A comparison of the predicted thrust and torque
from THPUF-3A, PSF-10, PSF10-FLAG
for propeller DTMB 5168 in fully wetted uniform flow
is shown at the right part of Fig. 8.
THPUF-3A and PSF-10 employ partial wake alignment[10]
while PSF-10FLAG employees complete wake alignment with tip vortex
roll-up [30].
The geometry of the paneled propeller is
shown at the left part of Fig. 8 together with the predicted
shape of the rolled-up
wake surface. The offsets for the propeller are given in [24].
Notice that the panel method predicts the blade forces
more accurately and for a larger range of operating conditions. There
is a ``mystery'' point at 
for which the panel method appears to
be ``inaccurate''.
Finally a comparison in terms of unsteady cavity shapes may be seen in Figures 9 and 10. Notice that the predicted cavities are very inaccurate without the inclusion of the tunnel wall effects[6].
Figure 10: Photographs of cavitating blade in CAPREX (Mishima et al, 1995).
Figure 9: Predicted cavity shapes with (top) and without (bottom) the tunnel effects;
same blade angles and conditions as shown in the next figure,
from Choi and Kinnas (1997).
Figure 8: Geometry of propeller DTMB 5168 and the trailing wake for
the key blade as predicted by PSF-10FLAG (left); measured and
predicted 
and 
by different methods (right).