SO2 Concentration retrieval algorithm using EMD and PCA with application in CEMS based on UV-DOAS

A concentration retrieval algorithm is presented, which is applied to in-situ measurement of the concentration of SO2 based on Ultraviolet-visible Differential Optical Absorption Spectroscopy (UV-DOAS). The algorithm uses the standard deviation of the differential absorbance to represents the gas concentration. Empirical Mode Decomposition (EMD) and Principle Component Analysis (PCA) are used to process the differential absorbance spectrum to improve the SNR. In the algorithm, the basis data for the concentration retrieval of SO2 is the combination of the EMD & PCA processing result and the standard deviation of the differential absorbance to improve the SNR. The algorithm is applied for a Continuous Emission Monitoring System (CEMS) whose optical path length is 0.3 m. The nonlinear calibration for the system was executed. The relative error of the retrieving concentration is less than 4.68%. The measuring result is −2.05 mg/m3 during the concentration of SO2 is zero. For testing the effect of in-situ application, the contrast experiment was implemented in one coil-fired power plant. And the result proof that the algorithm has high stability and durability.