Abstract :

Computational Fluid Dynamics (CFD) Has Become An Essential Tool For Analyzing Complex Fluid Flow Phenomena Across Aerospace, Automotive, Energy, And Industrial Applications. However, Traditional CFD Methods Often Require Significant Computational Resources And Extensive Numerical Discretization To Solve Governing Equations Accurately. Recent Advancements In Artificial Intelligence Have Introduced Physics-Informed Neural Networks (PINNs) As A Promising Alternative For Accelerating Fluid Flow Simulations While Preserving Physical Consistency. This Study Investigates The Enhancement Of CFD Analysis Through The Integration Of PINNs, Which Embed Fundamental Conservation Laws Directly Into The Neural Network Training Process. The Proposed Framework Combines Data-driven Learning With The Navier–Stokes Equations To Improve Prediction Accuracy, Reduce Computational Complexity, And Enhance Generalization Across Different Flow Conditions. The Methodology Involves Domain Discretization, Physicsconstrained Neural Network Training, Residual Error Minimization, And CFD Solution Refinement. Performance Evaluation Is Conducted Using Benchmark Fluid Flow Scenarios Including Laminar And Turbulent Regimes. Results Demonstrate That PINN-assisted CFD Significantly Reduces Computational Time While Maintaining High Levels Of Accuracy Compared With Conventional Numerical Approaches. The Study Highlights The Potential Of Physicsinformed Machine Learning For Next-generation CFD Applications, Enabling Faster Simulations, Improved Scalability, And Efficient Optimization Of Engineering Systems.

Published:

16-12-2022

Issue:

Vol. 22 No. 12 (2022)

Page Nos:

61-70

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite