Applying different types of artificial neural network for modeling thermal conductivity of nanofluids containing silica particles

Nanofluids are widely applicable in thermal devices with porous structures. Silica nanoparticles have been dispersed in different heat transfer fluids in order to increase their thermal conductivity and heat transfer capability. In this study, group method of data handling (GMDH) and multilayer perceptron artificial neural networks are applied for determining thermal conductivity of nanofluids with silica particles and different base fluids such as ethylene glycol, glycerol, water and ethylene glycol-water mixture. For cases with multilayer perceptron models, trained by applying scaled conjugate gradient (SCG) and Levenberg-Marquardt (LM) have been tested as two different training algorithms. The outputs of the applied models have good agreement with the values obtained in experimental studies. The values of R-2 in the optimum conditions of using GMDH, LM and SCG are 0.9997, 0.9991 and 0.9998, respectively. In addition, the MSE values of the mentioned methods are approximately 0.000010, 0.000032 and 0.0000078, respectively.

Automated summary

In this study, GMDH and multilayer perceptron artificial neural networks were used to determine the thermal conductivity of nanofluids containing silica particles and different base fluids. The models showed good agreement with experimental results, with R2 values reaching up to 0.9998.