Pumping from an overflow line of a water tank

[ponderer]

There is a tank(15m height) that is always overflowing during normal operation.

We plan to add a pump to pressurize the overflow stream so that it can be sent to a downstream hydrocyclone for particle separation.

We do not want to add a surge tank in between the tank and the pump for saving the capital.

But we are worry that air may be trapped in the water which may cause trouble for the pump.

Who can teach us about this issue.

 

[SASC]

Ponderer,

Have you guys thought about augmenting a small vacuum source, capable of high-end vacuum, to the peak of the tank? If the air content is relatively small, this vacuum pump can be very small, but fully capable of evacuating all entrained and peripheral free air in the atmosphere prior to the flow stream entry to the pump. This system works! Let me know what else you need to know.
Regards,
SASC

 

[quark]

Some thoughts.

1. If the vacuum maintained in the tank is less than the suction pressure (I am talking in absolute terms)of pump, will the water flow freely to pump?

2. If the vacuum maintained (again absolute) in the tank is more than that of suction pressure of pump, won't the air flow to the pump?

Is it a compulsory requirement that you should take pump suction from overflow? If not take the pumps suction from any of tank bottom nozzles and keep the pump flow rate equal to that of overflow rate. Otherwise my confined soul suggests a surge tank.

 

[SASC]

Quark,

1) Yes, the tank must be configured to have the overflow act as an internal weir....with the vacuum pump positioned at a point some distance above this weir. In other words the vacuum will provide an head, but free flow of water at the overflow will not be disrupted as the overflow is positioned below the vacuum pump.
2) I think you mean that in the instance that the vacuum source provides more vacuum force than the pump provides pumping force? If this is the case, there will still be no issues as the level of the overflow will still be at a point below the vacuum pump and thus the vacuum pump will only receive air, not water. Likewise, the pump will receive only water. No conflict.

I think particulants/debris means you can't pump from the bottom of the tank. This is a liquid-debris separation tank....correct?
Regards,
SASC

 

[Kawartha]

ponderer,

You have not advised whether the overflow is a continuous, steady flow or intermittent. Based on intermittent or varying flow rates, I like Quark's suggestion of taking flow from the bottom of the tank. If this is not possible, can you take flow from some point below the liquid level in the tank, perhaps as much as 1 m from the operating liquid level.

If this is possible, you would be able to control pump opertion with a simple level switch in the tank, with a delay timer to adjust pump operating time (to control starts/hour). An alternative would be to use a pressure switch mounted in the tank with output controlling a flow control valve on the discharge side of the pump. This system would also require a minimum pressure pump shut down point.

If flow is intermittent and you can not take flow from some point below the normal operating level, use of a vacuum pump will not ensure that the pump will have liquid at suction.

Did I miss something else on this SASC?

 

[SASC]

Kawartha & Quark,
A absolute means of combating entrained air content is the sole purpose of what I proposed here with the vacuum pump, not level development or ensuring a flooded intake to a centrifugal pump...thats not at all what I've suggested. O.K., maybe I've adulterated the problem(knowing me, thats very possible!), but I think you are making a few big assumptions: 1) that the "overflow is in the form that it constitutes a partially filled pipe.....in other words the flow is at the liquid level interface, or thereabouts, 2) that the overflow isn't occurring in a full pipe form at a point somewhere below the tank's normal liquid level, 3) that the liquid content of the tank is completely tame, with minimal turbulence and "no" entrained air content(particularly at the level below the overflow!)....this can be a variable to contend with, especially if the transit time is short into the overflow pipe. The point that Kawartha makes about the intermittent flow can be the biggest player here, and this issue has to be addressed by the means he outlined if it exists....there really is no other way if the flow isn't constant. However, if the overflow is either constant or intermittent with entrained air, the idea I've promoted here employing the vacuum pump, and married to Kawartha's outlined method, will be the only sure-fired means to deaerate the liquid to near completion. My idea works equally as well if we have constant flow, and the overflow piping is below the normal liquid surface and contains entrained air, and we won't have to reconfigure piping and we will hold onto the integrity of the tank as a means of separating debris from liquid...effective separation capacity in this tank is the function of the tank and this purpose should not be molested but rather preserved by all means. The method I've suggested works to great effect in many avenues of industry.
Regards,
SASC

 

[quark]

There is nothing like, if you have a system existing somewhere. I just took the analogy of barometric condensers for steam ejectors.

