A new correlation for estimating thermal conductivity of pure ionic liquids

Abstract

In the present study, a new correlation with both correlative and extrapolative capabilities is proposed for thermal conductivity of the pure ionic liquids (ILs) and the optimum values of its fitting parameters were obtained by a genetic algorithm. All of all, 209 data points of 21 ILs were collected from previously published literature. The collected data were divided into two different subsets namely training (143 data points) and testing (66 data points) subsets. Training subset was used to find the optimum values of the fitting parameters of the proposed correlation. After that, the extrapolative and correlative capabilities of the proposed correlation was tested by testing data subset was not utilized during the training stage. The average absolute relative deviation percent (AARD %) during the training stage was 5.22%, while the results revealed AARD % of 10.76% for testing stage. In general, the overall results revealed a rather good accuracy of the proposed correlation especially considering its extrapolative capability. Finally, the obtained results by the proposed correlation were compared with those obtained by different available correlations including group contribution and artificial neural network approach which shows a good functionality of the proposed correlation.