摘要
考虑到传统的水文模型串行计算方法无法满足当前复杂的水文模拟预报需求,为提高网格水文模型的计算效率,分析网格化水文模型计算中网格汇流特点,提出基于网格汇流流向的动态数据分区方法,不同分区数据可以并行计算,基于该方法提出基于Spark的网格水文模型分布式计算模型。通过在屯溪流域使用该模型与传统模型进行水文预报计算,验证了模型的高效性。
Considering that the traditional hydrological model serial calculation method cannot meet the current complex hydrolo-gical simulation and forecasting needs,to improve the computational efficiency of the grid hydrological model,the grid flow cha-racteristics in the calculation of hydrological models were analyzed,the dynamic data partition method based on grid confluence flow was proposed and the data of different partitions were calculated in parallel.Based on this method,a distributed computation model of grid hydrological model based on Spark was proposed.The efficiency of the model was verified by comparing with that of the traditional model in the Tunxi Basin.
作者
王继民
刘赛佳
李嘉玮
WANG Ji-min;LIU Sai-jia;LI Jia-wei(School of Computer and Information,Hohai University,Nanjing 211100,China)
出处
《计算机工程与设计》
北大核心
2023年第1期66-73,共8页
Computer Engineering and Design
基金
国家重点研发计划基金项目(2018YFC1508106)。
关键词
分布式水文模型
分布式计算
数据分区
通用数据格式
可扩展标记语言
分布式任务调度
流域水文模拟
distributed hydrological model
distributed computing
data partitioning
network common data format
extensible markup language
distributed task scheduling
watershed hydrology simulation
