The industry-standard MGB approach to predicting the noise generated by a g
iven aerodynamic Row field requires that the turbulence velocity correlatio
n be specified so that the source terms in the Lighthill acoustic analogy m
ay be computed. The velocity correlation traditionally used in MGB computat
ions is inconsistent with a number of basic qualitative properties of turbu
lent flows. In the present investigation, the effect on noise prediction of
using two alternative velocity correlations is examined. (C) 2001 Academic
Press.