A novel sensitivity analysis model of EANN for F-MWCNTs–Fe3O4/EG nanofluid thermal conductivity: Outputs predicted analytically instead of numerically to more accuracy and less costs

By Seyed Amin Bagherzadeh, Annunziata D'Orazio, Arash Karimipour, Marjan Goodarzi, Q. Bach. Physica A: Statistical Mechanics and its Applications, 2019

137 citations recorded by the source corpus.

Abstract The new approach of “enhanced artificial neural network” (EANN) is developed based on the new generated hybrid nanocomposite of F-MWCNTs–Fe3O4/EG which represents the Functionalized Multi Walled Carbon Nano Tubes together with Fe3O4 nanoparticles, dispersed in ethylene glycol (EG) as the base fluid. Moreover, a new suitable sensitivity analysis is presented which involves a novel proposed method for the sensitivity analysis via ANNs. In this method, the sensitivity of the outputs predicted by means of an ANN to the inputs is calculated analytically rather than numerically. The proposed method not only provides more perceptive, precise and accurate results, but also requires less computational time and cost. The proposed method can be used for all ANNs having various architectures, training algorithms and input–output data sets.

Subfields: Biomedical Engineering, Electrical and Electronic Engineering, Mechanical Engineering

View Paper: https://www.semanticscholar.org/paper/d5c0194781138be0aa028ba96c651107f4d1778b