US 11,933,724 B1
Device of complex gas mixture detection based on optical-path-adjustable spectrum detection and method therefor
Yin Zhang, Wuhan (CN); Xiaoxing Zhang, Wuhan (CN); Ran Zhuo, Wuhan (CN); Zhiming Huang, Wuhan (CN); Guozhi Zhang, Wuhan (CN); Dibo Wang, Wuhan (CN); Shuangshuang Tian, Wuhan (CN); Mingli Fu, Wuhan (CN); Yunjian Wu, Wuhan (CN); Yan Luo, Wuhan (CN); Shuo Jin, Wuhan (CN); Jinyu Pu, Wuhan (CN); and Yalong Li, Wuhan (CN)
Assigned to Hubei University of Technology, Wuhan (CN)
Filed by Hubei University of Technology, Wuhan (CN); and CSG Electric Power Research Institute Co., Ltd., Guangzhou (CN)
Filed on Dec. 1, 2023, as Appl. No. 18/527,214.
Claims priority of application No. 202310054615.1 (CN), filed on Feb. 3, 2023.
Int. Cl. G01N 21/3504 (2014.01); G01N 21/25 (2006.01)
CPC G01N 21/3504 (2013.01) [G01N 21/255 (2013.01); G01N 2201/0636 (2013.01); G01N 2201/101 (2013.01); G01N 2201/126 (2013.01)] 10 Claims
OG exemplary drawing
 
1. A device of complex gas mixture detection based on optical-path-adjustable spectrum detection, comprising:
a light source that is configured for generating an incident beam and emitting the incident beam into an optical gas cell:
the optical gas cell, comprising a cavity, and a reflection module group and a track arranged in the cavity, wherein the cavity is configured for accommodating a gas sample, the reflection module group comprises one or more reflection modules and is configured for reflecting the incident beam inside the cavity once or multiple times, and the track is consistent with a light path of the light beam in the cavity;
a detector module that is connected with the track in a relatively movable manner and is configured for receiving light beams and obtaining spectral data; and
a data acquisition unit that is configured for acquiring the spectral data obtained by the detector module and performing qualitative and/or quantitative analysis of a gas sample to be detected, wherein the steps specifically comprise: with the change in an optical path length as a dimension, constructing a three-dimensional spectrogram of absorbance, a wave number and an optical path length; then using a three-dimensional spectrum library to train a deep learning model;
and using the trained deep learning model for the qualitative and/or quantitative detection of the gas samples.