Research On Stochastic Simulation Algorithms For Aerosol Dynamics

The suspended particulate matter is one of the main factors which affect human health.Through ever increasing research findings,the hazards of fine and ultrafine particles(even with harmful substances absorpted on their surface)on human health has been gradually discovered.It is urgent to study aerosol dynamics to gain deeper understanding of the evolution of aerosol particles.Generally,aerosol dynamics is modeled by the General Dynamic Equation(GDE),which is a strong nonlinear differential quation,and analytical solutions are available only for a few limited cases.At present,the commonly used research methods include section method,moment method and Monte Carlo method,all have their own advantages and disadvantages.Using operator splitting technique is very precise and efficient.In this thesis work,the OSMC program was extented to include the Weighted Flow Algorithms(WFA)method,and Tau-Leaping method.The efficiency and precison of the original and extended methods have been compared thoroughly.First,the Weighted Flow Algorithms were investigated.Within WFA,every numerical particle is assigned a different weight to represent different number of real particles,which can increase the resolution of under represented particles in an un-weighted algorithm.The WFA algorithm is validated against known analytical solutions for a few testing cases.When compared with the original OSMC,it is found that the WFA demonstrates higher precision for predicting the higher order moments for the aerosol distribution.Then,the Tau-Leaping Algorithm was used to accelerate simulation speed of the OSMC.In the Tau-Leaping Algorithm,a leaping time step is used to adance the simulation.The simulation results indicate that the Tau-Leaping Algorithm can improve the simulation speed of OSMC effectively,but the but decrease the precision.With the increase of the leap size,the CPU cost time is reduced and the simulation error increases linearly.The CPU cost time is proportional to the reciprocal of the leap size.The findings provides insightful understandings on the efficiency and precision of various popular MC algorithms for aerosol dynamics,and they can be used to guide the choice of the proper MC method for a specific condition.