Using Venturi effect for air ventilation, or, can fluid be passively driven in a loop? My problem is practical - I have a boat that has a cabinet in it which has a persistent epoxy smell from recent work, causing all clothing stored in the cabinet to also smell of epoxy.  So it needs to be ventilated.  Inspired by carburetors, I was thinking of running a tube from inside this cabinet out the window of the boat, where there is typically wind - in the hopes that this wind would suck air out of the tube the same way fuel gets sucked up through the jet of a carburetor, thereby ventilating the cabinet [Editor's note, the cabinet is NOT airtight].  

The situation, with arrows indicating (desired) speed and direction of airflow.
Then thinking about it more, I realized this requires that in some cases (if wind is blowing from left-to-right, as in diagram) the wind would actually have to be travelling in a loop - which seems somehow unnatural, so I'm not sure if it would actually work. 
 It might actually cause air to flow from the cabinet into the cabin and make the whole cabin smell of epoxy.
Can anyone who understands something about fluid dynamics comment on whether this scheme makes any sense?  Does it only work when air flows right-to-left?  In general, in a passive system like this with no moving parts, is it possible to drive fluid in a loop?
 A: It is quite possible to drive fluid in a loop (in fact a loop or circulation component is essential to the lifting airfoil).
However your proposed solution depends on the relative pressures at the venturi and the cabin door. If the pressure at the door is lower then the circulation will go the wrong way. It will depend very much on wind speed and direction.
My suggestion would be to put up two or three ventilator tubes and turn their tops to point sideways in different directions, and down a bit (but not vertically) as well. This would have two benefits:


*

*Differential pressure in most any direction will drive circulation without impacting the cabin.

*Rain and spray would not soak your clothing so much.


If you might head for green water, do include caps to close the vents off!
A: Having thought about it, I'm pretty sure the answer is:
No, this scheme is impossible (when the wind is flowing left-to-right) if the airflow does not change with time.   We can imagine associating an air-pressure to every pixel in the diagram, creating a "pressure-field".  The flow (arrows) point in the direction of the gradient of this pressure-field.  Gradients point from high-to-low pressure.  It is impossible to follow gradients in a loop, and thus there can be no loops in your flow field. 
However, it is possible when the flow oscillates with time.  In this case, air could be pushed along a high pressure wave that takes it from an on-average-low to an on-average-high pressure, like how you could push a ball up a slightly sloped mattress by always pushing down on the bit of the mattress slightly uphill from it.  One way to have such an oscillating airflow would be to put one of these on the top of the tube: 

