does anyone have experience designing/sizing wet vacuum system piping? If so, I'd appreciate your comments on the system problem described below.

thanks
Brian
=======================================================
Our system consists of 3 liquid ring vacuum pumps (300m3/hr each) and a collection tank with 100mm (4 inch) header housed in the utilities building which is connected to a 100mm diameter (4 inch) header spread throughout the factory building. The factory header runs about 5 meters above ground level and the drops to drains are typically 25mm (1 inch) in diameter.
The pumps and tank in the utilities building maintain a vacuum at -80kPa and are also 5 meters above ground level. The header leaves the utility building, drops down 5 meters, travels under the road and rises 5 meters again to the factory header, ie. its forms a huge u-trap between the buildings.
The system has been running like this for 8 yrs but frequently locks up and frequently requires vacuum pump replacements. I was asked to spec a new vacuum pump and investigate the system.
I think the U-bend in the header is a problem but can't explain exactly why. At first I thought the U-bend is probably always flooded but if it is always flooded it wouldn't suck a vacuum downstream and yet it does. Therefore it must only be partially flooded in the horizontal section under the road allowing some air to pass by on the top of the pipe. But if its only partially flooded, then how does the water get sucked upwards to collection tank?
All I can think of is that the system is sucking water vapour caused by the vacuum pressure. This probably works fine when only small amounts of water are introduced to the system but when a large slug of water hits, the U-bend gets flooded and suction stops until water is evaporated off again. If this is the case, then the partially flooded pipe section is effectively reducing the 100mm header to only a fraction of that.
Does this make sense to anyone else?
could the U-bend be causing the lockups? Should the U-bend be eliminated? Should high level wet vac headers be drained by gravity to collection tanks
Web searches for wet vac system design turn up nothing so I'd appreciate any comments from other engineers.

I have seen a similar system (not quite the same) where the water trapped in a horizontal section was carried over by entrainment due to the high vapor velocity.  

The vertical leg on the vacuum pump side of the U-leg could also be functioning as an air lift pump. As the vertical leg has bubbles of low density vapor passing through it the average density of the fluid gets very low and you don't need 5mWG of head to lift it. This is the way a coffee percolator works.

AEBuck has summarised the ways to deal with the U-leg, but it would be better if you could eliminate it completely by going over the road rather than under it. And then try to slope the entire header down towards the vacuum pumps.

You don't say what sort of damage you get to the liquid ring pumps, but these pumps don't like excessive slugs of liquid coming down the suction line. Some sort of trap before the pumps would be good, but again you have the problem of how to drain the trap.

Steam can be a real problem in a liquid ring pump. Because it condenses in the pump the pressure in the pump can go very low, plus the latent heat in the steam raises the temperature of the ring water. The result of this can be cavitation. I have seen liquid ring pump impellers destroyed within a few days by this mechanism. Your pump supplier should be able to give you curves to calculate the effective pump pressure and the likelihood of cavitation.

Back to the problem of draining the U-leg. I haven't thought the 1" pipe system you describe through properly, but my first reaction is that I don't like it. Draw it out and HAZOP it thoroughly with your colleagues.

If you have sufficient head you may be able to install a small drain pot below the U-leg. A small (1 or 2 gallon?) sealed pot valved to allow the water to run into it, and then isolated from the U-leg while another valve is opened to drain it might work. This could be operated manually or put on a cycle timer.

MORE NEWS