Short Papers and Posters of the 19th Annual International Scientific Conference on Parallel Computational Technologies
Virabyan, Aram · 2025
This study presents the development of a predictive model for the specific heat capacity of nanofluids based on the Random Forest Regressor method, optimized using GridSearchCV.The final model parameters demonstrated high prediction accuracy (MSE = 4.16, R² = 0.99999), as confirmed by residual analysis and comparison between actual and predicted values.The model was successfully tested on data for isopropyl alcohol with nanoparticles, showing minimal deviations from the experimental values.Despite limitations associated with the clustered structure of the data, the model exhibits potential for application to other base fluids and nanoparticles, making it a valuable tool for studying the thermophysical properties of nanofluids and developing new materials.