Modeling thermal conductivity of Ag/water nanofluid by applying a mathematical correlation and artificial neural network

Abstract Due to the significance importance of effective thermal conductivity of heat transfer fluids in various renewable energy system, such as geothermal and solar thermal plants, using naofluids can result in augment in the efficiency. Metallic nano particles dispersion in a pure fluid leads to considerable enhancement in the thermal conductivity. The improvement in the thermal conductivity is dependent on various factors. In the present research, two machine learning algorithms, a correlation and Group Method of Data Handling, are applied to predict thermal conductivity of silver/water nanofluid. Temperature, concentration and size of solid particles are considered as the input data. According to statistical comparison of the models, employing GMDH artificial neural network results in more precise and appropriate model. The coefficients of correlation, R-squared values, for the proposed correlation and ANN-based models are 0.948 and 0.99 respectively.