Research On Comsbution Particles Characterization Sensing Based On Light Scattering Angular Spectrum

Industrial production and life are accompanied by burning of fossil fuels and biomass.In the process of combustion,a large number of combustion particles enter into the air and human body,and causing great harm to the environment and human health.The particle size determines the deposition efficiency and deposition location of combustion particles in the human body,and it is the core parameter to accurately evaluate the damage of combustion particles.With the rapid development of characteristic detection technology of combustion particle,accurate measurement of particle size distribution of combustion particle is no longer limited to the laboratory research,but also needs to meet the requirements of online,in-situ,convenient and other characteristics of measurement outside the laboratory,such as vehicle exhaust emission,coal combustion and other combustion partilces.Due to the advantages of non-contact and high sensitivity,the optical methods have been widely used to detect the particle size distribution of combustion particles.However,there are still exist some problems.Firstly,the existing detection technology for the characteristic parameters of combustion particles is insufficient to obtain optical signals,which makes it difficult to realize the mode-free and high-resolution sensing of particle size distribution.Secondly,many combustion particles are non-spherical structure,and it is difficult to accurately detect the particle size distribution of non-spherical combustion particles without knowing the morphology parameters.Finally,the refractive index cannot be accurately preset when the combustion substance is unknown.But the refractive index has a significant influence on the light scattering characteristics of combustion particles,which hinders the reconstruction of particle size distribution of combustion particles.In order to solve the above problems,this research was mainly carried out from the following three aspects:It is difficult to realize the mode-free and high-resolution sensing of particle size distribution due to the existing detection technology for characteristic of combustion partilces to obtain insufficient optical signals.The light scattering model of combustion particles was developed to simulate and analyze the relationship between the light scattering intensity and particle size of combustion particles.Then,the light scattering angular spectrum model of sensing particle size distribution is proposed and established.At the same time,a lens-free,compact and portable prototype sensor was designed to realize the high-dimensional measurement of light scattering angular spectrum of combustion particles based on the optical simulation optimization.The inversion algorithm of Tikhonov regularization was used to realize the sensing of particle size distribution of spherical combustion particles.The experiments of monodisperse DEHS aerosol and polydisperse combustion particles produced by smoldering cotton rope demonstrate that the proposed sensing method can accurately detect the particle size distribution of spherical combustion particles.The maximum relative error between the light scattering angular spectrum measured by the prototype sensor and the reference light scattering angular spectrum is 12.72%.The particle size distribution obtained by the prototype sensor is close to the reference particle size distribution,and the maximum Kullback-Leibler distance D_KL is 0.08.It is difficult for the existing optical methods to accurate measure the particle size distribution of non-spherical combustion particles due to the unknown morphology parameters.The structural model of non-spherical combustion particles was developed,and the relationships between the light scattering intensity,particle size distribution,morphology parameters and optical characteristics of non-spherical combustion particles under different morphology parameters are deeply studied.The light scattering angular spectrum model was established for sensing the particle size distribution of non-spherical combustion particles.Meanwhile,an adaptive particle swarm optimization algorithm combined with Tikhonov regularization algorithm was proposed to form a hybrid inversion algorithm.The prototype sensor was optimized and improved for further improving the information dimensions of the prototype sensor to measure the light scattering angular spectrum of combustion particles,and realize the sensing of the particle size distribution of non-spherical combustion particles.The experimental results of monodisperse DEHS aerosol and polydisperse combustion particles produced by smolding cotton rope showed that the improved prototype sensor can collect high-dimensional and low-error light scattering angular spectrum signals.The maximum relative error between the light scattering angular spectrum measured by the prototype and the reference light scattering angular spectrum is 14.78%.The experiment results of non-spherical combustion particles of n-heptane and random copolymerized polypropylene further proved that the proposed sensing method can accurately sense the particle size distribution of non-spherical combustion particles and decouple the morphology parameters of non-spherical combustion particles.The maximum D_KL of the particle size distribution that measured by the prototype sensor and the reference particle size distribution is 0.28.It is difficult to sense the particle size distribution of combustion particles without prior parameters of refractive index.In this dissertation,the influence mechanism of refractive index on the light scattering angular spectrum was deeply studied and analyzed,and the sensing model with high dimensinonal light scattering angular spectrum was developed to simultaneously characterize the particle size distribution and refractive index of combustion particles.Thus,the particle size distribution of combustion partilces can be successfully sensed in the case of unknown refractive index.The multi-output support vector machine model was optimized by particle swarm optimization algorithm for multi-paramter joint inversion.The experiments of monodisperse DEHS aerosols,smolding cotton rope,burning of n-heptane and so on have proved that the proposed sensing method can accurately sense the particle size distribution without the prior knowledge of refractive index.The maximum D_KL between the particle size distribution measured by the prototype sensor and the reference particle size distribution is 0.22.In summary,this dissertation developed a light scattering angular spectrum model for sensing the particle size distribution of combustion particles based on the light scattering theory,revealed the internal relationship between the light scattering characteristics and particle size distribution of combustion particles,completed the design and verification of the prototype sensor,and realized the sensing of particle size distribution.