PREDICTION OF STABILITY AND THERMAL CONDUCTIVITY OF SnO2NANOFLUID VIA STATISTICAL METHOD AND AN ARTIFICIAL NEURAL NETWORK
AUTOR(ES)
Kazemi-Beydokhti, A., Azizi Namaghi, H., Haj Asgarkhani, M. A., Zeinali Heris, S.
FONTE
Braz. J. Chem. Eng.
DATA DE PUBLICAÇÃO
2015-12
RESUMO
Abstract Central composite rotatable design (CCRD) and artificial neural networks (ANN) have been applied to optimize the performance of nanofluid systems. In this regard, the performance was evaluated by measuring the stability and thermal conductivity ratio based on the critical independent variables such as temperature, particle volume fraction and the pH of the solution. A total of 20 experiments were accomplished for the construction of second-order polynomial equations for both target outputs. All the influential factors, their mutual effects and their quadratic terms were statistically validated by analysis of variance (ANOVA). According to the results, the predicted values were in reasonable agreement with the experimental data as more than 96% and 95% of the variation could be predicted by the respective models for zeta potential and thermal conductivity ratio. Also, ANN proved to be a very promising method in comparison with CCD for the purpose of process simulation due to the complexity involved in generalization of the nanofluid system.
Documentos Relacionados
- An assessment of neural network and statistical approaches for prediction of E. coli promoter sites.
- Artificial neural network prediction of antisense oligodeoxynucleotide activity
- Photo-Induced conductivity of heterojunction GaAs/Rare-Earth doped SnO2
- Oxido misto de SiO2/SnO2,SiO2/SnO2/Fosfato : propriedades e aplicações
- Influence of pH of colloidal suspension on the electrical conductivity of SnO2 thin films deposited via Sol-Gel-Dip-Coating