You are hundred percent true that I considered liquid vapor interface as the level for overflow line because if the overflow line always lies below the water level, there are no great issues of air entry in to the suction pipe.

If the vacuum provided by the pump is more (if I say more, then it is gauge pressure) than that of pump suction, it will suck air definitely besides trying to raise(or hold) the water column in the tank. (a differential pressure of 10 mm of Hg raises water column to 132 mm height)

This may cause obstruction to flow.

But as you said you saw these type of systems, I may be wrong. But still I would like to know what is the flaw in my above reasoning.

Regards,

 

[saxon]

ponderer, Why don't you just use a sump pump with a seperate discharge to the hydr-cyclone. Pump control is based on level control in the tank. Remember, pump out flow must be greater than or equal to tank inflow. Leave the ovr-flow pipe alone, this is to be used as an emergency overflow device only.

Hope this helps.
saxon

 

[ponderer]

Thanks for everybody interested in this subject.

I would like to describe more clearly about our process system.

The tank which is always overflowing during normal operation actually is a settling tank.

The purpose of the tank is for particle separation from the water.

Now we find the separation efficiency of the tank is not enough.

So we plan to add a hydrocyclone to further treat the overflowing water.

This mean the water draw-off line must be from tank overflow line , not from tank bottom.

 

[Scipio]

Entrained air isn't the only problem, what kind of particles are we talking about? Will abraision resistance be a requirement in the pump?

One way out might be a progressing cavity pump, they can't be run dry, so some form of level protection on the suction piping would be required, but if I remember right they can handle entrained air fairly well, and can be built for abraision resistance - I've used them to transfer sand-laden water from the bottom of a holdup tank. As far as intermittent operation, if it becomes an issue, might be an idea to run continuously and just bypass if the tank level isn't enough to overflow, so you'll have constant flow through the hydrocyclone.

 

[SASC]

Scipio,

A progressing cavity pump just might be the ticket! Not that familiar with this pump type in actual operations myself, but how do they stack-up versus centrifugals insofar as out of the box price, operational costs, general maintenance, and capacity(which we all have forgotten to address here!)? Looks like the answer here may be a combination of parts from a lot of different initial ideas.....Now isn't that real world engineering in its purest form? Seems to always come out that way.

Regards,
SASC

 

[Scipio]

Out of the box costs aren't too bad, the last set I installed, which were 150#, 2.5 m3/hr flowrate, ran about $6000 to $7000 Canadian, across three bidders. Operational costs aren't much different from centrifugal, maybe a break on power since they're a positive displacement pump, usually can get by with a smaller motor than an equivalent centrifugal pump because of the better efficiency. Maintenance, don't know if there's any pros or cons here...about the same amount of moving parts, plus the elastomer stator, but PCP's have smaller parts, which are frequently harder to get at.

Flowrates are pretty flexible, just a matter of changing pump speed, mine were installed with a VFD, 2.5 m3/hr was at maximum speed, with 21 kPag suction, 248 kPag discharge. Physically the whole thing (pump + driver + baseplate) came in around 200 lbs weight, and came mounted on a 4 foot length of 6 inch channel for a baseplate.

As usual, best to check with some vendors, the two I see the most of are Seepex (

http://www.seepex.com)

and the Moyno's 1000 series (

http://www.moyno.com).

One thing to look out for, you need fairly stable temperatures. PCP's work with a metal screw that worms it's way through an elastomer (usually nitrile or viton) stator with a reverse-helix core. If the temperature fluctuates too much, the gap between the steel rotor and elastomer stator can either open too much and the pump won't pump, or close up too tightly and overload the motor. As long as your process is pretty steady within 20 or 30°C, shouldn't be a problem, vendors can select elastomers to minimize the problem.

 

[ponderer]

The settling tank is used to separate particle of petroleum coke from the water.

The overflow rate is 1000m3/hr.

The required pump head is 4 kg/cm2g.

 

[25362]

In the past we've used Bornemann gear pumps for slurries containing air to our entire satisfaction. Try

http://www.bornemannpumps.com/how.htm

or

http://www.hth.de/sino2000/projekte/bornemann/pcmount1.htm