Confusing problem with a Non-clog centrifugal pump

[Thaer Alashqar]

Hello
We have recently installed a centrifugal pump with screw open type impeller (curve attached) to deliver waste water (sludge) from thickener to drying beds (around 300m distance), and we are unfortunately facing confusing problem since the first start. The best way to describe the problem is as follows:

when the pump is turned on it doesn’t deliver any flow to the other side and the reading on the amperage is very low (half the rated). When the pump starts this way it stays like this until we turn it off.

When the pump is behaving that way, the only way to fix it is while the pump is running we open a valve on the discharge to divert the flow to a nearby tank (10m distance) for a short period (less than a minute) as soon we open the valve the amperage would increase (up to 70% of rated) and the pump sound changes. Then after the minute ends we close the valve and divert back the flow to its desired long path and the amperage reaches the rated and flow would be achieved at the other end and the pump could operate like this for many hours without any problems. But when turned off for an hour or more we face the same startup problem again.

Notes:
1-We installed air release valves on the discharge with no luck
2-we have two pumps in duty- standby configuration and the two behaves in the same way!.
3-the operator is refusing the option of opening the nearby tank valve on each startup.

I would like to hear your thoughts on this matter with many thanks.

 

[bimr]

Please post a diagram of the piping with the elevations.

 

[TugboatEng]

Is this pump self-priming or flooded suction? Either way, it sounds like your discharge pipe has a low point that is trapping fluid. The pump can't produce enough pressure to push through this trapped fluid while it's priming. A recirculation line back to the sludge tank will allow the pump to prime more easily.

 

[Thaer Alashqar]

Attached is a drawing of the piping inside the pumps room, the discharged line (where GV1 & GV2 are installed) is 2.25m above the suction line level. Thickener tank level is 2.5 m above pump’s suction.

The desired path is go drying beds through GV1 and the valve that we open for a minute is GV2

Discharge line is slopped upward to the drying beds.

 

[Thaer Alashqar]

@Tugboateng

Thank you for your reply. Could you please explain how fluid can be trapped in pipes ?

Just outside our room we have a downward elbow (L shape) 0.5m height. Could this be the are were fluid trapped and why the Pump doesn’t push it ?

Also, could you please explain the recirculation solution ? How can we connect it and operate it?

 

[TugboatEng]

I'm just thinking, if your long 300m run has any dips or a total rise that are greater than the diameter of the pipe these will trap fluid and possibly prevent the self-priming of the pump. A recirculation line that leads from the pump discharge back to the source tank would allow the pump to clear air and self-priming. If the recirculation line is small enough you won't need any valves as the volume lost to recirculation will be insignificant.

I see now that your pump is a flooded suction and that changes things. I'll have to think some more. There are no check valves in the system?

 

[bimr]

Are you having this problem with the existing pumps? What is the difference between the new and existing pumps.

Pump suction piping should have eccentric reducers. Your piping diagram is showing concentric reducers. If so, air may be trapped in the suction piping.

 

[Thaer Alashqar]

Tugboateng

A recirculation line back to the tank will be very difficult to install. Can we just connect it to the suction line instead ?

What i understood from your comments is that the pump during priming is not achieving the required head to push all fluid and therefore reducing the head by opening a nearby valve would assist the pump to prime. Is that right ?

Yes it is flooded suction type and we have two check valves at each pump discharge.

 

[Thaer Alashqar]

Bimr

No the existing pumps are not having any issues but they are connected to the holding tanks (the short path).

The reducers installed are eccentric type with flat side on top (we installed it upside down at first but we fixed it after noticing a cavitation sounds) now both pumps sound is normal but it is lauder when the pump is delivering flow.

Existing pumps docs are not clear but they are the same impeller type with differences in impeller and volute diameter (the existing is larger) and therefore there is also a difference in speed (new pump is 2940rpm).

 

[TugboatEng]

Yes, you are following my thinking correctly. The reason the recirc line needs to go to the tank is to allow the air to be separated from the fluid. If you tie it to the suction you'll just re-inject the air into the process. This is all for self-priming which isn't your case.

