Designing back propagation neural network to predict CO2 mass transfer enhancement factor of TiO2-MEA/MDEA blended amine nanofluids
Xue Li, Jiamin Qin, Xining Wu, Chaoyang Wei, Long Xu
Abstract
This study presented a novel methodology to predict the CO2 absorption enhancement performance of TiO2-Monoethanolamine/Methyl diethanolamine (MEA/MDEA) blended amine nanofluids using back propagation neural network (BPNN) model in artificial neural networks. The absorption enhancement factor of TiO2-MEA/MDEA nanofluid were determined experimentally by a two-step method with various nanoparticle solid contents (0.4-1.4 g/L). The results showed that the enhancement factor was firstly increased and then decreased with the rising nanoparticle solid content, and the extreme point appeared at 0.6 g/L. Based on the experimental data, a relevant empirical formula and a back propagation neural network (BPNN) model were used to estimate the enhancement factor of nanofluid and both exhibited good applicability. Additionally, an optimization model incorporating genetic algorithm, particle swarm algorithm and adaptive learning rate (C-BPNN) was also proposed to estimate the enhancement factor. Compared with the empirical formula and BPNN, C-BPNN exhibited a higher prediction accuracy (all data R2=0.9966) and a faster prediction rate. The weight analysis of key parameters (nanoparticle solid content, concentration of MEA and MDEA) showed that the relative importance of nanoparticle solid content was foremost (42.63%) in the absorption enhancement process. All these results indicate that the neural network can provide a guiding role for the research in the field of nanofluid transfer.