Advection-Diffusion BVP Data
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/ikalash/Schwarz-4-Multiscale
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了针对一维稳态对流扩散方程在以对流为主的区域中，理解物理信息神经网络（PINNs）训练的数值实验配置和结果。数据集还包括了配置点、损失函数参数以及通过二阶精确有限差分方法获得的真实数据。该数据集的规模为具有1024个配置点的一维输入层。任务旨在训练物理信息神经网络以解决偏微分方程问题。
This dataset contains numerical experimental configurations and training results for Physics-Informed Neural Networks (PINNs) applied to one-dimensional steady-state convection-diffusion equations in convection-dominant regions. It additionally includes collocation points, loss function parameters, and ground-truth data obtained using a second-order accurate finite difference method. The dataset has a one-dimensional input layer with 1024 collocation points. The task objective is to train physics-informed neural networks to solve partial differential equation (PDE) problems.
提供机构：
The authors of the paper
