Optimization Design Of Multi-component Analysis Algorithm For SO₂ And NO_X In Flue Gas

The negative impact of human beings on the highly developed industries has not been properly assessed and effectively prevented,leading to a very serious situation in the international community’s environmental pollution.China’s energy consumption has been dominated by coal burning for a long time.Thermal power generation is the main source of power generation in our country.SO₂,NO_X,solid particulate matter and other pollutants released by domestic coal-fired power plants endanger the health of the surrounding public.In order to ensure the implementation of the ultra-low emission standards of coal-fired power plants,on-line monitoring of pollutant emissions has become an essential technical basis.In this context,this thesis explores and improves the differential absorption spectroscopy technology for SO₂ and NO_X multi-component gas detection,build an experimental test platform,and study the theoretical basis,key technology and implementation method of DOAS technology to measure the concentration of flue gas pollutants.The main work done in this thesis is as follows:A measurement algorithm for low-concentration SO₂ gas was carried out.A new SO₂ measurement method combining DOAS and fast Fourier transform was proposed in291～312nm band.By analyzing the noise source of the instrument system,measures such as spectrometer temperature control,dark noise deduction,and wavelet domain denoising are taken.In order to reduce the spectral leakage,the Hanning window is used to extract the broadband component of the optical density by fast Fourier transform.The experiment results show that when the concentration range is 3～31ppm,the relative error of SO₂ concentration measurement is less than 1%in most cases.This method has high precision and linearity.The concentration measurement algorithm of NO_X was studied.Combined with the properties of NO and NO₂ and the mutual transformation between them,the spectral analysis methods of the two gases were studied according to the gas absorption characteristics.NO adopts polynomial fitting to obtain its differential optical density and differential absorption cross section.The experimental data show that the relative error of NO inversion by this method is less than 5%.The NO₂ absorption spectrum was extracted by the method of wavelet decomposition to extract the slow variation of the NO₂ absorption spectrum.The relative error of the concentration inversion was less thanThe method of quantitative analysis of SO₂and NO_X mixed gas was carried out.When measuring the concentration of two-component gas based on the classical least squares method,it is found that there is a serious aliasing interference between NO and SO₂ in the selected absorption band,and the measurement result deviates greatly from the actual set value,so a decomposition correlation spectrum is proposed.The optimization algorithm uses the concentration correlation of SO₂ in the range of200～230nm and 290～310nm to separate the NO absorption spectrum.The accuracy of concentration measurement is significantly improved,and the relative error of NO can be controlled below 10%,and the relative error of NO₂and SO₂ is less than 5%.