Exploration of Composition-Microstructure Relations in Polymer Nanocomposites: Intelligent Honey

Authors A.B. Pawar1, P. William2 , M.V. Kulkarni3, Sharmila4, D.K. Roy5, N. Yogeesh6
Affiliations

1Department of Computer Engineering, Sanjivani College Engineering, Kopargaon, SPPU, Pune, India

2Department of Information Technology, Sanjivani College Engineering, Kopargaon, SPPU, Pune, India

3Engineering Science and Humanities, Sanjivani College of Engineering, Kopargaon, SPPU, Pune, India

4Department of ECE, Raj Kumar Goel Institute of Technology, Ghaziabad, India

5Hyderabad Institute of Technology and Management, Gowdavelli Village, Medchal, Hyderabad

6Department of Mathematics, Government First Grade College, Tumkur, Karnataka, India

Е-mail william160891@gmail.com
Issue Volume 16, Year 2024, Number 4
Dates Received 10 April 2024; revised manuscript received 15 August 2024; published online 27 August 2024
Citation A.B. Pawar, P. William, et al., J. Nano- Electron. Phys. 16 No 4, 04004 (2024)
DOI https://doi.org/10.21272/jnep.16(4).04004
PACS Number(s) 81.05.Qk
Keywords Polymer nanocomposites (PNCs), Nanoparticle (NP), Radial distribution function (RDF), Intelligent Honey Bee-Fused Dynamic Random Forest (IHB-FDRF), Mean squared error (MSE).
Annotation

In polymer nanocomposites (PNCs), microstructure relationships encompass the intricate connections between nanofiller placement in the polymer matrix and resultant composite characteristics. The dispersion, size, shape and chemical interactions of nanofillers impact PNC performance, with thermo physical qualities varying based on composition. Establishing a universal composition-property relationship for PNCs is challenging due to their vast chemical diversity. This study proposes an innovative machine learning (ML) approach, the Intelligent Honey Bee-Fused Dynamic Random Forest (IHB-FDRF), to predict the composition-microstructure relationship of PNCs. Leveraging computational vision and image recognition, the IHB-FDRF predicts nanoparticles (NP) dispersion, validated through coarse-grained molecular dynamics simulations. The model forecasts NP arrangement in PNCs in latent space, translating to the radial distribution function (RDF) using the IHB-FDRF algorithm. The Mean Squared Error (MSE) in predictions, quantifying the average squared difference between predicted and actual values, is impressively low at 0.005 during the training phase, affirming the model's accuracy. The study's robustness is further confirmed by the overlap of latent values in both areas, signifying convergence between hidden characteristics and ensuring reliability across diverse contexts. In summary, this study gives significant findings on PNC microstructure relationships.

List of References