A new model for nanofluid conductivity based on the effects of clustering due to Brownian motion

Conductivity values of nanofluids calculated with the model proposed by Nan et al. 1 consistently underestimate the corresponding measured values for 20 sets of experimental data from 12 published studies; thus the conclusion of the recent International Nanofluid Property Benchmark Exercise 2 which stated that it is not necessary to resort to other theories (e.g., Brownian motion, liquid layering, and aggregation) for the interpretation of the INPBE database cannot be generalized. In view of this situation, a new model which takes into account clustering and micro‐convection is proposed and compared with experimental data for five nanofluids (with different particle sizes and a range of particle concentrations) as well as two previously published models. The maximum difference between the predictions of the proposed model and the measured values is 6.7% of the latter. © 2011 Wiley Periodicals, Inc. Heat Trans Asian Res; Published online in Wiley Online Library (wileyonlinelibrary.com/journal/htj). DOI 10.1002/htj.20350