This article describes alternate methodology for determining heat transfer correlations from experimental programs. In conventional methodology, the form of heat transfer correlations is determined before the experiment by deduction, and the experiment is performed in order to quantify constants in the deduced correlation form. In the proposed alternate methodology, the form of heat transfer correlations is determined after the experiment by induction.
Three examples in the text apply the alternate methodology to data in the literature, and compare the resultant correlations to widely accepted correlations based on conventional methodology and the same data. The correlations that result from the alternate methodology agree with the underlying data much more accurately than correlations that result from conventional methodology. The difference in accuracy reflects the fact that deduction is much more difficult than induction, and therefore much more likely to include errors, and less likely to accurately describe the underlying data.