| The flue-gas emissions from stationary sources are the major source of airpollution, such that to supervise and to control its concentration is important for airquality improvement. The basic method is online-monitoring of flue-gas emissionsfrom pollution source. In this research, based on Ultraviolet Differential OpticalAbsorption Spectroscopy (DOAS), an applied research of flue-gas online-monitoringsystem was carried out.Based on the principle of DOAS technique, an online monitoring system wasdeveloped to measure the flue-gas concentration. System wavelength resolution canreach0.101nm via optimization, which meets the requirements of spectralmeasurements of various gases, such as SO2, NO.Meanwhile, in order to solve the problem of nonlinear of DOAS system, acompensation method was proposed to predict the concentration of different gases viaexamining the sources of system nonlinear. The fast change spectrum, which presentsthe gas absorption, can be obtained using the Savitzky-Golay digital smoothing filter.Then, the least square method was applied to calculate the gas concentration ofstandard absorption cross section. Finally, the results were substituted into the fittingformula and calculated the compensated concentration of pollutants. Verificationexperiments were designed for compensation method on SO2and NO. The resultsindicated that the error of measurement after compensation was within1.55%, whichmeets the online monitoring requirements.The relationship between the absorption cross section and temperature changingwas analyzed in this paper. And the results show that the relationship of them isnonlinear in different wavelengths. On the basis of quantitative analysis, after using anonlinear compensation method for temperature correction, the corrected spectrumcan be adopted to inverse calculate the concentration, which can improve themeasurement accuracy of gas concentration. Verification experiments are employedon SO2and NO, and the results show that the measurement error of concentrationafter temperature correction is obviously decreased. The measurement errors of SO2and NO are both reduced from above10%to below5%. |