Physics-Informed Neural Network for Solving a One-Dimensional Solid Mechanics Problem DOI Creative Commons
Vishal Singh, Dineshkumar Harursampath, Sharanjeet Dhawan

et al.

Modelling—International Open Access Journal of Modelling in Engineering Science, Journal Year: 2024, Volume and Issue: 5(4), P. 1532 - 1549

Published: Oct. 18, 2024

Our objective in this work is to demonstrate how physics-informed neural networks, a type of deep learning technology, can be utilized examine the mechanical properties helicopter blade. The blade regarded as one-dimensional prismatic cantilever beam that exposed triangular loading, and comprehending its behavior utmost importance aerospace field. PINNs utilize physical information, including differential equations boundary conditions, within loss function network approximate solution. approach determines overall by aggregating losses from equation, data. We employed (PINN) an artificial (ANN) with equivalent hyperparameters solve fourth-order equation. By comparing performance PINN model against analytical solution equation results obtained ANN model, we have conclusively shown exhibits superior accuracy, robustness, computational efficiency when addressing high-order govern physics-based problems. In conclusion, study demonstrates offers alternative for solid mechanics problems applications industry.

Language: Английский

Physics-Informed Neural Networks for Predicting Internal Forces and Deformations of Structural Frames in a Single-Span Agricultural Greenhouse DOI
Solhee Kim, Tae‐Gon Kim, Jeongbae Jeon

et al.

Horticultural Science and Technology, Journal Year: 2025, Volume and Issue: 43, P. 1 - 19

Published: April 24, 2025

Language: Английский

Citations

0
Outlier-resistant physics-informed neural network DOI
David Duarte, Paulo Douglas Santos de Lima, João M. de Araújo

et al.

Physical review. E, Journal Year: 2025, Volume and Issue: 111(2)

Published: Feb. 20, 2025

Recent advances in machine learning have introduced physics-informed neural networks (PINN) as a valuable tool for addressing dynamics through governing equations and experimental observations. Outliers can be present measurements significantly affect the accuracy of solutions provided by PINN. To overcome this limitation, we construct an outlier-resistant PINN (OrPINN) based on Tsallis statistics. We investigate robustness OrPINN describing acoustic linear elastic wave under various outlier-level scenarios. find that improve even when data is highly corrupted.

Language: Английский

Citations

0
Hierarchical Design of Mechanical Metamaterials: an Application on Pentamode-like Structures DOI Creative Commons
S. Gómez, Emilio P. Calius, Akbar Afaghi Khatibi

et al.

International Journal of Mechanical Sciences, Journal Year: 2025, Volume and Issue: unknown, P. 110232 - 110232

Published: April 1, 2025

Language: Английский

Citations

0
Physics-Informed Neural Network for Solving a One-Dimensional Solid Mechanics Problem DOI Creative Commons
Vishal Singh, Dineshkumar Harursampath, Sharanjeet Dhawan

et al.

Modelling—International Open Access Journal of Modelling in Engineering Science, Journal Year: 2024, Volume and Issue: 5(4), P. 1532 - 1549

Published: Oct. 18, 2024

Our objective in this work is to demonstrate how physics-informed neural networks, a type of deep learning technology, can be utilized examine the mechanical properties helicopter blade. The blade regarded as one-dimensional prismatic cantilever beam that exposed triangular loading, and comprehending its behavior utmost importance aerospace field. PINNs utilize physical information, including differential equations boundary conditions, within loss function network approximate solution. approach determines overall by aggregating losses from equation, data. We employed (PINN) an artificial (ANN) with equivalent hyperparameters solve fourth-order equation. By comparing performance PINN model against analytical solution equation results obtained ANN model, we have conclusively shown exhibits superior accuracy, robustness, computational efficiency when addressing high-order govern physics-based problems. In conclusion, study demonstrates offers alternative for solid mechanics problems applications industry.

Language: Английский

Citations

2