preprocessing
the absorption spectra data, a deep belief network (DBN) model is established based on the training set data
to extract the low-dimensional features of the test data. Furthermore, the extreme learning machine (ELM)
is constructed by using the low-dimensional embedding characteristics of training data to realize the
calculation of the SO$_2$ concentration. The effectiveness of the proposed model is evaluated, and it
seems that the method is more suitable for accurate on-line detection of SO$_2$ concentration in industrial field.

Key words: gas concentration detection, SO2, differential optical absorption spectroscopy, deep belief network, extreme learning machine

黄鸿, 兰洪勇, 黄云彪. 基于深度信念网络和极限学习机的SO2浓度检测[J]. 大气与环境光学学报, 2020, 15(3): 207-216.

参考文献

[1]
Wang Hailing. Study on the Method of SO_2 Concentration

Measurement Based on the DOAS and
Support Vector Regression [D]. Chongqing: Master's Thesis

of Chongqing University of China, 2015 (in Chinese).

王海玲. 基于DOAS和支持向量回归的SO_2浓度检测方法研究 [D]. 重庆

: 重庆大学硕士论文, 2015.
[2]
Yu Changyi. The formation, harm and prevention of acid

rain [J]. Environmental Protection and
Circular Economy, 2017, 37(09): 42-44+47 (in Chinese).

于长毅. 酸雨的形成、危害及防治 [J]. 环境保护与循环经济, 2017,

37(09): 42-44+47.
[3]
Chen Mingyan, Jiang Xianzheng, Huang Ruhong. Forming,

distribution, damage and preventive measures
of acid rain [J]. Guangdong Trace Elements Science, 2009,

16(01): 15-20 (in Chinese).

陈明艳,姜显政,黄汝红. 浅析酸雨的形成、危害及防治措施 [J].广东微

量元素科学, 2009, 16(01): 15-20.
[4]
Wang Hongtao. Key Technologies of Industrial Gas

Monitoring Based on Absorption Spectroscopy [D].
Chongqing: Master's Thesis of Chongqing University of

China, 2014 (in Chinese).

王洪涛. 基于吸收光谱技术的工业气体监测关键技术研究 [D]. 重庆:

重庆大学硕士论文, 2014.
[5]
Han Xuebing, Zeng Yi, Zhao Minjie, et al. Calibration

method and storage format of data from space-borne
trace gas different optical absorption spectrometer [J].
Journal of Atmospheric and Environmental Optics, 2018,

13(04): 268-276 (in Chinese).

韩雪冰,曾议,赵敏杰,等. 星载大气痕量气体差分吸收光谱仪数据的定标

处理与存储方式 [J].大气与环境光学学报, 2018, 13(04): 268-276.

[6]
Gao Yue, Su Chong, Li Hongguang. A kind of deep belief

networks based on nonlinear features extraction with
application to PM2.5 concentration prediction and

diagnosis [J]. Acta Automatica Sinica, 2018, 44(02): 318

-329 (in Chinese).

高月,宿翀，李宏光. 一类基于非线性PCA和深度置信网络的混合分类器

及其在PM2.5浓度预测和影响因素诊断中的应用[J].
自动化学报, 2018, 44(02): 318-329.

[7]
Zhao Mingfu, Tang Ping, Tang Bin, et al. Application of

principal component analysis combined fisher
discrimination in water quality detection by UV-Vis

spectroscopy [J]. Journal of Atmospheric
and Environmental Optics, 2018, 13(06): 436-446 (in

Chinese).

赵明富,唐平,汤斌,等. 主成分分析联合Fisher判别在紫外-可见光谱法

水质检测中的应用 [J].大气与环境光学学报, 2018, 13(06): 436-

446.
[8]
Huang Hong, Shi Guangyao, Jin Yingying, et al. Random

Subspace Deep Regression and its Application in
Water Quality Analysis of TOC in UV Spectroscopy [J].

Application Research of Computers, 2017, 34(10): 3020-3023

(in Chinese).

黄鸿,石光耀,金莹莹,等. 随机子空间深度回归方法在紫外光谱水质分析

中的应用 [J].计算机应用研究, 2017, 34(10): 3020-3023.
[9]
Pan Junhong, Wang Yihuai, Wu Wei. Physical quantity

regression method based on optimized BP neural [J].
Computer Science, 2018, 45(12): 170-176 (in Chinese).

潘俊虹,王宜怀,吴薇. 基于优化BP神经网络的物理量回归方法 [J]. 计

算机科学, 2018, 45(12): 170-176.

[10]
Qin Guohua, Xie Wenbin, Wang Huamin. Detection and control

for tool wear based on neural network and genetic
algorithm [J]. Optics and Precision Engineering, 2015,

23(5): 1314-1321 (in Chinese).

秦国华,谢文斌,王华敏. 基于神经网络与遗传算法的刀具磨损检测与控制

[J]. 光学精密工程, 2015, 23(5): 1314-1321.

[11]
Shen Haoyang, Wei Anlei, Wang Xiaowen, et al. Research on

application of BP artificial neural network
in the prediction of the concentration of SO_2 in ambient

air [J]. Environmental Engineering, 2014, 32(06): 117-121

(in Chinese).

申浩洋,韦安磊,王小文,等. BP人工神经网络在环境空气SO_2质量浓度预

测中的应用 [J]. 环境工程, 2014, 32(06): 117-121.

[12]
Bai Peng, Li Yan, Zhang Bin, et al. Key technologies

research of mixture gas infrared spectrum
analysis based on SVM [J]. Acta Photonica Sinica, 2008,

37(3): 566-572 (in Chinese).

白鹏,李彦,张斌,等. 基于SVM的混合气体红外光谱分析关键技术研究

[J]. 光子学报, 2008, 37(3): 566-572.

[13]
Zhang Luda, Jin Zechen, Shen Xiaonan, etal. Applied study

on support vector machine (SVM) regression method in
quantitative analysis with near-infrared spectroscopy [J].
Spectroscopy and Spectral Analysis, 2006, 25(9): 1400-

1403 (in Chinese).

张录达,金泽宸,沈晓南,等. SVM回归法在近红外光谱定量分析中的应用

研究 [J]. 光谱学与光谱分析, 2006, 25(9): 1400-1403.

[14]
Liang Zhengtang. Theoretical Studies on Wind Power System

Forecasting and Decision-making Based on
Characteristic Mining of Uncertainty [D]. Jinan: Doctorial

Dissertation of Shandong University of China, 2017 (in

Chinese).

梁正堂. 基于不确定性特征挖掘的风电系统预测与决策理论研究 [D].

济南:山东大学博士论文, 2017.
[15]
Zhang Ning, Liu Tianjian. Kernel function ELM method for

short-term load forecasting considering influencing
factors [J]. Engineering Journal of Wuhan University,

2018, 51(08): 703-707+714.

张宁,刘天键. 考虑影响因素的短期负荷预测核函数ELM方法 [J]. 武汉

大学学报(工学版), 2018, 51(08): 703-707+714.

[16]
Xu Tingting, Ma Chenye, Zhang Jingmin, et al.

Classification of LAMOST spectra based on deep learning

[J].
Acta Astronomica Sinica, 2019, 60(02): 15-24.

许婷婷,马晨晔,张静敏,等. 基于深度学习的LAMOST光谱分类研究 [J].

天文学报, 2019, 60(02): 15-24.
[17]
Bian X H, Zhang C X, Tan X Y, et al. A boosting extreme

learning machine for near-infrared spectral
quantitative analysis of diesel fuel and edible blend oil

samples [J]. Analytical Methods, 2017, 9(20): 2983-2989.