Abstract
The thermal conductivity of continuous (or polydisperse), semi-continuous and discrete dense fluid mixtures has been investigated using the nonequilibrium molecular dynamics method of Evans [1]. Linear, gaussian and unsymmetric distributions have been studied. The results have been used to examine the n-component characterizations for such mixtures and the effect of the various distributions on these characterizations. We show that the 2-component characterization of these mixtures is significantly better than the 1-component characterization and that going to higher order characterizations does not improve results any further. In addition, equimolar characterizations are found to give better results than equiparameter (e.g. [sgrave], ε) characterizations, in the case of unsymmetric distributions.