An MPS algorithm accelerated by neural network on edge computing

WANG Zhuolin; JIANG Junyang; YANG Gengchao; YAO Qinghe; JIANG Zichao; ZHANG Yi

doi:10.13471/j.cnki.acta.snus.2022A064

您当前的位置：

首页 >

文章列表页 >

An MPS algorithm accelerated by neural network on edge computing

Research articles | 更新时间：2023-11-07

- An MPS algorithm accelerated by neural network on edge computing
- Acta Scientiarum Naturalium Universitatis Sunyatseni Vol. 62, Issue 5, Pages: 67-77(2023)
- 作者机构：
  
  1.中山大学航空航天学院，广东深圳 518107
  2.特文特大学机器人和机电一体化系，荷兰恩斯赫德 7522NB
- 作者简介：
- 基金信息：
- DOI：10.13471/j.cnki.acta.snus.2022A064
  CLC： O3
- Published：25 September 2023，
  
  Published Online：26 June 2023，
  
  Received：27 July 2022，
  
  Accepted：03 January 2023
扫描看全文
王卓霖,江俊扬,杨耿超等.面向边缘计算的神经网络MPS加速算法[J].中山大学学报(自然科学版),2023,62(05):67-77.

WANG Zhuolin,JIANG Junyang,YANG Gengchao,et al.An MPS algorithm accelerated by neural network on edge computing[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2023,62(05):67-77.
王卓霖,江俊扬,杨耿超等.面向边缘计算的神经网络MPS加速算法[J].中山大学学报(自然科学版),2023,62(05):67-77. DOI： 10.13471/j.cnki.acta.snus.2022A064.

WANG Zhuolin,JIANG Junyang,YANG Gengchao,et al.An MPS algorithm accelerated by neural network on edge computing[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2023,62(05):67-77. DOI： 10.13471/j.cnki.acta.snus.2022A064.

摘要

在无网格法中，移动粒子半隐式方法（MPS）利用流体的不可压性构造泊松方程，在获得了压力求解准确性的同时也带来了巨大的计算量，导致其不适合实现大规模流体模拟. 针对上述问题提出新型算法NN-MPS，将泊松方程的求解转化为利用神经网络求解回归问题. NN-MPS算法通过构建每一步的流场特征与压力的预测模型，实现泊松方程的快速求解. 本文进一步将基于NN-MPS算法的泊松方程求解过程移植到Atlas 200 DK上，实现边缘侧加速求解泊松方程. 本文采用多种溃坝模型进行数值实验，结果表明，本文的MPS加速方法具有低成本、高速度且较少精度损失的特点，求解速度实现了一定的提升. 本文同时也验证了边缘计算设备在计算流体力学领域应用的可行性.

Abstract

As a meshless method， the moving particle semi-implicit（MPS） method forms the pressure Poisson equation by using the incompressibility of fluid， which not only obtain the accuracy of pressure but also bring a high cost of calculation. Therefore， it is not appropriate for MPS to solve large-scale fluid simulations. IA new algorithm NN-MPS is proposed to solve the above problems， which transforms the solution of the Poisson equation into a regression problem by using neural network. The NN-MPS algorithm realizes the quick solution of the Poisson equation by constructing the prediction model of flow field features and pressure at each step. In this work， the procedure of solving the pressure Poisson equation is further transported to Atlas 200 DK device for a faster speed of solving procedure. Results show that the acceleration method of MPS mentioned in this work has the characteristics of low cost， high speed， and low accuracy loss， and the solution speed has been improved. We also verified the feasibility of applying the edge computing device in the field of CFD.

关键词

边缘计算神经网络NN-MPS计算流体力学

Keywords

edge computingneural networkNN-MPSCFD

references

崔勇，宋健，缪葱葱，等， 2017. 移动云计算研究进展与趋势［J］. 计算机学报， 40（2）： 273-295.

杜年春，沈向前，谢翔，等， 2022. Online SAR地质灾害监测和预警系统关键技术及应用研究［J］. 现代雷达， 44（10）：28-32.

何腾， 2020. 浅析边缘计算技术应用的现状及挑战［C］// 第三十四届中国（天津）2020’IT、网络、信息技术、电子、仪器仪表创新学术会议论文集：117-120.

华为技术有限公司， 2022. Atlas 200 DK开发者套件-概述［EB/OL］. （2022-12-29）［2023-02-23］. https：//www.hiascend.com/document/detail/zh/Atlas200DKDeveloperKit/1013/productdesc/atlas200_DK_pdes_19_0007.htmlhttps://www.hiascend.com/document/detail/zh/Atlas200DKDeveloperKit/1013/productdesc/atlas200_DK_pdes_19_0007.html

滕云豪， 2022. 基于边缘计算的室内空气质量监测技术研究［D］. 上海：上海第二工业大学.

