Machine Learning for Prediction of Heat Pipe Effectiveness

This paper details the selection of machine learning models for predicting effectiveness a heat pipe system in concentric tube exchanger. Heat exchanger experiments with methanol as working fluid were conducted. The value angle varied from 0° to 90°, values temperature 50 °C 70 °C, and flow rate 40 120 litres per min. Multiple conducted at different combinations input parameters was measured each trial. algorithms taken into consideration prediction. Experimental data divided subsets performance model analysed subsets. For overall analysis, which included all three parameters, random forest algorithm returned best results mean average error 1.176 root-mean-square-error 1.542.