Filtering Algorithm Research In Trace C2h2 Detection Based On Tdlas

Yubin Wei,Tongyu Liu,Jun Chang,Tingting Zhang,Yanfang Li

DOI: https://doi.org/10.1109/icocn.2017.8121233

2017-01-01

Abstract:Real-time and accurate detection of the flammable gas concentration is the key issue in spontaneous combustion prediction. A modified tunable diode laser absorption spectroscopy (TDLAS) technology has been used in this paper to achieve in-situ detection of trace Acetylene concentration. Noises and interference fringes are eliminated by wavelet adaptive threshold filtering, and then Kalman filtering has been used to smooth the estimated gas concentration. It is demonstrated that the sensitivity is improved from 2 to 0.5 ppm by wavelet denoising and Kalman filtering. The fluctuation of estimated concentration can be suppressed to 1/4 of the original value.