Craig,
While the more forgiving taper in the downstream section
does reduce noise by reducing turbulence, its primary purpose is to maintain
metering accuracy. At the most constricted point, the flow will be sonic (choked
flow), assuming there is adequate pressure differential. In the divergent
section, the flow will become supersonic in part of the divergent section.
The taper is more gradual on the down wind side for the same reasons as a wing,
to prevent flow separation and turbulence. If flow separation were to occur
early in the divergent section of the venturi, the pressure differential across
the venturi would be altered and the effective flow constriction could change as
well. The accuracy of the system would then approach that of a simple orifice
plate.
In a past life I used critical flow orifices (fancy name
for sonic venturi) to calibrate hydrogen flow monitors. If you install them
backwards, they don't work so well.
Rob
Yes,
your sketch is correct. Just for further information, the reason for the
diverging cone cross-section being downstream is it helps reduce
noise.
Craig
Berland