In collaboration with Professor Divya Nayar, this project aimed to develop a robust model for predicting the thermal conductivity of polymer-based composites. The focus was on understanding how varying mass fractions of fillers, particularly polyethylene and polystyrene, influence the thermal conductivity of the composite material.
- 📅Situation: The accuracy of existing theoretical models for predicting the thermal conductivity of polymer-based composites was found to be insufficient. The project commenced in February 2023 as part of a course project under the guidance of Prof. Divya Nayar.
- 🎯Task: Our task was to establish a predictive model capable of predicting more accurately than theoritical models, determining the thermal conductivity of polymer-based composites across different mass fractions of fillers.
- ⚙️Action:
- Utilized support vector regression (SVR) as the primary modeling technique.
- Employed particle swarm optimization (PSO) for parameter tuning to enhance the model's performance and predictive accuracy.
- 📈Result:
- Achieved a superior generalization ability with the SVR model, showcasing a remarkable R-squared accuracy of 0.954.
- The model outperformed existing theoretical frameworks, demonstrating its effectiveness in predicting the thermal conductivity of polymer-based composites.
- Presented the findings to a group of PhD researchers, professors, and colleagues, receiving positive feedback and valuable insights.