Plabon Paul
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Plabon Paul has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- A. K. M. Monjur Morshed
Department of Mechanical Engineering, Bangladesh University of Engineering and Technology, Dhaka, Bangladesh
1 shared publication Priom Das
Bangladesh University of Engineering and Technology
1 shared publicationSadia Tasnim
Bangladesh University of Engineering and Technology
1 shared publicationSakib Javed
Bangladesh University of Engineering and Technology
1 shared publication
Latest Publications
- Open Access
- 1 September 2025
- Published by Wiley
Data-Driven Machine and Deep Learning-Based Prediction of Interfacial Thermal Resistance Using High-Fidelity Molecular Dynamics Simulations
The van der Waals heterostructure (vdWH) of Stanene/Hexagonal Boron Nitride (Sn/h-BN) nanosheet has drawn attention due to its promising electro-thermal properties. Understanding the interfacial thermal resistance (ITR) between layers is essential for optimal nanodevice design. Various internal and external factors, such as in-plane tensile strain, defect percentage, interlayer coupling strength (contact pressure), and system temperature, significantly influence thermal transport at the Sn/h-BN bilayer interface. While molecular dynamics (MD) simulations offer insights, they are computationally expensive, especially for low-dimensional materials with complex thermal behaviors. To address this, machine learning (ML) and deep learning (DL) models were developed to rapidly predict ITR with high accuracy. A total of 1000 training datasets were generated using high-throughput MD simulations. Five deep learning models (ANN-10, ANN-20, DNN-10-10, DNN-20-20, and DNN_7) and five machine learning models (Linear Regression, Polynomial Regression, K-Nearest Neighbor, Support Vector Machine, Random Forest) were trained and hyperparameter-optimized. Performance was evaluated using R 2 score, MSE, and MAPE. Among DL models, DNN_7 achieved the best results with an R 2 of 0.97, a Mean Absolute Percent Error (MAPE) of 0.041, and a Mean Square Error (MSE) of 1.897. Among ML models, Random Forest performed best with an R 2 of 0.96. Correlation analysis revealed that interlayer coupling strength is the most influential factor in modulating ITR. This study demonstrates the effectiveness of data-driven models in predicting thermal properties at material interfaces, offering a fast and reliable alternative to traditional simulations. These findings provide foundational insights for designing high-performance nanoelectronic devices by enabling precise thermal management through interface engineering.
- 1 September 2025
- Published by Wiley
Open AccessData-Driven Machine and Deep Learning-Based Prediction of Interfacial Thermal Resistance Using High-Fidelity Molecular Dynamics Simulations
The van der Waals heterostructure (vdWH) of Stanene/Hexagonal Boron Nitride (Sn/h-BN) nanosheet has drawn attention due to its promising electro-thermal properties. Understanding the interfacial thermal resistance (ITR) between layers is essential for optimal nanodevice design. Various internal and external factors, such as in-plane tensile strain, defect percentage, interlayer coupling strength (contact pressure), and system temperature, significantly influence thermal transport at the Sn/h-BN bilayer interface. While molecular dynamics (MD) simulations offer insights, they are computationally expensive, especially for low-dimensional materials with complex thermal behaviors. To address this, machine learning (ML) and deep learning (DL) models were developed to rapidly predict ITR with high accuracy. A total of 1000 training datasets were generated using high-throughput MD simulations. Five deep learning models (ANN-10, ANN-20, DNN-10-10, DNN-20-20, and DNN_7) and five machine learning models (Linear Regression, Polynomial Regression, K-Nearest Neighbor, Support Vector Machine, Random Forest) were trained and hyperparameter-optimized. Performance was evaluated using R 2 score, MSE, and MAPE. Among DL models, DNN_7 achieved the best results with an R 2 of 0.97, a Mean Absolute Percent Error (MAPE) of 0.041, and a Mean Square Error (MSE) of 1.897. Among ML models, Random Forest performed best with an R 2 of 0.96. Correlation analysis revealed that interlayer coupling strength is the most influential factor in modulating ITR. This study demonstrates the effectiveness of data-driven models in predicting thermal properties at material interfaces, offering a fast and reliable alternative to traditional simulations. These findings provide foundational insights for designing high-performance nanoelectronic devices by enabling precise thermal management through interface engineering.