Investigation of void content in Borassus flabellifer fiber/epoxy bio-nanocomposite using hyperparameter tuned ANN and response surface methodology optimisation
Abstract Void development is one of the main problems faced by natural fiber polymer composites since it severely affects their physical and mechanical properties. It limits these composites for use in construction, aerospace, automotive and marine uses. Hence, this study takes on this issue by inco...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-06740-0 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Void development is one of the main problems faced by natural fiber polymer composites since it severely affects their physical and mechanical properties. It limits these composites for use in construction, aerospace, automotive and marine uses. Hence, this study takes on this issue by incorporating various nanosized Multi-Walled Carbon Nanotubes (MWCNT), hexagonal Boron Nitride (h-BN) and Alumina (Al2O3) nanofillers in epoxy-based Borassus flabellifer fiber (BFF) composites fabricated using the hand layup technique. The results show that increasing volume fraction of nanofiller gives way to decreasing fiber volume fraction, together with increase in composite density and void content. MWCNT-filled composites have the highest void content percentage among the different nanofillers investigated because of their lower theoretical density, which is inversely proportional to void content percentage. This research investigates the effects of the type of nanofiller, BFF mesh size and weight percent of the fiber upon the void content in fiber-reinforced composites. Design of Experiments approach is utilized to analyse the effect of these parameters and ANN model, employing advanced hyperparameter optimization strategy developed in Python, is used to elaborate upon specifics of the characteristics of void formation. Quantitative analysis of void content and particle distribution, analysed through SEM imaging and microstructural characterization through optical microscopy, further confirms these results, providing detailed information about void formation and filler dispersion. The optimized combination (1 wt% fiber content, 75 µm fiber mesh size and 1 wt% h-BN nanofiller) yielded 1.90% lowest void content after experimentation. This research provides fundamental understanding of the void mechanisms concerning bio-nano composites and presents an optimal predictor model that minimizes voids. This contributes and builds toward the directions of advancing materials for high-performance applications. |
|---|---|
| ISSN: | 2045-2322 |