Real time monitoring of air VOCs by artificial neural network and remote sensing FTIR

Real time monitoring of a five-component air pollution system was conducted in this study using remote sensing Fourier transform infrared spectroscopy(FTIR) and artificial neural network (ANN). Five volatile organic compounds (VOCs)-methanol, chloroform, hexane, acetone and butanol-were released from a point of pipe connection to function as the leaking source in the analyzed area, and there were strong overlaps between infrared characteristic peaks of each component. The wavenumber-absorbance-time 3D spectra of the leaking sources were plotted and the path integrated concentrations (PIC) were obtained by artificial neural network. Results showed that it is desirable and efficient to get good quantification results of five-component pollution systems with the method presented in this paper, and remote sensing FTIR system as well as ANN could be coupled together as a dependable methodology to get satisfactory results in various situations. Furthermore, this system can perform as a continuous and in-situ alert system for a variety of industrial and living areas.