Research On The Spectrum Data Processing Algorithms In Gas On-line Analyzer

With the rapid development of Chinese economic and continuous improvement of the industrialization degree, air pollution, industrial accidents and energy shortages become increasingly prominent. Therefore, an accurately real-time detection of the gas concentration has great significance and pressing needs for relevant departments to make effective regulations and control the production processes.The research thesis is derived from the on-line gas analyzer based on the TDLAS development project devoted to develop a real-time gas detection system with independent intellectual property in Jang Su Skyray-Instrument Co..The traditional methods usually need great time to analysis concentration which is a very complicated process. And the stability and sensitivity of the system is not ideal. The tunable diode laser absorption spectroscopy(TDLAS) technique has been regarded as the mainstream of the gas online detection technology and used in various measurements for its high accuracy, high sensitivity, selectivity, widely adaptability and rapid response.This dissertation, as a basic research of gas detecting, begins with the spectrum absorption principle, followed by the relevant theories of gas detection based on TDLAS technique and two basic measurement methods. And the corresponding TDLAS online analysis system designs are presented,all the parts and functions of which are also described briefly.Secondly the preprocessing methods of spectral data in the TDLAS system are introduced,which mainly include smoothing and baseline correction. The principle and performance of adaptive iteratively reweighed penalized least squares(Air PLS) method is emphatically analyzed,the baseline correction effect of which is also compared with the NLSA method.And a further discussion of the reconstruction and analysis methods of spectral data in the TDLAS system is then made. An advanced LM(ALM)method is put forward combining the characters of spectral absorbing line type. The ALM algorithm realizes an adaptive rapid reconstruction of spectral data by automatically capturing the searching starting point, which avoids defects that the traditional LM algorithm is sensitive to the initial value of parameters and inclined to fall into the local extremum causing wrong results. And two algorithms, the nonlinear spectral analysis algorithm(NLSA)aimed at directly absorption detection and the linear spectral analysis algorithm(LSA)aimed at harmonic detection are respectively studied based on the ALM algorithm, the correctness and efficiency of which are analyzed and verified by a large amount of experimental data.Finally, a valve system to configure the gas concentration and a monitor platform of TDLAS system are designed. Performance tests including linear error, output fluctuation and repeatability for the whole system are made based on the hardware system of the project team by measuring various concentrations of the oxygen in laboratory environment, which is compared with the result of the foreign TOLEDO gas analyzer. And the spectral analysis methods studied in the dissertation are then shown to be quite capable of implementing the real-time, accuracy and stability of the online detecting system, which are more reasonable and more efficient than the traditional ones.