Application of advanced neural network approaches to investigate the heat transfer features of Powell-Eyring hybrid nanofluids
Abstract
Possibilities from artificial intelligence are emerging swiftly in many fields, offering novel approaches as well as greatly enhanced ability to analyze complex situations and patterns of various facets. The focus of this paper is to extend the utilization of neural networks with the Levenberg-Marquardt Scheme (NNs-LMS) to explore entropy analysis of Powell–Eyring hybrid nanofluid, which considers the impacts of linear thermal radiation and viscous dissipation (EPNF-LTR-VD). The critical parameters like Prandtl number, suction/injection parameter, Eckert number, and material parameter are taken into consideration in order to understand the flow characteristics and heat transfer rates in this research. By employing the right similarity transformations, the set of nonlinear PDEs is reduced to set of nonlinear ODEs, thereby enabling the lowering the order of the system complexity. A similar dataset is then obtained with the help of the Adam numerical method for analyzing the impacts that these critical parameters can have on more extensive EPNF-LTR-VD scenarios. The dataset is utilized to test, train, and validate the EPNF-LTR-VD, proving its accuracy in predicting fluid system behavior. The empirical studies do not only provide evidence of effectiveness and accuracy of the proposed approach but also identify a high level of correspondence with reference data. Additional support for the model performance is provided by sophisticated performance graphs, error histograms, as well as regression evaluations. This research shows how AI can be used to advance the modeling of fluids; it helps to bring a new perspective into how fluid dynamics can be analyzed and differs from the previous works.
Author
Nashwan Adnan OTHMAN
ISSN
Publish Date: 2025-06-11