A Computational Intelligence Framework Integrating Data Augmentation and Meta-Heuristic Optimization Algorithms for Enhanced Hybrid Nanofluid Density Prediction Through Machine and Deep Learning Paradigms
This research presents a robust and comprehensive framework for predicting the density of hybrid nanofluids using state-of-the-art machine learning and deep learning techniques. Addressing the limitations of conventional empirical approaches, the study used a curated dataset of 436 samples from the...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10892114/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This research presents a robust and comprehensive framework for predicting the density of hybrid nanofluids using state-of-the-art machine learning and deep learning techniques. Addressing the limitations of conventional empirical approaches, the study used a curated dataset of 436 samples from the peer-reviewed literature, which includes nine input parameters such as the nanoparticle, base fluid, temperature (°C), volume concentration (<inline-formula> <tex-math notation="LaTeX">$\phi $ </tex-math></inline-formula>), base fluid density (<inline-formula> <tex-math notation="LaTeX">$\rho _{\text {bf}}$ </tex-math></inline-formula>), density of primary and secondary nanoparticles (<inline-formula> <tex-math notation="LaTeX">$\rho _{\text {np1}}$ </tex-math></inline-formula> and <inline-formula> <tex-math notation="LaTeX">$\rho _{\text {np2}}$ </tex-math></inline-formula>), and volume mixture ratios of primary and secondary nanoparticles. Data preprocessing involved outlier removal via the Interquartile Range (IQR) method, followed by augmentation using either autoencoder-based or Gaussian noise injection, which preserved statistical integrity and enhanced dataset diversity. The research analyzed fourteen predictive models, employing advanced hyperparameter optimization methods facilitated by Grey Wolf Optimization (GWO) and Particle Swarm Optimization (PSO). In particular, autoencoder-based augmentation combined with hyperparameter optimization consistently improved predictive accuracy across all models. For machine learning models, Gradient Boosting achieved the most remarkable performance, with R2 scores of 0.99999 and minimal MSE values of 0.00091. Among deep learning models, Recurrent Neural Networks (RNN) stacked with Linear Regression achieved superior performance with an R2 of 0.9999, MSE of 0.0014, and MAE of 0.012. The findings underscore the synergy of advanced data augmentation, meta-heuristic optimization, and modern predictive algorithms in modelling hybrid nanofluid density with unprecedented precision. This framework offers a scalable and reliable tool for advancing nanofluid-based applications in thermal engineering and related domains. |
|---|---|
| ISSN: | 2169-3536 |