Measurement Of Aerosol Velocity And Particle Size Distribution By Dynamic Light Scattering

Dynamic light scattering(DLS)technique is an effective method for measuring the particle size of submicron and nano particles in colloidal systems.This method obtains particle size distribution(PSD)by measuring and inverting the autocorrelation function(ACF)of the light intensity signal scattered by Brownian particles.In the inversion process,the Fredholm integral equation of the first kind needs to be solved,which is a typical ill-conditioned problem.For the flowing aerosol,the scattered light intensity received by the detector is affected by Brownian motion and directional translational motion.The directional translational motion of particles increases the difficulty of particle inversion,which not only intensifies the attenuation of the autocorrelation function of light intensity but also improves the ill-conditioned degree of the inversion equation,leading to the reduction of the accuracy of the inversion results.In order to improve the stability and accuracy of the inversion results and realize the accurate measurement of the flowing aerosol PSD,this paper has conducted a relatively in-depth study on the measurement of the flow velocity and PSD of the flowing aerosol.The main research contents are concluded as follows:1.An investigation of the applicability of three generally used regularization matrices in Tikhonov regularization inversion for the particle size inversion of flowing aerosol.When comparing the PSDs generated by inversion with three regular matrices at different noise levels and flow rates,it is found that when there is no noise,the three regularized matrices may precisely describe the particle size distribution.The results obtained by using the unit matrix are gradually distorted as noise and flow rates increase.The first derivative operator and the second derivative operator have nearly equal capacity to identify the particle peak position;however,the distribution error of the inversion results generated by using the second derivative operator is less and closer to the genuine distribution.For bimodal aerosol,the second derivative operator has higher bimodal recognition ability,but can only invert the unimodal distribution in the case of high noise level and high particle velocity.2.The method for obtaining the flow rate of flowing aerosol based on the method of cumulants was studied.In the DLS measurement for a flowing aerosol,an anemometer is needed to measure the flow velocity to invert the aerosol particle size.The deviation of the flow rate reading value caused by the aerosol fluctuation will lead to the error of the kernel matrix in the inversion equation,which will lead to a decrease of the accuracy in the particle size inversion.The method of cumulants has been used to obtain the aerosol velocity through the light intensity ACF based on the relation between the particle velocity and the second cumulant of the electric field ACF of flowing aerosol particles.Additionally,the PSD has been inversed by the regularization method,which enables the online measurement of flowing aerosol to be realized.The simulation results at different noise levels show that the accurate velocity and PSD for flowing aerosols can be obtained by the proposed method at low noise levels and high noise levels.It has been observed that higher noise causes the PSD broadening and the peak offset.For aerosol particles with a bimodal distribution,the increase of noise leads to the increase of peak height ratio and peak of small particles in bimodal distribution disappears,leaving only unimodal distribution.Furthermore,the increase of noise has no significant effect on the calculated velocity value both for unimodal and bimodal aerosol distributions.The experimental results of DLS measurements on unimodal and bimodal aerosols verified the above statements.3.The calculation method of flowing aerosol flow velocity and particle size distribution based on Bayesian inference was studied.In order to avoid the highly ill-conditioned inversion equation caused by the increase of flow velocity and improve the PSD inversion results,a Bayesian inference method was adopted and the Metropolis-within-Gibbs sampling was used to solve the problem.Particle flow rate and PSD parameters are regarded as random variables in this method,and the joint posterior probability density distribution of these random variables is calculated using Bayesian theory.By using Metropolis-within-Gibbs sampling,samples with unknown parameters that satisfy the posterior probability density function are created,and statistical inference on the samples yields the estimated values of the unknown parameters.In the iterative process,the method uses likelihood function and error to continuously optimize the prior information to obtain more accurate solutions.The PSD results obtained by simulation and measured data are better than those obtained by Tikhonov regularization inversion.Particle velocity and measurement noise are the main factors that restrict the measurement accuracy for the dynamic light scattering measurement technique of flowing aerosol.So far,in the few studies on dynamic light scattering measurements of flowing aerosols,the measurement results of flow velocity and particle size distribution are not satisfactory.The goal of this paper is to effectively improve the accuracy of the dynamic light scattering measurement results of flowing aerosol through research on the flow velocity acquisition method and particle size inversion algorithm,so as to promote the application and development of this technique in the growing field of aerosol measurement.