Deep learning multilayer architecture for analysis of three-dimensional Eyring-Powell nanofluid flow subject to viscous dissipation and joule heating
Artificial intelligence has enhanced complex systems modeling by achieving exceptional precision and effectiveness. The purpose of this paper is to use the Deep Learning Multi-Layer Soft Computing paradigm (DLML-SCP) to evaluate the model 3D flow dynamics of Eyring-Powell nanofluids subject to visco...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | Results in Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S259012302500917X |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Artificial intelligence has enhanced complex systems modeling by achieving exceptional precision and effectiveness. The purpose of this paper is to use the Deep Learning Multi-Layer Soft Computing paradigm (DLML-SCP) to evaluate the model 3D flow dynamics of Eyring-Powell nanofluids subject to viscous dissipation and joule heating (EPNF-3D-VJ). It can capture the complex behavior of a specialized fluid flow system with the changing of different physical parameters like Prandtl number, magnetic field parameter, radiation parameter, and Brownian motion parameter and the controlling parameters of non-Newtonian nature of the fluid, ε, δ1, and δ2, make up the source parameters of Eyring-Powell fluid. Similarity transformation in the governing partial differential equations (PDEs) ultimately leads to a reduced set of ordinary differential equations (ODEs) along with boundary conditions. Not only does it simplify the computational process, but it also boosts the model's predictive abilities. The Adam numerical technique is employed to generate a comprehensive dataset across the fluid-dynamic spectrum that covers diverse EPNF-3D-VJ cases, offering critical insights into the system's behavior. The dataset is utilized to test, train, and validate the DLML-SCP, proving its accuracy in predicting fluid system behavior. Results show the effectiveness and reliability of the model and are well aligned to the dataset. Validation results demonstrate that the DLML-SCP framework accurately predicts fluid system behavior, achieving mean square error (MSE) values between E-09 to E-10. Important findings from the performance evaluations, it can be noted that the increase in both the Prandtl number and magnetic field coefficient has led to a decrease in the temperature and velocity profiles. With an increase in the Eyring-Powell fluid parameter, the velocity profile is likely to increase. present study utilized an advanced machine learning framework that has been proven to outperform the state-of-the-art models in predicting complex dynamics of EPNF-3D-VJ. Providing strong support for a new class of physics-informed deep neural networks designed for solving fluid mechanics problems. |
|---|---|
| ISSN: | 2590-1230 |