关键词: 固定污染源, 傅里叶变换红外光谱, 多组分, 在线检测

Abstract: Objective　Stationary pollution sources refer to pollution sources that are emitted into the air through exhaust pipes from coal-fired, oil fired, gas-fired boilers and industrial furnaces, as well as production processes such as petrochemicals, metallurgy, and building materials. The types of pollutants from stationary pollution sources mainly include smoke, CO2, CO, SOx, NOx, and incompletely burned hydrocarbons. Scientific and accurate monitoring of stationary source flue gas is of great significance for protecting the environment, controlling air pollution, and ensuring human health. At present, the analysis methods used for stationary source gaseous pollutants mainly include electrochemical methods, ultraviolet differential absorption spectroscopy (DOAS), non dispersive infrared (NDIR) spectroscopy, and Fourier transform infrared (FTIR) spectroscopy. Among them, electrochemical methods have a short service life and a high detection limit, however, only a single component can be detected at a time. The DOAS method can measure SOx and NOx, but cannot measure gases such as CO2 and CO that have no obvious characteristic absorption in the ultraviolet band. As for NDIR spectroscory, when measuring nitrogen oxides in the exhaust gas, it is susceptible to interference from particulate matter, water vapor, and exhaust gas temperature. Comparatively, the FTIR method has the advantages of high sensitivity, wide measurement bandwidth, simultaneous measurement of multiple components, and no need for sample pretreatment. It is not only suitable for measuring conventional gases such as SO2, NOx, CO2, CO in flue gas emissions, but also for analyzing the concentration of VOCs in flue gas emissions. Therefore, the FTIR monitoring system is an ideal component of the continuous emission monitoring systems (CEMS) for atmospheric stationary pollution sources. Methods　We have developed an online FTIR detection system for multi-component gases in flue gas emissions using a direct extraction measurement method. The system consists of four parts: gas sampling probe, gas heated sample cell (a multiple reflection White type cell), FTIR spectrometer, spectral acquisition and quantitative analysis unit. The main parameters of the system are: (1) the resolution of self-developed FTIR spectrometeris 1 cm-1; (2) The spectral range is 800~ 4000 cm-1; (3) The maximum temperature that can be heated in the sample pool is 180 ℃; (4) The base length of the sample pool is 40 cm and a maximum optical path length of 64 m can be acchieved after multiple reflections. Based on IR spectra acquired by the FTIR detection system, quantitative inversion is performed on seven target gases including CO, CO2, HCl, NO, NO2, SO2, and NH3 using the Levenberg Marquardt nonlinear least squares algorithm. Results and Discussion　For SO2 with a concentration of 70 μmol/mol, the relative error of inversion is 3.7%, and the root mean square error of fitting is 0.19%. For NO2 with a concentration of 97.4 μmol/mol, the relative error of inversion is 2.47%, and the root mean square error of fitting is 0.69%. For NO with a concentration of 150 μmol/mol, the relative error of inversion is 3.79%, and the root mean square error of fitting is 1.41%. For CO2 with a concentration of 40000 μmol/mol, the relative error of inversion is 2.31%, and the root mean square error of fitting is 1.02%. For CO with a concentration of 50 μmol/mol, the relative error of inversion is 3.36%, and the root mean square error of fitting is 0.16%. For HCl with a concentration of 130 μmol/mol, the relative error of inversion is 3.70%, and the root mean square error of fitting is 2.52%. For NH3 with a concentration of 50 μmol/mol, the relative error of inversion is 3.00%, and the root mean square error of fitting is 2.00%. Conclusions　We propose a measurement scheme based on FTIR multi-component gas analysis system, and conduct preliminary laboratory research on the target gas samples. The measurement method adopts high-temperature heating sampling throughout the entire process combined with multiple reflection cells to ensure that the measured gas does not condense or accumulate in the gas path, while avoiding analysis errors caused by gas sampling and infrared absorption interference of water-soluble gases. The results show that the system can achieve accurate measurement of multi-component gases, with the relative error between the measured values and the actual values of each gas concentration within 5%. It is believed that the proposed method will play a more important role in future atmospheric environmental monitoring, especially in online detection of flue gas emissions from stationary pollution sources.

Key words: fixed pollution source, Fourier transform infrared spectroscopy, multi-component, on-line detection