Does the thicker tank get drained completely? That would be the time your suction piping fills with air. Is it possible to inject water into the suction pipe to prime the pump? Another thought is to place the check valves on the suction side of the pump and incorporating a vertical rise in the discharge piping will cause the fluid to flood back into the pump when the tank goes dry priming the pump for the next use.

 

[bimr]

It is easy to forget to take into account the pump starting conditions which are different than the pump design condition. Sometimes you have to start a centrifugal pump with a valve closed on the discharge side because of too much system resistance and the high torque and discharge pressure required to get flow moving and reach design head. If you have those conditions you need to start the pump against a closed block valve, or employ a control valve that opens just enough at all points along the pump's run up curve to keep it moving to design conditions within the stability margins. Starting against a closed discharge valve allows the discharge pressure of the pump to exceed the pressure of the downstream system resistance.

The solids that you are pumping settle out in the piping and probably add to the startup discharge pressure.

You can demonstrate this by closing GV1, starting the pump, and then slowly opening GV1 while monitoring the pump discharge pressure. You should install a discharge pressure gauge if you don't have one.

Example of automated pump control valve. The pump starts against a closed valve. When the pump is started, the solenoid control is energized and the valve begins to open slowly, gradually increasing line pressure to full pumping head. You will need to talk to the manufacturer to find a control valve that can also handle the solids.

pump control valve

 

[LittleInch]

I'm with bimr here and think there's something happening with your discharge line with respect to the fluid and something settling out or the same in the inlet line.

What's the purpose of the 2" water line? What foes it do?

What is this fluid?

How high above the discharge is the settling pond?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Thaer Alashqar]

We managed to try running the pump while closing the discharge valve for and then slowly opening it. The results were good for 2 out of 3 tries.

The 2” pipe connected to the suction is a pressurized water line that we installed previously when we thought that the problem is with sludge clogging the pump. However, running the pump while this line was open for a couple of minutes didn’t give a good consistent results (sometimes it worked and sometimes it didn’t).

The settling pond level (thickener) is less than 0.5m higher than the discharge. However whenever we open a valve at the top of the discharge pipe its doesn’t flow sludge!!??

Another point worth mentioning is that the discharge header is oversized with the pump flow rate (pipe is 8” and the flow is 1000L/min and pump discharge nozzle is 3”. Do you think this has any thing to do with it ?

 

[bimr]

I believe that you will have to open the valve very slowly.

The solids will be problematic over the long term. You will probably need a velocity of approximately 4 ft/sec to resuspend the solids that settle when you shut the pump down.

 

[Thaer Alashqar]

The valve we are taking about is a gate valve that takes about 5-10 sec to be fully opened, do you think we need to take more time ?

I think the velocity in our case is 1.5ft/sec, do you think this could be our problem specially that if one of the pumps was in good performance the other one will behave the same and the problem will appear again if we shut the pumps down for a couple of hours (sludge will have time to settle in pipes).

Also, do you have any reference for the minimum velocity that we must achieve in our application ?

 

[bimr]

If a force main is installed level, it takes about 3.5 ft/min to resuspend solids on startup. The reference is from Garrs Jones "Pumping Station Design". I think the solids issue is more of a longer term issue than the startup problem.

I think the startup may take longer than 5-10 secs. If you install a pressure gauge, you can monitor startup with the pressure reading.

 

[Thaer Alashqar]

i will keep trying this procedure and hopefully we will find a conclusion.

Thank you for your responses and support.

 

[Thaer Alashqar]

&Tugboateng

I just noticed your comment, the source tank is always full and we tried injecting water to the suction and it didn’t help.

 

[LittleInch]

Another one to try is when stopping the pump turn that water supply in for a minute or two to leave add much of the suction and discard pipe with water in it.

It's not ready to see what is happening from a distance but settle out of the sludge would seem to be a major suspect.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TugboatEng]

The fact that shocking it by opening and closing the discharge valve starts the pumping would make me think a check valve is sticking. The oversized discharge piping will exacerbate this because the valve will barely swing when fluid is flowing and you won't get any scouring from the fluid.