Established prediction models of thermal conductivity of hybrid nanofluids based on artificial neural network (ANN) models in waste heat system

Abstract The properties of water (W)/ethylene glycol (EG) mixtures vary significantly with the proportion of EG and temperature, so it is suitable to use such fluids as exchange heat mediums in a waste heat system with temperature fluctuations. The experiments were conducted with 1.0 wt% Cu/Al2O3- EG/W hybrid nanofluids at temperatures ranging from 20 to 50 °C, where the base fluid (EG/W) mixture ratio was varied from 20:80 to 80:20. To search individuals which contain optimal weights and thresholds, a genetic algorithm (GA) and a mind evolutionary algorithm (MEA) coupled with a back-propagation neural network (GA-BPNN and MEA-BPNN, respectively) were used to improve the accuracy in the predicted thermal conductivity. The results show that the thermal conductivity increases nonlinearly with the ratio of water to ethylene glycol and temperature, due to the higher thermal conductivity of water and stronger collision frequency between molecular and nanoparticles. Binary Polynomial Regression (BPR) was fit with (coefficient of determination) R2 = 0.9984 as functions of temperature and mixture ratio. Comparisons of the prediction performance and capability of BPR, the performance of R2 increases by 0.11% and 0.13% for GA-BPNN and MEA-BPNN. It indicates that the combined BPNNs both predicate more accurately, particularly MEA-BPNN has the highest prediction accuracy.