arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.04233physics.flu-dyncs.LGcs.NAmath.NAphysics.comp-ph
收藏

深度神经网络校正CFD中的亚精度误差

Deep Neural Networks to Correct Sub-Precision Errors in CFD

Indian Institute of Technology Madras(印度理工学院马德拉斯分校);Tokyo Institute of Technology(东京工业大学);Japan Science and Technology Agency(日本科学技术振兴机构)

Akash Haridas, Nagabhushana Rao Vadlamani, Yuki Minamoto

摘要:

在求解离散偏微分方程时,数值物理模拟中的信息丢失可能来自各种来源。特别地,与等效的64位模拟相比,当使用低精度16位浮点运算执行模拟时,与数值精度相关的误差(“亚精度误差”)可以在感兴趣的量中累积。另一方面,低精度计算比高精度计算更少占用资源。最近提出的几种机器学习技术已经成功地纠正了由于粗糙的空间离散化引起的错误。在这项工作中,我们扩展这些技术,以提高CFD模拟进行低数值精度。我们量化的精度相关的误差累积在Kolmogorov强迫湍流测试情况下。随后,我们采用卷积神经网络与执行16位算法的完全可微数值求解器一起学习紧密耦合的ML-CFD混合求解器。与16位求解器相比,我们证明了混合求解器对提高模拟的统计和逐点精度的各种指标的有效性。

英文摘要:

Information loss in numerical physics simulations can arise from various sources when solving discretized partial differential equations. In particular, errors related to numerical precision ("sub-precision errors") can accumulate in the quantities of interest when the simulations are performed using low-precision 16-bit floating-point arithmetic compared to an equivalent 64-bit simulation. On the other hand, low-precision computation is less resource intensive than high-precision computation. Several machine learning techniques proposed recently have been successful in correcting errors due to coarse spatial discretization. In this work, we extend these techniques to improve CFD simulations performed with low numerical precision. We quantify the precision-related errors accumulated in a Kolmogorov forced turbulence test case. Subsequently, we employ a Convolutional Neural Network together with a fully differentiable numerical solver performing 16-bit arithmetic to learn a tightly-coupled ML-CFD hybrid solver. Compared to the 16-bit solver, we demonstrate the efficacy of the hybrid solver towards improving various metrics pertaining to the statistical and pointwise accuracy of the simulation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