Performance of Al2O3/R134a nanorefrigerant based modified thermosyphon predicted by random forest algorithm is presented in this paper. Three types of thermosyphon namely conventional, thermosyphon with internal fin and thermosyphon with internal fin and internal cut were used in the study. The input parameters to the random forest algorithm are heat input, surface area, cooling water temperatures and the working fluid; while the output are in terms of truepositive rate, false positive rate, positive predictive value and accuracy. The results of the analysis indicate that accuracy of 95% has been obtained with regard to the experimental studies carried out. The predicted output is observed to be in good agreement with the experimental results. The temperature obtained at each section of the thermosyphon provides the true positive rate greater than 96% with less fault occurrence using random forest algorithmic approach.
Read full abstract