This is a post I put up the other day. Aeroponics can work, even on a large-scale, for-profit operation. If there is one piece of advise I could give is if using nozzles, never use less than 150 psi - you just don't get efficient atomization otherwise and will end up wasting a lot of water and nutrients.
I definitely recommend an open-loop system where the solution does NOT get re-used. There is too much opportunity for root disease if you recirculate, unless you add a large UV lamp in the system and increase DO more than you would otherwise have to. This system I built (19,000 SF over 2.5 years) left minimal exposure for the water to pick up anything. Well water was processed through a large DIY reverse osmosis system (>2,000 GPD), stored in multiple 86-gallon pressure tanks, then right when it was time to mist, the pressure tanks pushed it through Dosatrons @ 20 psi regulated, then to a booster pump that took me to between 150 - 225 psi, depending on the size of the circuit.
Once the tubes were misted, I had a "safety" ad the bottom of the tube drain rows in the form a tensiometer that would sense water coming down the drain trough and override the solenoid and booster pump relay to shut down so that I would not waste more than a gallon of nutrient solution per misting cycle (I know a gallon sounds like a lot, but in nearly half an acre of greenhouse, it ain't much).
The Linux program controlling it ran off an uninterrupted power supply and ran literally continuous for 5+ years without having to touch it (haha - take that, Bill Gates!!).
I haven't tried ultrasonic foggers yet, but plan on it in 2012. It will definitely be an open-loop system to reduce headaches. Is that more wasteful than closed-loop systems? Yes, but not by much if engineered properly and the problems one avoids tends to be worth it.
Here's that picture of my old aeroponic greenhouse I was talking about. I'm at the back end of the first structure. Total floor space amongst all three was ~19,000 SF. An old-school computer (it was a 486 PC) running Linux was fed environmental information through the joystick port (all in ohms - temperature, sunlight intensity, humidity, etc.) and the printer port was used to control a bank of 32 relays, all running 24 VAC that in turn controlled real world events (exhaust fans, evap cooling, tube fogging).
Inside those vertical tubes was........... nothing. Just air, with a <5 micron misting nozzle operating at 150-225 psi. Depending on the conditions, it could fire for 5 seconds once every few hours during cooler nights, or for 120 seconds every 5 minutes during sunny, 90-degree plus weather. A lot of air was available to the roots and growth rates were astounding. Plants on the left are 2 days from transplant in 3/4" oasis wedges, plants on the right are 9 days from transplant. In August, I could go from tiny seedling to 1/4=lb of leaf super-bushy basil plants in 19-20 days.
Sorry for not putting my face in the picture. I'm *done* with "face mining" software such as is prevalent on Facebook where emerging facial recognition programs are taking hold. Sophie the Mastiff's head is much more fun to look at anyhow....