Estimating Aqueous Nanofluids Viscosity via GEP Modeling: Correlation Development and Data Assessment

This study focuses on developing a new method that represents user-accessible correlation for the estimation of water-based nanofluids viscosity. For this, an evolutionary algorithm, namely Gene Expression Programming (GEP), was adapted based on a wide selection of literature published databanks including 819 water-based nanofluids viscosity points. The developed model utilized the base fluid viscosity as well as volume fraction and size of the nanoparticles as the inputs of the model. Several statistical parameters integrated with graphical plots were employed in order to assess the accuracy of the proposed GEP-based model. Results of the evaluation demonstrate fairly enough accuracy of the developed model with statistical parameters of AARD%=11.7913, RMSE=0.3567, and SD=0.1851. Furthermore, the trend analysis indicates that the GEP calculated points satisfactorily follow the trend of the nanofluid viscosity variation as a function of different model inputs. To provide more verification, the proposed GEP model was compared with some literature theoretical and empirical correlations leading to the supremacy of the developed model here. The applied sensitivity analysis reveals that the highest impact value is assigned to the volume fraction of the nanoparticle. Moreover, the outlier’s detection by Williams’ technique illustrates that about 96.5% of the GEP estimates are in the applicability domain resulting in the validity of the proposed model in this study. At last, the results of this study demonstrate that the new method here outperforms other literature-published correlations from the standpoint of accuracy and reliability.