Aerosol Characterization Parameters Sensing Method Based On Light Scattering Field

The particle size distribution and concentration(including number concentration,surface area concentration,volume concentration,and mass concentration)are the key factors affecting the physical and chemical characteristics of aerosols.Off-line measurement technology in the laboratory is often used for the analysis of aerosol characteristic parameters.However,these laboratory devices are generally complicated,heavy,and expensive,which are difficult to adapt to the increasing demand for in-situ and on-line sensing in many fields such as fire protection,energy,and public environmental health in recent years.Therefore,the analysis of aerosol characteristic parameters has changed from laboratory measurement technology to sensing technology focusing on in-situ,on-line,real-time,ultra-compact,and other characteristics.Optical methods are considered being the important way to realize in-situ and on-line aerosol sensing due to their natural advantages of fast measurement speed.However,there are still some key problems about the measurement mechanism of the current optical aerosol sensing methods.1.Existing optical sensing methods lack sufficient resolution for the discretized characterization of particle size distribution,which cannot meet the requirements of in-situ and on-line sensing for the particle size distribution.2.Existing optical sensing methods require prior information on aerosol refractive index,which seriously limits the application of the optical aerosol particle size distribution analysis methods.3.Existing optical sensing methods ignore the effect of changes in particle size distribution and refractive index on aerosol concentration sensing,which will lead to unstable accuracy of aerosol concentration sensing results.The above measurement mechanism problems can be scientifically summarized as how to comprehensively characterize aerosols based on optical signals and accurately extract the aerosol characteristic parameters from optical signals.Therefore,the theoretical model of light scattering field of aerosols is established in this thesis.The light scattering field is defined as the intensity of light scattered by aerosol particles in all directions.In this thesis,the light scattering field with rich optical features is used to comprehensively characterize the characteristic parameters of aerosols.Based on the established light scattering field,the specific research of this thesis focuses on three aspects as follows:Firstly,to solve the problem that the existing optical sensing methods have the insufficient resolution for the discretized characterization of particle size distribution,a sensing method based on the light scattering field is proposed.In this thesis,the particle size distribution of aerosols is characterized by the light scattering field,and a particle size distribution sensing model with a high-resolution based on the light scattering field is established.To solve the ill-posed problem in the inversion of particle size distribution,an improved Tikhonov regularization algorithm by adding 1-norm to the penalty term is proposed.To solve the acquisition problem of numerous spatial scattering signals,an acquisition scheme based on hemispherical projection imaging is proposed.And then a sensor prototype(size: Φ65×L135 mm,weight: 2.0 kg)is designed with simple structure,miniature portability,and low cost.The number of particle size measurement channels of the prototype is 81,equivalent to that of large laboratory instruments.The discrete particle size distribution with high-resolution can be measured by the prototype in 1 s.According to the results of the verification test,the particle size distribution output by the prototype is consistent with the results measured by the reference instrument,where the minimum Kullback-Leibler distance DKL is 0.06,the maximum DKL is 0.27,and the average DKL is0.18.Secondly,to break the limitation of refractive index on optical particle size distribution sensing,a joint sensing method of particle size distribution and refractive index based on light scattering field is proposed.In this thesis,the particle size distribution and refractive index of aerosols are simultaneously characterized by the light scattering field,and a joint sensing model of particle size distribution and refractive index is established.To solve the joint sensing model,an inversion algorithm based on deep learning is proposed,which can quickly decouple the particle size distribution and refractive index and output the results.In addition,the sensor prototype is improved to expand the collection range of the light scattering field,which can provide more optical features for the simultaneous characterization of particle size distribution and refractive index.Without any prior refractive index,the improved prototype can measure the particle size distribution with 139 channels in 1 s.Compared with the results measured by the reference instrument,the minimum DKL is 0.04,the maximum DKL is 0.43,and the average DKL is 0.21.Thirdly,to solve the influence of the changes in particle size distribution and refractive index on the existing optical concentration sensing,an aerosol concentration sensing method based on the joint correction of particle size distribution and refractive index is proposed.In this thesis,the light scattering field is used to characterize the concentration of aerosols finally,and an accurate sensing model of aerosol concentration parameters is established.The influence mechanism of particle size distribution and refractive index on concentration sensing is studied by simulation,and then a real-time correction algorithm of concentration conversion factor based on the particle size distribution and refractive index is proposed,which can accurately sense the concentration parameters of aerosols with different particle size distributions and different refractive indices.The experimental results show that,compared with the existing optical concentration sensing methods,the average measurement error of the number concentration is reduced from 85% to 9%,the average measurement error of the surface area concentration is reduced from 36% to 15%,and the average measurement error of the volume concentration is reduced from 86% to 16%.Based on the particle size distribution information,this method can also sense the concentration information in any particle size segment,and realize the digital cutting of concentration information.