|
Abstract
|
In this study, a model for estimating the nanofluids thermal conductivity using a feed forward
artificial neural network (FF-ANN) has been investigated. Nanofluids thermal conductivity was
modeled as a function of nanoparticle size, temperature, nanoparticle volume fraction and the
thermal conductivity of the base fluid and nanoparticles. A one-layer network with training
algorithm of Levenberg-Marquard (LM) has been applied. The obtained results have shown good
accuracy of ANN for estimating the thermal conductivity of nanofluids with sum square error
(SSE) 0.10218 for 24 systems containing 211 data points. Average absolute relative deviation
(?RD) for training, test and validation set data are 2.97, 2.88 and 2.99 %, respectively.
|