基于GAST模型的突发水污染事故中污染物输移模拟  

Simulation of pollutant transport in sudden water pollution accident based on GAST model

作　　者：石宝山 侯精明 王俊珲 康永德 孙亚联[2] 郭敏鹏 张兆安 SHI Baoshan;HOU Jingming;WANG Junhui;KANG Yongde;SUN Yalian;GUO Minpeng;ZHANG Zhaoan(State Key Laboratory of Eco-hydraulics in Northwest Arid Region of China,Xi'an University of Technology,Xi'an 710048,China;Northwest Engineering Co.,Ltd.,Power China,Xi'an 710065,China)

机构地区：[1]西安理工大学省部共建西北旱区生态水利国家重点实验室,陕西西安710048 [2]中国电建集团西北勘测设计研究院有限公司,陕西西安710065

出　　处：《武汉大学学报（工学版）》2022年第11期1112-1119,共8页Engineering Journal of Wuhan University

基　　金：国家十三五重大研发专项(编号:2016YFC0402704);国家自然科学基金项目(编号:51609199);陕西省水利科技项目(编号:2017SLKJ-14);陕西省科技厅项目(编号:2017JQ3043)。

摘　　要：突发水污染事故后,高效准确模拟预测是快速采取有效防控措施和制定应急预案的基础。采用基于非均匀网格的图形处理器(graphics processing unit,GPU)加速的水动力及污染物输移数值模型(GPU accelerated surface water flow and associated transport,GAST)对暴雨山洪或溃坝洪水引起的突发水污染事故中污染物输移规律进行了模拟预测。模型采用Godunov格式的有限体积法求解二维浅水方程和污染物输移方程,采用一种基于HLLC(Harten-Lax-Van Leer-contact)的近似黎曼求解器计算单元网格界面通量,采用二阶高精度格式(monotonic upstream-centred scheme for conservation laws,MUSCL)限坡线性重建和龙格-库塔时间积分法实现了二阶精度,同时采用GPU加速计算技术提高模型计算效率。算例结果表明:采用非均匀网格技术局部细化地形复杂区域生成精细网格,地势平坦区域粗化网格,大幅度减少了计算域网格数量,并且在保证模拟精度的同时,提升了模型计算效率。该模型可高效准确地模拟突发水污染事件中污染物输移过程,适用于大规模水污染事故的快速预警和评估。Efficient and accurate simulation prediction is the basis for quickly taking effective prevention and control measures and formulating emergency plans after sudden water pollution accidents.This paper establishes a graphics processing unit(GPU) accelerated hydrodynamic and pollutant transport numerical model GPU accelerated surface water flow and associated transport(GAST) based on non-uniform grids to simulate and predict the pollutant transport laws in sudden water pollution accidents caused by torrential mountain floods or dam-break floods.The Godunov format finite volume method is adopted to solve the two-dimensional shallow water equation and the pollutant transport equation,and the Harten-Lax-Van Leer-contact(HLLC) approximate Riemann solver is used to simultaneously calculate the cell grid interface flux. The second-order accuracy is obtained by using slope-limited linear reconstruction and Runge-Kutta time integration methods. At the same time,the GPU accelerated computing technology is adopted to improve the computing efficiency of the model.The results show that the non-uniform grid technology can locally refine the terrain complex regions to generate fine grids, and coarsely refine the grids in the flat terrain regions, greatly reducing the number of grids in the calculation domain,and improving the model calculation efficiency while ensuring the simulation accuracy.This model can efficiently and accurately simulate the transportation of pollutants in sudden water pollution accidents,and is suitable for rapid early warning and assessment of large-scale water pollution accidents.

关 键 词：GAST模型 突发水污染 GPU加速 非均匀网格 有限体积法 污染物输移 

分 类 号：TV131.2[水利工程—水力学及河流动力学]