Results in Engineering · 2024 · 21 citations · 33 references
Natural SciencesMaterials EngineeringMaterials ScienceMachine LearningMaterials CharacterizationNanomaterialsCarbon NanotubesMaterial ModelingMaterial PropertiesNanocompositesNanotubesTaguchi MethodsCarbon Nanotube CompositesCarbon Nanotube NanocompositesMaterial Performance PredictionMaterials Property PredictionTaguchi Designs
• Developed a novel microwave-assisted method to control inter-tube spacing in MWCNT/polymer nanocomposites. • Utilized the Taguchi method and ANOVA to identify key factors influencing impedance. • Applied five machine learning models to predict impedance with high accuracy. • Achieved top predictive accuracy with Random Forest and CatBoost (R² > 0.98). • Tailored impedance properties to enhance applications in electronics and energy storage. The impedance characteristics of multi-walled carbon nanotube (MWCNT)/polystyrene nanocomposites synthesized via microwave-assisted in-situ polymerization were systematically investigated to determine the effects of microwave power, exposure time, and frequency on impedance properties. The Taguchi method and analysis of variance (ANOVA) identified microwave power as the most significant factor, followed by exposure duration and frequency. A predictive model was developed, demonstrating high accuracy with a coefficient of determination (R²) of 0.96 between model predictions and experimental results. Additionally, response surface methodology (RSM) and contour plots were applied to explore optimal parameter combinations, offering valuable insights for achieving tailored impedance values. Machine learning model including Decision Tree, Random Forest, Extreme Gradient Boosting (XGBoost), Categorical Boost (CatBoost), and Light Gradient-Boosting Machine (LightGBM) were employed to enhance predictive capabilities. Among these, Random Forest and CatBoost demonstrated superior accuracy, achieving R² values of 0.9880 and 0.9811 on testing data, respectively, while Decision Tree and LightGBM exhibited lower performance. This study highlights the potential of machine learning methods to precisely adjust and tailor impedance properties of PS/CNT nanocomposites, supporting the engineering of materials for diverse applications across materials science and engineering.
33
T Arturo Fernández · conference · 2006 · 5.2K citations
Aanchna Sharma, T. Mukhopadhyay, Sanjay Mavinkere Rangappa et al. · Archives of Computational Methods in Engineering · 2022 · 249 citations
Natural Sciences, Materials Engineering, Materials Science +16
Bokai Liu, Xiaoying Zhuang, Xiaolong Fu et al. · Composites Science and Technology · 2022 · 184 citations
Natural Sciences, Materials Engineering, Materials Science +17