Accelerated simulation of stochastic particle removal processes in particle-resolved aerosol models

JOURNAL OF COMPUTATIONAL PHYSICS

卷:322

Article

Curtis, J. H.¹  Michelotti, M. D.²  Riemer, N.¹  Heath, M. T.²  West, M.³ 

[1] Univ Illinois, Dept Atmospher Sci, 105 S Gregory St, Urbana, IL 61801 USA

[2] Univ Illinois, Dept Comp Sci, 201 North Goodwin Ave, Urbana, IL 61801 USA

[3] Univ Illinois, Dept Mech Sci & Engn, 1206 W Green St, Urbana, IL 61801 USA

关键词: Aerosol modeling;    Stochastic simulations;    Particle methods;    Aerosol removal;   

DOI  :  10.1016/j.jcp.2016.06.029

来源: Elsevier

【 摘 要 】

Stochastic particle-resolved methods have proven useful for simulating multi-dimensional systems such as composition-resolved aerosol size distributions. While particle-resolved methods have substantial benefits for highly detailed simulations, these techniques suffer from high computational cost, motivating efforts to improve their algorithmic efficiency. Here we formulate an algorithm for accelerating particle removal processes by aggregating particles of similar size into bins. We present the Binned Algorithm for particle removal processes and analyze its performance with application to the atmospherically relevant process of aerosol dry deposition. We show that the Binned Algorithm can dramatically improve the efficiency of particle removals, particularly for low removal rates, and that computational cost is reduced without introducing additional error. In simulations of aerosol particle removal by dry deposition in atmospherically relevant conditions, we demonstrate about 50-times increase in algorithm efficiency. (C) 2016 Elsevier Inc. All rights reserved.

【 授权许可】

Free