Prediction of thermal conductivity and viscosity of water-based carbon black nanofluids based on GA-BP neural network model

Nanofluids have been widely used in various fields due to their unique enhanced heat transfer properties. The thermal conductivity and viscosity directly affect the applicability of nanofluids in practical engineering, so before examining the enhanced heat transfer characteristics of nanofluids, it is first necessary to analyze and study their thermal conductivity and viscosity. In this study, water-based carbon black collagen nanofluids were prepared by a two-step method using carbon black and collagen. The effects of carbon black and collagen concentration and temperature on the thermal conductivity and viscosity of nanofluids were analyzed. The weights of these parameters were mathematically calculated by the gray correlation method, and a BP neural network prediction model with three inputs and two outputs was established based on the experimental data, and the BP model was optimized by genetic algorithm (GA). The results showed that the BP neural network model optimized by the genetic algorithm had higher accuracy and better stability for the predicted output, and the regression coefficient and maximum deviation were 0.99918 and 0.002, respectively. This study was not only of great significance for understanding and controlling the thermophysical properties of water-based carbon black-collagen nanofluids, but also provided new ideas for the application of engineering design and materials science. © 2024 Chemical Industry Press Co., Ltd.. All rights reserved.