Share Email Print
cover

Proceedings Paper

On-line dynamic monitoring automotive exhausts: using BP-ANN for distinguishing multi-components
Author(s): Yudi Zhao; Ruyi Wei; Xuebin Liu
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Remote sensing-Fourier Transform infrared spectroscopy (RS-FTIR) is one of the most important technologies in atmospheric pollutant monitoring. It is very appropriate for on-line dynamic remote sensing monitoring of air pollutants, especially for the automotive exhausts. However, their absorption spectra are often seriously overlapped in the atmospheric infrared window bands, i.e. MWIR (3~5μm). Artificial Neural Network (ANN) is an algorithm based on the theory of the biological neural network, which simplifies the partial differential equation with complex construction. For its preferable performance in nonlinear mapping and fitting, in this paper we utilize Back Propagation-Artificial Neural Network (BP-ANN) to quantitatively analyze the concentrations of four typical industrial automotive exhausts, including CO, NO, NO2 and SO2. We extracted the original data of these automotive exhausts from the HITRAN database, most of which virtually overlapped, and established a mixed multi-component simulation environment. Based on Beer-Lambert Law, concentrations can be retrieved from the absorbance of spectra. Parameters including learning rate, momentum factor, the number of hidden nodes and iterations were obtained when the BP network was trained with 80 groups of input data. By improving these parameters, the network can be optimized to produce necessarily higher precision for the retrieved concentrations. This BP-ANN method proves to be an effective and promising algorithm on dealing with multi-components analysis of automotive exhausts.

Paper Details

Date Published: 24 October 2017
PDF: 8 pages
Proc. SPIE 10461, AOPC 2017: Optical Spectroscopy and Imaging, 104611A (24 October 2017); doi: 10.1117/12.2285325
Show Author Affiliations
Yudi Zhao, Xi'an Institute of Optics and Precision Mechanics (China)
Univ. of Chinese Academy of Sciences (China)
Ruyi Wei, Xi'an Institute of Optics and Precision Mechanics (China)
Univ. of Chinese Academy of Sciences (China)
Xuebin Liu, Xi'an Institute of Optics and Precision Mechanics (China)
Univ. of Chinese Academy of Sciences (China)


Published in SPIE Proceedings Vol. 10461:
AOPC 2017: Optical Spectroscopy and Imaging
Jin Yu; Zhe Wang; Wei Hang; Bing Zhao; Xiandeng Hou; Mengxia Xie; Tsutomu Shimura, Editor(s)

© SPIE. Terms of Use
Back to Top