Study Of Measurement Method Of Sulfur Dioxide Based On Data Fusion Technology

The sulfur dioxide (SO₂) is considered as one of the main pollution in atmosphere, which is a serious environment contamination to human beings. So apiece nations reinforce the control and managing SO₂.This dissertation begins with analyzing components of flue gas. And infrared spectrum absorption characteristics of SO₂ and CO₂ are explored in the middle infrared region. Waveband CO₂ effects SO₂ measure consequence is found out, that is, CO₂ has strong infrared absorption at wavelength of 4.28μm while SO₂ is 3.98μm. On condition that incidence light intensity and other terms are invariable, CO₂ absorbs some infrared ray, so SO₂ measure result increases. Secondly, single lamp-house and four detectors devices are designed to measure SO₂ concentration in order to dispel the across interference of CO₂. The concentration signals of SO₂ and CO₂ are surveyed by two couples of light filters and detectors. Then actual output signals are decomposed and disposed by weighted method. So cross sensitization of CO₂ to SO₂ is dispelled, and SO₂ measure precision is improved.Then SO₂ concentration signals collecting by infrared absorption method are disposed by principle component analysis and data fusion technology based on least square, D-S evidence, artificial neural networks and so on. Time domain and space domain are fused to cut down indeterminacy and uncertainty in SO₂ measure, eliminate impact of yawp, CO₂ and other disturbance to SO₂ measurement. So accuracy and precision of SO₂ are increased.Finally, fitting curves, curve diagrams of SO₂ concentration and signal response after algorithms fused are gained by data simulation experiment. Data are analyzed by correlation function of error theory and data processing. Experiment proves that CO₂ has impact on SO₂ measurement result, the disturbance mathematical model is changed with CO₂ concentration. And disturbance extent is stronger when CO₂ is lower concentration. And measurement precision increases widely after data fusion, that is, the maximum error of SO₂ is 3.684%, CO₂ is 6.708%, a available path to analyze gas by infrared absorption.