Mathematical modeling of air pollution dynamics by parallel computation

air polution;mathematical modelling

Dabdub, Donald ; Seinfeld, John H.

University:California Institute of Technology

Department:Chemistry and Chemical Engineering

关键词: air polution;    mathematical modelling;   

Others  :  https://thesis.library.caltech.edu/4987/1/Dabdub_d_1996.pdf

美国|英语

来源: Caltech THESIS

【 摘 要 】

The use of massively parallel computers provides an avenue to overcome the computational requirements in the study of atmospheric chemical dynamics. General considerations on parallel implementation of air quality models are outlined including domain decomposition strategies, algorithm evaluation and design, portability, modularity, and buffering techniques used in I/O operations. Results are given for the implementation of the CIT urban air pollution model on distributed memory multiple instruction / multiple data (MIMD) machines ranging from a cluster of workstations to a 512 node Intel Paragon.The central challenge in developing a parallel air pollution model is the implementation of the chemistry and transport operators used in the solution of the atmospheric reaction-diffusion equation. The chemistry operator is generally the most computationally intensive step in atmospheric air quality models. A new method based on Richardson extrapolation to solve the chemical kinetics is presented. The transport operator is the most challenging to solve numerically. Because of its hyperbolic nature non-physical oscillations and/or negative concentrations appear near steep gradient regions of the solution. Six algorithms for solving the advection equation are compared to determine their suitability for use in parallel photochemical air quality models. Four algorithms for filtering the numerical noise produced when solving the advection equation are also compared.A speed-up factor of 94.9 has been measured when the I/O, transport, and chemistry portions of the model are performed in parallel. This work provides the computational infrastructure required to incorporate new physico-chemical phenomena in the next generation of urban- or regional-scale air quality models.Finally, the SARMAP model is used to model the San Joaquin Valley of California. SARMAP is the updated version of RADM. It can be considered a state-of-the- art regional air pollution model. Like the CIT model, SARMAP incorporates the following atmospheric phenomena: gas-phase chemistry, advection and diffusion. In addition, SARMAP incorporates aqueous-phase chemistry and transport through cumulus clouds. Sensitivity studies performed show a significant dependence of ozone model predictions on boundary conditions.