杨超， 2018. 基于大涡模拟的移动粒子半隐式法研究流固耦合及冲击破坏问题［D］. 上海：上海交通大学.

张驰，张雨新，万德成， 2011. SPH方法和MPS方法模拟溃坝问题的比较分析［J］. 水动力学研究与进展A辑， 26（6）： 11.

张雨新，万德成， 2012. 用MPS方法数值模拟低充水液舱的晃荡［J］. 水动力学研究与进展A辑， 27（1）： 100-107.

CHIKAZAWA Y， KOSHIZUKA S， OKA Y， 2001. A particle method for elastic and visco-plastic structures and fluid-structure interactions［J］. Comput Mech， 27（2）： 97-106.

GARDNER M W， DORLING S R， 1998. Artificial neural networks （the multilayer perceptron）—a review of applications in the atmospheric sciences［J］. Atmos Environ， 32（14/15）：2627-2636.

GLOROT X， BORDES A， BENGIO Y， 2011. Deep sparse rectifier neural networks［J］. J Mach Learn Res， 15：315-323.

GOTOH H， SAKAI T， 2006. Key issues in the particle method for computation of wave breaking［J］. Coast Eng， 53（2/3）： 171-179.

JOUPPI N P， YOUNG C， PATIL N， et al，2017. In-datacenter performance analysis of a tensor processing unit［C］//Proceedings of 44th Annual International Symposium on Computer Architecture： 1-12.

KHAYYER A， GOTOH H， 2008. Development of CMPS method for accurate water-surface tracking in breaking waves［J］. Coast Eng J， 50（2）： 179-207.

KOSHIZUKA S， OKA Y， 1996. Moving-particle semi-implicit method for fragmentation of incompressible fluid［J］. Nucl Sci Eng， 123（3）： 421-434.

LEE B H， PARK J C， KIM M H， et al， 2011. Step-by-step improvement of MPS method in simulating violent free-surface motions and impact-loads［J］. Comput Methods Appl Mech Eng， 200（9/10/11/12）： 1113-1125.

LEE V W， KIM C， CHHUGANI J， et al， 2010. Debunking the 100X GPU vs. CPU myth： An evaluation of throughput computing on CPU and GPU［J］. Acm Sigarch Comput Archit News， 38（3）：451-460.

LIU D， CHEN T， LIU S， et al， 2015. PuDianNao： A polyvalent machine learning accelerator［J］. Acm Sigplan Notices， 50（4）：369-381.

MONAGHAN J J， 1992. Smoothed particle hydrodynamics［J］. Annu Rev Astron Astrophys， 30：543-574.

SHIBATA K， KOSHIZUKA S， 2007. Numerical analysis of shipping water impact on a deck using a particle method［J］. Ocean Eng， 34（3/4）： 585-593.

TANAKA M， MASUNAGA T， 2010. Stabilization and smoothing of pressure in MPS method by quasi-compressibility［J］. J Comput Phys， 229（11）： 4279-4290.

TSURUTA N， KHAYYER A， GOTOH H， 2013. A short note on dynamic stabilization of moving particle semi-implicit method ［J］. Comput Fluids， 82： 158-164.

TSURUTA N， KHAYYER A， GOTOH H， 2015. Space potential particles to enhance the stability of projection-based particle methods［J］. Int J Comput Fluid Dyn， 29（1）： 100-119.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Detail-aware discriminative feature learning model for vehicle re-identification

The Analysis and Forecast of Eco-environmental Quality ofWetland in Panyu of Guangzhou

A New Specific Combination Method of Wireless Communication Modulation Recognition Based on Clustering and Neural Network

Related Author

LI Xiying

QIU Mingkai

YANG Shixiong

ZHENG Yanwei

QIAN Lexiang

Zheng Zhuo

WANG Xiaojing

HUANG Kangyou

Related Institution

School of Intelligent Systems Engineering， Sun Yat-sen University / Guangdong Provincial Key Laboratory of Intelligent Transportation System / Key Laboratory of Video and Image Intelligent Analysis and Application Technology，Ministry of Public Security

Department of Earth Sciences, Sun Yatsen University，Guangzhou 510275, China

School of Geographical Sciences,Guangzhou University

School of Electronics and Information Engineering, Foshan University



Postal code：510275
Tel：020-84112585，84113223 Email：xuebaozr@mail.sysu.edu.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09064830号-19 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