Acceleration method of realizing stream-line simulation on basis of particle tracking algorithm

Abstract

The invention discloses an acceleration method of realizing stream-line simulation on the basis of a particle tracking algorithm, and belongs to the field of stream-line numerical-simulation. The method is run on a computer or a server with a plurality of GPUs, and includes: discretizing a calculation region into a plurality of grids; initializing basic information of a model, parameters of an aquifer and a coefficient matrix in velocity vector formulas; averagely dividing the grids into a plurality of blocks to pass the same to the plurality of GPUs; and opening a plurality of threads, for each thread, utilizing the forward and backward particle tracking algorithm to calculate a next position and a previous position of a particle according to a current position, carrying out repeated iteration on the algorithm to obtain a complete stream-line, passing back data to a CPU, and outputting the same into a result file. The method has the advantages that a place occupancy area is small, costs are low, energy consumption is low, an acceleration effect is remarkable, the stream-line can be quickly generated, and a requirement of actual application on real-time performance of stream-line simulation can be satisfied.

Description

technical field [0001] The invention relates to the field of streamline numerical simulation, in particular to an acceleration method for realizing streamline simulation based on a particle tracking algorithm. Background technique [0002] Streamline simulation has been increasingly used in various fields of groundwater research and oil and gas field development. Streamline visualization provides vivid and intuitive groundwater dynamic information for groundwater research, and can easily analyze the movement state of groundwater. In the study of groundwater flow, Pollock proposed that the semi-analytic particle tracking algorithm is widely used because of its flexibility and universality. The particle tracking method is also one of the most direct reasons for the application of the streamline method. In actual engineering, the scale of streamline calculation reaches millions, and ordinary machines run for a long time, which cannot meet the real-time requirements of groundwa...

AI Technical Summary

Problems solved by technology

In actual engineering, the scale of streamline calculations reaches millions, and ordinary machines run for a long time, which cannot meet the real-time requirements of groundwater research for data results.

[0003] The traditional CPU parallel acceleration framework is limited by the number of computing cores of the CPU. To achieve a good acceleration effect, a large number of CPUs are required, and correspondingly more server support is required, which will result in a large footprint and difficult maintenance , high energy consumption and other issues

Images