576 / 2019-01-18 23:56:12

The thermal conductivity and viscosity of Al2O3- water nanofluid were predicted by BP neural network - genetic algorithm (BP-GA)

BP neural network; Genetic algorithm; Density; Thermal conductivity

全体主题 > 纳米传热、纳米流体流动和纳米材料

In this paper, a hybrid model (BP-GA) including back propagation network and genetic algorithm is used to estimate the density and viscosity of nanofluids. The genetic algorithm is used to optimize the initial weights and thresholds of BP neural network, so that the optimized BP neural network can better predict the function output. The purpose of optimization of BP neural network by genetic algorithm is to get better initial weights and thresholds of network by genetic algorithm. The basic idea is to use the prediction error of BP neural network which represents the initial weights and thresholds of network and initializes individual values as the fitness value of the individual, and to find the optimal individual through selection, crossover and mutation operations, that is, the optimal initial BP neural network Weights. Based on the experimental temperature range of alumina-water nanofluids, the volume fraction of nanoparticles is in the range of 0.5% - 1%, and the mixing ratio of alumina at 20 and 50 nm is in the range of 1-9. The prediction results of BP neural network optimized by genetic algorithm with alumina particle size mixing ratio, temperature and volume fraction as input and thermal conductivity and viscosity as output show that the prediction model is in good agreement with experimental data, the relative deviation is less than 5%, and the determination coefficient (R^2≥0.96), the results also show that the prediction performance of (BP-GA) model is better than that of BP neural network and
radial basis function neural network (RBF), respectively, Finally, other experimental data of nanofluids are used as test data. The prediction model optimized by genetic algorithm is better in accuracy and stability of prediction results, which further proves the accuracy of BP neural
network prediction model optimized by genetic algorithm.