有限元法
稳健性(进化)
人工神经网络
计算机科学
水准点(测量)
灵活性(工程)
可靠性(半导体)
反向
适应性
领域(数学分析)
算法
区域分解方法
反问题
功能(生物学)
同种类的
多边形网格
时域
频域
逆方法
电磁学
数学优化
混合有限元法
有效域
计算电磁学
扩展有限元法
作者
Mohammad Mehdi Fakharian
出处
期刊:Compel-the International Journal for Computation and Mathematics in Electrical and Electronic Engineering
[Emerald Publishing Limited]
日期:2026-08-31
卷期号:45 (6): 1297-1315
标识
DOI:10.1108/compel-11-2025-0537
摘要
Purpose Accurate three-dimensional (3D) electrostatic (ES) simulation is crucial but computationally challenging using traditional finite element methods (FEM) due to meshing bottlenecks. This study aims to systematically evaluate physics-informed neural networks (PINNs) against FEM for solving 3D ES problems. Design/methodology/approach Two 3D benchmark problems were solved: a homogeneous cubic domain governed by Laplace’s equation, and a multi-material domain with discontinuous properties. Both methods were evaluated to assess accuracy, computational efficiency and adaptability to geometric and material complexities. Findings FEM demonstrated robustness in well-meshed domains, achieving a −4.8 dB mean error, though mesh generation created computational bottlenecks. PINNs achieved comparable accuracy without meshing, yielding −5.7 dB error in the homogeneous case and −7.0 dB in the multi-material case. PINNs excelled at enforcing interface continuity in multi-material problems, though they required careful loss function tuning. Originality/value FEM remains superior for static, well-defined problems, whereas PINNs offer significant advantages for mesh-free environments and inverse scenarios. The study proposes hybrid methodologies combining FEM’s reliability with PINNs’ flexibility to advance 3D electromagnetic simulations.
科研通智能强力驱动
Strongly Powered by AbleSci AI